Digital Logic Handbook 1967

mamaomo
LOGIC HANDBOOK
FLIP CHlp MODULES
DIGITAL EQUIPMENT CORPORATION
THE
LOGIC HANDBOOK
FLIP CHIP'TM MODULES
1967 EDITION
FLIP CHIP is a trademark of Digital Equipment Corporation
Maynard, Massachusetts.
Copyright 1967by
Digital Equipment Corporation
Digital Equipment Corporation makes no represen-
tation that the interconnection of its modular cir-
cuits in the manner described herein will not
infringe on existing or future patent rights. Nor do
the descriptions contained herein imply the grant-
ing of licenses to make, use, or sell equipment
constructed in accordance therewith.
FLIP CHIP is a trademark of Digital Equipment Corporation
Maynard, Massachusetts.
II
PART I: DIGITAL LOGIC PRIMER
PART II: FLiPCHIP MODULES
PART III: LOGIC LABORATORY
PART IV: HARDWARE, OCTAIDS AND PANELAIDS
PART V:. ANALOG - DIGITAL CONVERSION HANDBOOK
PART VI: COMPUTER CATALOG
III
IV
PREFACE
GENERAL PURPOSE MODULES
The first section defines the characteristics of the FLlpCHIP'M line of modules which cover
the spectrum of application in three series:
I. The R series which operates from DC to Two Megahertz.
2. The B series which operates from DC to Ten Megahertz.
3. The W series modules for interfacing with various types of external equipment
In addition to these various module lines, a totally new approach to subsystem design,
OCTAIDS and PANELAIDS are described in detail. These design aids provide the user with
a fast, accurate method of assembling such common elements as A to D converters, binary
counters, and real time clocks.
CONVERSION
. The second section of the Digital Logic Handbook is concerned with AnalogDigital
Conversion. A primer on conversion techniques prefaces this section and the complete line
of A series modules and A to D converters follow. The system designer is provided with the
necessary information to. select either a complete A to D converter or the various analog
modules. Application notes are included to further aid in the design of special systems.
HARDWARE
A section which describes Digital's complete line of hardware is included in the Handbook.
New in this issue are the 4096 x 13 bit memory and paper tape punch and reader units.
All of the hardware necessary for the fabrication of a complete system are included, and
each piece of hardware has been field proven in Digital's PDP series computers.
MONOLITHIC INTEGRATED CIRCUITS
The M series is a completely new line of functional TTL monolithic integrated circuit
modules for operation at speeds up to 10 megahertz. These modules are for use where
speed is of primary importance and packing density a consideration. Modules are provided
which perform the functions of shifting, counting, storage, binary to octal conversion.
and gating.
These modules are described indetail in separate literature which can be requested by
using the reply form included in this Logic Handbook.
v
INDUSTRIAL CONTROL MODULES
Digital's K Series Industrial Control Modules provide the answer to the problem of using
solid state logic in the high noise environment encountered in control systems. All silicon
semiconductors and monolithic integrated circuits have been designed into modules
providing all necessary functions formerly accomplished by awkward relay devices. K series
modules and hardware are designed for installation in standard NEMA enclosures.
Connectors for these modules are the field proven FLIP CHlp connectors which have
been used on two generations of Digital's computers and with-modules in every
conceivable application from steel mills to lathe controls. Coimection between terminal
strips and electronics are also pluggable, allowing the logic to be installed after field wiring
is complete. Standard functions of gating, storage, and counting are provided, plus
industrialoriented AC Input Converter, AC Switch, Timer, Interface Block, and Glow Tube
(Indicator) Driver. Sensing and output circuits operate at 115 volts AC for complete
electromechanical compatability. Solid state AC switches are fully protected against false
triggering, and provisions for interlocking are included.
Checkout and trouble shooting are easy with K series modules. Each system input and
output has a builtin indicator light and a special test probe provides its own local
illumination and built in memory fpr transient signals as well as steady states.
K series modules provide the solidstate advantages If size reduction, reliability, flexibility,
and low cost logic with an added bonus of easy interconnection. For a complete set of
specifications and helpful design examples send for the Industrial Controls Handbook by
using the reply form attached.
VI
TABLE OF'CONTENTS
PREFACE
PART I DIGITAL LOGIC PRIMER..
NUMBER SYSTEM
Counting in Different Number Systems.
Binary Number System.
BinaryDecimal Conversion
BinaryCodedDecimal Numbers.
Octal Number System.
Notation
BOOLEAN ALGEBRA .
OR Function
AND Function
Identities
Complement
De Morgan's Laws.
Boolean Algebra for use with Voltage Levels
BINARYCODED DECIMAL CODES
FourBit Codes ..
Arithmetic Operations with the 8421 or Excess 3 Codes
Codes Greater than Four Bits
PART II FLIP CHIP MODULES.
R SERIES - BASIC LOGIC MODULES (Dc TO 2 MHz) .
Basic Circuits ..
Logic Configurations
Specifications
V
1
2
3
4
5
5
6
7
8
8
9
10
11
11
11
16
16
17
21
23
25
26
38
57
B SERIES - NEGATIVE, HIGHSPEED LOGIC MODULES (Dc to 10 MHz)
Basic Circuits.
83
84
93 Logic Configurations
Specifications
W SERIES - DIGITAL LOGIC INTERFACE MODULES ..
Introduction
Specifications
APPLICATION NOTES
Estimating Propagation Delay
BCD Counting ..
General Purpose Digital Clocks.
HighSpeed Parallel Adders .
VII
..... 108
... 129
..... 130
..... 132
...175
...... 176
............. 177
. ......... 184
. ... 190
32Position Decoding. ......... ..................................... 194
Stepping Motor Drives (Translators) . . .......................................... 197
Generation of Pseudo Random Sequences .... ................. .202
PART III lOGIC lABORATORY.. . ........... .205
Introduction ........................................ .206
Specification ....... 212
PART IV HARDWARE, OCTAIDS AND PANELAIDS . . ................................... .221
.......................... 222 HARDWARE ACCESSORIES
Specifications ................ .... 227
OCTAIDS AND PANElAIDS - PRINTED CIRCUIT BOARD KITS ...... .251
Octaid Specifications. ... ... . . ... ................ 252
Panelaid Specifications ............................. .265
E and F Series - Printed Circuit Boards ............ .................. 273 .
PART V-ANAlOGDIGITAl CONVERSION HANDBOOK.
PREFACE
275
... .276
CHAPTER I BASIC ELEMENTS OF CONVERSION. . .................... 277
CHAPTER II MEASURES OF CONVERTER PERFORMANCE. . ........ .285
CHAPTER III SPECIAlANAlOGTODIGITAlCONVERSIONTECHNIQUES 297
CHAPTER IV TYPICAL CONVERTER lOGIC . . ............................ 306
CHAPTER V BASIC CIRCUITS ...... 317
CHAPTER VI INTERCONNECTION AND CALIBRATION ....................... 329
CHAPTER VII TESTING AN ANALOGTODlGITAl CONVERTER. 337
GENERALPURPOSE ANALOGDIGITALCONVERTER/MULTIPL EXER ........ 341
A SERIES MODULES - ANALOG LOGIC MODULES. . ...... 343
PART VI COMPUTER CATALOG . . ... 357
APPENDIX I MIL-STD.-806BandDECSymboIComparison ..................... 377
APPENDIX 2 Powers of Two Table.
APPENDIX 3 Abbreviations
APPENDIX 4 Definitions.
APPENDIX 5 Bibliography of Digital Logic ..
NUMERICAL INDEX.
PRICE LIST .......................................... .
VIII
. ........................... 385
.............. 386
. ............. 387
. ................... 388
. ..... 391
..393
RAND B SERIES MODULE FEATURES
COMPLETENESS OF LINE
A full selection of high-volume standard modules with sufficient types and accessory
hardware for building complete systems. 150 items in the current product line with 70
items in stock for immediate delivery.
FREQUENCY RANGE
Dc to 10 MHz in two fully compatible series.
SIGNAL LEVELS
o volts and -3 volts.
FAN-OUT
70 rna from pulse amplifiers
18 rna from diode gates
15 rna from typi"cal flip-flops
(typical input loads: 1-3 rna)
CLAMPED LOGIC
Signal voltages are independent of loading. All signals clamped to -3 volts.
WIDE TEMPERATURE RANGE
FLIP CHIP silicon modules operate over a temperature range of -20'C to +65C.
GOOD NOISE IMMUNITY
All low speed modules and some high speed modules have diode isolated inputs. Typical
noise rejection for diode gates: at Ov - l.Ov
at -3v-l.5v
LOW POWER DISSIPATION
Typically: 150 mw per flip-flop
40 mw per diode gate
SIMPLE POWER REQUIREMENTS
Two supply voltages required: +10v and -15v. All modules have standard power con-
nections. Complete line of power supplies available as standard items.
ULTIMATE IN LOGICAL FLEXIBILITY
Extremely flexible flip-flop configuration permits JK, RS, RST, or T memory elements to
be constructed without modifying modules.
Gates and power amplifiers may be paralleled for performing positive OR functions.
A wide range of interfacing modules are available for converting standard DEC levels to
external equipment reuirements.
IX
GENERAL CHARACTERISTICS- RAND B SERIES MODULES
FREQUENCY RANGES
Dc to 2 MHz (R Series)
Dc to 10 MHz (B Series) .
LOGIC LEVELS
Ovto -0.3,upper level
-3.2 v to ~ 3 . 9 v lower level
HIGH FAN-OUT
High driving capability for all modules.
Typically: 70 ma - pulse amplifiers
18 ma-diodegates
15 ma - flip. flops
HIGH FANIN
Low input current requirements. Typically 1 ma to 3 mao Diode gate inputs may be
expanded as high as 25 inputs with gate expander modules.
ALL LOGIC LEVELS DIODE CLAMPED
Signal voltages are diode clamped at -3v independent of fan-out and independent of
other input conditions. at the load.
LOW POWER DISSIPATION
Typically: 150 mw per JK flip:flop
40 mw per diode gate
CONSERVATIVE DESIGN
All circuits can tolerate at least 20% variations in power supply voltage.
NOISE IMMUNITY
Diode Gates
DCD Gates
10 MHz Inverters
At 0 volts
1.0v
O.7v
0.5v
TEMPERATURE RANGE
At -3 volts
1.5v
Totally insensitive
0.5 volts
-20C to +65C on all silicon FLIP CHIP modules. A few W-Series power driving
accessory modules include germanium semiconductors as indicated on their data
sheets, reducing their upper limit to +55C.
x
COMPATIBILITY
Many specialized interface modules available providing:
- Input voltage compatibility between 30v.
- Output voltage compatibility between 135v.
- Output currents as high as 10 amperes.
ANALOGDIGITAL CONVERSION
Complete selection of interface modules for building hybrid configurations including:
. Comparators
D to A converter modules
Analog switch modules
Reference supplies
POWER REQUIREMENTS
+10v nominal, module pin A
-15v nominal, module pin B
ground, module pin C
BOARD SPECIFICATIONS
Material- GIO FR4 Glass Epoxy
Copper Coating - 2 oz. (0.0028 inches)
Thickness - 0.055 .003 inches overall
Gold Plated Contacts - 0.00015 inch gold on copper
STANDARD FLIP CHIP WAVEFORM DEFINITIONS
R Series STANDARD PULSE
o V----------,..,.""_
.-3V---
I
I
I
I
I
--.j
10%
XI
L- 100 nsec -0,
t-'""" + 75 n sec
8 Series
ov
STANDARD PULSE
I
I
1
I
I
" I
100/0
i4- 40 :t 10 n sec
Edge Requirements for activation of R Series oeD pulse inputs
I I
o v - - - - - ~ - - - ~ - - . ~ ~ - - -
I 90%
-3V---- 10%
I
I
XII
I
I
I
I
1
14- 60 n sec max
;w ... ..,... .i.-Itlll
FLIP CHIP assembly line combines automated manufacturing steps with com
puter controlled checkout for lower cost, more reliable circuits.
X111
Printed circuits on FLIP CHIP boards are pr3duced in an etchingplating line
which includes an ISstep process to gold plate the plugin contacts.
XIV
Ten spindle automatic machine drills 40 boards si!l1ultaneously, provides pre,
cision component layout.
xv
Discrete components are positioned and crimped in place at rates up to 30
per minute on pantograph controlled inserting machine.
'XVI
High quality diffusion furnaces used for the generation of semkonductor
junctions.
XVII
Substrate screening and semiconductor mounting operations are performed
under clean room conditions.
XVIII
Thermal compression bonding ties semiconductors into hybrid networks.
XIX
To insure reliability, a wide variety of dynamic tests are performed online by
a computer controlled system.
xx
PART I: DIGITAL LOGIC PRIMER
NUMBER SYSTEMS
Early number systems were crude and awkward to use.A simple system, using a mark for
each unit, cannot be used to express large quantities such as a thousand. Later sys-
tems, such as Roman numerals, were a great improvement, but still extremely difficult
to manipulate in ordinary arithmetic. With the Arabic, or decimal, number system, com-
mon arithmetic operations, which correspond to true to life operations, can be defined
and easily used.
The decimal number system uses ten symbols representing the quantities a through 9.
Other numbers are constructed by assigning different values (or weights) to the position
of the symbol relative to the decimal point. For example, the number 008 (more com-
monly written simply 8) represents eight units, while the number 080 (again more
commonly written just 80) represents a quantity of eighty, and the number 800 repre-
sents a quantity of eight hundred.
Each position in a decimal number has a value which is ten times the value of the next
position to the right. In other words, every positional weight is a multiple of ten and can
be expressed by ten raised to some power: The tens position is 10" the hundreds position
is 10
2
, the thousands position is 10
3
, etc.
Simple' exponential arithmetic shows that the ones position is 100 = 1 (in fact, any
number raised to the a power, except 0, is equal to 1). This progression of increasing
exponents can be continued as far as desired to the left of the deCimal point. The
'same progression can also be extended to the right of the decimal point, but here the
exponents will be negative. For example, the first position to the right of the decimal
point is the tenths position, it has a weight of 10-
1
or 1/10
1

Figure 1 represents a general skeleton for any decimal number. The symbol which is
placed in any of the positions indicates how many multiples of that power of 10 are
in the total quantity represented by the number.
[3] 10
3
1
102
1
101
1
100
H 10
01
11002110031 ~ B
Figure 1
Ten is not a particularly magical number; there is no reason for the number of symbols
to be extended to ten (or limited to ten). It would be just as simple to have twelve
symbols or eight symbols or two symbols, or any other number of symbols. However,
one of the features of the decimal system is that there is only one way in which any
given number can be written, and on seeing a number written, there is only one value
which can be ascribed to it. I norder to keep this feature in a number system with a
different number of symbols, it is necessary to change the weights of the different
positions. The values which must be assigned, in fact, turn out to be powers of the
2
number of symbols available. The number of symbols used is called the radix of the
number system. Figure 2 shows the skeleton of a general number system with a radix R.
R\ I RO H R\ I R2 I R3 I J ~ ~
. Figure 2
Examples of counting in different number systems are illustrated in Figure 3. The duo-
decimal number system has a radix of .12 and the symbols A and B are used here to
represent the quantities 10 and II, respectively. The octal number system has a radix
of 8 and the binary number system has a radix of 2. The positional weights are given hi
decimal at the top of each column.
COUNTING IN DIFFERENT NUMBER SYSTEMS
DECIMAL DUODECIMAL OCTAL BINARY
5
i
w e x e
e i t i
0
I
0 g 0 e
g
0 t 0
e n
v
n h n e
h
u w n
n
e e
e t e n t 0 e
5 S s s S 5 5 5 5
---
0 0 0 0
1 1 1 1
2 2 2 1 0
3 3 3 1 1
4 4 4 1 0 0
5 5 5 1 0 1
6 6 6 1 1 0
7 7 7 1 1 1
8 8 1 0 1 0 0 0
9 9 1 1 1 0 0 1
1 0 A 1 2 1 0 1 0
1 1 B 1 3 1 0 1 1
1 2 1 0 1 4 1 1 0 0
1 3 1 1 1 5 1 1 0 1
1 4 1 2 1 6 1 1 1 0
1 5 1 3 1 7 1 1 1 1
1 6 1 4 2 0 0 0 0 0
1 7 1 5 2 1 0 0 0 1
1 8 1 6 2 2 0 0 1 0
1 9 1 7 2 3 0 0 1 1
2 0 1 B 2 4 0 1 0 0
2 1 1 9 2 5 0 1 0 1
2 2 1 A 2 6 0 1 1 0
2 3 1 B 2 7 0 1 1 1
2 4 2 0 3 0 1 0 0 0
Figure 3
3
BINARY NUMBER SYSTEM
Since the binary number system uses two symbols, it has a radix of 2 and the positional
weights are powers of' 2. Examination of the binary counting sequence, Fjgure 3,
shows that the binary number system follows the same number system skeleton which
was previously outlined. Because of this, the method of performing arithmetic operations
in binary numbers is the same as the methods used for decimal numbers. For example,
0+ 0 = 0, and 0 + 1 = 1. Since there is no symbol for two, however, 1 + 1 = 0 and 1
to carry. The tables for performing arithmetic operations are given in Figure 4. Since
there are only two symbols, the tables are considerably simpler than those required to
outline the same operations in the decimal number system. This, of course, leads to
considerably simpler computer circuitry also. Some typical examples -of arithmetic
operations in binary are shown in Figure 5.
Binary Addition
A + B = S (Sum)
0+0=0
0+1 = 1
1 + 0 ~ 1
1 + 1 = 0 & 1 to carri
Binary Multiplication
A X B = P (Product)
OXO=O
OX1=0
1XO=0
1 X 1 = 1
AdditiQn
-1--
101101
+ 1010
110111
Multiplication
101101
X 101
101101
00000
101101
11100001
11
101101
+1100
1H001
Figure 4
Figure 5
4
Binary Subtraction
A - B = D (Difference)
0-0=0
-0 - 1 = 1 & 1 to borrow
1-0=1
1 - 1 = 0
Binary Division
A -;-. B = Q (Quotient)
0-;-.0 =?
0-;-.1=,0
1-;-.0=?
1 -;-. 1 = 1
Subtraction
101101
-- 1100
100001
Division
1001
101 V101101
101
0001
0000
---10
00
101
101
101101
- 11001
10100
BINARY-DECIMAL CONVERSION
Numbers can be converted from binary to decimal and vice versa by hand by using
the methods outlined in Figure 6. As shown, a binary number is converted to decimal
simply by adding the positional weights of all those positions where a 1 appears.
Decimal to binary conversion is more a process of trial and error. First, subtract the
largest power of 2 which will go into the number that is being converted. This process
is repeated on the until the remainder is equal to o. The binary. number
then has ones in those positions with the values corresponding to the powers of 2
which were subtractedj all other positions are O.
BINARY TO DECIMAL CONVERSION
32 .!... JL..!.. .1.. ..l..
o 1 0 1 = 32 + 8 + 4 + 1 = 45
o 0 = 8 + 2 = 10
1 o 1 = 32 + 16 + 4 + 2 + 1 = 55
DECIMAL BINARY CONVERSION
45
32 16 8 4 2
' 1
0 1 1 0
13
t
I
-8
5"

"0
Figure 6
BINARY-COOED-DECIMAL NUMBERS
Since computer inputs and outputs must often be in decimal notation, a variety of
special codes .are used. These hybrid number systems are referred to as binary-coded
decimal or BCD.
An example of BCD is the 8421 code. This is often referred to as simply BCD since the
weights of the positions are the same as in the binary number system, as illustrated
below.
5
Decimal 8421 Code
0 0000
1 0001
2 0010
3 0011
4 0100
5 0101
6 0110
7 0111
8 1000
9 1001
The 8421 code employs four bits to represent each decimal digit. For instance, the
number 987 may be represented by the 12-bit number 1001 1000 0111. Although this
number contains only ones and zeros, it is not a true binary number since it does not
follow the rules previously established. Arithmetic operations with BCD would be quite
involved. However, it is relatively easy for the computer to convert to true binary,
perform the necessary calculations, and reconvert to BCD.
'BCD numbers do not always follow the pure binary number system. Special purpose.
number systems such as Excess Three Code, Gray Code, and Biquinary Code are often
used.
OCTAL NUMBER SYSTEM
As the name implies, the' octal number system has a radix of 8, i.e., it uses eight discreet
symbols: 0, 1, 2, 3, 4, 5, 6, and 7. The, positional weights in, the octal number system
are powers of 8.
The octal number system is widely used by digital engineers and computer programmers
since it can easily be converted to binary. At the same time, it is considerably easier to
work with, or to record, octal numbers than to use a long string of binary zeros and ones.
The binary-octal conversion may be performed 'quite simply due to the fact that 8 is the
third power of 2. This produces a direct correlation between the successive 3-bit
groups in a binary number and the octal digits. That is, an octal number may be con-
verted to binary digit by digit, while with a decimal number the entire number must
be converted to binary. The table for octal to binary conversion is shown in Figure 7.
OCTAL TO BINARY CONVERSION
Octal
O'
1
2
3
4
5
6
7
Figure 7
Binary
000
001
010
011
100
101
110
III
Using this table, the octal number 777, for example, could be easily and directly con-
verted to the binary number 111111111. Going in the opposite direction, the binary
number 110101110 can be converted directly to 656. (As in other number systems,
zeros are always assumed in the most significant bits. For example, the number
1110110 converts to 166 in octal.)
Arithmetic operations in octal are quite similar to the operations in decimal. A more
detailed discussion of this can be found in some of the reference books in the bibli-
ography.
NOTATION
When there may be some doubt as to the number system being employed, it is customary
to indicate this by writing the radix of the number system (in decimal) as a subscript
to the number. For example, 7778 indicates that this is the number 777 written in the
octal number system. The same number in the decimal system would be 511
11
),
In working with different number systems it is extremely important to be certain
which system is being used. For example, take that tricky little question, "what's two
and two?"
2 + 2 = (meaningless for radix of 2)
2+2=113
2 + 2 = 104
2 + 2 = 46 or more.
7
BOOLEAN ALGEBRA
Boolean algebra was introduced in 1847 by an English mathematician, George Boole.
The purpose of the algebra was to find a shorthand notation for the system of logic
originally set forth by Aristotle. Aristotle's system dealt with statements which were
considered to be either true or false, but never partially true or false. Boole's algebra
was based on a single valued function with two discrete possible states.
Boolean algebra lay almost dormant until recent times. Today, however, it is gaining
widespread recognition as an efficient method for handling any single valued function
with only two possible states. When it is applied to binary arithmetic, the two states are
o and 1. When discussing a switch, the two values are open and closed.
Figure 8 Switch Analogy
The convention used will be that the open state corresponds to the 0 state, while the
closed state corresponds to the 1 state.
OR FUNCTION
~ .
y
c
~
0+ 0= 0
0+ 1 = 1
1 + 0= ,
1 +' = 1
Figure 9 OR Function
If two switches, A and B, are connected in parallel to form a gate, inspection shows that
the gate can only transmit information if A or B or both are in the transmitting state,
8
i.e., closed. This is written in equation form as
A+B=C
(A or B equals C)
Figure 9 shows the parallel combination of two switches along with a table giving the
value of C for all possible values of A and B.
(A+B)+C A+(B+C) A+B+C
Figure 10 Compound OR Functions
By adding a third gate in parallel, as in Figure 10 it becomes obvious how the OR
function may be extended to any number of variables. This figure also serves to illustrate
that the communicative and associative laws are valid for the OR function, i.e.,
A+B=B+A
(A + B) + C == A + (B + C) == A + B + C
AND FUNCTION
Q
AB= c.
00= 0
0':;;; 0
, 0= 0
, ,:;;; 1
Figure 11 AND Function
If two or more gates are placed in series, the result is known as an AND gate. Inspection
of the arrangement in Figure 11 shows that the resulting gate will transmit only if
both A and B are closed, i.e., equal to 1. The equivalent equation in Boolean form is
AB = C
(A and B equals C)
9
ACBC) :: ABC 5' CBA
Figure 12 Compound AND Functions
Figure 12 demonstrates. how the AND function is applied to more than one variable.
The commutative and associative laws also hold.
AB = BA
A(BC) ::: (AB)C ::: ABC
IDENTITIES
1+A=1
O+A=A 1A =A OA=O
\ v\
f\o
\B
\c
y
y
A (B+C) AB+AC
Figure 13 Indentities
10
To enable the simplification of Boolean functions, there are many identities which are
helpful. In Figure 13 the combination of switches and corresponding equations
demonstrate these identities.
COMPLEMENT
If two gates are connected so that the same signal will open both of them or close both
of them simultaneollsly, the switches are given the same symbol. If two gates are
connected so that a single signal will open one gate while closing the other gate, and
vice versa, these gates are said to be the complement of each other. Thus, if one gate
is labeled A, the other gate will be labeled A (read "not A" or "A not").
An entire function may also be complemented. For example,
if
then
D = A(B + C)
5 = A(B + C)
The use of one label for more than one gate makes the following identities helpful:
A+A=A
AA = A
A + A =1
AA = 0
DE MORGAN'S LAWS
Two unique laws which can be applied only to Boolean algebra are known as De Morgan's
laws.
A + B + C + ... + N = ABC ... N
ABC ... N = A + B + C ... + N
These laws may be verified by constructing a table of various possible values.
BOOLEAN ALGEBRA FOR USE WITH VOLTAGE LEVELS
Since DEC voltage levels have only two possible values, Boolean algebra can also be
useful in the study of these levels.
11
OUTPUT
INPUT
PULSE -1
LEVEL
INPUT
Diode-Capacitor-Diode
Gate
COLLECTOR OUTPUT
-+
(NEGATIVE FOR POSITIVE BASE
OR NEGATIVE EMITTER
POSITIVE FOR NEGATIVE BASE
BASE AND POSITIVE EMITTER)
INPUT .
EMITTER
Inverter
Clamped Load
COLLECTOR
DIODE=qi
INPUTS
. EMITTER
NODE
INPUT
Diode Gate
Negative NAND
Positive NOR
ZERO OUTPUT ONE OUTPUT
(NEGATIVE WHEN FF IS ZERO I 1 (NEGATIVE WHEN FF IS ONE
GROUND WHEN FF IS ONE) GROUND WHEN FF-IS ZERO)
,-'0-----"--,
DIRECT CLEAR
FF
I I
ZERO INPUT ONE INPUT
Flip-Flop
LEVEL OUTPUT IS
NEG DURING DELAY
DIRECT SET
,,,,,--d
INPUT -1'-_____ . J ~ OUTPUT
Delay
(One,Shot) General Element
Figure 14 Device Symbols
12
Figure 15 Inverter
An inverter or single input diode gate may be used to perform a complement. If the emitter
is at ground and a signal is applied to the base, the resulting output is the complement of
the base input as shown in Figure 15.
* ~ * *--.--r--* * ~ * f G
A 8 C 0 E F ~ 5 V
NEGATIVE AND GATE
ASCDEF. G
Figure 16 Diode AND Gate
R, S, and W Series Standard levels are defined as -3 volts = 1, and 0 volts = 0, hence
a simple AND gate is formed by diodes shown in Figure 16. Only if A and Sand C and
D and E and Fare negative,will the resulting output G be negative.
Inspection of these diode units shows that the AND gate becomes an OR gate if the levels
are defined in the opposite manner; i.e., if ground is defined as a 1 and -3 volts is
defined as a O. This is a demonstration of De Morgan's law.
The use of symbols for signal definitions helps the designer keep track of these definitions
if he wants to change conventions in a system. The symbols used are shown in Figure 17.
Figure 18 shows the basic diode gate with its corresponding negative and positive
logic definitions. .
-----.
--<>

,
Negative Pulse
Negative Logic Level Positive Logic Level
or
Negative going Level Change
C>
~
~

Positive Pulse
or
Nonstandard Signal
Positive going Level Change
Branch Points
++
--+-1
JO
I
bd
0 I
~
FF FF
- -
Common Pulse Lines
Figure 17 Symbols for Standard Signals
13
C" AS" (AtS)
NEGATIVE NAND
POSITIVE NOR
Figure 18 Diode Gate Definitions
Regardless of level convention chosen, there is only one unique state which indicates
coincidence of inputs for the diode gate shown. That is, if, and only if both inputs are
-3 volts, the output will be at ground. This means that in a decoding situation, the only
active decoder output is ground in either definition of logic levels.
The circuit design is such that a number of gates may be connected together at a common
point to produce a positive OR function. In Figure 19, if either gate output goes to ground,
the output of the logic will be ground. Thus a convenient function is derived.
A -<>Dh--i:::iJ
S
F"AStCD" (AtBHCtD)
NEGATIVE
DEFINITION
C
YCI<I_D
F" (AtS) (CtD)" II B +eD
POSITIVE
DEFINITION
III--.--Kl<>- C
D
Figure 19 DR Gate
The DCD gate performs a positive AND function. If the DCD gate could be used as a separate
logic element (not in conjunction with pulse amplifiers or flip-flops) its positive and negative
definitions would be as shown in Figure 20
14
Figure 20 DCD Gate Definitions
When DeD gates are used in conjunction with an R602 Pulse Amplifier, for example, the
functions of the combined elements are defined as shown in Figure 21.
B o
poAB+cii
NEGATIVE
DEFINITION
poAB+CD
POSITIVE
OEFINITION
Figure 21 Definitions, DCD Gates with Pulse Amplifier
15
BINARY-CODED DECIMAL CODES
The digital computer can be thought of as an assemblage of two-state devices because
it manipulates the ones and zeros of the. binary number system. People, on the other
hand, are more accustomed to decimal numbers, and for this reason it is often desirable
to build a computing system which can be operated in decimal.
To build a decimal computer with two-state devices, it is necessary to encode the
decimal digits with binary bits. Four binary bits are needed. Although only 10 of the 16
permutations possible with the 4-bit decade will be used, all are available. The number
of codes that can be generated is calculated as follows:
l 6 ~ ! == 2.9 x laID
The choice of a code is obviously important. Desirable features of the code are: ease in
performing arithmetic operation, economy of storage space, economy of gating opera-
tions, error detection and correction, and simplicity. Several possible codes are shown
below, followed by a detailed explanation of arithmetic operations using two especially
convenient codes, the 8 4 2 1 and the Excess 3.
FOUR-BIT CODES
The 8 4 2 1 code is commonly referred to simply as binary-coded decimal because
the weights of the positions are the same as in the binary number system. Arithmetic
operations are easily performed using the same basic method as in binary since the
number sequence is the same.
In the Excess 3 code, a decimal number D is represented by the binary equivalent of
the number D + 3. The Excess 3 code is not a weighted code, but since it follows the
same number sequence as binary, it is useful in arithmetic operations. Addition is
facilitated since the need for a correction factor is easily detected and easily imple-
mented. Because it is self-complementing, the Excess 3 code is also useful in subtraction.
The 2 4 2 1 is a self-complementing weighted code which is commonly employed in
counting systems. Other examples of four-bit weighted codes include the 5 4 2 1, the
53 1 1, and the 7 4 -2 -1 code. All of these codes are shown in Figure 22 .
More than four bits may be used in each decade to provide additional special features
such as the detection of errors and the simplification of decoding.
16
DECIMAL 8421 Excess 3 2421
0 0000 DOll 0000
1 0001 0100 0001
2 0010 0101 0010
3 0011 0110 0011
4 0100 0111 0100
5 0101 1000 1011
6 0110 1001 1100
7 0111 1010 1101
8 1000 1011 1110
9 1001 1100 1111
DECIMAL 5421 5311 7421
0 0000 0000 0000
1 0001 0001 0111
2 0010 0011 0110
3 0011 0100 0101
4 0100 0101 0100
5 1000 1000 1010
6 1001 1001 1001
7 1010 1011 1000
8 1011 1100 1111
9 1100 1101 1110
Figure 22 Four-bit decimal codes
ARITHMETIC OPERATIONS WITH THE 8421 OR
EXCESS 3 CODES
Because the 8 4 2 1 and the Excess 3 codes follow the same number sequence as the
binary number system, standard binary methods may be used_ However, in binary
notation sitxeen states are represented with four bits_ In binary-coded decimal only
ten of these states are used; therefore, special correction factors must be added to
account for the six unused states_
Counting
In a binary-coded decimal (BCD) counter, the corrective action is very simple_ The
counter is divided into four-bit decades, and special gating is added to each decade_
This gating detects the number 9 and reroutes the next count pulse so that it will
reset the decade to 0 and generate a carry to the next decade.
In a down counter, the same approach is used_ Starting with a standard binary down
counter, the number 0 is detected, and the next count input resets the counter to the
appropriate 9 designation and produces a borrow.
A reversible BCD counter may be implemented by combining the techniques for the
individual up and down counters_ Such a counter, however, is more difficult to construct
than a single direction counter since provision must be made for isolating the carry
and' borrow chains and for assuring that count up and count down signals do not occur
simultaneously.
17
Addition
A common method of performing BCD addition is to add two numbers in the binary
adder and, if necessary, add or subtract an appropriate correction factor (see Figure
A19). When addition is to be performed in a decade by decade fashion (serial addition
with parallel decades), either code is useful. If addition is performed in parallel, however,
the Excess 3 code is superior to the 8 4 2 1 code.
In 8 4 2 1 code the sum will be correct if it does not exceed 9. If the decimal sum is
between 10 and 15, it is necessary to add +6 to the binary sum and generate a carry
to the next decade. If the decimal sun exceeds 15, a carry signal is generated by the
initial addition, but the correction factor +6 must still be added to the binary sum.
Addition of 8 4 2 1 coded numbers has the disadvantage that a carry signal can be
generated during the correction process. For this reason each decade in the adder
has to be corrected individually. Therefore it is nota desirable code in a parallel adder
(see Figure 23).
No correction necessary
(Sum ::s. 9)
0100 = 4 dec.
0011 = 3 dec.
0111 = 7 dec.
Initial sum in forbidden state
(10 ::s. sum :S15)
1000 = 8 dec.
0100 = 4 dec.
1100 = forbidden state
0110 = correction factor of +6
1 - 0010 = 2 dec. plus carry
Initial sum in incorrect notation
(16 ::s. sum ::s. 18)
1001 = 9 dec.
1000 = 8 dec.
1 ---- 0001 = 1 plus carry
0110 = correction factor of + 6
0111 = 7 dec. (plus carry from first
addition)
Multiple decade addition
0101 0101 1000
QQ!Q Q!.QQ 0011
0111 1001 1011
= 558 dec.
= 243 dec.
: d : ; ' ~ ~ ~ ~ } = 801 dec.
1--0000
1000
Figure 23 Addition with the 8 4 2 1 code
When two Excess 3 numbers are added, the sum will contain an excess 6; if the
decimal sum is 9 or less, it is necessary to subtract 3 in order to return to Excess 3
notation; if the decimal sum is greater than 9, the excess 6 contained in the sum
cancels the effect of the six unused binary states, but it is necessary to add 3 to
return to the Excess 3 notation.
18
Whether the correction factor is +3 or -3 is determined by whether or not a carry
signal appears during the initial addition. An initial carry requires a positive correction;
no carry, a negative correction. The correction process will never yield an additional
carry, thus simultaneous correction of all decades is possible.
The steps for performing Excess 3 addition are:
1. Add the two BCD numbers in binary fashion
2. Check each decade for a carry signal
3. Subtract 3 from each decade in which a carry has not occurred, while
simultaneously adding 3 to each decade in which the carry signal
has occurred.
The +3 correction is made by adding 0011 to the appropriate decade. Subtracting 3
from a decade is done by adding 1100 and using the end around carry from the most
significant bit of the decade. This is a method of 9's complement subtraction, described
under subtraction below.
Sum ::;'9
0111=4dec.
Q!.!Q = 3 dec.
1101 = uncorrected sum
- 00 11 = correction factor of - 3
1010 = 7 dec.
Sum 2: 10
1011 = 8 dec.
Qll!. = 4 dec.
1 - 0010 = uncorrected sum
0011 = correction factor of + 3
0101 = 2 dec .. plus carry from
initial addition
Multiple Decade Addition
1000
0101
1000
01ll
lll0 _ 0000
-0011 +001l
lOll 001l
1011
0110
0001
+001l
0100
Figure 24 Addition with the Excess 3 code
Subtraction
558 dec.
243 dec.
801 dec.
Since subtraction is the inverse of addition, the same circuits may be used for both
operations. Subtractions by this process is known as the system of adding complements.
The 9's complement of any number is that number which is obtained by subtracting each
individual digit from 9. With a self-complementing BCD code, such as Excess 3, the
9's complement of any number can be easily obtained by changing all zeros to ones
and all ones to zeros. Figure 25 illustrates Excess 3 code with 9's complement notation.
19
DECIMAL
EQUIVALENT CODE
sign tens decade ones decade
+99 0 1100
1100
+98 0 1100 lOll
+10 0 0100 0011
+3 0 DOll 0110
+2 0 0011 0101
+1 0 DOll 0100
+0 0 DOll 0011
-0 1 1100 1100
-1
1 1100 -1011
-2 1 1100 1010
-3 1 1100 1001
-10 1 lOll 1100
-98 1 0011 0100
-99 1 DOll 0011
Figure 25 Nines complement, Excess 3 code
To subtract with Excess 3, 9's complement code, the subtrahend is first complemented
then added to the minuend. During the first step of the addition process the individual
bits (in the decimal decades and in the sign bits) are added just as in binary. Carries
propagate from each digit to the digit of more significance and from the most significant
digit to the sign bit. If the sign bit produces a ca rry, it is added to the least significant
decade, a process known as end around carry.
0 1000 0100 51 dec. +51 dec.
.l
1001 1000
(complement of 34 dec.) -34 dec.
( 0
-0001 1100
__ ,_1 end around carry
0 0001 1101
0011 1100 correction factor
0100
( 1 0 ? ~
end around carry
0 0100 1010 17 dec. + 17 dec.
0 0110 0111 34 dec. +34 dec.
1 0111 1011 (complement of 51 dec.) - 51 dec.
1110 0010
1100 0011 correction factor
~ ~
0101
end around carry
1011 0101 -17 dec.
-17 dec.
Figure 26 Subtraction with the Excess 3 code, 9's complement notation
20
After this initial portion of the subtraction, a correction factor must be applied just as
in addition. That is, a binary 3 (0011) must be added to each decade in which a carry
signal has occurred, and binary 3 must be subtracted from each decade in which a
carry signal did not occur.
Subtraction of the correction factor may be performed in the same way as well as the
overall subtraction. However, the subtraction in this case operates only on the individual
digits. Thus, if a carry occurs from the most significant bit of the digit, it is not
carried out to the next digit; rather, it is added into the least significant bit of the
same digit.
A second method of per.forming subtraction is through the use of the 10's complement
notation. The 10's complement of any numbei may be obtained by adding 1 to the 9's
complement. Operations are similar to those used in the 9's complement notation
except that the end around carry is not required. Hence this is useful in systems where
a feed back loop would be particularly time consuming.
CODES GREATER THAN FOUR BITS
DECIMAL BIQUINARY RING COUNTER CODE
5. 0, 4. 3, 2, 1, 0 9, 8, 7. 6, 5, 4, 3, 2, 1, 0
0 0 1 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1
1 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 1 0
2 0 1 0 0 1 0 0 0 0 0 0 0 0 0 1 0 0
3 0 1 0 1 0 0 0 0 0 0 0 0 0 1 0 0 0
4 0 1 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0
5 1 0 0 0 0 0 1 0 0 0 0 1 0 0 0 0 0
6 1 0 0 0 0 1 0 0 0 0 1 0 0 0 0 0 0
7 1 0 0 0 1 0 0 0 0 1 0 0 0 0 0 0 0
8 1 0 0 1 0 0 0 0 1 0 0 0 0 0 0 0 0
9 1 0 1 0 0 0 0 1 0 0 0 0 0 0 0 0 0
DECIMAL SWITCH TAIL RING COUNTER CODE DECODING
A B C D E
0 0 0 0 0 0 A E
1 0 0 0 0 1 D E
2 0 0 0 1 1 C D
3 0 0 1 1 1 B C
4 0 1 1 1 1 A B
5 1 1 1 1 1 A E
6 1 1 1 1 .0 D E
7 1 1 1 0 0 C D
8 1 1 0 0 0 B C
9 1 0 0 0 0 A B
Figure 27 Codes greater than four bits
21
Codes greater than four bits are often used for error detection and simplicity in decoding
(Figure 27). The biquinary code is commonly used when error detection is required.
It is a 7-bit weighted code in which two ones and five zeros appear in the representation
of any number; thus it is always possible to detect single errors, and it is often possible
to detect multiple errors.
The ten-bit weighted code shown in Figure 27 allows any number to be represented
with a single 1 and nine zeros. This code is often used in counting operations; the
counter is a ten-stage shift register with the final stage connected to the initial stage.
This counter, often given the name of ring counter, requires no carrying propagate time
and the numbers may be decoded into ten lines without additional gates.
The switch-tail ring counter is a.five-stage ring counter with reversed feed-back from the
initial stage to the final stage. It requires fewer flip-flops than the ring counter and has
the same advantage that no carry propagate time is required. Any state may be decoded
by a two-input gate conditioned by two neighboring flip-flops.
22
, PART II: FLIP CHIP MODULES
23
24
R
SERIES
25
R-SERIES
BASIC CIRCUITS
DIODE GATE
The basic element of digital logic described in this chapter is the diode gate. The diode
gate is used in the R (2-megahertz) series to combine, amplify, invert, and standardize
the signals which represent various logic functions. Figure 1 is a schematic of a simple
diode gate with one input.
-15 VOLTS
CLAMP LOAD
DIODE RESiSTOR
-t5 VOLTS
_3qOLTS
BIASING
RESISTOR
BtASING
DIODES
BASE
INPUT o--C:+-+--l<J--i<:J-''''':;';+l
NODE
BIASING
RESISTOR
+tOVOLTS
OUTPUT
COLLECTOR
Figure 1 Single Input Diode Gate
When the input is negative, the node point is also negative and current flows from the
transistor emitter through the biasing diodes and the biasing resistor to minus 15 volts
(-15v). As a result, the PNP transistor is turned on,forming a short circuit between the
collector and the emitter. Thus, when the input voltage is negative, the output voltage
is ground. Since the output is from a saturated transi,stor, it has a low output impedance
and good driving power.
When the diode gate input voltage is ground, the biasing diodes and the resistor, which
is connected to the lOv supply, hold the transistor base more positive than the emitter,
and the transistor is turned off. The output is then an open circuit, and it will follow
the voltage of any other circuit connected to it.
26
R SERIES-BASIC CIRCUITS
If the load resistor and clamp diode are attached to the transistor collector, they serve as
a voltage source and hold the output at -3v while the transistor is off. When the tran-
sistor is on, the diode is cut off and the load resistor follows the output to ground.
The single-input diode gate therefore has three functions. First, it inverts the input
signal. Second, it standardizes the output voltage to -3v or ground (if the clamped
load diode and resistor are connected). Third, since the output current available from
the transistor is much greater than the required input current, the diode gate amplifies.
A fourth function, gating, is obtained by adding more diode inputs to the node point,
as shown in Figure 2.
-15 VOLTS
-3 VOLTS
'---i--<> OUTPUT
+10VOLTS
NODE
Figure 2 Multiple Input Diode Gate
The node terminal in this diagram will be at approximately the same voltage as the
most positive input. Thus, when any input terminal is grounded, the node terminal is
also at ground and the circuit output is at -3v. If all of the inputs are negative, the
node terminal will be negative and the circuit output will be at ground.
Gating also be by wiring together two or more diode gate
outputs and one load resistor, as shown in Figure 3. When any input is negative, it
saturates the corresponding transistor and forces the output line to ground. If all inputs
are at ground, all of the transistors are open circuits and the output voltage, determined
by the clamped load resistor, is -3v. .
27
INPUTS
-3 VOLTS -115 VOLTS

+IOVOLTS
-IS VOLTS
+ 10 VOLTS
-15 VOLTS
+IOVOLTS
Figure 3 Diode Gates in Parallel
The basic diode gate can be used to construct very complex logical functions. A drawing
that showed all of the circuit components, however, would be tedious to draw and
difficult to read. For this reason, the diagrams that follow use a shorthand notation
which represents one or more components as a single functional unit. Referring to
Figure 4, diodes are shown in the conventional way. The transistor circuit, including
the, biasing resistors and diodes, is shown as a simple rectangle with an arrowhead
indicating the direction of the transistor emitter. This part of the circuit is called an
inverter because of the function it performs. The load resistor is shown as a resistor
with a large dot at the top indicating that it is diode clamped to -,--3v. With these sym-
bols, one can easily and quickly draw complex logical functions.
INPUTS

DIODE GATE SYMBOL
CLAMPED LOAD
RESISTOR SYMBOL
Figure 4 Diode G<!te and Clamped Load Resistor Symbols
28
R SERIES - BASIC CIRCUITS
Assertion input and output voltage levels are shown by diamonds. A solid diamond
indicates a -3v level, and an open diamond indicates a ground level. In the 2input diode
gate of Figure 5, for example, if input A and in put B are both negative, the output will be
at ground. If either A or B is at ground, the output will be negative.
A AND B A OR B
A o--<>[>I-..... -CiJ
B o--<>[>I-.J
Figure 5 Diamonds Indicating the Voltage Levels
DIODE-CAPACITOR-DIODE GATES
The diode-capacitor-diode (OCD) gate is used to standardize the input to various units
such as flip-flops, delays, and pulse amplifiers. It provides logical isolation between
pulse and level inputs and produces a logical delay which is essential for sampling
flip-flops at the same time they are being changed. It also acts as a logical AND gate
since both pulse and level inputs must meet certain requirements for a signal to appear
at the output. Either positive pulses or positive-going level changes (both -3v to ground)
may be used as the pulse input.
OUTPUT
-'5VOLTS
LEVEL
INPUT
Figure 6 Diode-Capacitor-Diode Gate Circuit .
A schematic drawing of a DCD gate is shown in Figure 6. If the level input is held at
ground and the pulse input is held at -3v, the capacitor will become charged after the
setup time has passed. If the pulse input then suddenly goes to ground, a positive-going
pulse will appear at the output. There is delay at the level input, but the pulse input
goes to the output without delay. Even if the level input changes simultaneously with a
positive transition at the pulse input, the delay acts as a temporary memory: the pulse
input is gated according to the level input that existed during the interval before the
pulse.
29

PULSE
INPUT 1
0UTPUT
LEVEL
INPUT
Figure 7 Diode-Capacitor-Diode Gate Symbol
The symbol for the DCD gate (Figure 7) is distinguished from the diode gate by an X
in the rectangle. The output is at the top, the delayed (level) input is at the bottom,
and the differentiating (leyel change or pulse) input is on the side. The input signal to be
differentiated, whether a level change or a pulse, is indicated by an arrowhead, rather
than a diamond. The pulse symbols are hollow when positive-going and solid when nega-
tive'going. In the DCD gate, the pulse input must be positive-going.
Since many DCD gates may be driven by the same pulse, the side of the rectangle
opposite the pulse may be used to show a continuation of the same line, as in
Figure 8. The illustration on the left below is a simplified version of the identica'i logical
configuration on the right.
Figure 8 Pulse Lines to Multiple Gates
FLIPFLOPS
The flip-flop provides a convenient means of storing logical conditions within a digital
system. It has two stable states representing 0 and 1, and remains in one of these
states until an appropriate command to change state is received. Three commands
may be given: set, which puts the flip-flop in the 1 state; clear, which puts the flip-flop
in the 0 state; and complement, which changes the state of the flip-flop regardless of
its previous state. '
TABLE 1 FLIP-FLOP COMMANDS
Command State Before Command State After Command
;;t;e.t
1 1
0
>i)i::;;,c
1
~
.
~ :0:"
. ,
Clear
1 0
0 0
Complement
1 0
0 1
30
Figure 9 shows a schematic diagram and the. symbolic equivalent of an R series
flip-flop. The flip-flop consists of two diode gates, connected "back-to-back." When
transistor Ql is off, its output is negative. This holds transistor Q2 on, which in turn
maintains the off condition of transistor Q1. Direct set and direct clear inputs are
provided for operation of the flip-flops directly from external logical elements. When the
flip-flop is set to the 1 state, the 1 output is at -3v .
.---_-------,,--...... --I5V
-3V
L-__________ -+IOV
DIRECT CLEAR DIRECT SET
Figure 9 Flip-Flop Schematic and Symbol
Inputs to this flip-flop are often made through DeD gates as shown in Figure 10. The
gates provide sufficient delay so that information may be read out of one flip-flop and
into another at the same time that the first flip-flop receives a command to change
state. The OeD gate can also be used to perform additional logical operations, since
it is basically an AND gate. That is, both the pulse and level inputs of the OeD gate
must meet the proper input requirements for an output signal to occur.
31
o 1
OUTPUT OUTPUT
DIRECT DIRECT
CLEAR SET
o PULSE 1 PULSE
INPUT INPUT
-15 VOLTS
-15 VOLTS
-15 VOLTS
o LEVEL 1 LEVEL
INPUT INPUT
o 1
OUTPUT OUTPUT
DIRECT CLEAR DIRECT SET
o PULSE 1 PULSE
INPUT INPUT
o LEVEL INPUT I LEVEL INPUT
Figure 10 Flip-Flop with DCD Gates, Schematic and Symbol
As can be seen from the schematic in Figure 10, the DeD gate level inputs are con-
ditioned by the flip-flop outputs. Thus, a set signal will reach the flip-flop only if the
flip-flop was previously in the state. Similarly, a clear command reaches a flip-flop
only if it was previously in the 1 state. For simplicity, this conditioning is not shown
on the symbol but should be remembered because it is a very powerful element of
the flip-flop.
32
A complement terminal can be made simply by tying the set and clear pulse inputs
together as shown in Figure 11. The gate inputs are still available for external enables.
This technique allows a flip-flop with its DCD gates to be used in such varied applications
as up counter, down counter, up-down counter, shifting, multiple source buffering, jam
transfer register, ring counter, BCD counting, and special counts of 211 (2" + 1) all
without need for additional gate modules.
Figure 11 Complementing
Flip-flops may be collector-triggered by tying a flip-flop output to the output of one
or more diode gates. As illustrated in Figure 12, when the output of either diode gate
D1 or D2 is asserted (at ground), the corresponding flip-flop is cleared (put to 0). If the
gates are attached to the flip-flop through diode networks (Figure 12a), the diode gate
outputs (01 and D
2
) can be logically independent. If they are attached in parallel
without diode networks (Figure 12b), the ouputs will be dependent. That is, if any
output is at ground, the output of all gates connected in parallel with it will be at ground.
TO

LOADS
(A) WITH DIODE NETWORK--DIODE GATE OUTPUTS INDEPENDENT
TO

DIODE GATE
OUTPUTS
(B) WITHOUT DIODE NETWORK--DIODE GATE OUTPUTS DEPENDENT
Figure 12 Collector Triggering of a Flip-Flop
33
LOADS
DELAYS
The delay one-shot, or monostable multivibrator, is a basic timing element. The input
to the delay, like that of the flip-flop, is through a DeD gate (see Figure 13). When the
gate is properly enabled, and when its pulse input terminal is brought to ground by a
positive pulse or a positive-going level change, the output of the delay changes from
its normal ground level to a -3v level for a fixed, but adjustable, period of time. After
the fixed time has elapsed, the output returns to ground. This delayed output is suitable
for driving manyR series modules. A pulse output can be obtained with the addition of
a pulse amplifier to the delay output.
PULSE
INPUT
LEVEL
INPUT
OUTPUT
Figure 13 Delay One-Shot
Delay units are particularly useful in generating delayed pulses or signals of arbitrary
width. The network of delays in Figure 14a will produce the waveform shown in
Figure 14b.
START
PULSE
(A) SAMPLE CONFIGURATION
~ T, h ~ T 3 - + - T 4 - - - j T, I T2 ~ T 3 - j
(8) RESULTING WAVEFORM
Figure 14 Typical Application of Delay Units
34
PULSE AMPLIFIERS
Pulse amplifiers are extremely versatile elements since they not only amplify and
standardize various signals into standard lOO-nanosecond pulses (-3v to ground), but
they may also be used to carry the results of gating to many units. For example, when
the same gating is to be done on an entire register of flip-flops, it is most economically
performed at the input to the pulse amplifier which drives the register.
PULSE
INPUT
LEVEL
INPUT
Figure 15 Pulse Amplifier with Gates
PULSE
OUTPUT
Several pulse amplifier outputs may be ORed together by simply connecting their outputs
in parallel (Figure 16). Thus two levels of logic can be performed by pulse amplifiers.
LOGIC DIAGRAM
o VOLTS
-3 VOLTS --.1\'-__________ _
o VOLTS
-3 VOLTS ____ ..... f\ .... ________ _
o VOLTS
-3 VOLTS
TYPICAL WAVEFORM
Figure 16 ORed Pulse Amplifiers
35
TO
EXTERNAL
LOAD
P1 INPUT
P
z
INPUT
OUTPUT TO
EXTERNAL LOAD
Flip-flops may be collector-triggered by connecting the output of one or more pulse
amplifiers to the output terminals of the flip-flop (Figure 17). This connection may be
made directly or through the diode networks, as with the diode gates shown in Figure 12.
TO EXTERNAL LOAD
Figure 17 Collector Triggering of FlipFlops
CLOCKS
Variable R series clocks produce standard, 100-nsec pulses (-3v to ground) from stable
RC-coupled oscillators. These clocks are often used as a primary source of timing for
large systems (Figure 18).
CLOCK ~ OUTPUT
Figure 18 Clock
Where precise timing is required, a clock with a single frequency crystal oscillator
may be used.
LOADING RULES
When interconnecting basic circuits to perform logical operations, it is important to
keep the load on each circuit within its driving ability. The R series loading rules are
simple because all inputs draw current from the same direction and because all inputs
are either diode gate circuits or DCD circuits. .
Each diode gate input draws 1 ma (milliampere) and the output drives 20 mao The
load resistor draws 2 ma; .so a diode gate with a clamped load resistor tied to it can
drive 18 mao
A flip-flop is two slightly modified diode gates cross-connected. The direct set and clear
terminals draw 1 mao The output will drive 21 ma less 3 ma for the load resistor p,ermanently
connected in the flip-flop and less 1 ma for conditioning the opposite side of the flip-flop
for a remainder of 17 mao
36
The single shot delay has a similar output circuit with a built-in load resistor_ It will drive
20 -2 or 18 ma_
The pulse amplifiers will drive loads of up to 70 ma_
The DCD gates on flip-flops, delays, and pulse amplifiers draw 2 ma at the level inputs
and 3 ma at the pulse inputs_ When two DCD gates are driving both sides of the same
flip-flop, the load on both pulse inputs totals only 4 ma_ When the level inputs are
tied together as in a complement configuration, the total input load is only 3 ma, as
shown in Figure 19_
3 ma
2 ma 2ma
4ma 4ma
2 ma
3 m a o---()[::;;c::]
4 ma
6 ma 3 ma
Figure 19 Flip-Flop Input Loads
On flip-flops which have built-in DCD gates, the output driving ability is less because the
internal diode capacitor gate draws current from the flip-flop_ Table 2 lists the output
driving capability for each of the six types of flip-flops.
Table 2 Flip-Flop Output Driving Capability
Flip-Flop
R200
R201
R202
R203
R204
R205
o Output
17
11
15
15
17
13
37
1 Output
17
13
15
17
17
15
R-SERIES
LOGIC CONFIGURATIONS
COUNTERS
Counting, a basic digital operation, is performed by a wide variety of counter circuits.
Examples follow of how some of these circuits are built using R series modules.
The Binary Up-Counter
A typical binary-up counter is shown in Figure 20. When a flip-flop changes from the
to the 0 state, its 1 output complements the next flip-flop in the counting chain. Flip-
flop C, the first in the chain, complements on each input pulse. Flip-flop B complements
when C changes from 1 to 0, and so on through the counter. Note that a flip-flop
complements only if all preceding flip-flops are in the 1 state when the next input pulse
arrives. The time required for a "carry" to propagate up the chain is 70 nsec per stage.
The input load for the complete circuit is only that required to complement flip-flop C.
COUNTING SEQUENCE
DECIMAL BINARY
FF
ABC
0 000
00 I
2 010
3 o I I
4 I 00
~ I o I
6 I I 0
7 I I I
0 000
INPUT 0---..... --------'
Figure 20 Three-Bit Binary Up Counter
(R201, R202, or R205 Flip-Flops)
38
R SERIES - LOGIC CONFIGURATIONS
The Direct Clear input resets the counter to O. This i'nput requires a 400-nsec pulse,
-3v to ground, (rather than a standard, 100-nsec pulse) or a level that remains at
ground for more than 400 nsec. This time is needed to hold all flip-flops in the 0 state
while the carries die out.
Binary Down-Counter
The binary down-counter is identical to the binary up-counter,except that the comple-
menting level change comes from the 0 terminal of the preceding flip-flop rather than
the I terminal. In the down-counter the direct set rather than the direct clear requires
400-nsec pulses. The counting sequence is the reverse of the up counter, that is
111, 110, .. ,001,000 or 7, 6, ... , 1, O.
Binary Up-Down Counter
The binary up-down counter is a combination of the up counter and the down counter.
All pulse inputs are standard positive pulses, -3v to ground. (Both direct set and direct
clear require 400-nsec pulses.)
In order to avoid counting in both directions simultaneously, the DCD gate level inputs
must be used. The two control lines must not be grounded simultaneously. Input pulses
must not follow control line changes sooner than 400 nsec to allow for DCD gate set up
time. Control line changes must not follow input pulses closer than 70(M1) nsec, to allow
an M bit counter time for carry propagation. Count sequency is shown in Table 3, and an
example of a fourbit binary up-down counter is illustrated in Figure 21.
TABLE 3 COUNTING SEQUENCES
Up Counting Sequence Down Counting Sequence
Decimal Binary Decimal Binary
Flip-Flop Flip-Flop
A B C D A B C D
0 0 0 0 0 15 1
1 . 1
1
1 0 0 0 1 14 1 1 1 0
2 0 0 1 0 13 1 1 0 1
3 0 0 1 1 12 1 1
o 0
4 0 1 0 0 11 1 0 1 1
5 0 1 0 1 10 1 0 1 0
6 0 1 1
Q
9 1 0 0 1
7 0 1 1 1 8 1 0 0 0
8 1 0 0 0 7 0 1 1 1
9 1 0 0 1 6 0 1 1 0
10 1 0 1 0 5 0 1 0 1
11 1 0 1 1 4 0 1 0 0
12 1 1 0 0 3 0 0 1 1
13 1 1 0 1 2 0 0 1 0
14 1 1 1 0 1
00
0 1
15 1 1 1 1 0 0 0 0 0
0 0 0 0 0 15 1 1 1 1
39
R SERIES-LOGIC CONF!GURATIONS
DIRECT DOWN
CLEAR
( ... 0000)
Figure 21
COUNT UP DIRECT
SET
( ... 110
Four-Bit Binary Up-Down Counter
(R201 Flip-Flops)
40
R SERIES-LOGIC CONFIGURATIONS
Special Counts
It is often desirable to build a counter that produces a signal after a particular number
of events, N, has occurred. If N is an integral power of 2, the output is automatically
produced by the final digit of a counter of the appropriate length. If N is not a power of 2,
gating must be performed to detect the desired number, produce a signal, and reset the
counter to O.
Diode gates may be used to sense the number N-l, gate off the input to the counter,
and reroute the Nth input.pulse so that it will clear the counter and generate an output
signal. This method may be used for any value of N ( Figure 22).
PULSES
TO BE ~ ~ ~ ____________________ 4-____________ -l
COUNTED
Figure 22 Arbitrary Count - Count of 12 Illustrated
(R201, R202, or R205 Flip-Flops, R601 P.A.l
41
Because of the carry propagation time through the counter, the maximum input frequency
depends on the number of bits. For M bits, the maximum frequency is
f = 1000 mc
70 M + 500
This allows 400 nsec to enable the DCD gate and 100 nsec delay through the diode gate and
inverter.
Special counts of N = 2p + 1 (that is, 3, 5, 9, 17 .. .) are simple to build using R201,
R202, or R205 modules. Figure 23 illustrates how a count-of-5 counter is implemented.
The input of flip-flop A is connected directly to the counter input rather than the output
of flip-flop B. Thus the last flip-flop is cleared on the first pulse after it has been set;
that is, the states of the last flip-flop will be ... 0, 1, 0, 0, '.' ., 0, 1, 0, .... The
level inputs of flip-flop C are connected to the 1 output of flip-flop A. This prevents the
counter from counting when flip-flop A is in the 1 state (has been set on the previous
pulse). This counting sequence can easily be extended by adding several binary stages
in place of flip-flop B. The 2"th pulse which sets the last flip-flop clears all previous
flip-flops (see "Binary Up Counters"). The next pulse (2" + 1) will. clear the last flip-flop,
returning the counter to zero.
OUTPUT
COUNTING SEQUENCE
DECIMAL BINARY
FF
ABC
0 000
I 00 I
2 o I 0
3 o I I
4 100
0 000
INPUT PULSES
TO BE COUNTED
Figure 23 Count-ofFive Counter
42
The maximum input frequency for counts of 5 or more is
f = 1000 mc
70 (P+1) + 400
This allows 400 nsec set-up time for the DeD gates on the input stage, plus 70 nsec propa-
gation ti me per stage.
Special counts of N = 2R (2
P
+ 1) are produced by adding a 2M up counter in front of
a 2
P
+ 1 counter. The input of the 2
P
+ 1 counter is the output of the 2R counter.
Thus the 2
P
+ 1 counter counts every 2Rth pulse for a fina I count of 2M (2
P
+ 1).
Binary-Coded Decimal Counter
A count of 10 = 2' (2
2
+ 1) can be produced as illustrated in Figure 24. This BCD
counter may be used wherever decimal results are desired. Longer duration decimal
counters are made by using many BCD counters in series. '
OUTPUT
COUNTING SEQUENCE
DECIMAL BINARY
FF
ABC D
0 0000
I 000 I
2 00 I 0
3 001 I
4 01 00
5 01 o I
6 o I I 0
7 o I I I
8 I 0 0 0
9 I 00 I
0 0000
(PULSES TO BE
Figure 24 CountofTen Counter
(R201, R202, or R205 FlipFlops)
43
COUNT OF 5
COUNT OF 2
BUFFER AND SHIFT REGISTERS
Buffer Register
A buffer register is a flop-flop network used for temporary data storage. With no additional
gating, the R201 accepts inputs from five external sources, the R202 from two sources,
the R203 from one source, and the R205 from two sources.
Figure 25 illustrates a simple buffer register. To correctly read in a number, the
register is first cleared. Then each flip-flop is set in accordance with the state of the
level input to its DCD gate. This operation can also be performed by setting all flip-flops
and then reading in through clear gates.

READ-IN 0-----------------+--------------4------------_
Figure 25 Three-Bit Buffer Register
JR20l, R202, R203, or R205 Flip-Flops)
Shift Register
A shift register (Figure 26) shifts the contents of each flip-flop in a series into the next
flip-flop in the series. If any flip-flop is in the 1 state, its 0 output is at ground enabling
the 1 input of the next flip-flop. Hence the application of a shift pulse puts the second
flip-flop into the 1 state. Similarly, a shift pulse puts. the flip-flop into the O' state when
the previous flip-flop has been in the 0 state. At the time of the shift pulse, flip-ftop C
is cleared if the read - in level has been at ground. If the. read - in level has been at
-3v, flip-flop C is set to 1.
44
.P 0..
, ,
Figure 26 Three-Bit Shift Register
Ring Counter
A ring counter can be made from a shift register by connecting the 0 output of the last
flip-flop to the 1 level input of the first flip-flop and the 1 output of the last flip-flop to
the 0 level input of the first. Thus on each count pulse (shift pulse) the ring counter
will shift the contents of the last flip-flop to the first flip-flop (Table 4).
Switch-Tail Ring Counter
A switch-tail ring counter can be made from a shift register by connecting the 1 output
of the last flip-flop to the 1 level input of the first flip-flop and the 0 output of the last
flip-flop to the 0 level input of the first. On each count pulse (shift pulse). the switch-
tail ring counter will shift the inverse of the contents of the last flip-flop into the first
flip-flop (Table 4).
45
TABLE 4 COUNTING SEQUENCE FOR RING COUNTERS
(A Shift Pulse Equals One Count)
3 Flip-Flop Ring Counter
3 Flip-Flop Switch Tail
Ring Cou nter
001 (Initialized) 000 (Initialized)
010 001
100 011
----0'01-------
111
110
100
------------
000
Jam-Transfer Buffer
The jam transfer buffer (Figure 27) is like a series of two-bit shift registers. When a shift
pulse (or positive-going level change) occurs, the contents of A, B, and C will be shifted
to E, G, and H. If A and B contain ones and Ccontains 0, a shift pulse sets E and G to
. ones and H toO. A [lumber is also read into flip-flops A, B, and C at the same time as the
shift operation. When any of the three level inputs are at -3v,a.l is read.into the cor-
. responding flip-flop. A is read in if the corresponding terminal is at ground.
Figure 27 Three-Bit Jam Transfer Buffer
(R20l, R202, or R205 Flip-Flops)
46
ARITHMETIC ELEMENTS
Parallel Adders
Parallel adders may be used to add two binary numbers. The augend is called the resident
number and is stored in the accumulator register. The addend, or incident number, is
stored in the incident register. The sum appears in the accumulator. R series Type R201,
flip flops can be used as shown in Figure 28.
(21
til
L
(3'
ACCUMULATOR ~
(RESIDENT NUMBERI
BUFFER REGISTER
!iNCIDENT NUMBER I
Figure 28 Parallel Adder
(R201 Flip-Flops in Accumulator and
Any Flip-Flops in Incident Register)
HALF ADD
--COMPLEMENT
END AROUND CARRY
"}
121 OPTIONAL
(3)
CARRY
INITIATE
Addition is performed in two steps. The first step is a half-add .. Each digit of the
accumulator is complemented if the corresponding digit of the incident number is 1.
The second step is a carry. A carry is generated if a digit in the accumulator is 0 and
the corresponding incident number is 1. A carry is also propagated if an accumulator
digit is 1 and it receives a carry pulse from the next less significant accumulator digit.
Each stage will propagate one carry at most. After all carries have been propagated,
addition is complete and the accumulator contains the sum of the incident and resident
numbers.
Carry initiate must lag half add no less than 525 nsec to allow sufficient set up time. Next
half add must lag carry initiate by at least 50M nsec (where M is word length in bits) to
allow for M bits of carry propagation.
47
The final step in the addition produces a carry pulse if the sum is a larger number
than the accumulator can hOld. This pulse may be discarded, added to the first stage (by
means of the end-around-carry), or stored externally by means of an additional flip-flop.
Three examples are shown in Table 5.
TABLE 5 ADDITION STEPS
-SUM-
Accumulatot After Carry
Original Accumulator End Externally
Incident Accumulator After Around Discarded Stored
Number Number Half-Add Carry Carry Carry
001 010 011 011 011 0011
001 001 000 010 010 0010
101 011- 110 001 000 1000
Subtracters
An adder may be also used for subtraction. To subtract a number from the accumulator,
the number is made negative and added to the accumulator. The steps involved in
performing a subtraction depend on whether the l's compll'ment or the 2's complement
number system is used to represent a negative number.
The l's complement number system is easiest to implement. To subtract a number
from the accumulator, the steps are (1) complement the incident number, (2) half-add,
and (3) carry. With this number system it is necessary to use the end-around-carry
shown in Figure 28. One's complement subtraction may also be performed by (1) comple-
menting the accumulator, (2) half-add, (3) 'carry, and (4) recomplementing the accumu-
lator.
Serial Adder
Figure 29 illustrates a serial adder. The contents of register A are added to register B,
and the sum is stored in register B. The two numbers to be added are read into registers
A and B with the least significant bits stored in flip-flops AN and BN. The carry flip-flop
may be cleared before addition is begun. N + 1 pulses are allowed to enter the shift
pulse input. When a shift pulse is received, the number in each flip-flop is advanced
one place. The least significant bits are added, together. with the carry, and read into
flip-flop BO. If there is a carry, it is stored in the carry flip-flop. After N + 1 shift pulses
have occurred, register A has been cleared and register B contains the N + 1 least
significant bits of the sum. The most significant, or overflow, bit is stored in flip-flopC.
48
INCIDENT
REGISTER
SHIFT
(lIO)
~ ~ ~ ~ + - - - - - - - - - - 4 - - ____________________________________________ ~
Figure 29 Serial Adders
NOTE: Parentheses on output gates indicate digits from flip-flops AN, BN, and C which each
gate detects.
COMPARATORS
Comparing A Toggle Switch Register
And A Flip-Flop Register
The simplest comparator is a counter that stops or resets itself after a preset number of
counts. This type of comparator (Figure 30) has a multiple-input diode gate with each
diode connected through a toggle switch to a flip-flop. Each toggle switch corresponds to
a single bit; a closed switch represents a I, an open switch represents a O. In Figure 3D,
the counter goes from the binary number 0000 to the number lOll, then resets to 0000.
Note that with the single-throw toggle switches, number 1111 could also generate an out-
put. However, this number is never reached because the counter is reset after 1011.
49
PULSES
TO BE
. COUNTED
r----'
, ,
, ,
L. ____ .J
SWITCH
Figure 30 CountotTwelve Comparator
(R201, R202, or R205 FlipFlops, R601 P.A.) )
When the comparator output shuts off the counter input, it also reroutes the next count
'pulse to perform the clearing action. Since this action requires an additional input pulse
to operate the pulse amplifier, the toggle switches must be adjusted to detect one less
than the number of pulses to be counted, or the counter's least significant bit must be
set while the rest of the counter is cleared. Table 6 shows the resulting counting
sequences for each method.
50
TABLE 6 COUNTING SEQUENCES
All Flip-Flops Cleared A, B, C Cleared, D Set
Decimal Binary Decimal Binary
Flip-Flop Flip-Flop
A B C D A B C D
a a a a a 1 a 0 0 1
1 0 0 0 1 2 0 0 1 0
2 0 0 1 a 3 0 0 1 1
3 0 0 1 1 4 0 1 0 0
4 0 1 0 0 5 a 1 0 1
5 0 1 0 1 6 0 1 1 0
6 0 1 1 0 7 0 1 1 1
7 0 1 1 1 8 1 0 0 0
8 1
0
0 0 9 1 0 0 1
9 1 0 0 1 10 1 0 1 0
10 1 0 1 a 11 1 0 1 1
11 1 0 1 1 12 1 1 0 0
0 0 0 0 0 1 0 0 0 1
This technique can also be used with a down-counter. The diode gates are connected to
the flip-flops so that a closed switch corresponds to a O. The first comparator output is
used to preset the counter to the all 1 state,or to turn off the counter input.
Comparing a toggle-switch register with a flip-flop register that does not necessarily
start at 0 requires a single-pole, double-throw switch for each bit. One side of a switch
is connected to the 1 output of a flip-flop, and the other side, to the 0 output. The rotor
of the switch is connected to an input of the diode gate. When the 0 state of any bit
is to be detected, the toggle switch is connected to the 0 terminal of that bit. Similarly,
when the 1 state is to be detected, the toggle switch is connected to the 1 terminal of
that bit. Thus, the diode gate detects the desired register state regardless of the regis-
ter's tounting sequence.
Because of the carry propagation time through the counter, the maximum input frequency
depends on the number of bits. For M bits, the maximum frequency is
f = 1000 mc
70 M + 500
This allows 400 nsec to enable the DCD gate and 100 nsec delay through the diode gate
and inverter.
Comparing Two Flip-Flop Registers
Comparison of two flip-flop registers, A and B, requires an exclusive OR for each corre-
sponding pair of bits since two possible conditions produce an inequality: bit Al equals
o and bit Bl equals 1; or Al equals 1 and Bl equals O. The entire set of exclusive ORs
must be ORed together to determine when any pair of bits is unequal. Figure 31 illus-
trates the use of the Type R 141 Diode Gate for comparison of two 3-bit registers. When
the two registers are equal, the output is negative; when the -two registers are unequal,
the output is ground.
51

B I <>---<II
A I o-"DI----.
Bio-"I>I---'
liZ 0-.. 1>1----,
B 2.0-".:>1---'
A 2 0--.11>1----,
__ A"""-
A3 0--.11>1----,
__ A"""-
A 3 0-.. 1>1----,
830--.11>1---'
UNEQUAL
EQUAL
Figure 31 Comparing Two Flip-Flop Registers
(R141 AND-NOR Gate)
Type R131 Exclusive OR can also be used for this purpose, and requires only one input
per bit.
Sign Of An Inequality
A more sophisticated comparator is used to compare the numerical value of two
registers to determine equality or the sign of an inequality .. Pairs of bits must be investi-
gated in the order of their significance. Once an inequality has been discovered, all
further investigation is stopped so that the differences in bits of lesser significance
do not affect the output.
EQUALITY LINE FROM
PREVIOUS BITS
-3 VOLTS
INDICATES
EQUAL so FAR AI 0-... 01--..
EQUALITY
LINE
EQUAL so FAR
MEANING OF OUTPUTS

AI o-"'{)!--4
SIGN
OUTPUT
-3
-3
L ____________________________
o
o
(0 VOLTS INDICATES A)BI
Figure 32 Comparator with Sign Output
52
EQUALITY
OUTPUT
-3
o
o
-3
MEANING
A=B
A<B
A>B
IMPOSSIBLE
Figure 32 shows the it!' stage of such a comparator chain. To begin investigation, the
equality line is brought to -3v. This negative signal is ANDed with an exclusive OR
for the first pair of bits. If there is equality, the negative signal is propagated down the
equality line. If all corresponding pairs of both registers are equal, the negative signal
appears at the end of the final stage. However, if an inequality is reached, further in-
vestigation is stopped and the equality line output remains at ground. If this inequality
is such that register A is greater than register B, a ground signal is generated at the sign
output. The. comparator illustrated may contain up to 100 stages.
SYNCHRONIZERS
Synchronizers are important not only to determine what happens when asynchronous
or incompatible signals occur, but also to protect against catastrophic errors resulting
from such signals. Asynchronous signals are often commands such as start, stop, or
clear. Incompatible signals must be guarded against whenever a single fJ.ip-flop register
is connected for more than one possible mode of operation.
Synchronization is required ina shift register when it is being changed from parallel
mode to serial mode (or vice versa). Care must be taken that a split or partial pulse is
not allowed to enter the system, because it could result in a partial shift, partial read-in,
or partial read-out. Likewise, a bidirectional shift register should never be commanded to
shift left and shift right at the same time. In a bidirectional counter, particular care
must be taken to assure that the add and subtract inputs are not pulsed simultaneously.
A synchronizer is built by ANDing a random input signal and a clock (or other primary
pulse train) and using the results of this AND gate to set a single flip-flop. If the
random signal partially enables the pulse gate of the synchronizer flip-flop, the single
synchronizer flip-flop decides whether the signal will be accepted or rejected. Provision
must also be made for the rejected signal to be accepted by the next clock pulse;
thus, the input signal may be slightly delayed by the synchronizer flip-flop, but there
will be no middle state to activate only a portion of the main circuit.
53
Level Synchronizer
Figure 33 shows a method of synchronizing a command level, such as a mode control,
with a clock. When the input level is ground, the clock is gated through terminal A.
When the input level is negative, the clock is gated through terminal B. If the input
level has changed too recently to allow the DeD gates to fully set up before the clock
pulse occurs, the decision to accept or reject the signal change is made by the syn-
chronizer flip-flop only. Thus, no pulses are gated through both terminals A and B at
the same time, and no split pulses are gated through either terminal. This method, or
a variation, is often used in controlling the direetipn of count in an up-down counter.
CLOCK
LEVEL
FLIP-FLOP 1
TERMINAL
OUTPUT A
OUTPUT B
OUTPUT B
OUTPUT A
\ ~ - - - - - - - - - - - - - - - - ~
Figure 33 Level Synchron izer
54
Start-Stop Synchronizer
Figure 34 shows a method for synchronizing start and stop commands with a clock.
The random start and stop pulses are converted to stop or start levels by flip-flop A.
Flip-flop B converts this randomly changing level into a level that changes synchronously
with the clock. This level, in turn, conditions the output pulses. The action may be de-
layed by one clock pulse; however, a split pulse will not be allowed to enter the
main system.
CLOCK
START
STOP
FLIP-FLOP A-'
OUTPUT
FLIP-FLOP B-'
OUTPUT
OUTPUT
OUTPUT
CLOCK
STOP PULSE START PULSE

Figure 34 Start-Stop Synchron izer
55

Single-Pulse Synchronizer
Figure 35 shows a ,method of synchronizing a single pulse with a clock. The operation
of the synchronizer is similar to that shown in Figure 34 except that the' stop command
is automatically generated by the first pulse to leave the synchronizer. Thus, only one
pulse passes into the main system, and this pulse is syl)chronized with the clock.
1--...... __ [> OUTP UT
C L O C K ~ ~ - - r - - - - - - - - - ~ ~
CLOCK
TRIGGER --1\ _____________________________________ _
FLIP-FLOP A-1
OUTPUT
FLIP-FLOP 8- 1
OUTPUT
,------'
- - - - - - - - - - - ~ ~ ~ - - - - - - - - - - - - - - - - - - -
Figure 35 Single-Pulse Synchronizer
56
DIODE NETWORKS
TYPES R001,R002
INPUT 0 <>--{>I----o E OUTPUT
F <>--{>I----o H
J <>--{>I----o K
L <>--{>I----o M
N <>--{>I----o P
R <>--{>I----o S
T o--{:::!---o U
ROOl DIODE NETWORK
Diode networks can expand the logic capability of
any R-Series, W-Series, or A-Series module which
has one or more node inputs, such as the RIll
diode gate. They can also make it possible to OR
into an R-Series flip-flop output terminal for setting
or clearing from several sources.
Diode networks cannot be cascaded to perform
57
INPUT 0::=:!J-
F OUTPUT
E .
R002 DIODE NETWORK
other logic operations.
Propagation delay of R-Series gates expanded by
ROOI or R002 diodes will increase typically 15-30
nsec when gate output rises from -3v to av, but
will not change noticeably when gate output falls
from Ov to -3v. Diodes used are similar to type
IN3606.
ROOl-$4.00
R002-$5.00
I
INVERTER
TYPE R107
,-l-. ,.j-, ,-l-'
- - -
- - -
. .j-, ,.j-, ,.J-.
- - -
- - -
INPUT U
v
. ! ''"''"'
NODE ~
R10T 1NVERTER
The R107 Inverter contains seven inverter circuits
with single-input diode gates. Six of the circuits are
used for single-input inversion; the seventh circuit
can be used for gating by tying additional diode
input networks to its node terminal. Clamped load
resistors of 2 rna are a permanent part of each in-
verter.
OUTPUT: Standard levels of -3v and ground. Each
inverter can drive 18 rna of load at ground. Output
terminals of inverters may be connected in parallel.
Some typical propagation delays are shown below.
High frequency logic designs may benefit from the
application note "Estimating Propagation Delays."
INPUT: Diode - Standard levels of -3v and ground,
lOO-nsec minimum duration. Input load is1 rna,
shared among the inputs that are at ground. Node
Terminal- Accepts only ROQ1 or R002 Diode Net-
works or their equivalent. The combined length of all
leads attached to the node terminal must not exceed
6 in. Input signal and load characteristics for diode
networks are the .same as those given for the diode
input above.
58
Fan-out 4 10 16
Output Rise 30 nsec 35 nsec 40 nsec
Output Fall 60 nsec 100 nsec 140 nsec
POWER: + 10 v(A)/O.7 rna, -15 v(B)/30 rna.
R107-$24.00
EXPANDABLE NAND/NOR GATE
TYPE R111
CLAMPED m
L.OAD
RESISTORS
J P V
~
U T P U T H
DIODE INPUTSJD _
\: -
NODE TERMINAL f
RIll
EXPANDABLE NAND/NOR GATE
The Rill contains three diode gates, each connected
to a transistor inverter. The gate operates as a NAND
for negative inputs, and as a NOR for ground inputs.
Each gate has three input terminals: two are con
nected to diodes, a third is connected directly to the
node point of the diode gate. The third terminal
allows the number of input diodes to be increased by
adding external diode networks such as the ROOI or
R002. External diodes must be connected in the
same direction as the diodes in the Rill. Unused
inputs may be left open.
INPUT:. Diodes - Standard levels of -3v and
ground, 100-nsec minimum duration. Input load is
1 ma, shared among the inputs that are at ground.
Node' Terminal- Accepts only ROOI or R002 net-
works or their equivalent. The combined length of all
leads attached to the node terminal must not be
greater than 6 in. Input signal and load character-
istics for the diode networks are the same as those
given for the diode above.
59
output can drive 20 ma of load at ground. Clamped
load resistors are included in the module. Each
clamped load resistor represents 2 rna of load. The
outputterminals of diode gates may be connected in
parallel. Two gates in parallel (driven by the same
signal) can drive 38 rna at ground (20 rna each, less
the 2-ma clamped load)_ If they are not driven by the
same signal, gates in parallel drive 20 rna at ground
minus 2 rna for each clamped load used. Some
typical propagation delays are shown below. High
frequency logic designs may benefit from the appli-
cation note "Estimating Propagation Delays."
Fan-out 4 10 16
Output Rise 30 nsec 35 nsec 40 nsec
Output Fall 60 nsec 100 nsec 140 nsec
POWER: +10 v(A)IO.3 rna, -15 v(B)/18 rna.
Rill - $14.00
NAND/NOR GATE JIRl
L - - ~ ~ = ~ __ TY_P_E_R_11_3 ______ ----l ~
R113 NAND/NOR GATE
The R113 contains five diode gates, each connected
to a transistor inverter. The gate operates as a NAND
for negative inputs, and as a NOR for ground levels.
INPUTS: Standard levels of -3v and ground, 100
nsec minimum duration. Input load is 1 ma, shared
among the inputs at ground. Unused inputs may be
left open.
output can drive 18 ma of load at ground. Output
terminals may be connected in parallel. Clamped
60
loads included in the module are 2 ma each. Some
typical propagation delays are shown below. High
frequency logic designs may. benefit from the applica.
tion note "Estimating Propagation Delays."
Fanout
Output Rise
Output Fall
4
30 nsec
60 nsec
10
35 nsec
100 nsec
16
40 nsec
140 nsec
POWER REQUIREMENTS: +10V (A) 0.5 ma, -15V
(8) 23 mao
R113 - $20.00
NAND/NOR GATE 11;1'
'--______ T_Y_P_E_R_1_2_1 ______ ~ ~
,- ~ ; ~
F ~ K ~
L
R
M::rF. 2MA
N =
p
2 MA
v <>-1>1--
R121 NAND/NOR GATE
The R121 contains four R111-type circuits with 2-ma
loads internally connected to each output. This mod-
ule increases density at the expense of flexibility,
since gate expanders ROOI and R002 cannot be
used.
INPUT:' Standard levels of -3v and ground, 100
nsec,minimum duration. Input load is 1 ma,shared
among the inputs that are at ground.
61
output has a permanently attached 2 rna clamped
load resistor. Each output can drive 18 rna of load
at ground. Delays are similar to R111 delays. See
application note "Estimating Propagation Delay"
for more information.
POWER:' + lOv(A)/ 0.4 rna., -15v/20 rna.
R121 - $17.00

NOR/NAND GATE II;[
L-______ T_Y_PE_R_1_2_2 ______ --..J ~
2 MA 2 MA
2 MA 2 MA
R122 NOR/NAND GATE
Provides the logical complement to the R121 NAND
Gate at some sacrifice of speed and economy.
INPUT: Standard levels of -3v and ground. Mini-
mum duration: 400 nsec at ground, 100 nsec at
-3v. Input load is 1 ma at each input.
62
output has a permanently attached 2 ma clamped
load resistor. Each output can drive 18 ma of load
at ground. Propagation delays for output rise are
similar to Rll1 delays. Propagation delays for out-
put fall. are typically 75 nsec longer than RIll
delays.
POWER: + 10v(A)/3 ma, -15v/31 mao
R122 - $26.00
INPUT BUS GATE 11---;1
L--______ T_Y_PE_R_1_23 ______ -----l ~
R123 INPUT BUS GATE
This module contains six RIll-type diode gates ar-
rayed for convenient driving of the PDP-8 computer
input bus, and for .other matrix-like applications.
Clamped loads are not provided on this module, and
must be obtained from some module in the asso-
ciated logic.
INPUT: Standard levels of -3v and ground, 100
nsec, minimum duration. Each of the six gates is a
1 ma load shared among its grounded inputs; thus
inputs F, M, and T may be loaded with up to 2 ma
at ground.
63
output can drive up to 20 ma at ground. A 2 ma or
heavier clamped load must be used at each group
of paralleled collector outputs, chosen to provide
fall times fast enough for the use intended. Delays
are similar to RIll delays. See application note
"Estimating Propagation Delay" for more informa-
tion.
POWER: + lOv(A)/.6 ma; -I5v/15 mao
R123 - $19.00
I
EXCLUSIVE OR
IG;;]
TYPE R13t
0
: ~
XOR
H
J
12MA f2MA f2MA
~ ~
XOR
H
N
H M
:::J
H
XOR
T
: ~
XOR
H
RI31 EXCLUSIVE OR
This module provides a convenient way to compare
two binary numbers or patterns. The output of each
circuit is negative if its inputs are the same, and
ground if they are different. If the outputs of several
circuits are tied together, the common output line
will be negative if every input pair matches, ground
if any pair doesn't match.
During the transition from one input pattern to
another with the same output, there is an interval
during which the R131 output may be wrong for
both patterns. Transitions between unequal inputs
have a relatively short settling time, but transitions
between equal inputs may produce transients to
ground lasting 250 nsec or more.
INPUTS: Standard levels of -3v and ground. Each
input is a 2 ma load at ground.
64
OUTPUTS: Standard levels of -3v and ground.
Each output can drive 18 ma at ground. Propagation
delay for output rise is similar to Rlll delay. Propa-
gation delay for output fall is typically 300 nsec
longer than RIll delay.
POWER: + 10v(A)/0.8 ma; -15v(8)f36 mao
TRUTH TABLE
Input Output Input
E(K, P, Ul D(J, N, Tl F(L, R, Vl
Ov Ov -3v
-3v Ov Ov
Ov .-3v Ov
-3v -3v -3v
R131 - $35.00
AND/NOR GATE IIRl
L--______ T_Y_P_E_R1_4_1 ______ ----l ~
INPUTS E
R141 AND/NOR GATE
The R14l AND/NOR Gate performs two levels of
gating. The module contains a multiple-input diode
gate with a transistor inverter for signal amplifica-
tion. For negative input signals the R141 is seven
2-input AND gates which are NORed together. For
ground inputs, it is seven 2-input OR gates NANDed
together. This module is frequently used to mix
multiple inputs to a pulse amplifier, or to compare
the contents of two flip-flop registers.
The back-to-back diode circuits are possible because
of an internal bias resistor connected to the input
of each second stage diode. The bias holds the input
of the second stage at -3v unless one of the first
stage inputs is grounded. Propagation delay for out-
put rise is similar to RIll delay. For output fall,
delay is typically 100 nsec longer than Rlll delay
65
under similar loading conditions, assuring sufficient
pulse stretching to allow 70 nsec inputs. Output is
typically too wide, however, to allow 2 mc rates.
Maximum rate depends upon R14l loading, and
may be as low as 1 mc.
INPUT: Standard 100-nsec pulses, standard levels
of -3v and ground, or 70-nsec negative. pulses such
as those generated by the W607 Pulse Amplifier. In-
put load is 1 ma per input pair shared by the
grounded inputs. When any pair of inputs is not
being used, at least one of the two must be grounded.
OUTPUT: Standard levels of -3v and ground. The
output can drive 20 ma of external load at ground.
It has no internal load.
POWER: +10 v{A)/0.5 ma, -15 v{B)/19 mao
R14l-$13.00
BINARY -TO-OCTAL DECODER
TYPE R151
TRUTH TABLE
INPUTS OUTPUTS
BINARY
INPUTS
K
M OUTPUTS
J
-3v
-3v
-3v
-3v
Ov
Ov
H E
Ov -3v
Ov -3v
Ov Ov
Ov Ov
-3v -3v
-3v -3v
Ov -3'1 Ov
Ov -3v Ov
-3v -3v -3v
ENABLE
-3v -3v -3v
-3v -3'1 0'1
R151 BINARY-IOOCTAL DECODER
-3v -3v
Ov -3v
-3v Ov
The R151 decodes binary information from three
flip-flops into octal form. When the enable input is
at ground, the selected output line is at ground and
the other seven outputs are at -3v. Wherl the enable
input is at -3v, all outputs are at -3v. The internal
gates are similar to those in the RIll. The enable
input is the common emitter connection of the out-
put inverters. Typical total transition times are
75 nsec for output rise and 60 nsec for output fall.
INPUT: Binary - Standard levels of -3v and
ground, 100 nsec minimum duration. Input load is
2.3 ma per grounded input. when the inputs are
binary, as in the first 8 lines of the truth table. The
input current is 4 ma at ground when only one input
is grounded, as in the last 6 lines of the truth table.
Enable - Standard levels of -3v and ground, 100
nsec minimum duration. Input load at ground is 3
rna plus the current required by the load on the
selected output when the inputs are binary. as in the
-3v
-3v
-3v
66
F K L D M N P R S T U V
Ov -3v Ov Ov Ov -3v -3v -3v -3v -3v -3v -3v
Ov Ov -3v Ov -3v Ov -3v -3v -3v -3v -3v -3v
-3v -3v Ov Ov -3v -3v Ov -3v -3v -3v -3v -3v
-3v Ov -3v Ov -3v -3v -3v Ov -3v -3v -3v -3v
Ov -3v Ov Ov -3v -3v -3v -3v Ov -3v -3v -3v
Ov Ov -3v Ov -3v -3v -3v -3v -3v Ov -3v -3v
-3v -3v Ov Ov -3v --3'1 -3v -3v -3v -3v Ov -3v
-3v Ov -3v Ov -3v -3v -3v -3v -3v -3v -3v Ov
-3v -3v -3v -3v -3v -3v -3v -3v -3v
-3v Ov -3v Ov -3v Ov -3v Ov -3v Ov -3v Ov
-3v -3v Ov Ov Ov -3v Ov -3v Ov -3v Ov -3v
-3v -3v -3v Ov -3v -3v Ov Ov -3v -3v Ov Ov
Ov -3v -3v Ov Ov Ov -3v -3v 0'1 Ov -3v -3v
-3v -3v -3v Ov -3v -3v -3v -3'1 0'1 Ov Ov Ov
-3v -3v -3v Ov Ov Ov Ov Ov -3v -3v -3v -3v
first 8 lines of the truth table. For other inputs, the
load is 3 ma per selected output plus the loads on
those selected outputs. The maximum input current
is 10 ma when driven from an inverter collector. No
more than one inverter can be placed in series with
this pin and ground. If any external circuit brings
an R151 output to ground, any gate being used to
enable pin D must not drive anything else.
OUTPUT: Standard levels. Each octal output has a
permanently attached 2-ma clamped load resistor.
Each output can drive 7 ma of load at ground. If the
enable input is permanently grounded, each output
can drive 18 ma of load at ground. The length of the
wire used to ground the enable input (pin D) should
be kept as short as possible. Note: Simultaneous
switching of R151 outputs is not assured. If adjacent
R151 outputs are ORed together for example, the
gate output may contain spikes.
POWER: +10 v(A)/0.9 ma, -15 v/32 mao
R151 - $33.00
DC CARRY CHAIN
TYPE R181
l[;g
L-__________________________________ ~
R181 DC CARRY CHAIN
The R181 DC Carry module is designed for building
counters with no carry propagate delay. A 2-mc
counter of any size, with all flip-flops switching
simultaneously, can be constructed using the dc
carry modules interconnected as in Figures 1 and 2
on the next page. The pulse amplifier interconnec-
tion of Figure 1 should be used between the first
pair of dc carry modules. The dc carry interconnec-
tion of Figure 2 may be used between all following
pa i rs of stages.
If the time between input count pulses is greater
than 400 + 100(N ~ 1) nsec (where N is the number
of dc carry modules), the pulse amplifier connection
is not necessary and the dc interconnection may be
used between ill I dc carry modules.
The carry modale contains an independent I-input
diode gate and six interconnected diode gates with
67
two, three, four, five, six, and seven inputs respec-
tively. The outputs are all similar to the Type RI07.
INPUT: Inputs are DEC standard levels. The input
loads on Pins M and Hare 6 ma each. The load on
Pin F is 5 ma; on Pin K it is 4 ma; on Pin E, 3 ma,
etc. The loads presented by Pins D and L are 1 ma
each. All loads are at ground, there is no load at
~ 3 v .
OUTPUTS: Each output is at ground only when the
input to the common gate and all inputs to gates of
lesser significance are at ~ 3 v. Each output circuit
can supply 18 ma at ground and has an internal load
of 2 mao
POWER: + 10 v(A)/0.7 ma, ~ 15 v(B)/26.2 mao
R18l - $35.00
68
"
"
FLIP-FLOP IIRl
L...-______ TY_P_E_R_20_0 ______ ---' ~
R200 FLlpFLOP
The R200 is a basic flip-flop for use in set-reset ap-
plications. It can.be set and cleared at any frequency
up to 2 mc. A set input makes the 1 output go to
-3v and the 0 output to ground: a clear input makes
the 0 output go to -3v and the 1 output to ground.
INPUT: Direct Set and Clear - A standard 100-nsec
pulse or a ground level of 100 nsec minimum dura-
tion aC,tivates the input: the load at ground is 1 mao
When not in use, the direct set and clear terminals
must be at -3v. If both inputs are held at ground,
both outputs are at -3v. Collector Triggering - The
flip-flop can also be set or cleared through its output
by a diode gate or a diode network. The triggering
circuit load is the external load on the output termi-
nal beiog driven plus the internal load.
69
OUTPUT: Standard levels. Each output can drive
17 ma of external load at ground. The internal load
is 4 mao If more than 18 in. of wire is attached to
an output, additional clamped loads (see the W002,
W005) should be connected to decrease the output
fall time. The load is sufficient if the positive transi-
tion at the opposite terminal reaches -lv within
80 nsec after the flip-flop is pulsed.
Note: Additional driving capability at -3 v is re-
quired by some circuits outside the R series. Auxil-
iary clamped loads W002 and W005 are available for
this purpose.
POWER: +10 v(A)/0.3 ma, -15 v(B)/16 ma.,
R200- $9.50
FLIP-FLOP IIRl
'---______ T_Y_P_E_R_2_0_1 _____ ----l ~
A201 FLlPFLOP
The R201 FlipFlop has direct set and clear inputs
and five diodecapacitordiode (DCD) gates. Because
of this large number of inputs, the R201 can be used
in any of the following applications without addi-
tional gating:
.1. Any two of the following as well as conditional
readin from an external source: up counter,
shift register, jam transfer buffer, ring counter,
and switch tail ring counter. Down counters
or up-down counters can also be implemented
if conditional read-in is not reQuired
2. BCD counter with read-in from two sources.
3. Buffer register or control flip-flop with readin
from five sources.
4. Special Counts of 2" (2
1
+ 1)
INPUT: Direct Set and Clear - A standard 100nsec
tion activates the input, the load at ground is 1 rna.
When not in use, the direct set and clear terminals
must be at -3v. If both inputs are held at ground,
both outputs will be at -3v. If the flipflop is used in
an up counter with carry gates enabled, the direct
clear pulse must be at least 400 nsec long to sup-
press carry propagation. Similarly, if the down
counter gates are enabled, the direct set pulse must
be 400 nsec long. DCD Gates, Level- Standard
levels of -3v and ground. Because DCD gates are
internally conditioned by the state of the flip-flop,
complement inputs may be formed by tying 1 and 0
DCD gate inputs together. A DCD gate is enabled by
a ground level and disabled by a -3v level. The con-
ditioning level must be present for at least 400 nsec
before the gate is pulsed. The level input represents
70
2 ma of load at ground. When 1 and 0 DCD gates are
connected in parallel to form a complement input,
the total level load is 3 rna at ground. Pulse-
Standard 100-nsec pulses (-3v to ground) at any
frequency up to 2 mc. It can also be driven by
positive-going level changes (-3v to ground) with
rise times of 60 nsec max and duration of 100 nsec
min. Prior to operation the input must have been at
-3v for at least 400 nsec. The pulse input repre-
sents 3 ma of load at ground. When a pair of 1 and 0
DCD gates have a common pulse input, as in com-
plementing or shifting, the total pulse load is 4 rna
at ground. Collector Triggering - The flip-flop can
also be set or cleared from its outputs by a diode
gate circuit or a diode network. The triggering circuit
load is the external load on the terminal being driven
by the circuit plus the internal load on that terminal.
OUTPUT: Standard levels of -3v and ground. The
carry propagate time is 70 nsec. The 0 terminal can
drive 11 rna of externarload at ground. The internal
load is 10 mao The 1 terminal can drive 13 rna of
external load at ground. The internal load is 8 mao
If more than 18 in. of wire is attached to an output,
additional clamped loads (see the W002, W005)
should be connected to decrease the output fall
time. The load is sufficient if the positive transition
at the opposite terminal reaches ~ l v within 80
nsec after the flipflop is pulsed.
Note: Additional driving capability at -3v is re-
iary clamped loads W002 and W005 are available
for this purpose.
POWER: +10 v(Al/0.2 ma,-15 v(B)/27 mao
R201 - $22.00
DUAL FLIP-FLOP I !;l
L--______ T_Y_P_E_R_2_0_2 ______ ---' ~
DIRECT SET
M ~ ______________ -L ____ ~ ~ ____________ -"
R2D2 DUAL FLlpFLDP
The R202 Dual Flip-Flop contains two identical flip-
flops. Each has a direct clear input, a common set
input, and two DCD gates. The R202 can perform
in anyone of the following applications without addi-
tional gating: up counter, down counter, shif.t regis-
ter, ring counter, jam transfer buffer, and switch tail
ring counter.
INPUT: Direct Set and Clear - A standard 100 nsec
.> tion activates tlie input; the load at ground is 1 rna
for each dear input, and 2 rna for the set input.
When not in use, the direct terminals must be at
-3v. If the flip-flop is in an up counter with carry
gates enabled, direct clear pulses must be at least
400 nsec long to suppress carry propagation. In
like manner, a 400 ns set pulse must be used when
the flip-flops are arranged as a down counter. If
both inputs are held at ground, both outputs are at
-3v. DCD Gates, Level- Standard levels of -3v
and ground. Because DCD gates are internally con-
ditioned by the state of the flip-flop, a complement
input may be formed by tying the 1 and 0 DCD gate
inputs together. A DCD gate is enabled by a ground
level and disabled by a -3v level. The conditioning
level must be present for at least 400 nsec before
the gate is pulsed. The level input represents 2 rna
of load at ground. When 1 and 0 DCD gates are
connected in parallel to form a complement input,
the total level load is 3 rna at ground. Pulse-
Standard 100-nsec pulses (-3v to ground) at any
71
frequency up to 2 mc. It can also be driven by
rise times of 60 nsec max and durati.on of 100 nsec
min. Prior to operation the input must have been
at -3v for at least 400 nsec. The pulse input repre-
sents 3 rna of load at ground. When a pair of 1 and
o DCD gates have a common pulse input, as in
complementing or shifting, the total pulse load is
4 rna at ground. Collector Triggering - The flip-flop
can also be set or cleared through its outputs by a
diode gate circuit or. a diode network. The triggering
circuit load is the external load on the terminal
being driven by the circuii: plus the internal load on
that terminal (6 rna each).
OUTPUT: Standard levels. The carry propagate time
is 70 nsec. Each terminal can drive 15 rna of exter-
nal load at ground and has an internal load of 6 rna.
If more than 18 in. of wire is attached to an output,
additional clamped loads (see the W002, W005)
should be connected to decrease the output fall
time. The load is sufficient if the positive transition
at the opposite terminal reaches -1 v within 80
nsec after the flip-flop is pulsed.
for this purpose.
POWER: +10 v(A)/0.5 rna, -15 v(B)/34 rna.
R202 - $25.00

TRIPLE FLIP-FLOP
TYPE R203
L. M 5 T
DIRECT
CLEAR
o
-9-
, -9-'
FF FF
K R
H N U
PULSE
INPUT
R203 TRIPLE FLlpFLOP
The R203 Triple Flip-Flop contains three identical
flip-flops. Each flip-flop has a direct clear input and
a DCD gate for conditional read-in.
INPUT: Direct Clear-A standard 100-nsec pulse
or a ground level of 100 nsec minimum duration
activates the input; the load at ground is 1 mao When
not in use, the direct clear terminal must be at -3v.
DCD Gates, Level- Standard levels of -3v and
ground. A DCD gate is enabled by a ground level
and disabled by a -3v level. The conditioning level
must be present for at least 400 nsec before the
gate is pulsed. The level input represents 2 ma of
load at ground. Pulse - Standard 100-nsec pulses
(-3v to ground) at any frequency up to 2 mc. The
flip-flop can also be driven by positive-going level
changes (-3v to ground) with rise times of qO nsec
max and duration of 100 nsec min. Prior to operatioo
the input must have been at -3v for at least 400
72
nsec. The pulse input represents 3 ma of load at
ground. Collector Triggering - The flip-flop may also
be set or cleared from its outputs by a diode gate
circuit or a diode network. The triggering circuit Iqad
is the external load on the terminal being driven by
the circuit plus the internal load on that terminal.
OUTPUT: Standard levels of -3v and ground. The 0
terminal can drive 15 ma of external load at ground.
The internal load is 6 mao The 1 terminal can drive
17 ma of external load at ground. The internal load
is 4 mao If more than 18 in. of wire is attached to an
output, additional clamped loads (see the W002,
W005) should be connected to decrease the output
fall time. The load is sufficient if the positive transi-
tion at the opposite terminal reaches -Iv within
80 nsec after the flip-flop is pulsed.
POWER: +10 v(A)/0.7 ma, -15 v(B)l40 mao
R203 - $28.00
QUADRUPLE FLIP-FLOP
TYPE R204
QIRECT
CLEAR
O ~ - L ______ ~ ~ ~
il
K :
Ff
J "
R204 QUADRUPLE FLIP-FLOP
The R204 Quadruple Flip-Flop contains four flip-
flops. Each has direct set and direct clear inputs.
Two of the flip-flops share a common direct clear
input. The R204 is used in general control applica-
tions. A set input makes the 1 output -3v and the
o output ground; a clear input makes the 0 output
-3v and the 1 output ground.
INPUT: Direct Set and Clear - A standard 100nsec
tion activates the input; the load at ground is 1 ma
per flip-flop. When not in use, the direct set and clear
terminals must be at -3v. If both inputs are held at
ground, both outputs will be at -3v. Collector Trig-
gering - The flip-flop can also be set or cleared
through its outputs by a diode gate circuit or a diode
network. The triggering circuit load is the external
load on the terminal being driven by the circuit
73
plus the internal ["ad on that terminal. The internal
load is 4 ma for each terminal.
terminal can drive 17 ma of external load at ground,
and has an internal load of 4 mao If more than 18 in.
of wire is attached to an output, additional clamped
loads (see the W002, W005) shou Id be connectec( to
decrease the output fall time. The load is sufficient
if the positive transition at the opposite terminal
reaches --Iv within 80 nsec after the flip-flop is
pulsed.
Note: Additional driving capability at --3v is re-
for th,s purpose.
POWER: I 10 v(AI/O.9 ma, -15 v(Bl/42 mao
R204 - $28.00
DUAL FLIP-FLOP jlRl
'--_____ -,--T_Y_P_E_R_2_0_5 ______ --' ~
DIRECT
CLEAR
H
LEVEL
INPUT
M
F ~ - - - - ~ , - - - - - ~
R205 ~ U A L FLlpFLOP
The R205 contains two identical flip-flops with a
common direct clear input. Each has three DCD
gates, and can be collector-triggered at either output
by a diodetransistor gate or a diode network. The
R205 can be used in any of the following applica-
tions without additional gating: up counter, down
counte.r, shift register, ring counter, or jam transfer
register.
INPUT: Direct Clear - A standard 100nsec pulse
or a ground level of 100 nsec minimum duration acti-
vates the input; the load at ground is 1 mao When
not in use, the direct clear terminal must be at-3v.
If the flip-flop is used in an up counter with carry
gates enabled, direct clear pulses must be at least
400 nsec long to suppress carry propagation. DCD
Gates, Level- Standard levels of -3v and ground.
Because DCD gates are internally conditioned by
the state of the flipflop, complement inputs may be
formed by tying 1 and a DCD gate inputs together. A
DCD gate is enabled by a ground level and disabled
by a -3v level. The conditioning level must be
present for at least 400 nsec before the gate is
pulsed. The level input represents 2 ma of load at
ground. When 1 and a DCD gates are connected in
parallel to form a complement input, the total load
is 3 ma at ground. Pulse - Standard 100-nsec
pulses (-3v to ground) at any frequency up to
2 mc. It can also be driven by positive-going level
changes (-3v to ground) with rise times of 60 nsec
74
max and duration of 100 nsec min. Prior to opera-
tion the input must have been at -3v for at least
400 nsec. The pulse input represents 3 ma of load
at ground. When a pair of 1 and a DCD gates have a
common pulse input as in complementing or shift-
ing, the total pulse load is 4 ma at ground. Collector
Triggering - Triggering circuit load is the external
load on the terminal being driven plus the internal
load on that terminal. Internal load for the 1 termi-
nal is 6 ma; for the a terminal, 8 mao
OUTPUT: Standard levels. Carry propagation time is
70 nsec. The a terminal can drive a 13-ma external
load at ground; the 1 terminal, 15 ma at ground.
Internal load on the 1 terminal is 6 ma; for the a
terminal, 8 mao If more than 18 in. of wire is at-
tached to an output, additional clamped loads (see
the W002, W005) should be connected to decrease
the output fall time. The load is sufficient if the
positive transition at the opposite terminal reaches
-Iv within 80 nsec after the flip-flop is pulsed.
for this purpose.
POWER: +10 v(A)10.5 ma,-15 v(B)/36 mao
R205 - $29.00
DELAY (ONE SHOT)
TYPE R302
K l
F L i J p
---; : POTENTIOMETER
R 5
CAPACITOR L ___ ;I\.\'-_..J
EXTERNAL
POTENTIOMETER
R302 DElAY (ONE SHOT)
The R302 contains two delays (one-shot multivi-
brators) which are triggered by DCD gates. Each
delay is independent 'and can be externally or in-
ternally controlled. When the input is triggered, the
output changes from its normal ground level to
-3v for a predetermined. adjustable period of time
and then returns to ground. The length of the delay
is determined by the capacitor and potentiometer.
External capacitors can be attached between termi-
nals Hand J (or Rand S), J (S) being the more posi-
tive terminal. The 20-kilohm internal potentiometer
can be used by putting a jumper between terminals
J and K. (or Sand T). External potentiometers can
be attached between terminals J and L (S and U).
The total resistance between these terminals must
not exceed 20 kilohm. A 20% change in power
supply voltage will change the delay less than 2%.
Delay jitter due to power supply ripple is less thar,
0.2%.
The expected delay of any combination (with more
than a 500-pf capacitance) can be estimated by the
following formula:
Delay =RC
where the delay time is in nsec, R in kilohm and C
inpf. The total capacitance, C, equals 220 pf of in-
'ternal capacitance plus any external capacitance
used, The resistance, R, is equal to the resistance
of the potentiometer plus 1 kilohm .of internal re-
sistance. The minimum delay is 400 nsec. The min-.
75
imum delay in nsec for a given external capacitor
is C where C is equal to the external capacitance
in pf plus a 220-pf internal capacitance. The re-
covery time is twice the minimum delay.
The delay range for typical capacitors used with the
internal potentiometer is given in the table that
follows:
DELAY RANGES
Total
Capacitance Minimum
Recovery
Used Delay
Time
(External -I- 220 pf Range
Internal)
Internal
400-4000 nsec 800 nsec
220 pf only
2000 pf 44O I<SeC 8 !J.sec
20 nf 40400 "sec 8OI,sec
200 nf 0.4-4 msec 0.8 msec
2000 nf 440 msec 8 msec
2OI,f 40400 msec 80 msec
2OOI,f 400-4000 msec 800 msec
R302 - $44.00
Large electrolytic capacitors can have internal leak-
age enough to substantially modify time delay. For
best results, use wet-slug tantalum electrolytics for
delays of several seconds or more. Four volt ratings
are adequate in most cases, but 6 or 8 volt ratings
may be desirable to further reduce leakage in some
cases.
Remote Control Wiring: Noise picked up on wires
leading to remote timing capacitors' or rheostats
tends to synchronize the end of the delay period (or
it could cause false triggering in extreme cases).
Even for 1 ft control wires, a grounded shield may
be advisable if smooth control and freedom from
jitter are essential.
INPUT: Level- Standard levels of -3v and ground.
A DCD gate is enabled by ground level and disabled
by a -3v level. The conditioning level must be
76
present for at least 400 nsec before the gate is
pulsed. The level input represents 2 ma of load at
ground. Pulse - Standard 100-nsec pulses (-3v
to ground). It can also be driven by positive-going
level changes (-3v to ground) with rise times of
60 nsec max and duration of 100 nsec min. Prior to
operation the input must have been at -3v for at
least 400 nsec. The pulse input represents 3 ma of
load at ground. The delay cannot be set from its out-
put terminal.
OUTPUT: Standard Level of -3v for the duration of
the delay time. The output can drive 18 ma of ex-
ternal load at ground. The internal load is 2 mao
POWER: +10 v(A)/0.6 ma; -15 v(B)/88 mao
INTEGRATING ONE SHOT
TYPE R303
r-
EXTERNAL -...1..
CAPACITOR +""r
I
I
I
I
L_".,"y",r--1
't __ J
EXTERNAL
RHEOSTAT
l20 K nJ
NOTE'
V DLRECT
SET
ALL PIN CONNECTIONS
REFER TO UPPER HALF
OF MODULE
R303 INTEGRATING ONE SHOT
The R303 contains a zero recovery time multivibra-
tor and complementary output buffers. Its unusual
characteristics include the ability to respond to in-
puts even while in the ONE state, so that successive
inputs above a preset frequency can postpone the
return to ZERO indefinitely. This characteristic can
be used, for example, to detect gaps in an otherwise
continuous pulse train, or to determine whether an
input pulse rate is above or below a preset fre-
quency threshold. The module can also be used to
establish initial conditions after system power is in-
terrupted, since it always goes to the ONE state
when the power" is first applied.
Delay is 3.5 microseconds to 0.9 second. Jitter is
less than 1.4% peak-to-peak. Precision: Delay time
will change less than 2% for a change of 20% in
supply voltage.
INPUTS:'Direct Set-A standard 100 nsec pulse or
a ground level of at least 100 nsec duration starts
the delay. The load at ground is 1 rna. At least 90%
to 99.5% of total delay (for 0 and,20 kilohm rheostat
setting, respectively) will not be measured out until
-3v is restored, a fact which may be important if
this input is grounded for longer than 300 nsec.
DCD Gate - Same as R302.
77
OUTPUTS: Each output can drive 18 rna at ground,
o rna at -3v. Extra 10 rna clamped loads may be
connected to change the driving capability at each
output to 8 rna at ground, 7 rna at -3v. The ONE
output will be at -3v during the delay period. and
ground otherwise. The ZERO output is grounded dur-
ing the delay period and -3v otherwise.
POWER: + lOv(A)/6 rna; -15v/75 rna.
CONTROLS: To choose desired range of delay,
ground the appropriate capacitor pin K through N
(for minimum delay range, ground none of these).
Ranges are separated by approximately a factor of
ten. For extra long delays, connect an external ca-
pacitor from pin J to ground. To use the internal
rheostat, connect pin P to pin R. For external control,
connect a variable resistance no larger than 20,000
ohms from pin P to pin S.
EXTERNAL CONTROL: Delay times may be con-
trolled by external Rand C in the same manner as
described for R302. Substantially the same Rand
C are required in the R303 as in the R302 for a
given delay, taking into account the ten times larger
minimum capacitance built into the R303. If elec-
trolytic capacitors are used, at least a 6-volt rating
is required.
Capacitor Values (MFD): Internal - 0.0022, Pin K -
0.027. Pin L - 0.39. Pin M - 3.9. Pin N - 39.0.
R303 - $45.00
I
VARIABLE CLOCK
TYPE R401
~
~
,--
,
,
,
,
,
I
EXTERNAL r--i
RHEOSTAT I c'"
L-r.-:
~ :
R
EXTERNAL 1_
CAPACITOR ... 1 ...
-ri-
oe
INTERNAL
RHEOSTAT
R401 VARIABLE CLOCK
The R401 Variable Clock is a gateable clock that
produces standard 100-nsec pulses from a 'stable
RC-coupled oscillator. The variable. clock is often
used as a primary source of timing for large systems.
The frequency of the R401 Clock is variable from
30 cps to 2.0 mc. Five capacitors provide coarse
frequency control, and a built-in 20,000-ohm poten-
tiometer permits fine adjustment. Terminals for an
external potentiometer or capacitor are available.
The maximum size of the external potentiometer to
be used is 20,000 ohms.
FREQUENCY SELECTION
Select Pin R C, - 82 pf
Pin P C, = 1200 pI
Pin N C, = 0.015 MFD
Pin M C, = 0.15 MFD
Pin L C, = 2.2 MFD
300 kc to 2.0 mc
30 kc to 375 kc
3.5 kc to 40 kc
300 cps to 4.5 kc
30 cps to 340 cps
Lower frequencies may be obtained by adding an
external capacitor between pins Rand C. A 20%
change in power supply voltage wifl change the prf
less than 1 %. The pulse-to-pulse jitter is less than
0.2%.
INPUT: The clock is enabled by a -3 v level or an
'open circuit at its enable gate input. The total transi-
tion time from the time the gate is enabled until the
first pulse reaches 90% of its amplitude is approxi-
mately 45 nsec. The pulses that follow appear at the
frequency selected. The clock may be disabled by
applying a ground level at the enable gate pin S.
The enable gate loading is 4 rna at ground. Disable
duration must exceed the period to which the clock
is set.
OUTPUT: The output i ~ a standard 100 nsec pulse,
-3v to ground. The output can drive 70 rna of
, external load at ground. The internal load is 3 rna.
POWER: +10 v(A)/1.3 rna; -15v(B)119 rna.
R401 - $45.00
78
CRYSTAL CLOCK
TYPE R405
CLOCK ~ OUTPUT
R405 CRYSTAL CLOCK
The Type R405 employs a series resonant crystal
oscillator, . squaring circuit, and output pulse am-
plifier. The crystal clock's output frequency remains
within 0.01% of specified value between OC
and + 55 C. The clock frequency is specified any-
where iii the 5 kc to 2 mc range by the customer
79
and is stamped on the crystal can.
OUTPUT: 100-nsec pulse, -3v to ground. The out-
put can drive 70 rna of external load at ground.
Internal load is 3 rna.
POWER: +10 v(A)J5.4 rna; -15 v(B)J50 rna.
R405 - $100.00
PULSE AMPLIFIER
TYPE R601
H
LEVEL.
INPUT
M
400 -NANOSECOND
PULSES
R601 PULSE AMPLIFIER
The R601 is a pulse amplifier that standardizes
pulses in amplitude and width. Outputs may be
either standard 100- or 400-nsec pulses (-3v to
ground). It has six DCD gates so that inputs from
as many as six sources may be mixed. Input pulses
can occur at any frequency up to 2 mc for 100-nsec
pulse outputs and up to 1 mc for 400-nsec outputs.
Delay through the pulse amplifier is approximately
50 nsec.
DeD GATE INPUTS: Level- Standard levels of -3v
and ground. A DCD gate is enabled by a ground level
and disabled by a -3v level. The conditioning level
must be present for at least 400 nsec before the
gate is pulsed. The level input represents 2 rna of
load at ground. Pulse - 40-nsec or longer pulses,
-3v to ground, at any frequency up to 2 mc: It can
also be driven by positive-going level changes (-3v
to ground) with rise times of 60 nsec max, and dura-
80
tion of 40 nsec min. The input must have been at
-3v for at least 4QO nsec prior to operation of any
input. The pulse input represents 3 rna of load at
ground.
OUTPUT: With terminals E and F connected to-
gether, the output is a standard 400-nsec pulse
(-3v to ground). With E and F open, the output is a
standard 100-nsec pulse, -3v to ground. The output
(for either 100- or 400-nsec pulses) can drive 70 rna
of external load at ground. The internal load is 3 rna.
Pulse amplifier outputs may be paralleled for a logi-
calOR.
Pulse lines and ground lines should be kept as short
as possible.
POWER: +10 v(A)/l.l rna; -15 v(B)/33 rna.
R601 - $25.00
PULSE AMPLIFIERS
TYPES R602, R603
'ic-------::qt-L-i --.--J
PA
~ ~
T
The R602 and R603 contain pulse amplifiers for
power amplification and for standardizing pulses
in amplitude and width. Each amplifier produces
standard 100-nsec pulses and one section of the
R602 can also produce 400 nsec pulses. DCD gates
and a single diode input permit inputs "from many
sources to be mixed. Input pulses can occur at any
frequency up to 2 mc for 100 nsec pulses, and Up
to 1 mc for 400 nsec pulses. Delay through the
pulse amplifier is approximately 50 nsec.
INPUTS: Level and Pulse - Same as R601. Diode-
Standard 100-nsec pulses (-3v to ground) or
a rise time of 60 nsec max. The input level must be
LEVEL
INPUT
P?A "V:p=hA T
1" l.
u
I( -=- R -;:
L 5
81
returned to -3v for at least 400 nsec before another
input may occur at either the diode or the DCD gate
input. The diode input represents a 1-ma load at
ground.
OUTPUTS: Outputs are standard 100 nsec pulse,
-3v to ground (except pin K of R602, which may be
changed to 400 nsec pulses by connecting pin D to
pin M). Each output can drive up to 70 rna load at
ground. The internal load is 3 rna. Pulse amplifier
outputs may be paralleled to obtain a logical OR.
Pulse lines and grounds should be kept as short as
possible.
POWER: R602: + 10 v{A)/2.2 rna; -15 v{B)/45 rna.
R603: + 10 v{A)/3.3 rna; -15 v{B)/57 rna.
R602 - $22.00
R603 - $28.00
BUS DRIVER
TYPE R6S0
GROUND FOR
SLOW RISE TIME
R650 BUS DRIVER
The R650 contains two inverting bus drivers for
driving heavy current loads to either ground or nega-
tive voltages. The four input terminals make the
R650 a versati Ie logic element as well. The diode
inputs D and E (N and P) are the principal inputs.
. They form a NAND gate for negative inputs or a
NOR gate for ground inputs. Gate inputs, such as
the R001 or R002, can be added through the node
terminal F (R). Other gating sources may be mixed
with the gate inputs by using collector terminal L
(V).
The bus drivers operate at frequencies up to 2 mc
with typical rise and fall times of 25 nsec. The
typical total transition times are 60 nsec for output
rise and 65 nsec for output fall.
By grounding pin H (S) the rise and fall time can be
increased to avoid ringing on exceptionally long
lines. The driver then operates at frequencies up to
500 kc with typical rise delay of 50 nsec, fall delay
of 50 nsec, and total transition time of 800 nsec for
82
output rise and 700 nsec for output fall. Terminal
K (U) can be used for driving coaxial cable.
INPUT: Standard levels at frequencies up to 2 mc
(up to 500 kc with H or S grounded). The diode in-
puts, including any diodes attached to the node
terminal, represent 1 rna of load, shared by all
grounded inputs. Collector terminal L (V) represents
10 rna of load at ground. External clamped loads
should not be connected to this terminal. The com-
bined length of all leads attached to the node termi-
nal should not exceed 6 in. The combined length of
all leads attached to the collector terminal should
not exceed 18 in.
OUTPUT: Direct - Standard levels. The output can
drive 20 rna of external load at either ground or
-3v. Resistor - Standard levels. The resistor output
drives 90-ohm coaxial cable such as RG-62-U. The
output can drive 5 rna of external load at either
ground or -3v.
POWER: +10 v(A)/50 rna; -15 v(B)/81 rna.
R650 - $23.00
B

SERIES
83

BASIC CIRCUITS
INTRODUCTION
B-series FLIP CHIP modules operate at frequencies from dc to 10 mc. They are electri-
cally and mechanically compatible with all other FLIP CHIP modules, including the
W-series accessories described in the previous section.
The B-series circuits described below are somewhat different from the R-series circuits
described in earlier sections of this catalog.
The Transistor Inverter
Many logical operations with B-series modules are performed with saturating PNP
transistor inverters. When a transistor is turned completely on or saturated, the
collector is practically a short circuit to the emitter of the transistor. If the emitter is at
ground in this condition, the output from the collector will also be at ground.
When a transistor is turned completely off (opened) the collector-to-emitter path is
practically an open circuit. If the collector is connected to a clamped load resistor, the
collector will be at -3 v.
Figure 1 depicts a B'series inverter and clamped load resistor. The capacitor shunting
the input resistor provides overdriving current to the transistor during input level
changes, thus switching the transistor much more rapidly. The load resistor is clamped
at -3 v with a diode so that when the transistor is off, the output signal is always at
-3 v, regardless of the loading on the inverter output.
-3 -15
tt
TRANSISTOR I
COLLECTOR :
COLLECTOR
INPUT ! INVERTERo--ef
BASE BASE
EMITTER
I
EMITTER I
: 1-
1
SYMBOL
CIRCUIT
Figure 1 Inverter Circuit and Symbol
84
B SERIES- INTRODUCTION
To simplify logic drawings, symbols are used, as shown on the righthand side of Figure l.
When the input is negative, the output is shorted to ground. When the input is at ground,
the transistor is open-circuited and the output, if connected to a clamped load resistor,
is at -3 v.
The inverter switch is analogous to the mechanical switch, as shown in Figure 2. The
logical designer can often build networks of inverters in the same manner as he would
if he were using mechanical relays or simple switches.

1. INPUT
l OUTPUT

Figure 2 Switch-Inverter Analogy
Figure 3 shows two switches clJflnectedin series to form a NAND circuit. Mechanical
switches A and B must be closed in order to ground the output. Similarly, in the series
inverter network, transistor inputs A and B must be negative in order to short the output
to ground. If either input is ground, that transistor will be an open circuit, and the out-
put will be -3 v. Therefore, the network also acts as a NOR circuit for ground levels.
Figure 3 Series Circuit
A similar analogy can be made with a parallel circuit,. as shown in Figure 4. If any of
the switches' close, the output will be shorted to ground. If A or B or C is negative, the
output will be shorted to ground. Only if all inputs are at ground will the network output
be negative. The parallel arrangement is a NOR gate for negative levels and a NAND
gate for ground levels.
85
A
Figure 4 Parallel Circuit
Inverters can be stacked in complicated series-parallel combinations, like that of
Figure 5, to perform sophisticated logical operations. There are some rules, however,
since the inverters are not quite ideal switches. Tliese rules are discussed under
"Special Instructions for B-series Logic Design."
A
B
-= -=
C=AB+AB
C=AB+AB
Figure 5 Series-Parallel Circuit
Diode Logic
All logical systems could be constructed with only parallel and series combinations of
transistors in inverter networks. However, as the number of inputs to a particular gate
increases, it becomes more economical to perform the gating action with diode gates.
The outputs of diode gates, which are permanently connected to a transistor inverter,
have the reference level re-established after every gate. Thus, such diode gates may
be connected in tandem indefinitely.
The circuit in Figure 6 illustrates a diode gate which can ?erve as a NAND gate for
negative signals and a NOR gate for ground signals. When both input pin A and input
pin 8 are negative, the current in the resistor to -15 v will overwhelm the small positive
bias and turn on the transistor, bringing the collector to ground. When either input A
or input B is held at ground, the positive bias on the base assures that the transistor
will be cut off and that the collector of the inverter will be at a negative voltage. If
either pin A or pin B is not connected, the circuit will act as a simple inverter. These diode
gates have the same circuit geometry as R series gates.
86
8 SERIES - INTRODUCTION
-t5V
A o--C>I--'
Bo--f)r.-...... -Kl--+C1--.--H
1
: ~
CIRCUIT
SYMBOL
i
I
+tOV
i
Figure 6 Diode Gate
Unbuffered Flip-Flops
If two grounded inverters are interconnected as shown in Figure 7, a bistable flip-flop
is obtained, whose symbol is shown on the right. When one side is cut off, its output is
negative. This condition holds the other side on, which, in turn, holds the first side off.
A
-=
Figure 7 Flip-Flop
~
~
If a grounded inverter, C, is connected to the off side, A, of the flip-flop in Figure 8,
the state of the flip-flop can be changed by pulsing the base of transistor C with a
negative signal. This extra transistor will short the collector of transistor A to ground
and will cut off transistor B, which will turn on transistor A. The flip-flop will then stay
in that state even after the signal is removed from C.
C
INPUT
-=
Figure 8 Setting A FlipFlop
87
The flip-flop is said to be in the 0 state when the 0 output terminal is at -3 v and the
1 output terminal is at ground. It is in the 1 state if the 1 output terminal is at -3 v
and the 0 output terminal is at ground.
A flip-flop input can be gated as shown in Figure 9. If the level input is negative when
the pulse arrives, the flip-flop will be set. If the level input is at ground, the flip-flop
will not be set. Negative pulses may be used in this way only if they are at least 70 nsec
wide. To set or clear B .series unbuffered flip"flops with 40 nsec pulses, a single inverter
must be used, as in Figure 8. On any inverter or diode gate connected to a flip-flop out-
put, the emitter terminal must be tied directly to ground. It must not be gated.
PULSE __ _
INPUT
Figure 9 Gated Setting of a Flip-Flop
Buffered Flip-Flops
In logic configurations described near the beginning of this catalog for R-series FLIP CHfP;;:'
modules, a diode-capacitor-diode (DGD) gate is always used to read into a flip-flop if the
conditioning level is changing. The delay in the DCD gate allows the gate to be sampled
while its conditioning level is changing.
o
OUTPUT
CLEAR
LEVEL
INPUTS
{
-3V

""h
r

":' ":'
5mo
+IOV +IOV
5mo
=
".,
'"
".,
PULSE o--Cu '"
INPUT
".,
"-
'"
....
Figure 10 B200 Buffered Flip-Flop
88
OUTPUT
}
ET
EVEL
NPUTS
A similar p ~ i n c i p l e is used in one type of 10 mc flip-flop. At the higher frequency, however,
the implementation of gate delay has to be modified to avoid excessive loading. Figure 10
shows how the 8200 general-purpose 10 mc flip-flop is arranged, with two diode gates
controlling the delayed inputs.
The connection from each output to the opposite input gate makes the level inputs condi-
tional. This means that if both set and clear inputs are at -3v together, the flip-flop will
be complemented by a pulse input. Flip-flops with this characteristic are sometimes called
"JK" flip-flops. All complementing FLIP CHIP flip-flops are "JK" flip-flops.
In addition to conditional inputs, direct set and direct clear inputs are provided. The direct
set input of the 8200 permits read-in from several sources, as shown in Figure 11. The direct
clear input allows a negative 40 nsec pulse to clear several flip-flops simultaneously.
r--------------,
L ______________ .J
Figure 11 8200 Read-I n
Since flip-flops with buffers have isolation between their outputs and the internal circuitry,
their outputs can be used as the lower of two series inverters. One of the sources shown
from which ONES are read in makes use of this property.
Some applications of 10 mc flip-flops would involve using the 8200 input gates for two
purposes. For example, it may be necessary to shift and count, or to receive a jam transfer
and shift. If one flip-flop must do two such operations the 8201 is called for.
The internal structure of the 8201 flip-flop differs from that of the 8200, to allow access
directly at the flip-flop inputs themselves. While this permits gate expansion by ordinary
10 mc inverters, it also requires the addition of delay to the output buffers. The resulting
arrangement still allows the flip-flop to be pulsed and sampled simultaneously.
89
B SERIES-INTRODUCTION
<>----.--+
o
INPUT
o
t
INPUT
Figure 12 B201 Buffered Flip-Flop
o t
OUTPUT OUTPUT
o
INPUT
t
INPUT
The 8201 flip-flop and the 8620 carry pulse amplifier sometimes used with it are limited
to two input inverters in series, and they require the upper inverter to be pulsed. In Figure 13,
read-in configurations are shown using the normal 1 input, rather than the direct set input
used for read-in on the 8200.
PULSE
INPUT
LEVEL

Figure 13 B201 Read-In
Pulse Amplifiers
-Flip-flop action is usually initiated by pulses. These pulses may be generated within the
system by a clock (see below) or brought in from "an external source and standardized by
a pulse amplifier.
The pulses which initiate the action in 8 series modules are 2.5 v volts in amplitude, In
most cases they are negative-going with a positive overshoot similar in amplitude to the
basic pulse. The duration of the pulse is 40 nsec.
Pulse amplifiers are powerful logical elements because they not only amplify and standardize
the shape of the pulses but also gate pulses, When the same logical gating is to be done on
a whole register of flip-flops, it can often be done once before the pulse amplifier which
drives the register.
90
B SERIES - INTRODUCTION
INPUT ~ . PA
Figure 14 Pulse Amplifier
8602 pulse amplifier outputs are pulse transformers that are capable of driving many
units of load. Normally both terminals of the transformer are available so that pulses
may be transmitted over twisted pairs, with a ground connection at the receiving end
only. Negative pulses are used for clearing, setting, and complementing through inverters.
Delays
The 8301 Delay contains a flip-flop which has only one stable state. When the input
terminal is shorted to ground by a pulsed inverter, the level output terminal will switch
from the normal ground level to the -3 v level for a fixed period of time.
At the same time the level output returns to its normal ground condition, a standard
pulse is produced at the pulse output terminals. Pulse outputs are transformer coupled
and have both terminals available at the module connector, like a 8602 Pulse Amplifier.
The duration of the delay is adjustable with both coarse and fine controls. Typical wave
forms for a delay unit are shown.
"1.
LEVEL OUTPUT
INPUT
r
J
LEVEL OUTPUT
-
-
Ie
r-
-
11
PULSE OUTPUT
fr--
IU
Figure 15 One-Shot Delay
The 8301 requires a short amount of time to recover after the delay is over. For
applications in which this recovery time is inconvenient, or where only a short delay is
needed, the 8360 Delay Line is also available. It accepts a pulse input or level transi-
tion input and produces a corresponding pulse output or level transition output a short
time later. The pulse amplifier built into the module allows pulse timing chains to be
built conveniently, as shown in Figure 16
91
B SERIES-INTRODUCTION
TIME t
(I NPUT PULSE I
f
TIME 2
TIME 3
Figure 16 Delay Line Ti,ming Chain
Clocks
If pulses are to be generated internally, a variable dock may be used. These units
produce standard pulses from Re-coupled oscillators. The 'clocks have a coarse and
fine control for setting the desired frequency up to the maximum of the speed line.
If precise timing is required, crystal clocks which contain a single frequency crystal
oscillator are available.
I CLOCK ~
Figure 17 Clock
92
B-SERIES
LOGIC CONFIGURATIONS
10 Me Shift Registers
Figure 18 shows two shift registers, both equally capable of 10 mc operation. Aside from
their use in serial parallel converters, shift registers can be used as downcounters when
fed back to form ring counters or switchtail ring counters. This possibility takes on special
importance at frequencies approaching 10 mc because loop delays can easily exceed 100
nsec in conventional binary counters modified to count at unusual radices.
,-----
Figure 18 Two 10 Me Shift Registers
(upper register: 8200 FlipFlops; lower register: 8201 FlipFlops)
The counter shown in Figure 19 forms the first three bits of a 10 mc scaler whose loworder
stages are R series flip flops such as type R202. This is the simplest 10 inc counter.
TO R SERIES FLIP FLIPS 1.25 Me
I
COUNT
OO------------uU+-----------4:Dt-----. i ~ ~ ~ E S
Figure 19 10 Me Scaler
(8200 FlipFlops)
93
8 SERIES - LOGIC CONFIGURATIONS
Pulse-Carry 10 Me Counter
Sometimes it is desirable to make each stage the same way in a long counter. The counter
in Figure 20 uses one 8201 and onehalf of a.8620 for each stage. The nine input inverters
built into the 8201 are not used in this application, and are therefore available for other
operations such as shift, jam transfer, set, clear, or complement. The carry propagation
delay in this counter will be about 10 nsec per bit.
Figure 20 FastCarry 10 Me Counter
(8201 FlipFlops, 8620 Pulse Amplifiers)
. Counter with Simultaneous Transitions
INPUT
PULSE
Counters with simultaneous transitions are sometimes required for on-the-fly read-out or
scanning applications. When the input pulse rate is low, the counter circuit of Figure 21
may be used. In this counter, a dc carry network is made with high speed diode gates and
inverters. Immediately after an input pulse has been received, a dc carry begins to propa-
gate from the least significant bit. When the carry is completely propagated, all of the gates
will be setup so that the next input pulse can simultaneously jam all flip-flops to the proper
state. Of course, it is important that input counts do not come in before the carry has propa-
gated all the way down the counter.
In the counter illustrated, the set-uptime must include the transition time of the flip-flop
as well as the transition times of an inverter and diode gate for each group of three flip-
flops el(cept the last group. Thus, the time between input pulses must be at least as great
. as 50 + N(60 + 15), where N is the number of diode gates used for carry.
94
'"
'"
CARRY
OUT
CARRY
IN
TO

Figure 21 Counter With Simultaneous Transitions
(8200 Flip-Flops)

a:J
CJ)
....,
::c
iii
CJ)
I
5
Cl
(')
C"l
o
Z
::!l
Cl
C
::c

i5
z
CJ)
'"
en
125MC
OUTPUT
PULSES
dJ. '. ~ p ~ ~
Figure 22 10 Me Counter With Simultaneous Transitions
(8200 Flip-Flops, 8602 Pulse Amplifiers)
PULSES
OJ
CJ)
rn
::0
m
CJ)
I
5
(;)
a
8
z
"T1
15
c:
~
(5
z
CJ)
B SERIES- LOGIC CONFIGURATIONS
Counters with Both High Input Rates
and Simultaneous Transition
A sophisticated counter is required for high resolution of the time between fast input signals
or for high speed scanning. When only a few states are needed, it is practical to use a ring
counter or switch tail ring counter such as described in the R series Logic Configurations
section at the beginning of this catalog. These two techniques are approximately equal in
complexity. The ring counter has the advantage that it is already decoded to single line out
puts. The switch tail ring counter has the advantage that it requires fewer flipf1ops.
For high resolution counters, the techniques outlined in Figures 19 and 21 can be com
bined. The least significant bits should use the technique shown in Figure 19. If a pulse
amplifier is at the input stage and another is added for carry pulses out, as shown in Figure
22, the delays of the two pulse amplifiers will approximately cancel each other. The lower
frequency stages still need to have simultaneous transitions but the input is now only 1.25
mc. Thus, the method of Figure 21 can be used for later stages.
A 10mc counter with a total of 15 stages can be made this way, and more if the 1.25 mc
P.A. is duplicated. The time difference between the earliest and the latest f1ipflop output
transitions for anyone input pulse will be simply the variation in delays of the flip flops and
pulse amplifiers, about 30 nsec or less.
Overflow Detector
The simplest, most common counters have nonsimultaneous output as would be obtained
with the counter of Figure 19 followed by R series flipflops. Sometimes it is necessary to
detect the all ones state and initiate action such as disconnecting the input source or
setting an alarm flipflop when this state is reached. The arrangement shown in Figure 23
can be used for this purpose. 8y treating the fastest bit of the counter separately, this
system attains sufficient speed to allow input pulse rates up to 10 me.
The first input to come to its final 1 state will be the most significant bit, which
receives carry pulses after the greatest delay. After each of the other flip-flops comes
to its final 1 state, eventually the first bit in the counter does so. This least significant
flip-flop receives input pulses directly, without delay. The next input pulse, which clears
the counter, also produces a pulse output from the overflow detector. This pulse can
be used to clear a control flip-flop, cutting off pulses into the counter.
With proper regard for timing considerations it is possible to use a similar technique
as the basis of counters and scalers for arbitrary numbers and ratios at high speeds.
Synchronizer
8series FLIP CHIP modules form synchronizers in much the same way as Rseries modules
do. A startstop synchronizer is shown in Figure 24. Notice that a 8201 flipflop must be
used, since it is being sampled by an inverter gate rather than by a 8200 delaying input
gate.
97
B SERIES - LOGIC CONFIGURATIONS
C I
C2
C3
C4
C5
up TO
C6
12 MOST-
SIGNIFICANT
FLIP-FLOPS
C7
ce
C9
CIO
CII
C12
PULSE
OUTPUT
ON ALL
ONES
~ " } o I ~ - - c o FASTEST
INPUT
PULSES
Figure 23 Overflow Detector
(8171 Diode Gate, 8602 Pulse Amplifier)
r-- -- -,
I
I
I
I
L
START
PULSE
Figure 24 Synchronizer
I
I
I
I
J
STOP
PULSE
(8201 Flip-FI.op, 8602 Pulse Amplifier)
98
FLIP-FLOP
B SERIES - SPECIAL INSTRUCTIONS
Gateable Clock
Below is shown a method of producing a stable gateable clock from a B360 Delay Line
and two external inverters. The clock is capable of operating at frequencies from 10 me
down to approximately 4 mc. The enabling level must be at -3 v for the duration of
operation. Clock pulses are begun by a 40 nanosecond negative pulse at the "start"
input. Lower frequencies may be obtained by inserting one or more additional B360
sections in the loop.
START
PULSE
ENABLING
LEVEL
~ A Y
...LLiNE -=
-=
Figure 25 Gateable Clock
(B360 Delay Line)
SPECIAL INSTRUCTIONS'FOR B SERIES
LOGIC DESIGN.
Inverter Usage'
A maximum of two inverters. may be put in series if the output is to drive another
inverter (Figure 26) . .If a flip'flop output is driving an inverter emitter, this. flip-flop must
be counted as an inverter (Figure 27).
=-
Figure 26 Level Gating Figure 27 Flip-Flop Buffer as an Inverter
99
I
8 SERIES - SPECIAL INSTRUCTIONS
When the output of a series of transistor inverters is driving the input of a pulsed unit,
as in Figure 28, an inverter pulse gate may be. added in series with the two, level gates.
This pulsed inverter must be placed at the bottom of the series string; i.e., the end
farthest from the load. The emitter may be driven by a flip-flop instead of being
grounded as shown, replacing one <if the inverters.
Figure 28 Pulse Gating
The Flip-Flop B201 is a special case. No more than two inverters in series may be used
to drive the B201, and the inverter at the top' must be pulsed with' a standard40-nsec
pulse as shown in Figure 29. The same applies to the B620, which is used with the
B201 for counting,applications.
Figure 29 8201 Gating (8620 alsol
The collector of an' inverter driving an emitter in a network .of transistors must also
supply the base current leaving the inverters higher in the chain. This number is
normally small, but in complex networks it must be considered.
An inverter can drive no more than one clamped load resistor and six bases of on-
inverters. Since transistors are almost symmetrical, this on base current can also flow
through the collector of a transistor whose emitter is open, as shown in Figure 30. In this
case, the collector of the bottom on-transistor must carry the current A from the load
resistor and the base currents from B, C, and D. Nominally, the current required is 10 ma
to a negative voltage return; however, this is increased by the base current required.
100
Figure 30 Loading
Because inverters are not really ideal switches, each collector of a series string of dc
inverters supplying a pulse inverter will go somewhat negative during the pulse. This
means that if a series of inverters is supplying both pulse current and a dc signal, care
must be taken because a signal will occur in the dc output during the pulse.
CiJ+-----;I...I\..)4--PULSE
E
Figure 31 Illegal Inverter Use
In Figure 31,when the input C is negative, flip-flop A should be set by,the pulse, but
flip-flop B should not be set. However, during the pulse, collector E ,of the dc inverter
feeding flip-flop A will go slightly negative. It will partly turn on the dc inverter feeding
flip-flop B, and sometimes it will set flip-flop B as well as flip-flop A. This network will
work only if the two pulsed inverters are driven by separate, non-simultaneous signals.
101
B SI'l'It<S - SPECIAL INSTRUCTIONS'

E
c
Figure 32 Recommended Inverter Use
The network shown in Figure 32 will work whether or not the 'pulses are simultaneous,
Two inverters have been added so that collector E will no longer be pulled negative by
the pulse, since the pulse current will now come directly from ground. instead of
collector E. (If both outputs are driving the same flip-flop, the network shown in
Figure 31 can' be used safely because the output of the A-side will be much greater than
that of the B-side.) .
B-SERIES' MODULE INTERCONNECTIONS
Because the operating speed of B-series modules is high, the effects of wiring reactances
must be taken into account. The following' sections contain guidelines for the logic
designer. By following them in his planning, he will keep the unwanted effects to a
minimum.
Level and Pulse Tr.ansmission
A level can be transmitted through a single wire if the length is less than one foot. A
series termination of about 100 ohms at the driving end is recommended for longer
lin!ls. It isolates the line capacitance but reduces driving capability. The loss in driving
capability can usually be overcome by slight logic changes.
Multi-conductor strip cable can also be used to transmit levels, in which case every
other wire should be grounded to provide isolation. Long lines transmitting levels should
not be bundled together. Particular care must be taken with unbuffered flip-flops, since
they can be set or cleared bythe capacitive coupling of fast changes to their output leads.
102
The total current path from the pulse gate emitter to ground (including the path through
any level gate) should be as short as possible. The recommended maximum length is
6 in. If this path is closed to ground by a mechanical switch the lead length may be
longer, provided a 0.001 to 0.01 I,f capacitor is connected from the emitter to ground
to isolate the lead inductance. The switch should be single-pole, double-throw with
one side to ground and the other to an unused clamped load.
Pulse transmission is optimized by using a twisted pair of wires, one grounded at the
first load. Resonance effects may cause the pulse amplitude to increase, especially
with heavy loads and long wires. A carbon damping resistor of about 100 ohms may be
required at the receiving end to reduce the amplitude to 2.5 v. Lightly loaded pulses
may be transmitted about 6 in. on a single wire.
Noise in Emitter Gating
The benefits in logic speed and simplicity obtained by the use of gated emitter logic
can be lost if interconnections are not adequately planned. Figure 33 illustrates h o ~ a
wire which is conditioning two pulsed inverter emitters can act as part of a resonant
circuit. When one inverter is turned off, resonance may produce a positive pulse on the
emitter of the other inverter, turning it on at the wrong time. To avoid this, each emitter
must be driven by a separate wire, unless the tota( wire length is 4 in. or less.
Figure 33 Effects of Wiring Reactances
"I
I
I
I
Another consideration is shown in Figure 34. Here the hazard is that stray capacitance
may provide an effective ground for the pulsed emitter, even though no logical path to
ground is provided. To avoid trouble of this sort, a W002 or W005 clamped load should
be connected to the gated emitter to discharge the stray capacitance if the total of such
capacitance will exceed 35 pf (see data under "Fall Time" below>. Not only flip-flops,
but delays and pulse amplifiers too are susceptible to noise generated by these
mechanisms. .
103
B SERI ES - SPECIAL INSTRUCTIONS
r ---- ----_-'----+
I
35 pI TOTAL IS ...1_
MAXIMUM. ALLOWABLE -,-
WITH NO CLAMPED _L
LOAD -:
OV
=
Figure 34 Stray Capacitance
Delay
ov
On certain occasions a logical design requires a minimum delay around a loop. However
none of the standard logic circuits except the 8201 buffered flip-flop are built to maintain
a minimum delay. Actually, efforts are regularly made' to reduce inherent logic delays in
existing circuits. Wherever a small logical delay is needed, a 8360 Delay Line should
be used.
Rise Time
B series modules are designed so that rise times (not including delay before rise) at any
output will be at least as fast as fall times (not including delay before fall) at the same
output, under most conditions. Thus an estimate of fall ti.me Is also an estimate of worst-
case rise time, so that rise times need not be considered in design.
Fall Time
Slow fall times of logic levels sometimes may limit the repetition rate of systems made
with 8 series modules to less than 10 mc. The diagram below .shows how the fall time
of _a 8 series logic level can be estimated from a knowledge of the size of the clamped
load and the total resistive and capacitive loading. Since a standard clamped load can
drive up to 7 ma at -3 v, the fall time when a standard clamped load is doing the
driving will be roughly half as many nsec as there are pf of capacitance to be driven.
The calculation is given in Figure 35 .
,
r----,
I
I
I
I
I -= I
_J
r----------,
I
C I
I!
-= I -= -3V I
___ J
FALL TIME (NSEC) = CAPACITANCE (pflXVOLTAGE (V)
CURRENT (ma)
Figure 35 Calculation of Fall Time
104
This rule will predict fall times with adequate accuracy for many conditions, but as
the number of inverter base inputs driven by the clamped load approaches the limit
of seven, the predicted results become optimistic due to the static i ma drive absorbed
by each inverter input. To restore accuracy to the calculation for heavy loading by
inverter base inputs, the table below provides several bench marks for estimating.
FALL TIME FACTORS vs. INVERTER LOADING
Clamped Load
Factor by Wilich
Dr i ve Cu rrent Total Capacitance
Drive Current
Absorbed by Static Loading Should be Multiplied
Available
at -3 v To Estimate
at -3 v
Fall Time in Nsec
0- 2 ma 0.4
7 ma
4 ma 0.5
6 ma 0.7
0- 4 ma
0.2
14 ma 8 ma
0.25
12 ma
0.35
To obtain reliable predictions from this table, it is important to add together all the
capacitances being driven by a given clamped load. Below is a summary of pertinent
capacitances in B series modules:
Standard Inverters (8104,8105,8123,8124,8620)
Base Input: 60 pf if either the emitter or collector is grounded.
4 pf if neither the emitter nor the collector is grounded.
Emitter Input: 75 pf if the base is at -3 v.
8 pf if the base is at ground.
Collector Output: 8 pf regardless of input conditions.
Diode Gates (8113, 8115, 8117, 8130, 8155, 8171)
Diode Inputs (except B155):
Diode Inputs (B155 only): .
Emitter Input (B117 only):
Emitter Input (B155 only):
Collector Outputs:
Emitter Outputs (8130 only):
8 pf if any other diode input is grounded.
12 pf if no other diode input is grounded.
20 pf at each of the six inputs, regardless
of the conditions at the other five inputs.
8 pf regardless of diode input conditions.
20 pf regardless of diode input conditions.
8 pf regardless of input conditions,
except pin D of B171: 12 pf.
8 pf regardless of input conditions.
105
8 SERIES - SPECIAL INSTRUCTIONS
8uffered Flip-Flops (8200,8201)
Inverter Base Inputs:
Input Inverter Collectors:
Buffered Flip-Flop Outputs:
60 pf for inverters with emitters grounded.
8 pf each.
8 pf each. Estimate output fall times on
the basis of 7 ma available drive at -3v
for B200, 14 ma for B201.
Unbuffered Flip-Flops (8204 or other flip-flops constructed from two 81 OS-type inverters)
Flip-Flop Outputs: 68 pf each.
Pulse Amplifiers (8602)
Pulse Amplifier Input: 8 pf. Estimate input fall time on the basis
ot 7 ma available drive at -3 v when
calculation is necessary to assure 50
nsec dwell before going to ground.
Base Input:
Power Inverters (8681)
100 pt.
Collector Output: 12 pt.
Input:
8us Drivers (8684)
8 pt.
Clamped Load (W002, WOOS, or any 10-ma standard clamped load with its own
pin connection)
Each Clamped Load:
Single Wire:
Twisted pair with one wire
grounded, or single wire in a
loose bundle:
Single wire or twisted pair,
in laced cable:
8 pt.
Hookup Wire
1 pf' per inch.
1112 pt per inch.
2 pt per inch.
106
Example: Calculate the maximum number of inverter collectors at a 8602 input for 10 mc
operation.
Pulse-to-pulse interval
Dwell required at -3v
Input pulse width
Time left for fall
100 nsec
50 nsec
40 nsec
10 nsec
Factor by which capacitance should be multiplied to obtain fall time (from
table): 0.4
Total allowable capacitance: 10/0.4 = 25 pf
Number of 8 pf connections: 25/8 = 3 +
Thus in addition to the input capacitance of the P.A. itself, 2 inverter collectors
may be connected.
107
INVERTERS
TYPES 8104,8105,8123,8124
CLAMPED
LOADS
ClAMPED I
LOADS
f I
H M S
ClAMPED f I I
LOADS
I I
t L L
COLLECTOR
OUTPUT
COLLECTOR

f-o.
B4SE E
INPUT
EMITTER 0
INPUT

8104 INVERTER
CL.AMPED!
LOADS
o
E
BASE
OUTPUT
INPU: J
T4J

to f-..
8124 INVERTER
F

N S V

OUTPUT
BASE 0
INPUT
BASE
INPUT a--1: a--1:
lo-ef

-= -=
To--r
8105 INVERTER
108
8123 INVERTER
The B104 contains three standard 10-ma clamped
loads and four transistor inverters, each with' its
base, emitter, and collector brought to connector
,
The Bl05 has five standard lD-ma clamped loads
and five transistor inverters" with each emitter
grounded, and with each base and collector brought
out.
The B123 has three standard lO-ma clamped loads
and eight transistor inverters. The inverters are tied
together in series groups of two.
The 8124 has three standard 10-ma clamped loads
and nine transistor inverters, each with emitter
grounded, and with each base and collector brought
to terminals. The collectors are tied together in
groups of three.
Each inverter is analogous to a switCh. If the in-
verter base is at -3 v and the inverter emitter is at
ground, the transistor is saturated and a conducting
path is between the emitter and' col-
lector of the inverter. If the base is at ground, or if
both base and emitter are at --3 v, the emitter-
collector path is open circuited (Le., will not allow
current to flow).
Delay through the inverter is approximately 12 nsec
for lightly loaded inverters driven by a pulse.
I NPUT: Inverter-Base - Whenever the base input is
at -3 v and the emitter input is at ground, the
static input load is 1 ma. In any other case (i.e.,
emitter at -3 v or both base and emitter at
ground) there is no static load. The base can reject
0.5 v of noise. Pulse - Pulse inputs are standard
40-nsec pulses at any frequency up to 10 mc.
(Pulses of longer duration may also be used.) Level
- Level inputs are standard levels of ground and
-3 v. Emitter - An inverter whose base input is
at -3 v will saturate if its emitter is brought to
ground by any conducting path. The circuit that
establishes this path (another inverter, flip-flop,
direct connection to ground, etc.) will be loaded
with whatever external load may be present at the
inverter collector, plus the internal load of 1 mao
If the base is at ground or both base and emitter
are at -3 v, there will be no static load. Clamped
109
Load - Each clamped load draws 10 ma from any
circuit that brings it to ground.
OUTPUT: Inverter - The maximum output driving
capability at ground is 16 mao This current is avail-
able if the emitter is connected directly to ground
(as is always the case with the BI05 and B124).
If the emitter is not directly grounded, the maximum
output load is 1 ma less than the maximum input
available to the emitter. A lO-ma clamped load at-
tached to the output (collector) will provide a maxi-
mum output driving capability at -3 v of 7 ma.
Clamped Load - Each clamped load can supply up
to 7 ma at -3 v.
Note: The saturation voltage drop of the inverter
places a limit on the number of inverters which
may be connected in series. For more information
see jjlnverter Usage."
POWER: B104: +10 v(A)/O ma -15 v(B)/38 mao
8105: + 10v(A)/0 ma -15v(B)/58 mao
8123: +10v(A)/0 ma -15v(B)/38 mao
8124: + 10v(A)/0 ma -15v(8)/38 mao
8104-$17.00
8105 - $2l.00
8123 - $31.00
8124-$31.00
NAND/NOR GATES
TYPES 8113,8115,8117,8171
CLAMPED 1 1 1
LOADS b b i
H M S
DIODE
o pc:J
INPUTS
E
_ _ K
:=:tf fC::
8113 NAND/NCR GATE
DIODE
OUTPUT
E L
F
H _
J
K

P V
R
S M
: EMITTER
8117 NAND/NCR GATE
The B113, B115, B117, and Bl71 are pcsitive NOR
diode gates; they form NOR gates for ground inputs
and NAND gates for -3 v inputs. The .outputs of the
diode gates drive inverters similar to the BlO5, for
power amplification. The typical total transition time
is 40 nsec for output fall and 60 nsec for output
rise. (Because the rise and fall delays differ, these
diode gates may shorten negative input pulses
markedly; see below.)
The B113 provides three standard 10-ma clamped
Icads and four diode gates, each with two diode
inputs and the collector brought out.
The B115 has three standard lO-ma clamped loads
CLAMPED 1 1 1
LOADS 1 b b
J P V

DIODE
INPUTS: _
8115 NAND/NOR GATE
8171 NAND/NOR GATE
and three diode gates, each with three diode inputs
and the collector brought out.
The B117 has two diode gates, each with six diode
inputs and the collector brought out. In addition, the
emitter of one of the inverters is available.
The Bl71 is a single gate with twelve diode inputs.
In addition tc the positive NOR output, another in-
verter has been added at the output; using the in-
verted output makes the B 171 an OR gate for ground
inputs and an AND gate for -3 v inputs.
INPUT: Diode Inputs - The static load is 1V4 ma,
shared by all inputs which are at ground. Pulse-
Standard 70-nsec negative input pulses may be
used for setting or clearing flip-flops only. Due to
110
pulse shortening by the gate, 40-nsec negative
pulses may not be used as inputs. Level- Level in-
puts are standard levels of ground and -3 v. R
series 100-nsec pulses must be regarded as levels
when used with B series modules. Emitter (Bl17
only)-If all the base inputs are at -3 v, the B117
will saturate if its emitter is brought to ground by
any conducting path. The circuit that establishes
this path (another inverter, flip-flop, direct con-
nection to ground, etc.) is loaded with whatever
external load may be present at the inverter col-
lector, plus the internal load of 11/4 mao If the base
is at ground (any 1 diode input at ground) or the
base and emitter are at -3 v, there is no static load.
Clamped Load - Each clamped load draws 10 rna
from any circuit that brings it to ground.
111
OUTPUT: Collector Outputs - The collector outputs
have a maximum output .drive of 16 ma at ground.
This current is available if the emitter is connected
directly to ground. If the B 117 emitter is not directly
grounded, the maximum output load is 11/4 ma less
than the maximum input available to the emitter. A
lO-ma clamped load attached to the output (col-
lector) drives 7 rna at -3 V. B171 Only - For the
B171, terminal 0 drives 5 ma at ground and 7 ma at
-3 V. Terminal E can drive 16 ma at ground.
Clamped Load - Each clamped load can supply up
to 7 ma at -3 v.
POWER: Bl13: +10 v(A)/0.7 ma; -15 v(B)/43
mao Bl15: +10 v(A)/0.5 ma; -15 v(B)/42 rna.
Bl17: +10 v(A)/0.3 ma; -15 v(B)/2.5 mao B171
+10 v(A)/0.3 ma; -15 v(B)/31 mao
B113 - $23.00
8115 - $21.00
8117 - $14.00
B171 - $18.00
THREE-BIT PARITY CIRCUIT
- TYPE 8130
0.
0
"I"
B13D THREE-BIT PARITY CIRCUIT
This special logic module has two levels of high
speed logic and complementary outputs_ It is de
signed to compute the parity (o(jd or even) of the
contents of a flipflop register with a minimum of
time delay, but it can be used wherever there is a
need for four 3-input negative diode AND gates feed-
ing a 4-input OR gate.
Delay is typically 15 nsec from 50% of the input
transitio'n to 50% of the output transitions when
output capacitive loading is very small.
INPUT: Each of the gates has a 2-ma load shared
among the ground inputs Qf that gate. When the
112
inputs are connected to compute parity, the total
load on each of the input lines at ground is 2% mao
OUTPUT: Each of the complementary outputs can
drive 10 ma at ground in addition to the built-in
clamped load. The clamped loads each can drive
7 ma at. -3 v. Due to the special nature of this
circuit, inverter emitters may not be driven.
An indicator output is provided for applications
where parity is to be displayed by an indicator-with-
amplifier (4902, 4903) through an Indicator Con-
nector Board W020.
POWER: +10 v(A)/49 ma; -15 v (B)/92 mao
B130 - $50.00
HALF BINARY-TO-OCTAL DECODER
TYPE 8155
K D
ENABLE EMITTER
INPUTS INPUT
B155 HALF BINARY-TO-OCTAL DECODER
The 8155 module is used alone as a 2-bit decoder
with two enable inputs, or it is used with another
8155 to form a full 3-bit (binary-to-octall decoder,
using one combined enable line_ Either way, each
binary input combination results in one selected
output held at ground if the decoder is enabled_ No
output will be selected if an enable input is held at
ground_ The decoder consists of four 4-input diode
gates with appropriate input interconnections_ All
of the output transistor emitters are connected to
pin D, providing a third enabling point Also included
are four standard 10-ma clamped loads_
INPUT: Diode Inputs - The input load is 1'14 ma
per negative output, shared among the grounded in-
puts. When the inputs are binary, as in the first 4
lines of the truth table, the current at ground is
2 (111.) -0- 2, or 2 mao When a pair of 8155s are
connected as a binary-tooctal decoder the input
load would be 7 (1 Y4) -0- 3, or 3 ma per grounded
input. When four 8155s are connected as a sixteen-
state decoder, the input load is 15 (1%) -0- 4, or
4% ma per grounded input. Similarly, the input load
for the last 4 lines of the truth table is 3% ma for
the grounded input. If the only grounded input is an
enable input such as pin Lor K, the load will be 5 mao
Emiller Input - If both enable inputs are at -3v,
one of the output transistors will saturate if the
emitter input is brought to ground by any conducting
TRUTH TABLE
INPUTS OUTPUTS
H D K M N R
-311 011 -311 0" all -311 -311 all -311 -311 -3v
-311 all 011 -311 all -311 -311 -311 011 -311 -3v
011 -311 -311 011 Oil -311 -311 -311 -3,.. all -3v
all -311 0" -311 -311 -311 -311 -311 -311 -311 Ov
-3v -3v -3v -3v -3v
Ov -3v -3v -3v -3v
Ov -3v -3v -3v -3v
-3" -311 all -311 011 -311 -311 -311 0" -311 0 ....
-311 -311 -311 all 011 -311 -311 all -311 011 -3v
Ov -311 -311 -311 all -311 -311 -311 -311 all Ov
-311 011 -311 -311 all -311 -311 all 0.... -311 -3v
113
path. The circuit that establishes this path (an in
verter, flip-flop, another diode gate, etc.) is loaded
with whatever external load may be present at the
selected output, plus the internal load of 11/4 mao
Clamped Load - Each clamped load draws 10 ma
from any circuit that brings it to ground.
When the two 8155 modules are used together, each
of the four binary inputs on one module is tied to the
corresponding input on the other module and these
four lines are driven by the least significant two bits.
One enable input on each module is driven by the
third bit so that only one of the two modules has an
output selected. The second enable input allows four
modules to be combined to form a 16state decoder,
and the emitter input allows further expansion to 32
states using 8 modules.
OUTPUT: Each decoder output can drive 16 ma at
ground. Each clamped load can drive 7 ma at -3 v.
If the emitter input is not directly 'grounded, the
maximum output load is limited to 1% ma less than
the maximum current available at the emitter input
Note: Simultaneous switching of 8155 outputs is not
assured. If adjacent 8155 outputs are ORed together,
for example, the gate output may contain spikes.
POWER: +1O(A)/0.6 ma; -15(8)/53 mao
8155 - $25.00
FLIP-FLOP I ra-
'--______ T_YP_E_B_20_0 __ -'-___ ---' ~
D
V o - - - - - - - ~ - - - - - - ~ J - - - - - - - O T
t----KJ----<> p
Ro--C:!--'
8200 FLlPFLOP
Most 10 mc registers can be built with 8200 buffered
flipflops. The delay from pulse input to flip flop out
put is short, suiting the 8200 for unidirectional
counting and shifting applications in which com
parators are used to stop the action. Delayed level
inputs are conditional, providing JK characteristics.
Some typical operations the 8200 can perform at
10 mc input rates are: gated shifting, parallelserial
conversion, jam transfer, and simultaneoustransition
counting. Typical delay: 30 nsec. Typical output rise
time: 35 nsec. See "8 Series - Logic Configurations"
for examples of 8200 applications.
INPUTS: Pin V is a direct coupled clear input. Though
it will force the zero output negative as long as it
is held negative, its normal use is as a pulsed clear
input. When driven from a 40 nsec pulse amplifier
such as 8602, pin V is equivalent to an inverter
base input. If a clamped load used to bring it nega
tive 2 milliamperes should be alloted to each pin
V driven. Pin T is a direct coupled set input. Though
it will force the one output negative as long as it is
grounded, its normal use is as a pulsed preset. When
used, pin T must be connected to an external clamped
load; for 10 mc operation a 10 ma clamped load is
required. This input requires the same drive as a
5 ma clamped load. Pulsewidth minimums are 40
nsec with inverters and 70 nsec with diode gates.
Pins J, K, N, P, R, and S drive diode gates whose
114
outputs are delayed. The delayed gate outputs car
be sampled by a 40 nsec pulse at pin H. Each gatE
is a 5 ma load shared among its grounded input!
including an internally connected input from thE
flipflop output that makes it conditional. These diodE
inputs are normally driven by outputs from other
8200 flipflops sharing a common pulse source for
pin H inputs. The internal delay is sufficient to per
mit full 10 mc operation connected this way, as lone
as fall time at gate inputs or flipflop outputs doe!
not overlap next input pulse at pin H. See table below
Pin H must be driven by negative 40 nsec pulse!
only. When pin H is pulsed, the flipflop will respona
to a "1" stored in either gate delay during the pre
vious 100 nsec. Pin L is normally grounded to pin
M. It is possible to use pin L as a gate for pin H
inputs if no more than a 3 inch path to ground i,
provided by the gating inverter. Two inverters in
series may not be used, but several parallel inverters
may be used. This input constitutes a 15 milliamperE
load on the inverter (8104, 8105, or 8125) or buff
ered flipflop (8201) that drives it.
OUTPUTS: 0 and 1 outputs - Each output can
supply 16 ma at ground, or 7 ma at -3v. Each out
put is connected internally to one of the diode gate
inputs. The 1 output (pin D) therefore shares a 5
ma gate load with any grounded inputs at pins J, N,
or R. The 0 output (pin E) shares a 5 ma gate load
with any grounded inputs at pins K, P, or S. Thus
in some cases flipflop output driving ability may be
reduced to as little as 11 ma at ground. At high
frequencies fanout from B200 outputs to 10 mc
inverter bases is reduced. The table below gives
some results from "Special Instruction for B Series
Logic Design."
B200
Maximum Inverter
Inverter Bases With
FlipFlop
Base Inputs Short Leads
Input
With Short and 10 ma
Frequency
Leads Clamped Load
10 mc 2 4
6 mc 4 9
3 mc and 7 14
below
115
Indicator Output - a separate output at pin F drives
a 4910 indicator through Connector Board W020
without loading the flipflop output excessively with
stray wiring capacitance. When used, this output
reduces the 1 output drive by 1 ma when used with
the W020, or by 2 ma when used to drive a 4910
indicator amplifier directly. Use direct outputs for
driving W050 inputs.
POWER: 'II0v(A)/11 ma; -15v(B)/45 mao
B200 - $25.00
FLlpFLOP
TYPE 8201
u
0------.---+-7
v
1>-----<p--o T
8201 FLIP-FLOP
Some 10 mc operations require a greater variety of
pulse inputs than the 8200 can provide. For ex-
ample read in from several sources, bi-directional
shifting, and arithmetic operations all require the
greater flexibility of the 8201.
In order to allow the outputs to be sampled by the
same pulse that' is changing the state of the internal
flip-flop, the buffers include a controlled amount of
delay. The diagram below shows the internal con-
struction of the 8201.
o
OUTPUT
1
OUTPUT
o
INPUT
1
INPUT
The 8201 has nine built-in inverters for accomplish-
ing such operations as set, clear, jam-transfer, shift,
and complement without the need for additional
gating. The 8201 can also be used in counters.
Logic diagrams for these operations are shown
under "8 Series Logic Configurations."
116
INPUT: All built-in inverters are similar to 8104
and 8105 circuits. Each conditional input is equiva-
lent to a 15-ma clamped load. Counting internal
and external inverter collectors as 8 pf each, not
more than 40 pf capacitive loading is allowable at
either input T or input U. Each input gate must
consist either of one inverter, or of two series-
connected inverters with the top inverter pulsed (see
"Inverter Usage", page 99). If a level or pulse wider
than 40 nsec is used to set or clear the flip-flop,
50 nsec must be allowed between the trailing edge
of that signal and the leading edge of the next input
that changes the state of the flip-flop. The 8201 is
complemented by simultaneously bringing both in-
puts to ground with pulse inverters (40-nsec pulses
only). Since the inputs are conditioned by the state
of the flip-flop, the total load is equivalent to one
15-ma clamped load.
OUTPUT: 0 and 1 Outputs - Each.output can supply
16 rna at ground or 12 rna at -3 v. When the flip-flop
input rate exceeds 3 mc, dynamic loading must be
considered. The table below shows the maximum
number of inverter bases that can be driven at
selected frequencies, both with and without an
extra 10 rna clamped load that increases drive at
-3 volts while reducing to 6 rna the drive available
at ground (See "Special Instructions for 8 series
Logic Design" for more information about fall times
and dynamic loading.)
B201
Maximum
Inverter
Flip-Flop
Base Inputs
Input
With Short
Frequency
Leads
10 me 3
8 me 5
6 me 7
5 me 9
3 me and 12
below
Inverter
Bases With
Short Leads
and 10 ma
Clamped Load
4
7
10
13
18
117
Indicator Output - A separate output at pin F
drives an indicator-with-arnplifier (4910) through
Connector Board W020 without loading the flip-flop
output excessivelY with stray wiring capacitance_
When in use, this output reduces the 1 output drive
at -3v by 1 rna when used with W020, or by 2 rna
when used to drive a 4910 directly_
POWER: +1O(A)/5 rna; -15(B)/63 rna.
B201 - $56.00
I
QUADRUPLE FLIP-FLOP
TYPE 8204
C M V
4J
8204 QUADRUPLE FLlpFLOP
Module 8204 contains four bits of unbuffered flip-
flop memory. Each flip-flop comprises two 8105-
type inverters, two lO-ma clamped loads, a common
clear input, and an indicator driver resistor.
INPUT: Each flip-flop may be individually set by
grounding the 0 output and may be cleared by
grounding the 1 output. The collector of an inverter,
whose emitter is tied directly to ground, may be
used to ground a flip-flop output. Diode gates such
as 8113, 8115, 8117, etc., may also be used, but
due to their slower operation they must be condi-
tioned "on" for at least 70 nsec to provide drive
equivalent to an inverter driven by a 40-nsec pulse.
A negative pulse at least 40 nsec wide applied to the
input of the pulse inverter will clear all four flip-
flops. Clear input loading is equivalent to one in-
verter base input.
OUTPUT: Each flip-flop output can drive 6 rna at
118
-3 v. Flip-flops driven by inverters with 40-nsec
pulse inputs can have up to 200 pf of total dynamic
loading at each output, counting internal loading of
68 pf, and up to 2 rna of static load at -3 v. When
the input conditions are present for longer than
40 nsec, output loading may be increased until the
estimated fall time is I1f2 times as long as the
grounded output is held at ground externally. (See.
"Fall Time" under "SpeCia'l Instructions for 10 mc
Logic Design"). No gated-emitter inverters may be
driven. '
The 1500-ohm resistor outputs allow a control panel
indicator with amplifier (4910) to be driven through
Indicator Connector W020 without excessive loading
of the flipflop by the capacitance of the intercon
necting wire. When used, this output may be counted
as 1 rna of static base load on the appropriate 1
output, if the wire to the W020 is short.
POWER: +10(A)/0 rna; -15(8)/94 rna.
8204 - $29.00
DELAY (ONE SHOT)
TYPE 8301
DELAY INPUT -f\, ______ -+I ____ _
(PIN RJ
LEVEL OUTPUT

PULSE OUTPUT : U
I I
- 0 NEGATIVE PULSE
OUTPUT
INVERTER
COLLECTOR
+ E POSITIVE PULSE
OUTPUT
INVERTER T
BASE
INVERTER
N EMITTER
CAPACITOR
TERMINALS
r- IL IV
EXTERNAL ...l
CAPACITOR1-=- =--e
'
2 ,
I K
L_
EXTERNAL
r--t'oc: RHEOSTAT
p _____ ,
INTERNAL 5 I "I 5000n
RHEOSTAT 1-------_J. __ .J
5000n
-'V
8301 DELAY (ONE SHOT)
A delay (one shot) is a monostable multivibrator.
When the input terminal is grounded, either through
the inverter or externally, the level output switches
from its normal ground level to -3 v for a pre-
determined, but adjustable, period of time, then it
switches back to ground. Simultaneously with the
final transition, a standard 40-nsec pulse is gen-
erated at the pulse output.
The 8301 contains three capacitors for delay range
selection, and a screwdriver-adjustable rheostat for
fine control. Typical level output duration ranges are
60 to 700 nsec, 0.5 to 10 pSec, and 7 to 150 f'sec
using pins J, L, and V respectively. To increase the
range further, connect an external capacitor be-
tween pins J and K. When pins U and Pare jump-
ered together, fine adjustments are made with the in-
ternal control. For external control, a rheostat of
about 5000 ohms can be connected between pins
Sand P.
The circuit recovery time using a given timing ca-
pacitor is approximately 10% of the maximum delay
available with that capacitor. This limits the maxi-
mum input frequency to about 6.5 mc.
119
A 20% change in power supply voltage will chance
the delay typically 1 %. Delay jitter (due to power
supply ripple) is less than 0.3%.
EXTERNAL CONTROL: The use of timing resistances
larger than 5000 ohms is not recommended. A
5000-ohm rheostat will give approximately 20:1
variation with any but the smallest timing capacitors.
In choosing external timing capacitors, allow about
1 nf for every 3 f'sec of delay desired at 5000 ohms.
Noise on remote control wires tends to synchronize
the end of the delay period (or it could cause false
triggering in extreme cases), consequently, the con-
trol wires should be kept short. Even for 1-ft control
wires, a grounded shield may be advisable if smooth
control and freedom from jitter are essential.
INPUT: Pin R - The delay begins when this point is
brought to ground by a pulsed inverter. Either the
internal inverter or external inverters may be used.
This input is equivalent to a lO-ma clamped load.
Pins T and N - These are the base and emitter
terminals of a standard inverter. See descriptiOn of
8104.
I
OUTPUT: Level- When the input is pulsed, a nega-
tive standard level occurs for the duration of the
delay interval. The output supplies 12 ma at ground
in addition to the 20-ma internal clamped load. The
clamped load supplies 14 ma at -3 v. Dynamic load
at the output is 8 pf. Pulse - At the end of the delay
interval, a DEC standard 40-nsec pulse occurs. The
120
negative output will be active if the positive output
terminal is grounded, and vice versa. This signal
can drive up to eight inverter bases and an appro-
priate terminating resistor.
POWER: +10 v(A)/2 ma; -15 v(B)/110 mao
B301 - $73.00
DELAY
TYPE 8310
Double Height Board
AL
AU
{
T
DELAY AS
""""0= ..
PULSE __
INPUT
BL
{
K
DELAY 8J
'""lJ:: ..

INPUT
BU
{
T
DELAY BS
,"rumQ:: ..
PULSE __
INPUT
8310 DELAY
The 8310 contains fou( delay lines, each producing
maximum delay of 50 nsec in 12.5 nsec steps. The
output of each line is connected to a transistor in-
verter whose emitter is grounded. The collector
terminal is available for logical gating. The 15 nsec
delay through the inverter must be added to the
delay of the line.
INPUTS: 40 nsec negative pulse or equivalent. One
unit of load.
121
OUTPUTS: Collector should go to P.A. input of
8602 or other unit being pulsed. However, up to
two inverters may be placed in series between col-
lector and P.A. input for additional logic gating.
POWER: None required.
8310 - $66.00
I
DELAY WITH PULSE AMPLIFIER
TYPE 8360

DELAY 'NPUT G----,;,-----+-----
-- -- UPULS:EVELI Iln--
UVEL
-",
1'--------' I \
DELAY OUTPUT PULSE \
I
PA OUTPUT --+-------h
POSITIVE
OUTPUT
L. ___ -'. NEGATIVE
PULSE OlJTPUT

DELAY
INPUT -=

p:
--r-- 20-250 NSEC
-= DELAY LINE
8360 DELAY WITH PULSE AMPLIFIER
The 8360 contains a delay line which may be varied
from 25 nsec to 250 nsec, and a standardizing pulse
amplifier similar to one half of a 8602. The length
of the delay is adjusted by means of a slotted screw
accessible from the handle-end of the module. The
high resolution of the delay line (approximately lf4
nsec) makes it ideal for high-speed timing chains.
8y connecting the delay and pulse amplifier together
with suitable logic in a feedback loop, a stable gate-
able clock may be obtained (see Application Sec-
tion).
INPUT: Delay - Pulse - The pulse input to the
delay is a standard 40-nsec 2.5-v negative pulse.
Loading is equivalent to four inverter bases. Level
- The static input load is 12.50 ma at -3 v. Dy-
namic load is negligible. The input may be driven
by three W005 Clamped Loads in parallel, or by one
W005 Load in parallel with a standard 10-ma
clamped load:
Pulse Amplifier - The input of the pulse amplifier is
equivalent to a 10-ma clamped load, and is driven
by an inverter. Pulse - At the base of the inverter
the pulse input which drives the pulse amplifier, is
normally a standard 4D-nsec negative pulse. How-
ever, any negative pulse having an amplitude be-
tween 2 and 5 v, a rate of change greater than 1 v/12
nsec and width (at -2 v) greater than 25 nsec, can
be used instead of the 40-nsec pulse. Inverter input
122
pulses of less than -0.5 v will not generate an out-
put pulse. Level- The input to the pulse amplifier
may also be a standard positive-going level change
(-3 v to ground). A negative-going level change
will not produce an output signal from the amplifier.
The input must have been at -3 v for at least 50
nsec before going to ground.
OUTPUT: Delay - Pulse - When the input of the
delay line is a 40-nsec pulse, the maximum output
driving capacity at ground is 16 ma as with a stand-
ard inverter driven by a pulse. Level- When the
input to the delay' is at level, the maxmium output
driving capability at ground is 36 ma as with the
output transistors of a buffered flip-flop. Each 10-ma
clamped load attached to the output (collector) pro-
vides a maximum output driving capacity of 7 ma at
-3 v. The dynamic load of the collector is 8 pf.
Pulse Amplifier - The output of the pulse amplifier
(for either a negative pulse or a negative-going level
change at the base of the inverter driving the ampli-
fier) is a standard 2.5-v 40-nsec pulse which occurs
at the output terminal every time the input signal
meets the input requirements. The negative output
will be activated if the positive output terminal is
grounded. Each output drives twelve 10-mc bases
and appropriate terminating resistors.
POWER: +10 v(A)/5 ma; -15 v(B)/50 mao
8360 - $84.00
-CLOCKS IIBl
L-_____ TY_P_E_S_B_4_0_1,_B_4_0_5 _____ --'

I

'RHEOSTAT

o:t ..

RHEOSTAT
20ICSl
B401 VARIABLE CLOCK
The B401 Variable Clock produces standard pulses
from a stable, RC-coupled oscillator with a wide
range of frequencies. The variable clock is often
used as a primary source of timing for large systems.
Where very precise timing is needed, the B405
Crystal Clock, which contains a single-frequency
crystal oscillator, may be used.
B401 FREQUENCY RANGE
PIN Capacitance Approx. Range
J 100 pf 1 mc-lO mc
K 1000 pf 100 kc-1 mc
L .01 mfd 10 kc-lOO kc
EXTERNAL 0.1 mfd 1 kc-lO kc
EXTERNAL 1 mfd 100 cps-1
The frequency of the B401 is variable from 10 kc to
10 mc. Three capacitors determine the frequency
NEGATIVE
_0 PULSE
OUTPUT
+ E POSITIVE
PULSE
OUTPUT
NEGATIVE
PULSE

L. ___ -' POSITIVE
PULSE
OUTPUT
B405 CRYSTAL CLOCK
range, and a potentiometer provides fine control.
For lower frequencies, an external capacitor may be
used. When terminals U and P are connected to
. gether, the internal rheostat provides fine control.
123
If desired, an external rheostat can be connected
between terminals P and C. A 20% change in power
supply voltage will change the frequency less than
1 %. Pulse-to-pulse jitter is less than 0.3%.
The B405 contains a series resonant crystal oscil-
lator circuit and a pulse-shaping buffer amplifier
which produces standard 40-nsec pulses. The fre-
quency, specified by the customer, can be between
2 and 10 mc. The frequency is stamped on the
crystal. Stability is 0.01 % over the temperature
range of -20 to +55C.
OUTPUT (EITHER CLOCK): Standard 40-nsec pulses
at the preselected frequency. The negative output
is active if the positive output terminal is grounded;
the positive output is active if the negative output
terminal is grounded. Each output can drive eight
10-mc inverter bases and an appropriate terminat-
ing resistor.
POWER: 8401: + 10 v(A)fO ma; -15 v(B)170 mao
8405: +10 v(A)/51 ma; -15 v(B)/25 mao
B401 - $57.00
B405 - $100.00

PULSE AMPLIFIER
TYPE 8602
NEGATIVE

PULSE
OUTPUT
PA
POSITIVE
PULSE
OUTPUT

PA

8602 PULSE AMPLIFIER
The 8602 contains two pulse amplifiers which are
used for power amplification, for standardizing
pulses in amplitude and width, and for transforming
a level change to a pulse. Delay from the input of an
inverter that drives the PA to the PA output is ap-
proximately 20 nsec. Input pulses may occur at any
frequency up to 10 mc.
INPUT: The input of the pulse amplifier is equivalent
to a lO-ma clamped load and may be driven from a
transistor collector. Pulse - The pulse input to the
,base of the inverter whose collector drives the pulse
amplifier is normally a standard 40-nsec negative
pulse. However, any negative pulse having an
amplitude between 2 and 5 v, a rate of change
greater than 1 v/12 nsec and width (at -2 v) greater
than 25 nsec, can be used. Input noise of 0.5 v or
less will not generate a PA output pulse. Several
pulse gate collectors can be connected together to
mix pulse sources. The inverter emitter must be at
124
ground for assertion at the output. Level- The input
to the pulse amplifier may also be a standard
positive-going level change (-3 v to ground). A
negative-going level change will not 'produce an
output signal from the amplifier. The PA input must
rest at -3 v for at least 50 nsec before goi ng to
ground.
OUTPUT: The output of the pulse amplifier, when
either a pulse or a level brings the input to ground,
is a standard 2.5-v, AO-nsec pulse which occurs at
the output terminal every time fhe input require-
ments are met. The negative output is produced
when the positive output terminal is grounded, the
positive output is produced when the negative termi-
nal is grounded. Each output can drive twelve 10-mc
inverter bases or their equivalent and an appropriate
terminating resistor.
POWER: +10 v(A}/2 rna, -15 v(8}/75 rna.
8602 - $36.00 .
CARRY PULSE
TYPE 8620
lB
SEFilES
t.; I '/E
PULSE Q:.r;PUl
8620 CARRY PULSE AMPLIFIER
Module 8620 supplements the 8201 for 10-mc
counting applications. It supplies the circuitry to
complement two 8201 Flip-Flops and propagate
their carry pulses. One 8620 and one 8201 can also
be combined to form one bit of an up-down counter.
The 8620 contains two pairs of inverters for comple-
menting lO-mc flip-flops with conditional set and
clear inputs, and two standardizing pulse amplifiers
each capable of driving three inverter bases. The
propagation delay is approximately lO nsec. Maxi-
mum pulse repetition frequency is 5 mc for pulse
amplifiers.
INPUT: The eight inverters of the 8620 are similar
to other 10-mc inverters, for example, the inverters
125
of the 8lO4 and 8105. To obtain minimum propa-
gation delay, input H or S should be pulsed. Use
inputs E, F, P, and R for gating inputs. Pulse am-
plifiers will not produce standard pulse outputs
unless a standard 40-nsec pulse input is used.
OUTPUT: When 40-nsec pulses are used at their
inputs, the pulse amplifiers produce 2.5-v, 40-nsec
standard negative pulses capable of driving one to
three inverter bases. Maximum length of wire used
to connect pulse amplifiers to the inverters is 6 in.
Noise pulses may occur if the ground pin next to
each output pin is not connected directly to the local
ground of the inverters that the output drives.
POWER: +10 v(A)/6 ma; -15 v(8)/20 mao
8620 - $47.00
POWER INVERTER
TYPE 8681
20MA 20 MA
:r :=r-
B681 POWER INVERTER
The 8681 Power Inverter contains four high-
current inverters, each _ with separate emitter con-
nections. A 20-ma clamped load is permanently con-
nected to each collector. Four additional 10-ma
. clamped loads are supplied. Input and output cur-
rent ratings are double-those of a standard inverter.
INPUT: Each input requires a steady-state 2-ma
current at -3 v. Each independent clamped load
requires 10 rna from a transistor collector to hold it
at ground.
126
OUTPUT: Each power inverter transistor can supply
32 rna at ground. Each internal clamped load absorbs
20 rna, and 12 rna remain for an external emitter
(level or pulse), or loads to -15 such as odd-
numbered diode gates, etc. Each of the clamped
loads that are _ internally connected to power in-
verter collectors can supply 14 rna at -3 v. Each of
the additional clamped loads can supply 7 rna at
-3v. -
POWER: +10 v(A)/O rna; -15 v/130 rna.
8681 - $25.00
BUS DRIVER IIBl
~ ______ T_Y_PE_B_6_8_4 ______ ---l ~
~ c
B684 BUS DRIVER
The 8684 contains two dual-purpose, non-inverting
bus drivers and a -3 v supply. Each bus driver pro-
vides standard levels either to a large number of
inverter base and diode loads, or to a terminated
90-ohm cable. All logic terminals are available at
the connector. Delay through a bus driver is approxi-
mately 30 nsec.
INPUT: Standard levels or equivalent; 1 ma of load
when the input is at -3 v. Input capacitance is 8 pf.
The input circuit will tend to oscillate if it is driven
by a wire more than 3 feet long.
OUTPUT: Direct - Standard levels capable of driv-
ing 40 ma at 10 mc. Each bus driver level control
pin must be connected to -3 v when using the
direct output. The separate ground terminal pro-
vided for each circuit should be the ground point
for loads driven by that circuit. This output is the
one to use for driving heavy loads over short dis-
tances.
127
Resistor - Ground and -6 v levels at the resistor
output provide standard levels at the end of a 93
ohm cable when it is terminated with a 100-ohm
resistor to ground. The terminated cable will drive
10 ma at 10 mc. The level control pins must be
left open when driving a terminated cable from the
resistor output. The separate ground terminal pro-
vided for each circuit should be the ground return
point for loads driven by that circuit.
For driving 5 ma loads or less, an unterminated
cable or open wire may be driven from the resistor
output using 3v levels. This connection allows
heavy local loads to be combined with light distant
loads on one circuit.
POWER UNLOADED: +10 v(Al/80 ma; -15 v(8)!
120 mao
POWER FOR LOAD: Current to bring loads to ground
must be added to the total demanded from + 10(A);
current to bring loads negative must be added to the
total current from -15(8).
8684 - $52.00
128
W
SERIES
129
W SERIES
INTRODUCTION
The W series provides inputoutput compatibility between FLIP CHIP Modules and other
digital devices. The range of W series inputs and outputs is wide. Inputs up to 48v,
outputs up to 135v or up to 10 amp, inputs from source impedances as high as 100 KQ,
floating systems, bouncing contacts, long transmission lines, all can be handled easily
by one or more of these modules. A few of their uses are listed below.
Clamped Loads
The inputs of all R series circuits draw current only when the input voltage is at ground.
Other DEC circuits may require input current to be supplied at the negative voltage level.
The use of additional clamped load resistors allows an R series module to be preloaded
so that it is able to supply higher currents at the -3v level.
It is also possible to obtain higher drive capability at ground from DEC circuits which
normally supply current at -3v. In this case the normal clamped Idad resistor would be
replaced by one which is alighter load at ground. The Type W002 contains clamped load
resistors of this type. Care should be exercised when such a lighter load is used since an
increased fall time will result
A clamped load can also be used as a shortcircuit proof, -3v supply for toggle switches
or patchboards.
Cable Connectors
Cable connectors provide a convenient means for going from one section of logic to
another. They may also be used to hold series isolation resistors or shunt termination
resistors. They allow large systems to be broken into pluggable subsystems for efficiency
in checkout and field maintenance.
Drivers
The Types W050 and W051 provide signals for indicator lights, relays, and solenoids.
W051 outputs go to ground and can therefore be used to drive RIll or similar diode
inputs. W061 can drive NIXIE tubes, or other loads returned to positive voltages. W061
outputs supply sufficient current for most projection type numerical displays as well, and
for relays and solenoids too large for W051 circuits. W040 is the next larger driver, and
has sufficient output capability for most electromechanical devices. Stepping motors and
other exceptionally high current devices may require the use of a W042. Line voltage can
be controlled by the W080 module, and if a Triac is added to its output this module allows
up to 1 kilowatt of AC power to be solidstate controlled by FLIP CHIP logic.
Input Converters
The W501 allows external low frequency or noisy signals to be converted to DEC standard
levels. The Type W510 converts positive signals of a few volts amplitude into DEC levels.
Other external levels that are noisefree and have a swing greater than 3v can usually be
biased so that they can be brought directly into the logic modules. Pulses, which are noise
130
free and fast, can usually come directly into the pulse amplifiers. Small signals, with
amplitudes of a few millivolts or more, can be brought into the logic system using W520
or A50I comparators. Floatingbattery signals such as those in process controls can be
converted by W502 modules. The W500 buffers inputs from photocells, vacuumtube
equipment, and other high impedance sources. Some types of computer can
use the W590 module. W700 filters bounce from contact signals not requiring W50I
standardization.
Output Converters
The W600 and W601' provide a means of driving low power digital devices requiring
signals in the range of-15 to : 20v. Most drivers can be used as output converters
under some circumstances. Some types of computer interface can use the W690 module.
Many types of output pulse requirements can be met by W607 or W640 pulses.
Commu11ications Interface Modules
The W706, W707, and W708 are functional modules which are used to provide Teletype
communications system interfaces to computers and other parallel devices. These units
contain serial to parallel and parallel to serial conversion for 5-bit and 8-bit communication
systems.
Accessory Modules
The accessory modules include blank boards for constructing special circuits, a module
extender, which ,allows access to the module while it is connected to the mounting panel,
and a system module adapter, which allows DEC system modules to be connected into a
FLIP CHIP mounting panel. Related are the A series amplifier boards A990 and A992, for
mounting operational amplifiers of various types. .
Pulse Conversion'
See the application note with the above heading for a table of pulse conversions between
FLIP CHIP and earlier System Modules and Laboratory Modules made by DEC_ All modules
use logic levels of ground and -3v, so in many cases no conversion at all is required.
131
CLAMPED LOADS
TYPES W002, W005
2 MA

W002' CLAMPED LOAD
The W002 contains 15, 2-ma clamped loads. These
can be used for clamping voltages at the output of
inverter collectors in R-series modules, or for con-
verting B-series modules to work with R-series,
OUTPUT: When the clamped load output is
grounded, it draws 2 rna. When it is at -3v, it can
supply up to 1.4 rna.
POWER: -15 v(B)/46 rna.
132
5 MA

vo-AIVv-e
WOOS CLAMPED LOAD
The W005 contains 15, 5-ma clamped loads, These
can be used for clamping voltages at the output of
inverter collectors in B-series modules, or for con-
verting R-series modules to work with B-series, Two
of these clamped loads in parallel are equivalent to
one B-series clamped load,
OUTPUT: Whe'n the clamped load terminal is
grounded, it draws 5 rna. When at -3v, it can supply
up to 3,5 rna,
POWER: -15 v(B)/91 rna,
W002 - $13.00
W005 - $15.00
Type
W018
W023
CABLE CONNECTORS
FOR INDICATOR AMPLIFIERS
TYPES W018, W023
The W018 and W023 provide 18 line ribbon cable connections to FLIP CHIP
mounting panels. In the W018 connection to each pin is through a series low
leakage silicon diode. The W023 provides unbroken signal lines from the cable
to the connector pin.
When these cables are used with 4917 or 4918 indicators, the W018 must be
located at the FLIP CHIP panel and the W023 inserted in the indicator socket
connector. Cables may be ordered with connector modules on both ends or on
one end only. Cable length may be specified in increments of 1 inch.
For ordering information, see W021, W022, and W028 on next page.
Care should be taken when using the W023 for other purposes, since the
Power Pins (A, B) are unprotected.
Price with Cable Attached Type Price without Cable
$19.00* W018U $18.00
$13.00* W023U $ 4.00
*Ribban Cable: Add $.60 per foot, or fraction
133
CABLE CONNECTORS FOR LEVELS
AND PULSES
TYPES W021, W022, W028
The W021, W022, and W028 provide cable connections to the FLIP CHIP
mounting panel. The cable is a 19-conductor ribbon with nine signal leads and
ten shields. The signal leads are connected to pins D, E, H, K, M, P, S, T and V.
The shields are internally connected together and to fJins C, F, J, L, N, R, and U.
In the W021, the signal leads are connected directly to the signal pins. In the
W028, jumpers are available for series or shunt terminators. The Type W022
has a lOO-ohm ~ h u n t terminator from each signal wire to the shield.
Connectors may, be ordered in like or unlike pairs. They may also be ordered
separately, in which case the other end of the cable is left free. Cable length
may be specified in increments of 1 in.
134
ORDERING INFORMATION: To insure clear communication, use the format below.
Type
W021R
W022R
W028R
W02X XXX
7t/1
connector
Iype of
inches R ~ c ~ ~ ~ e B O N
of cable C ~ CO.AXIAL
(Zo ~ 93")
W02X
-.--
second
connector
(if any)
EXAMPLE: W02824C Iwo feel of coaxial cable attached, no second connector
Ribbon Cable Connectors Coaxial Cable Connectorst
Cable No Cable No
Attached Type Cable Type Attached Type Cable
$13.00* W021RU $4.00 W021C $31.00** W021CU $ 4.50
$13.50 * W022RU $4.50 W022C $33.00** W022CU $6.50
$13.00* W028RU $4.00 W028C $31.00** W028CU $4.50
*Ribbon Cable: Add $.60/ foot. **Coaxial Cable: Add $1.50/foot.
135
I
SOLENOID DRIVERS
TYPES W040, W043
DIODE INPUTS
o
TRUTH TABLE
f--__ --oR
INPUTS OUTPUTS
NODE TERMINAL
Vo---------------------t
EXTERNAL NEGATIVE VOLTAGE
u
0
IJI
Ov
Ov
-3v
-3v
E R
I KI lSI
Ov OFF (NEG)
-3v OFF I NEGI
Ov OFF INEGI
-3v ON 1-2VI
Lf-1
W043 ONLY
(SEE "GROUNDING")
W040. W043 SOLENOID DRIVER
These high currentdriverscan drive relays, solenoids,
stepping motor windings, or other similar loads. The
output levels are -2 volts and a more negative voltage
determined by an external power supply. One term
inal of the load device should be connected to the
external power source, the other to the driver output.
There are two drivers per module and both modules
usethe same pin connections.
Pin V of the driver module must be connected to the
external supply so that the drivers will be protected
from the back voltage generated by inductive loads.
If the wire to the power supply is more than 3 feet
long it may have to be by passed at the module with
an electrolytic capacitor to reduce the short over
shoot caused by the inductance of the wire. If pin V
is connected to the supply through a resistor, the
recovery time of inductive loads can be decreased
at a sacrifice in maximum drive voltage capability.
Maximum rated supply voltage (see below) less
actual supply voltage should be divided by load
current to find the maximum safe resistance. When
both circuits on a module are used, the load current
for the above calculation is the sum of the currents.
INPUTS: Standard DEC levels or equivalent. The
maximum input load is 3 rna. per driver, shared by
all grounded inputs. Additional diode inputs may be
added by connecting diode networks such as ROOI
or R002 to the node terminal.
OUTPUTS: The table below shows maximum ratings
for individual circuits. No more than two circuits
should be paralleled to drive loads beyond the cur
rent capabilities of 'single circuits. For larger loads
use the W042. When both circuits on a W043 are
used with a duty cycle exceeding 35%, use the
136
current derating curve shown below.
GROUNDING: The W043 has three ground terminals
C, T, U, which should be wired together externally to
limit current through each connector pin to safe levels.
(High current loads should be grounded at the W040
or W043 modules to avoid noise due to high pulse
currents in ground conductors.)
MODULE
W040
W043
MAXIMUM
VOLTAGE
1.5
1.0
0.5
-70V
-35V
CIRCUIT A
AMPS
MAXIMUM
CURRENT
0.6 amp.
2.0 amp.
CIRCUIT B
AMPS
0.5 1.0 1.5
W043 DERATING CURVE
TYPICAL
DELAY
5.usee
10 "sec
POWER: W040: -t 10v (A) I 0 MA, -15v (8)/24 MA.
The external voltage supply must supply the output
current of the two drivers (1.2 amp max.)
W043: -+ lOv (a)/0.25 MA, -15v (8) 16 MA. The
external voltage supply must supply the output cur
rent (2.0 amp max.)
W040 -- $36.00
W043 -- $35.00
10 AMP DRIVER
TYPE W042
(DOUBLE-HEIGHT, DOUBLE-WIDTH MODULE)
r----------<>ES
r-l>l-,--<> ~ ~ ~ ~ ~ ~ [ ~ ~ P l
f-.... --+----<> OUTPUT HI
AD
AE
AF OUTPUT #2
SIX
AH TABS
AJ
AK OUTPUT # 3
AL
AM
AN
OUTPUT #4
(BOTTOM)
AP
AR
AS
W042 10 AMP DRIVER
This module has four germanium transistor
drivers each capable of providing up to ten am
peres of DC drive at ambients up to 40'C for
heavy loads such as paper tape punches, card
punches, hydraulic servo valves, or hightorq ue
stepping motors like Responsyn (T.M. United
Shoe) or SloSyn (T.M. Superior Electric). In 55'C
ambients, up to 8 amps total current may be
obtained. AMP "Faston" tabs at the handle end
of the module provide high current connections
for ground, ES ,and the four outputs. Loads are
to be connected between the outputs and ex
ternal ground. Due to the fact that this module
may dissipate as much as 20 watts when operated
at rated output, special consideration should be
given to an unobstructed flow of cooling air. It is
recommended that no modules be mounted
directly above any W042 operating at more than
4 amps average current. Typical delay: 20 micro
seconds for the circuit alone. Load current decay
time may be much longer, if its inductance is large.
INPUTS: Each input requires 2 ma at ground.
Negative input brings corresponding output to
ES . Input gates may be expanded with ROOI or
R002 diode modules, if no more than 6" of wire is
connected to each node. All connections are made
to the A half (upper) of the module.
137
OUTPUTS: Total DC current from the W042 may
be up to 10 amp, and may be distributed at will
among the four outputs. Even higher currents
may be obtained briefly by taking into account the
4 minute (approx.) time constant of the heat sink.
For example, four 10 amp solenoids can be actio
vated together, as long as they are on only a few
seconds and at low duty factors. Outputs are not
short circuit protected if shorted to ground. Shorts
to output supply voltage are harmless. Clamp
diodes are provided from each output to ground
to damp transients when turning off inductive
loads. Dampingdiode ground and all ES power
is connected only by tab terminals. Logic ground
and + lOv power use standard connector pins.
Output circuit power supply must be grounded to
digital system power externally. Power jumpers
Type 914 may be used to make connections. Out
puts may not be paralleled to increase short duty
peak current, unless O.ln currentsharing reo
sistors are connected in series with each output.
The negative supply voltage (ES) must be between
-12 and -15 volts.
POWER: + 10v(A)/180ma; ES/270ma plus output
current.
W042 - $80.00
I
30 MA INDICATOR DRIVER
TYPE W050
INPUT OUTPUT
E F
"4
~ J
K4 ~ L
M4 ~ N
P4 ~ R
'4
~ T
"4
~ v
W050 INDICATOR DRIVER
The W050 contains seven transistor amplifiers that
can drive miniature incandescent bulbs, such as
those on an indicator panel. It is used to provide
remote indicators for R- or B-Series flip-flops. If the
input is at -3v, the output is at -2v.
INPUT: Standard levels of -3v and ground. Each
input represents 1 rna of load at ground.
OUTPUT: The output is capable of supplying 30 rna
138
at -2v. The external load may be connected to
any voltage between -2 v and -20 v. The output
is capable of driving an indicator light, such as Drake
11-504, Dialco 39-28-375, or Eldema CF2-WT-1762,
returned to -15 v.
POWER: +10 v(A)/1.1 rna; -15 v(B)f7 rna.
NOTE: An additional 210 rna is drawn from the in-
dicator supply when all remote lights are on. This
power is not drawn from the module.
W050 - $13.00
100 MA INDICATOR AND RELAY DRIVER
TYPE W051
INPUT
E4
DRIVER
'------'
H F
'4
"4
P4
54

NODE
WOSI DRIVER
The' W051 contains seven inverter amplifiers suit-
able for driving indicators, relays, and other medium
power devices. The amplifiers can supply up to 100
ma at ground, and each output is diode clamped
to -15v to prevent overvoltage when the current is
interrupted in an inductive load. If the input is at
-3v, the output is at ground.
INPUT: Diode - Standard levels of -3v and ground.
The input load is 3 ma at ground. Node - Other in-
puts may be connected here through diodes such as
those in the R002. The 3-ma load is shared among
the grounded inputs.
139
OUTPUT: The output is capable of supplying 100 ma
at ground. The external load may be connected to
any voltage between 0 and -15v. The negative out-
put is diode clamped to prevent it from going more
negative than -15v. Typical delay for circuit alone:
1 microsecond. Decay time of current in inductive
loads may be much longer.
POWER: +10 v(A)/3 -15 v(B)/23 mao
NOTE: An additional 700 ma will be drawn from
-15v when all circuits are on. This power is not
drawn from the module.
W051 - $22.00
I
RELAY DRIVER
TYPE W061
W
SERIES
TYPE WOGl RELAY DRIVER
The W06I Relay Driver has four all-silicon 250 ma
drivers with gateable inputs; it can drive relays and
solenoids with positive voltage supplies up to 55v.
INPUTS: 2 ma at ground, no load at -3v. Use diodes
for ungated input. Multiple-input AND gate may be
obtained by connecting ROOI or R002 diodes to
node inputs.
OUTPUTS: The loads are to be connected be-
tween the outputs and an external positive supply.
Each output can supply a quarter-amp load at
140
ground when the input(s) is (are) negative. Pin V
must be connected to the positive voltage supply
so inductive loads will not cause output transistor
breakdown. Supply voltage should be between 2
and 55v positive. Typical delay for circuit alone:
1 ~ s e c . Decay time of current in inductive loads
may be much longer.
POWER: + 10v(A)/70 ma; -15v/8 mao
OTHER POWER: Inductive loads cause clamp cur-
rents at pin V tending to drive supply voltage more
positive when drivers are turned off.
W061 - $35.00
I
L
ISOLATED AC-DC SWITCH
TYPE W080
WOBO ISOLATEO ACDC SWITCH
This module contains two photoncoupled transistor
switches with bridge rectifiers. Both turnon and
turnoff are slow enough to minimize output noise.
Output tabs are at handle end of module for maxi
mum isolation. Drives relays, solenoids, panel lamps,
small motors directly. Larger AC loads can be driven
by the use of SCR or Triac* buffers. For example,
one SC45B Triac with a woao circuit tied from gate
to anode 2 and a lOOn. resistor from gate to anode 1
can switch AC loads up to one ki lowatt.
INPUT: Each input is a 30 rna load returned to
+ lOv. A W061 driver is suitable. Switching rate
must not exceed one hertz per second. Grounding an
input turns on the switch, and an open circuit at
the input turns off the switch.
OUTPUT: Each circuit can switch up to 1/4 ampere
from supplies up to 135 volts DC or AC (RMS) into
resistive or inductive loads, or 30 va maximum at
120 volts. Can drive up to 40 va intermittently; up
to 5 seconds on 50% duty factor. Derate by half for
driving incandescent lamps. Typical "on" voltage
drop: a volts. Typical switching time: 1110 second.
*G.E. trademark
141
Not designed for series or parallel operation.
WIRING: Three AMP "Faston" tabs replace module
handle. Type 914 Power Jumpers can be used to
c o ~ n e c t these to a nearby terminal block, etc. Use
caution on high voltage.
NOISE: woao is designed to generate little or no
switching noise. However, power lines often carry
noise from distant sources. Some types of loads
generate noise, such as bush-type motors and power
relays. Even SCR and Triac circuits generate fast
transients on each turnon cycle. For these reasons,
it is important to locate woao modules and their
output leads away from logic and logic wiring. If a
woao module must be used close to a logic module,
put a W992 or W993 copper clad board with pins A
and B cut away between them to form an electro-
static shield. In some cases line filters may also be
necessary.
POWER: + 10v(A)/60 ma; -15v(B)/0.
woao - $60.00

PDP-8 DEVICE SELECTOR
TYPE W103

BF
BH -----o>CI------1
BJ -----Oo{)f----1
BL -----Oo{)f----1
BM -----0'1>1----1
BN
BP -----o>CI------1
BR
BS
B T -----o>CI------1
BU -----oo{)f----1
BV -----o>CI------'
5m'
AF

5m,
AM
1---1>Qu----i> AL
5m,
AT
1----<>CiJ--<> AS
WI03 PDPS DEVICE SELECTOR
This module is used to decode the six device selector
bits transmitted in complement pairs on the PDp8
or PDP8/S 1/0 bus, and it provides standard pulses
to the selected device. The device code is selected
by cutting one diode of each pair, BE or BF, etc., off
the board. Device coding can also be accomplished by
selective wiring of the bus inputs to the diode pairs.
INPUTS: Diodes - Standard levels of -3v and
ground. Input load is 1 ma shared among the
inputs that are at ground. Pulse Input - Standard
100 nsec negative pulse or any pulse at least
142
100 nsec wide with an amplitude of 2.5 volts.
OUTPUTS: With terminals AH and AJ or the
corresponding terminals on other sections con
nected together, the output is a standard 400
nsec pulse. With these terminals open the output
is a standard 100 nsec pUlse. Both positive
(-3v to ground) and negative pulses are available.
Each positive output can drive 65 ma of external
load at ground and each negative output 15 mao
POWER: +10 v(A)/6.4 ma; -15 v/57 mao
W103 -- $52.00
DECODING DRIVER
TYPE W108
(DOUBLE HEIGHT)
8P BR BS 8T
TAB TERMINAL -1/ (TAB)
-7.5 TO -12 VOLTS
WIOB DECODING DRIVER
This driver provides up to 300 ma bipolar drive
currents for use in memory systems including core
memories, such as the H201, and magnetic tape
systems. There are eight drivers on a module, each
of which can be selected either by one of eight
address lines or by using the builtin binarytooctal
decoding matrix. The drive current direction is
selected by one of two input select lines and will be
the same for all drivers on a module. Drive current
can be varied from 160 ma to 300 ma by adjusting
. the external negative voltage reference connected to
a tab terminal on the module handle.
INPUTS: Address selection diode inputs have a 2
ma at ground load shared between all inputs used
on each driver. Positive output drive can be selected
by a -3 volt signal at pin BU. When pin BU is brought
to ground, the input driving source must supply 35
, ma per addressed driver. The negative output drive
current is selected by a -15 volt signal at Pin BV.
This input requires a 25 ma drive at ground inde-
pendent of driver addressi ng.
For a complete description of drive current direction
selection see the truth table_
The negative voltage reference applied to the tab
terminal input may have a voltage range of -7.5 to
-15 volts with the current amplitude approximately
+AE
-AF
+AJ
-AK
+A.
-AN
+AR
-AS
+AU
-AV
+"
-8F
+8J
-8K
INPUT
VOLTAGE
OUTPUT CURRENT
~ ~
DRiVE DIRECTION
au BII
0 0 NO DRIVE
-3 0 NEGATIVE DRIVE
0 -15 POSITIVE DRIVE
-3 -15
POSITIVE AND
NEGATIVE DRIVE
TRUTH TA.BlE
determined by the formula
Tab Voltage
I (ma) = 0.048 + External Series Resistance in Kilohms
143
This external source must be able to sink the maxi
mum negative drive currents used in the W108. For
tab voltages more negative than -12 volts the duty
cycle of each negative driver should be such that
the average drive current per driver is less than 200
mao Momentary shorts to the negative drive outputs
will cause no module damage.
OUTPUTS: All negative current outputs sink current
as controlled by the tab voltage. Typical turn-on and
turnoff times for the negative current are, respec
tively, 100 nsec and 1.5 "sec. The positive current
driver output must be used in series with a negative
current driver, the latter ,then controls the positive
drive current. Each positive output is diode-protected
to -15v against inductive backswing.
POWER: -15(B)/16 ma plus 35 ma per selected
driver
-V(TAB)/maximum of 250 ma per negative current
driver used.
W108-$75.00
HIGH IMPEDANCE
FOLLOWER
TYPE W500

.
INPUT D
1
3000ni A
Y'C",l---< __ -oOUTPUT E
3000n
-15 VOLTS
W500 HIGH IMPEDANCE FOLLOWER
High impedance signal sources such as photo-
cells and low-current instrumentation amplifiers
can drive Schmitt Trigger W501 or logic gates
through a W500 circuit. The module contains 7
fault-protected circuits, each comprising two
cascaded emitter-follower amplifiers_ Input volt-
age excursions up to 30v or short-circuits from
output to ground are harmless_ Outputs can go as
negative as -15v with very light loading, but will
not exceed -lOv when driving a W501 input.
INPUTS: Excursions Between -0 and -3v: Input
currents of 100 Ma or less (typically 50) flow to-
ward the driving source, tending to bring it more
positive_ Low frequency equivalent input resis-
tance exceeds 10Kn even while the output voltage
is passing through the input threshold region of a
Schmitt circuit or diode gate_ Voltage offset
between input and output: less than 1/3v_
Larger Excursions: A diode shorts the active com-
ponents of the follower circuit if the input voltage
goes more positive than ground or more negative
144
than -15v, and the input equivalent circuit
changes to 3000n returned to the limiting voltage_
If the output is connected to a clamped load for
driving grounded loads such as B-series inverters,
the limiting negative voltage changes from -15
to -3v.
OUTPUTS: Excursions Between 0 and -3v: Each
circuit can drive up to 15 ma at ground. Driving
capability at -3v is 3 ma more than that of any
clamped load attached. If the output is brought to
ground by a paralleled transistor collector, not
only the internal 5 ma load and the external load
must be driven, but also the current demanded
by the input 3000n resistance returned to the
negative input voltage present. 10 mc emitters
may not be driven. Larger Excursions: If no
clamped load is attached, each output will follow its
input as far negative as its internal3000n resistor
to -15 v will drive the load. Output voltage cannot
go more positive than ground.
POWER: +10v(A)/18 ma; -15v/35 mao
W500 - $25.00
NEGATIVE INPUT CONVERTER
AND SCHMITT TRIGGER
TYPE W501
4.7Kn
Ui-
t5V
, OUTPUT RISE
\ ON CLOSING
\
I
: INTEGRATOR T INPUTS

: LOWER
\ OUTPUT RISE _ THRESHOLD
\ ON OPENING - CONTROL
t (CONNECT S TO U)
Vi
W501 SCHMITT TRIGGER
The W501 contains a Schmitt trigger circuit which
produces standard levels as a result of some outside
activity such as the closure of a switch or relay. A
ground level input produces a -3v level output,
and a negative level input produces a ground level
output. Nominal switching thresholds of - 2.2v and
-0.8v are obtained by connecting terminal L to
terminal K and terminal M to terminal N. The switch
ing thresholds can be varied over the range of
o to - 2.5v by applying external voltage levels to
terminals M and L. Terminal M controls the lower
level threshold, and terminal L controls the upper
level threshold. The module also contains an inte
grating circuit to filter contact bounce when a switch
or relay is used to generate the levels.
INPUTS: Diode - Any signal at pin R between 10v
will not cause damage to the circuit. The input im
pedance is 7500 ohms to + 10v when the input is
more negative than the lower threshold, and is an
open circuit when the input is more positive than the
upper threshold. The output will switch from -3v
to ground if the input voltage goes more negative
145
than the lower threshold after 'having been more
positive than the upperthreshold.The output switches
from ground to - 3v if the input voltage goes more
positive than the upper threshold after having been
more negative than the lower threshold. Upper and
lower thresholds must be at least lj2V apart. The 2 ma
clamped load at pin D cannot be used to bring this
input to -3v since it sinks insufficient current.
Direct: - Pin P provides a bypass of the diode con
nected at pin R. This node input can be used with
R001 diodes to form a NANDed input to the W501
as shown in Fig. 1 below. In addition, this input can
be used to obtain an integrated input when many
contacts or switches are connected as shown in Fig.
2 below. This latter scheme gives an output rise
when contacts close.
Integrating - The input to the integrating circuit is
a switch or relay contact. To obtain output rise when
contacts close connect contacts between pin Sand
U and connect pins Rand T. To obtain output rise
when contacts open, connect contacts between pin
S and ground, connect pin V to pin S, and connect
pin R to pin T.
W501 - $13.00
~
1< .J- i
R
I
I
Figure 1. NANDed Input
2mo 2mo
R
S T P
~ -
~
Figure 2. Integrated Input
146
PHOTON-COUPLED TRIGGER
TYPE W502
UPPER
TABi-__ -f----l
:!: 48 VOLT INPUT 5MA
:!: 14 VOLT INPUT N
COMMON 1 - ~ . . . . - - - o V
ISOLATED
TRIGGER
ISOLATED
TRIGGER
1--L.--oD
OUTPUT
5MA
1---'---oE
502 PHOTON COUPLED TRIGGER
This special interface module allows process con-
trol system 48 volt signals to be transmitted to a
digital control over long distances and despite float-
ing supplies and noise. Additional low voltage inputs
permit use as a pushbutton input device. Inherent
delay in the photon-coupled isolation circuit filters
out contact bounce. High voltage connections are at
the handle end of the module to optimize isolation.
INPUT: Each input requires 40 rna at "nominal input
voltage. Input voltage tolerance: 200/0. Voltage re-
versal harmless.
147
OUTPUT: Each output can drive 15 rna at ground.
Output goes to ground when input is energized.
Typical switching delay: 20 msec.
WIRING: Three AMP "Faston" tabs replace module
handle. Type 914 power jumpers may be used for
input connections. Keep input wires separated from
logic wiring to avoid interference from transients
on 48 volt wires.
POWER: + lOv(A)/7 rna; -15v(B)/44 rna.
W502 - $38.00
POSITIVE INPUT CONVERTER
TYPE W510
'E
2m.
D
OUTPUT
STANDARD
LEVELS
OV -3V
2m.
2m.
W510 POSITIVE LEVEL CONVERTER
The Type W510 Positive Level Converter contains
three circuits that convert positive levels to DEC
standard levels of ground and -3v. Each circuit
consists of a grounded-emitter inverter with a diode
string between its input and the base of the in-
verter. By shorting out sections of the diode string,
the switching threshold may be varied to either +2v,
Threshold Connections
+2v none
+lv H.& F, N & M, U & T
Ov H&E,N&L,U&S
In jumpering pins together to obtain the desired
switching point, it is very desirable to use the short-
est possible wire. Under no condition may anything
else be tied to these pins.
Maximum frequency is 2 mc. Maximum delay for
output fall is 100 nsec. Maximum delay for output
rise is 60 nsec.
INPUTS: Voltage levels must not exceed +25v or
go below -15v. For inputs more negative than the
+lv, or Ov (see the following table). When the input
is more positive than the switching threshold by lv,
the inverter is cut off and the output is at -3v.
When the input is more negative than the switching
threshold by lv, the inverter is saturated and the
output is at ground.
Output = -3 V Output = 0 v
Input"" + 3.0 v Input"" + 1.0 v
Input"" + 2.0 v
Input ""
0.0 v
Input", + 1.0 v Input"" -1.0 v
148
switching threshold by Iv or more, the input load is
equivalent" to 3900 ohms returned to +lOv. For
inputs which are more positive than the switching
threshold by + 1.5v the input leakage is 100 p'a or
less.
OUTPUTS: The output is an inverter with a 2-ma
clamped load. It can drive 18 ma at ground.
POWER: + 10 v(A)/8.0 ma; -15 v(B)/17 mao
W510-$17.00
NEGATIVE INPUT CONVERTER
TYPE W511
2 MA
E ~ F H--.--,L&.J
INPUT
D
OUTPUT
STANDARD
DEC LEVELS
OV,3V
2MA
v
N o---{>f---o P R<>---<l-1&1
TVPE W511 NEGATIVE INPUT CONVERTER
The Type W511 Negative Level Converter contains
two circuits that convert negative levels to DEC
standard levels of ground and -3v. Each circuit
consists of a grounded emitter inverter with a string
of bias diodes between its base and the input pins.
A separate input diode is also provided. 8y connect-
ing the input diode to various points on the diode
Threshold Connections
Ov F & H, P & R
-Iv F & J, P & S
-2v F&K,P&T
-3v F&L,P&U
In connecting input diodes to the bias string, use
short, direct wire. Under no conditions should any-
thing but the input diode be connected to a bias
string pin. Inputs must be connected only to pins E
and N.
INPUTS: Voltage levels must not exceed + 25v or
go below -50v. Input current required is approxi-
mately 1 ma when the input is slightly more positive
string, the switching threshold can be set at Ov,
-lv, -2v, or -3v (see the table below). When the
input is more positive than the switching threshold
by lv, the inverter is cut off and the output is at
-3v. When the input is more negative than the
switching threshold by lv, the inverter is saturated
and the output is at ground.
Output = -3v Output = Ov
Input;,; + 1.0v Input;,; -1.0v
Input ;,; O.Ov Input;,; - 2.0v
Input;,; -1.0v Input;,; -3.0v
Input;,; -2.0v Input;,; -4.0v
149
than the threshold, rising to a maximum of 4 ma
when the input is at + 25v. Input leakage is 100 ~ a
or less when the input is more negative than the
threshold.
OUTPUTS: The output is an inverter with a 2 ma
clamped load. It can drive 18 ma at ground.
POWER: + 10v(A)/3 ma; -15v(8)/24 mao
W511 - $17.00
POSITIVE LEVEL CONVERTER
TYPE W512
(Single height and width)
. tmo tmo tmo . I

+0.8.
1
v
W512 POSITIVE LEVEL CONVERTER
Positive logic systems, such as those using mono-
lithic integrated circuits, can use the W512 to make
available standard accessory modules in the Wand
A series.
Input threshold voliage to each converter is normally
1.6 volts for compatibility with DTL and TTL ievels.
This theshold can be set at 0.8 volts by grounding
pin V for RTL level conversion.
INPUTS: Input current -1 ma or less for input volt-
ages between 0_3 volts and ma for
150
inputs above the threshold. Input voltages must not
exceed 6.0 volts with pin V open, or 5.3 volts with
pin V grounded. Inputs must exceed nominal thresh-
olds by at least 0.4 volts for full switching with
minimum noise rejection.
OUTPUTS: Each output can supply up to 8 ma at
ground. Grounded inputs provide grounded outputs
and positive inputs provide negative outputs.
POWER: +10v (A)/104 ma; -15v (8)/30 mao
W512 - $25.00
COMPARATOR IIWl
L--______ T_Y_P_E_W_5_2_0 ______ ---' ~
~
3 M A
0+ I H
- 2
E ~
~
M A
I N
K
2
L ~
~
3 M A
R I U
2
S ~
-3V I _
L_S_U_PP_L_Y---.JI V
W520 COMPARATOR
This module is useful as an inexpensive comparator
for AID work, or as a generalpurpose input level
converter. The W520 contains three fourtransistor
difference amplifiers which give DEC Standard levels
at the output. The state of the output is determi ned
by the relative polarity of the input voltages.
provided. The more positive input appears as 100 K
ohms to -15v; the more negative input supplies a
maximum of 0.5 I,a from + 10v at room tempera
ture. For proper operation of the module, input 2
voltage must remain between +5v and -10v, and
input 1 between +lOv and -15v. Input excursions
beyond + lOv or -15v will cause damage.
Max delay for output fall:
75 ns, 50% to 50%
Max delay for output rise:
150 ns, 50% to 50%
Typical rise time 10%
t090%: 25 ns
for 200 mv square
wave about a fixed
reference voltage
The W520 is tested for 100 mv differences minimum.
It is not a replacement for the more precise A502.
INPUTS: High impedance inputs to the amplifier are
151
OUTPUTS: Standard DEC levels of 0 and -3v; cap
able of driving 17 ma at ground. The internal load
is 3 mao If input 1 is more positive than input 2, the
output will be zero. The -3v output should be lightly
loaded (less than 1 ma current) by any external path
to ground unless an external resistor is added be
tween pin V and -15 volts. Up to 50 ma may be
drawn by external path from pin V to -15v.
POWER: + 10v(A)/37 ma; -15v/32 mao
W520 - $43.00

DUAL AC-COUPLED
DIFFERENCE AMPLIFIERS
TYPE W532
W532 DUAL AMPLIFIERS
The W532 contains two ACcoupled differential am
plifiers for use with many magnetic sense systems,
including the H201 core memory. These amplifiers
provide the high differential gain and common mode
noise rejection necessary to amplify information
signals in 'a system using a single sense line per
plane for a memory, or per channel for a tape system.
INPUTS: Pins E, F, T, U require an input current of
0.15 ma or less and must be terminated to ground
through the internal lOOn resistor or an external
resistor or transformer of nominal impedance 1000n
or less. These terminations bias the inputs at ground.
In the absence of common mode signals, the differ
ence signals must not exceed 80 mv for linear ampli
fication. For positive common mode signals the
maximum differential input must be reduced by 5mv
per half volt of common mode input. Negative com
man mode voltages allow an increase in maximum
differential input by the same ratio. See Table 1
below for additional specifications.
OUTPUTS: The output voltage with no input signal is
nominally at -11.5 volts so that a W533 can be used
152
with a W532 to detect differential signals above a
preset threshold. Output impedance is 1000n. Due
to power supply ripple it is recommended that the
output be ACcoupled to other modules. See Table 1
for additional specifications.
TABLE 1. MODULE SPECIFICATIONS
Specifications Minimum Maximum
Output Voltage (no signal) -11.0 -12.0
Common Mode Input Voltage -5 +5
Common Mode Voltage Gain - 0.37
Difference Mode Voltage Gain 88 96
Output Rise Time Square - 250 ns
Wave Input
Output Fall Time Square -
400 ns
Wave Input
3db Bandpass 1 kHz 0.8 MHz
POWER: + 10(A)/40 ma, -15(B)/40 ma
W532 - $30.00
DUAL RECTIFYING SLICER
TYPE W533
J
K
P
R
D
E 0------'
F 0-------'
V
STROBE INPUT
W533 DUAL SLICER
This module is used to detect amplified magnetic
system sense signals from a W532 (see above dia
gram) and convert them to positive DEC pulses.
Detection of signals as narrow as 100 nsec is pos
sible over a wide range of detection thresholds. There
are two slicer circuits on each W533. Two input
terminals per circuit permit rectification so that
bipolar difference signals can be sliced and
standardized.
INPUTS: ACcoupled inputs J, K, P, and R have an
input impedance of lOkn and a 1 "sec coupling
network time constant. When any negative going
input exceeds its threshold and a -3 volt to ground
strobe pulse is applied at pin V, the output M or T
corresponding to the input used will rise to ground.
Provided that the strobe pulse is narrower than the
time during which the detection threshold is ex
ceeded, the output pulse will be of the same width as
the strobe pulse. Input load on pin V is 1 ma to
ground. The threshold level is preset at -1 volt, but
can be varied by paralleling resistors as shown below.
The maximum threshold is - 1.5 volts. Pins Land S
are test points which will go negative when the
153
W532 I W533
W533 AS USED WITH A W532
IN A H20l MEMORY
M
OUTPUTS
T
threshold is exceeded. No connections should be
made to these points.
MODULE INTERNAL
PINS CIRCUITRY
W533
r-
I E
~
165
ADJUSTING
I F +
RESISTORS
f--
1.5-=-
I -
-
~
V
T
330
I
+
L_
D
THRESHOLD ADJUSTMENT
OUTPUTS: Identical to those of a R1l1 or R123.
Drive capability is 20 ma to ground and outputs may
be paralleled. Delay from input to output is 50 nsec,
from strobe to output 70 nsec.
POWER: + 10(A)/4 Oma., -15(8)/28 mao
W533 - $30.00
IBM N LINE TO DEC CONVERTER
TYPE W590
2m.
OUTPUT
F
INPUT
Eo-...,....-I:Au
3000
+0.9V
JI500
2m. 2m.
3000 3000
.+0.7V
JI500
2m. 2m.
3000 3000
W590 IBM N LINE TO DEC CONVERTER
Each of the 5 inverting amplifiers on this module
provides input characteristics 'compatible with three
types of IBM N Lines. Input impedance is nominally
300 ohms, with 100 ohm available by
connecting 150 ohm shunts proJided. Each circuit
has a switching threshold near zero volts, with input
biasing included to maintain a definite output state
when the input is open-circuited.
Unshunted inputs will tolerate input excursions up
to +4v and -6v, so these circuits may also be
used to convert IBM T, 0, or Q lines if the IBM
circuits involved can safely drive the W590 input
loads.
154
INPUTS:
IBM Floating Lower Input
Line Input Level Impedance
N
+0.8v
(@-23ma) (shunted)
Transmission -1.5v 100 n
C
+0.7v
(@-12ma)
300 n
Line -2.6v
N
+0.7v
(@ -6 rna)
300 n
Logic -O.BY
Maximum input voltages: Unshunted (300 n) in-
puts: +4v, -6v; Shunted (100 n) inputs: 4v.
OUTPUTS: 18 rna at ground. 1 rna at -3v.
POWER: + lOv(A)/40 rna; -15v(B)123 rna.
W590 - $26.00
NEGATIVE OUTPUT CONVERTER
TYPE W600
Fh
J P V
NEGATIVE
~ ; ~ 1,
7.5K
OUTPUT
H
r ~ g ~ ~ 0
TERM ~ ~ ~ ~ E -=
W600 NEGATIVE LEVEL AMPLIFIER
The W600 contains three inverting amplifiers that
convert standard levels to outputs of ground and
an externally supplied negative voltage. The ex-
ternal clamp voltage is applied to terminal F (M, T)
and must be between -1 and -15 v. Additional
inputs may be added by tying diode networks, such
as those contained on the R001 or R002, to the
node terminal. These inputs form a NOR gate for
ground levels and a NAND gate for negative levels.
That is, if any input diode is at ground, the output
is at the external clamp voltage; and if all inputs
are at -3 v, the output is at ground.
INPUT: Standard levels. The input load is 1 rna
shared by all grounded inputs, including those at-
tached through diode networks to the node terminal.
CLAMP VOLTAGE: The external voltage Ec applied
to terminals F, M, and T may be any voltage be-
tween -1 and -15 v. The load is 2500 ohms to
155
-15 v without the extra load resistors or 500 ohms
to -15 with the extra loads.
OUTPUT: The output voltage levels are ground and
the negative external clamp voltage, Ve. Driving
depends on the external clamp voltage and is given
in the following table.
Maximum Output Current
At Ground At Clamp Voltage
Without add itional
18 rna 15+E"
resistor ~ m a
With additional
resistor
10 rna 15+E
c
(H, N, or U connected
~ m a
to J, P, or V)
Eo = external clamp voltage, -1 to -15 v.
It is therefore a negative number in the above equations.
POWER: + 10 v(A}/0.3 rna; -15 v(B}133 rna.
W600 - $12.00
POSITIVE OUTPUT CONVERTER
TYPE W601
SUPPLY VOLTAGE ES
POSITIVE
F CLAMP
1.8 K 7.5 K
H
OUTPUT
fmpuEr Do--I>I--r-Cil
NODE E
TERMINAL.
VOLTAGE 1.8)(
EC
7.5K I.SK
W601 POSITIVE LEVEL AMPLIFIER
The W60( contains three amplifiers for converting
DEC standard levels to outputs of ground and an
externally supplied clamp voltage level, E, .. This
external clamp voltage is applied to terminal F (M)
and mllst be between +1 and +20v. Additional in-
puts can be added by tying diode networks, such as
the ROOl or the R002, to the node terminal. These
inputs form a NOR gate for ground levels and a
NAND gate for negative input levels. That is, if any
input diode is at ground, the output will be at
ground, and if all inputs are at -3v, the output will
be at E, .. A positive supply voltage E. greater than
E" should be tied to terminal V. If E,. is less than
+IOv, the +lOv supply on terminal A may b'e used
at the supply voltage on terminal V. '
INPUT: Standard levels. The input load for each
amplifier is 2 ma shared by all grounded inputs in-
cluding those attached through diode networks to
the node terminal.
EXTERNAL VOLTAGE: Terminal F (M) - The ex'
ternal clamp voltage Eo applied to terminal F (M)
can be between + 1 and +20v. The load is 500
ohms to +E. (If the 1aOO-ohm resistors are used)
156
or 2500 ohms to +E. (If the 1aOO-ohm resistors are
not used). Terminal V - The supply voltage ER on
terminal V should be greater than E,. but not greater
than +20v. The load is 500 ohms (if the laOO-ohm
resistors are used) or 2500 ohms (if the laOO-ohm
resistors are not used) to ground.
OUTPUT: Output levels are ground and a positive
external clamp voltage, E, .. Output drive depends on
the two external voltages. They are given in the
table below.
EXTERNAL DRIVING CAPABILITIES OF W601
OUTPUT CURRENT
AT AT CLAMP
GROUND VOLTAGE
Without additional
(20 _s. )ma
(E, - E,,)
resistor
7.5
----r.r:;- rna
With additional resistor,
(20 - 1 ~ 5 ) r n a
(E< - E,)
i.e. H (N, T) connected
to J, P, or U.
~ r n a
POWER: +10 v(A)f3 ma; -15 v(B)/6 mao
W601 - $13.00
BIPOLAR OUTPUT CONVERTER
TYPE W602
v
W602 BIPOLAR LEVEL AMPLIFIER
For driving EIA standard communication lines and
other applications demanding levels both positive
and negative with respect to ground, the W602 pro-
vides up to 15 rna at up to 6v. There are three
inverting amplifiers on the module. To control noise
on long transmission lines the output rise and fall
times are intentionally slowed to roughly 50 nsec/v,
and at low repetition rates capacitance may be con-
nected externally from outputs to ground to further
increase rise and fall times. Output upper levels can
be set at + 6v, + 3v, or Ov, and lower levels can be
set at -6v, -3v, or Ov using clamp voltage supplies
provided.
INPUTS: Signals - Diode inputs require 1 rna drive
at ground. Input gating can be achieved by connect-
ing ROOl or R002 diodes at node inputs. Clamp
Voltages - Voltages from Ov to + 6v may be applied
to pin P to establish the upper. output level for all
157
three amplifiers. Each circuit whose output is high
supplies 4 rna tending to make pin P more positive.
Voltages from Ov to -6v may be applied to pin R to
the lower output level for all three ampli
fiers. Each circuit whose output is low supplies 2 mn
tending to make pin R more negative.
OUTPUTS: Signals - Maximum output current from
each circuit is 15 rna. Outputs are high when in
puts are low, and vice versa. Momentary shorts from
outputs to ground will not cause damage. If switch-
ing speed must be reduced to reduce noise genera-
tion, capacitance to ground sufficient to extend
transient times to 50% duty factor may be added
from outputs to ground. Clamp Voltages - Voltage
supplies are adequate for use with the three circuits
on the module.
POWER:+ 10v(Al/32 rna.; -15v/31 rna.
W602 - $40.00
I
POSITIVE LEVEL AMPLIFIER
TYPE W603
+IOV +IOV +IOV

. v v. V
E 0 H F K J
+IOV +IOV +IOV
. 2.2Kn v t 2.2Kn v 2.2Kfi V

+IOV
v

W603 POSITIVE LEVEL AMPLIFIER
Positive logic systems. such as those using RTL,
DTL, or TIL monolithic integrated circuits can be
driven from FLIP CHIP systems through the W603.
Clamped load resistors at the output of each circuit
permit output levels to be adjusted to the type of
circuit being driven. Normally the clamp voltage at
pin V is provided by the logic supply voltage used
with the monolithic circuits. This clamp voltage is
common to all seven converters on the module.
INPUTS: 1 rna at ground.
158
OUTPUTS: Each output can supply up to 5 rna at
ground. Drive capability at the positive output voltage
is provided by internal 2200ohm resistors returned
to + 10 volts. The upper positive level will be no
more than 0.8 volts above the clamp voltage.
Grounded inputs provide grounded outputs; negative
inputs produce positive outputs. Typical rise and fall
times are respectively 100 ns and 200 ns.
POWER: +10(A)/35 rna, -15(8)/7 rna.
W603 - $23.00
PULSE OUTPUT CONVERTERS
TYPES W607 AND W640
K!1
PA
L
W607 PULSE CONVERTER

n:

rr
CONNECT
FOR '#lIte
PUL.SES
W640 PULSE CONVERTER
These pulse converters were designed primarily to
facilitate the use of Flip Chip modules in conjuncUon
with Digital Laboratory and System Modules. In
addition. the W607 can be useful in setting or clear
ing B series unbuffered flipflops via inverters such
as BI04 or gates such as B113.
Outputs from these pulse converters are taken from
floating pulsetransformer windings. In addition to
allowing data transmission independent of ground
system integrity, this feature permits two or three
outputs to be seriesconnected for larger pulse ampli
tudes when inputs are driven simultaneously.
For purposes other than driving Digital Laboratory
and System Modules, it may be important to consider
the effect of pulse transformer backswing at the end
of each pulse. When the load is light, this trans
former recovery spike approaches the amplitude of
the pulse itself ..
INPUT: Standard Digital positive pulses or a level
change from -3 volts to ground. See table below
Module
Input
for other characteristics. Unless wider than standard
output pulses are acceptable, W607 inputs must not
be paralleled with clamped loads, such as those
internally tied to RI07 outputs. W640 output pulse
width is not affected by input loads. No connections
should be made to W640 pins E or F (L or M, S or T)
other than shorting them together to obtain 1 I'sec
output pulses.
OUTPUT: A standard 2.5v pulse. To obtain a negative
output, ground the positive output terminal. To
obtain a positive output, ground the negative "output
terminal. Each output can drive up to 10 ma of load,
in addition to a terminating resistor. A terminating
resistor must be used. Its value should be about
470 for lightest to.ads, and about 1500 for heaviest
loads. Reactances in loads and leads cause pulses
to grow in transmission, and optimum values must
be determined empirically. See table for other

POWER: W607: + 1Ov(A)/O ma; -15v(B)/35 mao
W640: + 10v(A)/O ma; -15v(B)/25 mao
Output
Laboratory System
Load Ground Rise Times -3 volts Delay Width Max. Freq.
Series Series
W607 10 ma nsec nsec S;330 nsec
W640 2ma fl;70 nsec nseC
W640
2 ma osee nsec
E to F. etc.
.ii::3
159
20 nsec 70 nsec
40 nsec 400 nsec
40 nsec 1
2.5mc 100 1000
SOD kc 3000 4000
200 kc 3000 4000
W607 - $42.00
W640 - $42.00

DEC TO IBM N LINE CONVERTER
TYPE W690
-15V
-15V
N o-Dl--.--uu
M R
L
-15V
-15V
T
p
W69D DEC TD IBM N LINE CONVERTER
Each of the four inverting drivers on this module
provides outputs compatible with the three types of
IBM N lines, depending upon what output currents
are programmed by grounds or open circuits at pins
T and U. Node points are provided at each input.
Maximum delay: 100 nanoseconds driving N trans-
mission lines.
Outputs will drive loads returned to voltages as high
as + 12, so this module will also drive T, D, or Q
lines with suitable biasing networks added.
INPUTS: 2 ma at ground, 0 ma at -3v in. Node
input provided for connection to ROOl or R002
diode gate expander cards. .
OUTPUTS: Outputs are open for negative inputs.
Table below shows nominal output currents for
160
grounded input. Maximum negative output excur-
sions: -3.5v (clamped). Shorting outputs to ground
will not cause damage.
POWER: +lOv(A)/O ma; -15v(B)/150 mao
IBM NEGATIVE OUTPUT
LINE OUTPUT PROGRAMMING
N
26 ma
T Open
Transmission U Open
---
20 ma
T Ground
U Open
C
14 mao
T Open
'Line U Ground
N
8 ma
T Ground
Logic U Ground
W690 - $36.00
SWITCH FI L TER
TYPE W700
I ~
~ - - - - - - - - - - - - - - - - - - - - ~ - - - - - - - - - - - - - - - - - - ~
W700 SWITCH FILTER
The W700 contains six switch filters for reducing
contact closures to standard levels. The output
drive of the switch filter is determined by the voltage
to which the switch contact is returned. For maxi-
mum output drive at ground level, terminal 0 should
be connected to -15v and the external contacts
should be returned to + lOv. In this case, open con-
tacts produce a -3v output, and closed contacts
produce a ground output. For maximum output drive
at the -3v level, terminal 0 should be connected to
+ 10v and the external contacts should be returned
to -15v. In this configuration, open contacts pro-
duce a ground output and closed contacts produce
a -3v output. Typical rise and fall times are given
in the table below.
INPUT: When terminal 0 is connected to -15v, 6 rna
flows through a closed contact. When terminal 0 is
connected to +lOv, 7 rna flows through a closed
contact.
OUTPUT: See table below.
. POWER; Terminal 0 connected to -15v; +10v(AlI-
o rna; -15v(Bl/31 rna. Terminal 0 connected to
+10v; +10v(Al/8 rna; -15v(Bl/22 rna.
Typical Typical
Contact
Output
Drive
Rise Fall
Current
Time Time atgnd at 3v
D connected to 5 msec 20 msec 4ma 1 ma
-15 v, switch (switch (switch
6 ma
(switch (switch
contact returned contacts contacts contacts contacts
to + 10 v closing) opening) closed) open)
D connected to 25 msec 3 msec 1 ma 5ma
+ 10 v, switch (switch (switch
7 ma
(switch (switch
contact returned contacts contacts contacts contacts
to--15v opening) closing) open) closed)
W700 - $20.00
161

POWER 5UPPLY(+3.6 VOLT)
TYPE W705
(SINGLE HEIGHT, TRIPLE WIDTH)
POWER
SUPPLY
J +3.6
1---0 VOLTS
W705 POWER SUPPLY
This inexpensive power supply is of primary use in
conjunction with the W706 and W707 teletype
modules. The output can supply up to 1 amp at. a
nominal voltage of 3.6 volts. Voltage regulation for
variable loading is not provided, however, under
approximate constant loads the internal filters will
minimize supply voltage noise.
It is recommended that this supply be located where
air flow is not restricted. Power dissipation is 610
watts, depending on load current. This supply should
have a minimum output load of 200 mao
POWER: +10v(A)/1 amp.
W705 - $15.00
162
TELETYPE RECEIVER
TYPE W706
(DOUBLE HEIGHT)
POWER CLEAR BV
ENABLE BN
8S
8J
CLOCK
ENABLE
SERIAL INPUT
READ BUFFER
W706 TELETYPE RECEIVER
The W706 Teletype Receiver is an integratedcircuit,
serialtoparallel Teletype code converter, self con
tained on a doubleheight module. This unit includes
all of the serial to parallel conversion, buffering,
gating, and synchronizing necessary to transfer in
formation between an incoming asynchronous serial
teletype line and a parallel binary device. Either a
5bit serial character consisting of 7.0, 7.5, or 8.0
units or an 8bit serial character of 10.0, 10.5, or
1 LO units can be assembled into parallel form by
the W706 through the use of selective jumpers on the
module. The serial input for one character is ex
pected to be in sequence; a one unit -3 volt start
signal, the five or eight character bits, a ground level
stop signal of LO, L5, or 2.0 units. When the con
version is complete, the start and stop elements
accompanying the serial character are removed. A
logical 1 for a character bit is a ground level and a
logical 0 is -3 volts. The first bit received on the
serial line is Bit 1 at the parallel output.
To perform the serial to parallel conversion, the reo
INPUTS: Standard Digital levels of -3 volts and
ground or 400 nsec pulses as generated by module
types R602 and W103. Input pins are shown on the
diagram above.
CLOCK: 400 nsec positive pulses with a maximum
receiver input frequency of 200 kHz. The clock fre
163
ceiver continuously examines the serial input line,
and when a start element is recognized, the receiver
enables the external clock through the Clock Enable
Output and synchronizes with the incoming signal.
When the last character bit, either bit 5 or bit 8, is
received, the flag is set and a ground level appears
at the Flag Output At this time, the Parallel Data
Outputs of the W706 can be examined by a Read
Pulse, and if desired, the flag can be cleared by a
pulse on the Clear Flag Input A new serial character
must not be put on the Serial Input until the stop
time of the previous character is counted out and so
indicated by a ground level on the Clock Enable
Output. For additional timing information see Figure
1. The W706 may be connected to devices other than
a Teletype, providing that their serial output is sim
ilar to a Teletype code. Start element noise rejection
of the W706 is approximately one Vu,t from ground,
requiring a line filter or use of the W708 on noisy
teletype lines. To obtain additional Teletype appli
cations data, write for Applications Note APWL
quency must be twice the required serial input fre
quency thus defining one unit of character time as
two clock periods. Input loading is 2.8 ma at -3
volts. The clock used must be externally gateable and
similar to a R401 unless the W706 is used with
a W708.
W706 - $150.00

I I
o
CLOCK. I
,I i I I I I I I I I I I I
, t t t t t t t t t t t t t t f f t t f t f t t '
I I I I I I
fT'''; I I
-3 II I
I I r
I
F L A ~ STAqBE INPUT
. I I I
_: I I II
I I I
STR?6EDF,LAGO,UTPU
I
I
BEGINNING OF
CHARACTER
I I
END OF
CHARACTER
Figure I. Typical Timing Diagram - Parallel Output, 8-Bit
(01,111,111) - 2 Unit Stop Time. Clamped Loads
(W002 or W005) on Outputs.
ENABLE: A diode input, which if brought to ground,
will disable the clock through clock enable. Disabling
(or enabling) the clock during a serial input character
can result in incorrect character reception.
CLEAR FLAG: A ground level or Digital standard 400
nsec positive pulse will clear the flag. If a level is
used, it must be returned to -3 volts before the flag
can be set. Loading is 1.4 rna at -3 volts. Typically
the flag is sensed through one of the flag outputs
and then cleared.
FLAG STROBE: Digital standard 400 nsec negative
pulse or a -3 volt level. Loading is 1.4 rna at -3
volts. This input is NANDed with the flag and pro-
vides a ground level Strobed Flag Output signal when
the flag is set.
READ BUFFER: A 400 nsec positive pulse provides
parallel information from the W706. During this
pulse, any bit which is a logical 1 will generate a 400
nsec positive pulse at the corresponding bit output.
This input can be held at ground for continuous mono
itoring of bit outputs. Typically, this pulse is gener-
ated after a Flag Output has been sensed so that no
incorrect character will be received on the parallel
lines. Loading is 2.8 rna at -3 volts.
POWER CLEAR: Same input signals and loading as
for Flag Strobe. Initialization of module components
by a Power Clear signal is not necessary if the first
character received after power turnon is insignifi
cant. When not used, Power Clear can be left
disconnected.
. SERIAL INPUT: Digital standard levels of -3 volts
and ground. A ground level during a bit input repre
sents a logical-I. The first character bit to come in
on this input appears at Bit 1 output. Loading is
2.8 rna at - 3 volts.
164
OUTPUTS: All outputs are capable of supplying 20
rna at ground. The external load may be connected
to any voltage between ground and - 20 volts.
Clamped loads such as W002 and W005 can also
be used.
BITS 1 THRU 8: Buffered outputs generated by
NAN Ding the internal bit and the Read Buffer. A
ground level or positive pulse output represents a
logical 1 for that bit. Unused outputs can be left
open.
FLAG OUTPUT: Ground level output when the flag
is set.
STROBED FLAG OUTPUT: Ground level output or
pulse output when Flag Strobe is at -3 volts and the
flag is set.
CLOCK ENABLE: Used with R401 clock or equivalent
to synchronize the clock to incoming serial data. The
output is an open circuit whenever a serial input is
present and at ground at all other times. When used
with a R401, this output is connected to the enable
input of the R40l.
READER RUN: Of use in teletypes equipped with
relay controlled paper tape readers. The Reader Run
Output is enabled (ground level) by a Clear Flag
pulse and disabled by the W706 circuitry when a
start pulse is received on the serial input. For ad-
D ~ ~ D
TT
0
9 00
~ bTt ~
~
I 8 I
I bit I
0
6
0
~
ditional information see Figure l.
JUMPERS: Jumper positions are indicated on the top
view physical sketch shown in Figure 2. The W706 is
shipped with all jumpers in position.
POWER: -15 (6)/12 ma: +3.6 volts/400 mao This
power is available from a W705 or any commercial
supply that has an output regulation of 5 %.
0 0
....-vvv-e
~
~
~
~
0
~
0
......-vvv-.
......-vvv-.
o 0
......-vvv-.
......-vvv-.
~
00
~
~
~
0
0
~
...-vvv---e
...-vvv---e
~
~
~
~
0 0
0 0
9,
5
0
~
~
lunits
~
0 0
~
<j'1.0
1.5?
~
I
I and
and I
D
2.0
0
bunits
0
2.0
0
D
units
1.01
units 6
~
~
<>-I>f--O
~
~
e--vvv--e
o----t>t---o
.-vvv--e
Figure 2. W706 Jumper Diagram
165
TELETYPE TRANSMITTER
TYPE W707
(DOUBLE HEIGHT)
CLEAR FLAG
FLAG STROBE
-'W"'A"-IT'---____ --l CONTROL
POWER CLEAR au
CLOCK INPUT BO
FLAG
STROBED
FLAG
SERIAL
OUTPUT
PARALLEL { ~ : ~ ~ ~ ~ SHIFT
I N ~ ~ + ~ B"'I""TS'---'B'!!K-<>I REGISTER
BIT6 BP
BIT7 8M
BITe BN
ENABLE 85
LOAD BUFFER BE
W707 TELETYPE TRANSMITTER
The W707 Teletype Transmitter is an integrated
circuit parallelto.serial teletype code converter, self
contained on a doubleheight module. This unit in
cludes all of the parallel to serial conversion, buffering,
gating, and timing necessary to transfer information
in an asyncronous manner between a parallel binary
device and a serial teletype line. Either a 5bit or
8bit parallel character can be assembled into a 7.0,
7.5, or 8.0 unit serial character or a 10.0, 10.5, or
11.0 unit serial character, respectively, by the W707
through the use of selective jumpers on the module.
When the conversion is complete, the necessary one
unit negative voltage start signal and a ground level
stop signal of 1.0, 1.5, or 2.0 units have been added
to the original parallel character and transmitted over
the serial line. The serial character is transmitted
with the start signal first, followed by bits 1 through
8 in that order, and completed by the stop signal.
Onehalf unit after the stop signal is put on the serial
line, the flag is set indicating that the previous.
character has been transmitted and that a new paral
lei character can now be loaded into the W707.
Transmission of this new character will not occur
until the stop time from the previous character is
completed. See the timing diagram (Figure 1) for
additional information.
The W707 may be connected to devices other than a
Teletype. For example, two computer systems can be
166
connected using a serial line as shown in Figure 2.
To obtain additional Tel.etype applications data write
for Applications Note AP-W1.
INPUTS: Standard Digital levels of -3 volts and
ground or 400 nsec pulses as generated by modules
types R602 or W103. Input pins are shown on the
module diagram above.
CLOCK - 400nsec positive pulses with a maximum
transmitter output frequency of 200 kHz. The clock
frequency must be twice the required serial output
frequency thus defining one unit of character time
as two clock periods. Input loading is 2.8 ma at
-3 volts.
LOAD BUFFER - A 400nsec positive pulse which
loads the parallel character into the W707. Typically
this pulse is generated after a Flag Output has been
sensed so that no incorrect characters will be trans
mitted. Loading is 2.8 ma at -3 volts.
BITS 1 THROUGH 8 AND ENABLE - Digital stand
ard levels or equivalent with input loading 1.4 ma
at -3 volts. When an 8bit character is to be trans
mitted, all bit inputs are connected to data lines with
bit 1 the least significant bit. For 5bit characters,
bits 1 through 5 are connected to data lines with bit
1 the least significant bit. Bit 6 is. now used as the
W707 ~ $150.00
LOAD BUFFER INPUT
ot t: . t
-3 ttl
CLOC.K INPU1T I I I
1 I I I I I I I I I I I I
_: t t t t t t t t t f t t t f t t t f f f f f f f t
S E R I ~ L OUr'PUT II II I I 'I I I I I
I I I I I I ,I
.START
1
BIT! I 8'T21 81T3 81T4,81T5 BIT6 BIT.7
1
BIT8
1
STIDP I
o I
-3
BEGINNING OF
CHARACTER
t
I
END OF
CHARACTER
Figure 1. Typical Timing Diagram - Parallel Input, 8Bit
(01,111,111) - 2 Unit Stop Time. Clamped Loads
(W002 or WOOS) on Outputs.
Enable input, and bits 7, 8, and Enable are tied
together and either returned to -15 volts through
a 2.7K resistor or individually connected to W002
clamped loads. If the Enable input (Enable or bit 6
depending on character length) is at -3 volts during
a Load Buffer Pulse, the parallel character informa
tion is loaded into the W707, but no serial transmis
sion will occur. The Emible input mustbe at ground
during a Load Buffer Pulse for serial information
transmission. Ground levels on bit inputs represent
a logical 1 or a Teletype "mark," and generate a
ground output on the serial line at the corresponding
bit times.
CLEAR FLAG: A ground level or DEC standard 400
nsec positive pulse will clear the flag. If a level is
used, it must be returned to -3 volts before the flag
can be set. Loading is 1.4 ma at -3 volts. Typically
the flag is sensed through one of the flag outputs
and then cleared.
FLAG STROBE: DEC standard 400 nsec negative
pulse or a -3 volt level. Loading is 1.4 ma at -3
volts. This input is NANDed with the flag and pro
vides a ground level Strobed Flag Output signal when
the flag is set.
POWER CLEAR: Same input signals and loading as
for Flag Strobe. Initialization of module elements by
a Power Clear signal is not necessary if the first
&erial character transmitted.after power turnon need
167
not be correct. When not used, Power Clear can be
left disconnected.
WAIT: This input is available for use with the W708
in half duplex operation. Internal logic levels of +3.6
and ground appear at this input. It must not be
connected to any signal but the WAIT output of the
W708. If not used, this input must be left discon
nected.
OUTPUTS: All_outputs are capable of supplying 20
ma at ground. The external load may be connected
to any voltage between ground and - 20 volts.
Clamped loads such as W002 and W005 can also
be used.
SERIAL OUTPUT: Provides the teletype code serial
output during character transmission. A logical 1
output is a ground level. If inductive loads are driven
by this output, diode protection must be provided
by connecting the cathode of a diode to the output
and the anode of this diode to the negative supply
used at the output. .
FLAG OUTPUT: Ground level output when the flag
is set.
STROBED FLAG OUTPUT: Ground level output when
Flag Strobe is at -3 volts and the flag is set.
INVERTER: Pins BJ and AP are the input respectively
of an inverter that can be used for any needed buffer
ing. Input load is 1.8 ma at -3 volts.
JUMPERS: Jumper positions are indicated on the top
view physical sketch shown in Figure 3. The W707 is
shipped with all jumpers in position.
PARALLEL
INFORMATION
COMPUTER
NUMBER I
~ -,---------,
POWER: -15(8}/3 ma; +3.6 volts/400 mao This
supply that has an output regulation of 5%.
COMPUTER
NUMBER 2
Figure 2. ComputertoComputer Serial Transmission.
2.0uni!s
.1.. ()o----o()
D D
D b ~ D
0
Q
0
:1.0
: units
~
0
..-vvv-e
D
6
D
0
~
Q ..-vvv-e
'La
~
0
:units
..-vvv-e
6
~
D
~
~
~
D D D
~
..-vvv-e
~
~
~
...-vvv-e
~
D
~
~
D
9
D D
...-vvv-e
:5
...-vvv-e
..-vvv-e
I bit
..-vvv-e
9?
..-vvv-e
...-vvv-e
o-jf-o 1.0, 1
8
.-vvv-e
andl .bit
D
D " " l i ~ l i, D
~
D D
~
~
~
~
~
~
~
~
~
Figure 3. W707 Jumper Diagram.
168
TELETYPEINTERFACER
TYPE W708
8 X CLOCK
TO W706 CLOCK INPUT
TELETYPE SERIAL INPUT
TO W707 CLOCK INPUT
CLOCK ENABLE FROM W706
TELETYPE
TO POWER CLEAR W706
POWER CLEAR
INTERFACER
TO CLEAR FLAG W706
CLEAR FLAG 1
CLEAR FLAG 2
TO WAIT W707
W70B TELETYPE INTERFACER
The W708 provides special gating controls and clock
synchronization for Teletype and data communica-
tions systems, when used with the W706 and W707
Teletype modules. For additional applications data
and more complete specifications on the W708, write
for Applications Note AP-Wl.
ADDITIONAL FUNCTIONAL CAPABILITY
SPIKE ELIMINATOR: This gating structure provides
onehalf unit start pulse noise rejection so that long
noisy teletype lines can interface with the W706.
WAIT: Provides the necessary control gating so that
the W706 and W707 can be used in halfduplex
operation. This control section is used to hold W707
transmission while the W706 is receiving a character.
169
CLEAR FLAG CONVERTER: This control provides ad
ditional gating so that the W706 flag can be cleared
in either of two modes during halfduplex operation.
SINGLE CLOCK CONVERTER: Reduces the number
of clocks needed in a W706 and W707 Teletype sys
tem from two to one. This control also allows the use
of a nongateable clock such as the R405 with such
a system. The frequency of the required clock is
eight times the serial transmission frequency. All
startstop synchronization of the W706 clock input
is provided.
POWER: -15(8)/25 ma; +3.6 volts/200 mao This
supply that has an output regulation of 5%.
W708 - $55.00
RELAY
TYPE W800
CONNECT TO
SUPPRESS NOISE
r-;I
~
Il;i. ft;;l,
TIT I TIT I
~
' > v : :
I I
i I
D I I
L __ J
E
F ~
~
" v l :
I I
I I
I I
H L __ J
,
. ~
TYPE WBOO RELAY
The Type waoo Relay consists of two separate Form
A reed relays, each with an optional protecting cir-
cuit. When the protecting circuit feature is desired,
Nand P (T, U) should be connected together, and
the external circuit Connected to P and R (U, V).
To use the relay without the pr.otecting circuit, the
external circuit should be connected between M and
R (S, V). The protecting circuit consists of a capaci-
tor and a parallel combination of an inductor and a
resistor. The protection circuit slows down current
and voltage rise time at the time of contact closure
and voltage rise time at the time of contact closure
or opening in order to minimize undesirable effects
on sensitive logic in the vicinity of the relay. The
170
Type waoo is used to drive heavy loads on com-
puter or logic command. The frequency limit is
100 cps. Maximum relay operating time is 2 msec.
INPUT: A Standard Level of -3v operates the relay.
Input load is 1 rna at ground, 0 rna at -3v, shared
by inputs at ground. Pins F and "K are for use only
with diodes such as R001 and R002. A maximum of
6 in. of wire may be attached to these points.
OUTPUT: The relay contacts close when the input
requirements are met. Maximum contact ratings are
250v, 500 rna, 10 watts maximum.
I>OWER: -15v/124 rna; + 1 Ov(A)f 0.6 rna.
W800 - $45.00
RELAY MULTIPLEXER
TYPE W802
(double height module)
W802 RELAY MULTIPLEXER
The W802 Relay Multiplexer contains eight double-
pole, normally open reed relays. One of its uses is
to address memory lines in memory testers. It can
also be used as a low-speed multiplex switch where
the grounded, low-noise performance of the A111
multiplexer is not required. Maximum closing time:
1.5 msec; typical opening time: 500 "sec.
INPUTS: Each driver is a 1 rna load shared among
171
its grounded inputs. Contacts close when inputs are
negative.
OUTPUTS: Relay contacts rated at 250v, 500ma, 10
walts maximum. Contact resistance typically 250
milliohms.
POWER: +lOv(A)/2 rna; -15v(B)/20 rna plus 25
rna per energized relay (220 rna max for all relays
energiied).
W802 - $160_00
Type
W990
W991
\V992
W993
W994
W995
W970
W971
W972
W973
BLANK MODULES
TYPES W970-W973, W990-W995
These 10 blank modules offer convenient means of integrating special circuits
and even small mechanical components into a FLIP CHIP system, without loss
of modularity. Both single and doublesize boards are supplied with contact
area etched and gold plated. The W990 Series modules provide connector pins
on only one module side for use with H800 connector blocks. W970 series
modufes have etched contacts on both sides of the module for use with double
density connectors Type H803. .
Pins Description Handle
18 Bare board, split-lug terminals attached
36 Bare board, split-lug terminals
18 Copper clad, to be etched by user separate
36 Copper clad, to be etched by user separate
18 Perforated, 0.067" holes, 18 with attached
etched lands. The holes are on 0.2"
both horizontally and vertically.
36 Perforated, 0.067" holes, 36 with attached
etched lands. The holes are on 0.2"
centers, both horizontally and vertically.
36 Bare board, no split lugs, similar attached
to W990
72 Bare board, no split lugs, similar
to W991
attached
36 Copper clad, similar to W992 separate
72 Copper clad, similar to W993 separate
172
Price
$2.50
$5.00
$2.00
$4.00
$4.40
$8.80
$4.00
$8.00
$3.00
$5.00
MODULE EXTENDER
TYPE W980
The W980 Module Extender allows access to
the module circuits without breaking connections
between the module and mounting panel wiring.
SYSTEM MODULE ADAPTER
TYPE W985
The W985 is an adapter which permits DEC system
modules to be plugged into a FLIP CHIP mounting
panel. It requires a block of four (two high and two
wide) FLIP CHIP module spaces. The pin connec-
tions are made in two vertical slots with all pins
being used on the upper connector and only pins S,
T, U, and V being used on the lower connector. The
two connectors on the component side of the module
(to the left as you face the wiring) must be left
vacanUo accommodate. the extra width of the sys-
tem mo.dule. Built on a W991 board.
173
DUAL STANDARD SYSTEM
MODULE RECEPTACLE FLIP CHIP PLUG
::ro
AA
CO---OAB
o 0--0 AC
EO---OAE
FO---OAF
Ho--oAH
JO---OAJ
Ko---oAK
LO----OAL
MO--OAM
No----oAN
PO---OAP
RO----OAR
SO---OAS
TO---OAT
UO----OAU
v 0---0 AV
UPPER
CONNECTOR
WO---OB' }
)( 0----0 8T LOWER
y 0---0 au CONNECTOR
ZO---OBV
W980 - $14.00
W985 - $34.00
174
175
APPLICATION
NOTES
ESTIMATING PROPAGATION DELAY
(R107, R111, R113, R121, R123, R151)
APPLICATION
NOTE
The chart below shows the effect of logical complexity of R-Series g?te delay for output fall.
Fall time is often the main source of delay, even though two-thirds of the rise time and only
one third of the fall time elapses before the input threshold is reached. This is because the
2 ma load alone must charge wiring and input capacitance during output fali. Ample excess
gate output current accelerates rise times, so propagation delay for output rise remains
typically 30-40 nsec regardless of fan-out. Gate expansion at node inputs adds 15-30 nsec
delay.
If speed is important, excess fanout capability may be traded to obtain it. The dotted line
on the chart shows how adding a Wq05 clamped load affects propagation delay.
There are many factors that influence propagation delay: wiring capacitance, node expansion,
enabling of connected gates, even ambient temperature. Since ali of these factors together
can account for substantial variations, the chart is a guide only.
-3V
PROPAGATION OELAY VS NUMBER OF OUTPUT CONNECTIONS
nsec
50
~
...-w;;:;; BUILT-IN
2 MA LOAD
ONLY
/
V
--
::::::::.--
f-w;Ti; ADDED
5MA LOAD
150
100
5 10 15 FAN-OUT
PROPAGATION
DELAY
Figure 1 Propagation Delay
176
BCD COUNTING
APPLICATION
NOTE
It is sometimes preferable to represent binary numbers as decimal, especially when ease
or speed of recognition by an operator is'an important consideration. Familiar examples
are elapsed time measurements, event recording, digital 'volt meters, and digital servo
systems. The Application Note, General Purpose Digital Clock, describes one specific use
of BCD counting.
The Application Note on Binary Coded Decimal Codes lists various popular codes in use
today. Four of the most commonly used codes are implemented in this application note as
single digit (decade) elements. These elements can be cascaded to provide long counting
chains by simply connecting the proper carry out of the most significant bit in the decade
to the proper carry input of the least significant bit of the next decade. I n using the decades
as elements in a system, it must be remembered that the longest carry propagation time
may not be the time for propagating carries between successive digits (decades) in the
counter. In a three digit, 8421 up counter, there are only two such carries between digits
and one carry within each digit, as shown.
Present
099
'0
COUNTER STATE
Carries to propagate
0000 1001 1001 0001
'9'----'"- __ 1'"---
Next
10010
0000 0000
When one decade is changing from a decimal 7 to a decimal 8 there are three carries which
must propagate within the decade itself.
DECADE STATE
Present
Carries to propagate
0111 1000
.... ~ .....
Next
810
One of the most common BCD codes is 8421 code. Figure 1 is an up counter for this code.
The inhibit decade input may be used to turn a counter off from a control level rather than
shutting off the pulse train. Note that the level should change at least 400 nsec prior to the
next pulse occurring at the pulse input in order for that pulse to be excluded from being
counted. Figure 2 is an 8421 down decade. It is more expensive than an up counter be
cause it requires a pair of gates in addition to the DCD gates on the R202's.
Figure 3 is a bidirectional8421 counter decade that requires level control of the direction
of count. The least significant decade of such a counter requires only a single pulse source,
so on this digit the up and down pulse inputs may be tied together. It is important that
direction levels do not conflict with one another. A ground level enables the direction, so
the opposite direction must be at a -3v level. Changes in the direction level should occur
at least 400 nsec prior to the next pulse. A single flipflop may be used to furnish the direc
tion levels in order to prevent conflicts from occurring.
Figure 4 is an 8421 up down counter with pulse control of direction. Here, two separate
pulses must be furnished to the least significant digit. Changes in direction should occur
no closer than 500 nsec.
177
Another popular decimal code is 2421. Figure 5"is a 2421 up decade. Figures 6 and 7 are
up decade counters for the 5421 and X53 (excess 3) decimal codes.
When interconnecting decades in a counter configuration it is the positive going edge of a
pulse or level that causes the decade to increment or decrement. Figure 8 shows block
diagram interconnections for up, down, and up/down decades.
Figure 1 8421 Up Decade
8
R202
o
Figure 2 8421 Down Decade
178

1-1lCk>-4>---o
STATE
9
8
7
6
5
4
8 4 2 ,
, 0 0 ,
, 0 0 0
o , , ,
o , , 0
o , 0 ,
o , 0 0
o 0 , ,
00' 0
, 0 0 0 ,
o 0 0 0 0
1
Lc
o!<",

.; <" ci
.... ""
6J
'-----< 0---
-1-
2
01
Q:
1
.,.,<" o!<", .... <" ..
''''r' "'r" '"
'----< 0---
NI_
01
0
N
Q:
L
I

o! .....
"'r"
....<" c!

6J
'-----< 0---

0
01
N
Q:
,-J
-r"'
ci 6.J cl
'------1 0---

0
01
N
Q:
179
J I'

i-
rt
L
>
r
L I'

E
-
c:
o
U

Q)
-'
ar
"
fl
Q)
o
c:
:;:
o
o
-..
0.
::>
.....

rtl

::I
tlIl
u::
-
(Xl
o
UP
ENABLE
UP
PULSE
DOWN
PULSE
DOWN
ENABLE
<J--'"1A
fV1
~
R201
~ - ~ ~
~
1lI
r;-c 1 1 ~
ir-;s-
I
R201 R201
0 ~ . 0
'---
:".
~ r n
Ivt
? ?
""
""
~
~
~
-=- -=-
~ ~ I r ~
~ v W
Figure 4 8421 Up/Down Decade, Pulse Control
~
~ 1 n
1 I ~
R201 ~
0 I ~
~ ~
DOWN
PULSES
UP
PULSES
UP
ENABLE
DOWN
ENABLE
4
R202 R201
o o
R202
o
181
STATE
0
4
2 4 1
0 0 0 0
0 0 0 1
0 o 1 0
0 o 1 1
o 1 o 0
1 o 1 1
1 1 o 0
1 1 0 1
1 1 1 0
1 1 1 1
..... -1--0 PULSE IN
STATE 5 4 2 1
o 0 0 0 0
o 0 0 1
o 0 1 0
o 0 1 1
4 0 1 0 0
5 1 0 0 0
1 0 0 1
1 0 1 0
1 0 1 1
1 1 0 0
+---,--0 PULSE IN
I
STATE A B C 0
0 0 0 1
1 0 0 0
2 0 0
3 0 1 0
4 0 1 1 1
5 0 0 0
A
R202 .--t--o PULSE IN
o
Figure 7 XS3 (Excess 3) Up Decade
MSo 4 4
ETC
I
:
11:
LSD
~ }
PULSE OR
MSo '4 UP DECADES 4 POSITIVE EDGES
ETCJ: U:
LSD
~ TO BE COUNTED
DOWN DECADES
MSD 4
UP UP
UP
LSD
0 0
DOWN
DOWN DOWN
UP/DOWN DECADES (LVEL CONTROL Of olRECTIONl
MSD ..
UP
UP PULSE
PULSE
UP UP
UP ENABL
UP ENABLE
DOWN ENABLE
DOWN ENABLE
OOWN
0 DOWN
DOWN PULSE
0 DOWN
-=DECREMENT
PULSE
PULSES
UP/DOWN DECADES (PULSE CONTROL Of DIRECTIONl
Figure 8 Interconnecting Decades
182
PARTS LIST
Type Quantity Description
8421 Up Decade
R202 2 2 Flip-Flops
(Figure 1)
8421 Down Decade
R202 2 2 Flip-Flops
(Figure 2)
Rlll 1 3 Diode Gates
ROOl 1 Diode Network
8421 Up/Down Decade,
R201 4 Flip-Flop
Level Control (Figure 3)

ROO2 1 Diode Network
8421 Up/Down Decade,
R20l 4 Flip-Flop
Pulse Control (Figure 4)
ROO2 1 Diode Network
2421 Up Decade
R202 1 2 Flip-Flops
(Figure 5)
R20l 2 Flip-Flop
RIll 1 3 Diode Gates
5421 Up Decade
R202 2 2 Flip-Flops
(Figure 6)
XS-3 Up Decade
R202 1 2 Flip-Flops
(Figure 7)
R20l 2 Flip-Flop
183
GENERAL PURPOSE DIGITAL CLOCKS
APPLICATION
NOTE
In most digital systems which collect data or record. events, it is often desirable to have a gen-
eral purpose digital clock from which real time of day may be read for recording along with
measurements. The clock is .also useful for generating periodic time signals which may be
used to interrupt a computer when performing periodic scans of inputs, etc.
Such clocks most often derive their base frequency from the line frequency. The rest of the
clock is usually a decimal divider chain which keeps track of the hours, minutes, seconds,
and any smaller desired increments. The block diagram of the clock is sQown in Figure 1.
Figure 1 Block Diagram of Clock
Desirable features of such a clock are that it provide 24 hour time, may be preset to any
time, may be read as a binary or BCD word, and that single pulses may be derived at periodic
intervals for use as an auxiliary signal.
Two designs are offered here are:
120v, 60 cps line, BCD clock, and
220v, 50 cps line, BCD clock ..
The only differences between the clocks are whether or not the first stage is a count of 60 or
a count of 50. The count of 60 produces output changes at a 1 pps rate from a 60 cps source,
while the count of 50 derives a 1 pps rate from a 50 cps source.
The input network consists simply of a 6.3v rms filament transformer, a small integrator to
minimize high frequency noise, and a Schmitt Trigger (W50l) to shape the negative portion
of the cycle between -0.75v and -2.25v into a pulse which drives between -3v and
ground. The positive edge of the output pulse occurs when the input reaches -2.25v. Fig-
ure 2 shows the shapero
184
t20VAC
60 CPS -----"311
2 5 g ~ A C ~
50 CPS
Figure 2 Line Frequency Shaper
60 (50) CPS
PULSE TRAIN
The first counter is either for a 60 cps source (Figure 3), or a 50 cps source (Figure 4). The I
output of this chain is 1 pulse per second. Following this are two counts of 60 providing
minute and hourly pulses, and finally, a count of 24 providing daily pulses. (Figure 5).
The preset logic is the same for each counter unit. Figure 6 shows typical preset logic for
two digits. Only the bits required for the counter unit are implemented. A block diagram
showing the pulse logic for presetting the clock is shown in Figure 7. The one shot (R302)
provides a delay between clearing and presetting the bits of the clock, which should be ones.
The PA's must generate 400-nsec pulses to insure proper clock clearing.
Type
R202
W002
RIll
R601
R302
W501
W501
PARTS LIST
Quantity
14
1
5
1
1
1
1
Description
Counter
Preset Logi c
Diode Gate
Pulse Amplifier
Delay (One Shot)
Schmitt Trigger
Line Frequency Shaper
Additional Hardware
1 6.3v til XFMR
11K -1f4 watt res
1 .01 J.lf cap
4 decade switches
185
OUTPUT
OUTPUT
,..----1",---"'1-+-0 P40 P40
n

0
c
P20

P20
...
\l
\l
..
+----I>'-r---T-'<I---o PIO +-+----i)iL.,----,-J.:J--+-o PIO

+-+---i)iL.,----,-J.:J--+-oP2
+----I>'-r---T-'<I---oPI +-----i)iL.,----,-J.:J----oPI
CLEAR
INPUT
CLEAR
INPUT
Figure 3 BCD Count of 60 Logic Counts 0 through 59
186
P20
O--+-------<h-------,--'<l-l--I--o PlO

O-+------<:4--------,...l:J.-I--I--o PB
O--------<:4--------,...l:J.-------oPI
CLEAR
OUTPUT
187
DECIMAL SWITCHES
PRESET
W002 CLAMPED LOADS
.. NY--.----t> PI
.. NY--.----t> P2

PB

"Nv---.---{> P20
e___Nv-.---{> P40
7 DIODE GATES
(2 Y3 Rill's)
Figure 6 Preset Switch Logic (Hours or Minutes)
188
PRESET ~
BUTTON I
7 LINES
Figure 7 Clock Preset Logic
189

6 LINES
HIGH-SPEED PARALLEL ADDERS
APPLICATION
NOTE
Parallel adders can be used to add two binary numbers. The augend is called the resident
number and is stored in the accumulator register (ACO-7). The addend, or incident number,
is stored in the incident register (10-7). The sum appears in the accumulator. R series Type
R201 Flip-flops can be used as shown in the logic diagram, Figure 1. Addition is performed
in two steps. The first step is a half-add. Each digit of the accumulator is complemented
(made negative) if the corresponding digit of the incident number is 1 (inner pair of DCD
gates on ACO-7). The second step is a carry. A carry is generated if a digit in the accumulator
is 0 and the corresponding incident number is 1 (RIll gates). A carry is also propagated
if an accumulator digit is 1 and it receives a carry pulse from the next less significant
accumulator digit (upper DCD gate on each pulse amplifier). Each stage will propagate one
carry at most. After all carries have been propagated, addition is complete and the accumu-
lator contains the sum of the incident and the resident numbers.
When the most significant bit is used as a sign bit, a carry pulse amplifier is provided on the
input to AC7 (least significant bit) to provide for End Around Carry (EAC). EAC only occurs
when dealing with negative numbers (ACo and 10 are considered the sign bits). If the full
8 bits of the AC are to be used as a positive accumulator, the EAC pulse amplifier must be
disabled so that end carries will not affect bit AC7. .
Overflow can occur only in adding two positive numbers or two negative numbers together
(in the latter case it sometimes is referred to as underflow). Overflow is characterized as a
carry out of the most significant bit when adding two positive numbers, or a lack of a carry
out of the most significant bit when adding two negative numbers. In the case of the unsigned
adder (where AC" is the most significant bit) the overflow is the EAC pulse. In this case the
output of the unused EAC pulse amplifier may be used to set an overflow flip-flop.
Subtraction
An adder may also be used for subtraction. To subtract a number from the accumulator, the
AC is complemented and added to the incident number. The result is then recomplemented
as the final step.
The steps involved in performing a subtraction depend on whether the l's complement or
the 2's complement number system is used to represent a negative number. Since the l's
complement number system is easiest to implement, it is the one described here. To subtract
a number from the accumulator, the steps are (1) complement the AC by means of the
complement pulse input, (2) half-add, (3) carry, and (4) recomplement the AC. With this
number system, it is necessary to use the end around carry into AC7 as shown in Figure 1.
One's complement subtraction may also be performed by (1) complementing the incident
number, (2) half-add, and (3) carry. Note that in l's complement arithmetic, there are two
possible zeros. This is because the l's complement of +0 is -0 (11111111).
When subtraction is included in the arithmetic capability of the unit, the AC and incident
numbers must be considered as signed numbers (bits ACo and 10 being the sign bits).
Overflow and underflow detection is quite important here, for one of the properties of 1 's
190
@ @) @ @@
If
-:--"
I
jLJ
i
I'

I'
0

L-- I
If
-f-o
I
jLJ
i
I'

I'
n

.

-f-o
I
U

i
I'

I
t1

-f-o
I U
"M
1
I'

Ll
00

'--I
----0
0
191
(;

..
0
:N
:N
e
... -
... -
(;
:0
I
!!
in
-
c
'"
u
:;=
"2
bI)
iii
1;;
0
:;

..
"C
"C

]i
f!
'"
a.
-
in
00
III
...

"
bI)
ii:
....
IS
@
@
@
@
@

@)

I I I I
I' AC. 1
I' 1 I' 1 I' 1
AC. AC. AC ,
RZot R201 RZOI R201
.--
..t:-
.--
1::-
r-
J:-:-
r-
1:-

L3

L-J

'--

'--

'--

.-----.

- I--
'"""":J....
L L L L
.--
-- --
ENO __

,
.
Rill L;;-,

I I
1 1 1
14tOJ 1,(1) Ia(l) 1
6
tOI 1
7
(11 1
7
(0)
Figure Ib aBit Parallel Adder (4 Least Significant Bits)
HALF
ADD
COMPLEMENT
AC
CARRY
INITIATE
complement arithmetic is that the sign of the result is incorrect if overflow or underflow
has occurred. Detection of overflow is simply derived from the previous definition of l's
complement overflow. If the result of the sign bit after the half-add is a 0, then the carry
into bit AC" is allowed to set the overflow flip-flop (see Figure 2). Precautions must be taken
to prevent this carry from setting overflow when performing a complement of the AC. This
may be done by gating AC" with a status level to indicate that a'n addition is being performed.
RESET OVERFLOW --1>'-,...-.,.....-1
CARRY INTO ACo --i>C8I--ofSS:l
ACo l1l--+---I
HALF
10 (1J -":>I-...J
Figure 2 Overflow logic
Underflow is detected by implementing the rule that no carry into the sign bit occurs. We
allow the overflow to set on the half-add step if 10 and ACo are both ones. When the carry
into ACo occurs if ACo is a 0 (after half-add), the overflow flip-flop will reset. If there is no
carry into ACo, the overflow flip-flop remains set indicating an ungerflow. The overflow flip-
flop must not be looked at until after the addition is completed (after carries have rippled).
The overflow bit must be reset before the next addition.
The timing for the adder is as follows:
Type
R201
R602
RIll
Type
R203
1. Time between half-add pulse and carry initiate: nsec
2. Time for settling after carry initiate: 560 nsec maximum (EAC initiating a ripple
down the length of the counter)
Adder and Control System
Quantity Description
8 Flip Flop
4 2 Pulse Amplifiers
3 3 Diode Gates
Buffer Register for Incident
Number where required
3 3 Flip Flops
PARTS LIST
193
Type
R201
RIll
Overflow Detection
1 Flip Flop
1 3 Diode Gates (use
spare in adder)
32-POSITION DECODING
APPLICATION
NOTE
A five-bit binary number can be decoded into its 32 permutations by using 32 five-input
NAND gates made from: eleven RIll, five ROOl, and six R002, a total of 22 modules.
Various classical techniques of decoding can be used to achieve different degrees of
simplicity and/or economy. In applying these techniques to the decoding problem the
degree of simplicity and/or economy that may be attained is a function of the module
configurations available, such as packing density, optimum number of inputs per gate,
special matrix configurations of gates, etc.
One of the classical techniques employs 11 matrix. This method may be used to minimize
the number of inputs per gate, but at the expense of increasing the number of gates. In
logic configurations which are input limited, this technique may be the only economical
solution to the problem. In pure diode logic systems (true AND functions) the technique
actually produces more economical results.
In the DEC FLIP CHIP line, the R151 module performs a binary to octal conversion. That
is, given both sides of three flip-flops, the module produces eight mutually exclusive outputs
which represent the eight permutations of the three flip-flops. Figure 1 shows the truth
table for the R151 Binary to Octal decoder.
Truth Table for R151 Binary to Octal Decoder (1 = -3v; 0 = Ov)
Inputs Outputs
H J F E L K
22 21
2
0 1 2 3 4 5 6 7
0 1 0 1 0 1 0 1 1 1 1 1 1 l'
0 1 0 1 1 0 1 0 1 1 1 1 1 1
0 1 1 0 0 1 1 1 0 1 1 1 1 1
0 1 1 0 1 0 1 1 1 0 1 1 1 1
1 0 0 1 0 1 1 1 1 1 0 1 1 1
1 0 0 1 1 0 1 1 1 1 1 0 1 1
1 0 1 0 0 1 1 1 1 1 1 1 0 1
1 0 1 0 1 0 1 1 1 1 1 1 1 0
In addition to the six input lines for the complementary outputs of three flip-flops, there is
one additional input which is a control line. When the control line is at 0 (0 volts), the
decoder is enabled and functions according to the truth table. When at 1 (-3 volts), the
decoder is disabled and all output lines produce a 1 regardless of input. Note that input
pairs (HJ, FE, and LK) represent the 1 and 0 sides of a flip-flop respectively and that the
zeros of the truth table on the output lines may be used for the assertion of each of the
eight states of the three flip-flops.
194
.....
ID
c.n
0 2 3 4 5 6 9 10 11 12 13 14 IS 16 17 18 19 20 21 22 23 24 25 26 27 28 29 31 32
01234567 01234567 01234567 01234567
1 2
0
I 2' I 22 r1 2
0
I 2' I 22 I 1 2
0
I 2' I 22 I r1 2
0
I 2' I 22 I
0-1' 01' 0-1' 0-1' 01' 0-1' 0-1' 01' 0-1' 0-1' 01' 0-1'
ttl t T T t T T

-,.
'3-R1tl"

r--- -
Ir II
o 'I

I
Figure 1 32-Position Decoder
The minimization technique applied in this case to make optimum use of the R151 module
involves breaking the five bit binary number into two groups of three bits and two bits each.
The least significant three bits may be considered one octal digit. If we now consider the
two most significant bits we find that there are only four permutations of these bits. For each
_of these permutations, the least significant three bits may assume eight unique states for
a total of 32 permutations which, of course, is the correct number of combinations of five bits.
To implement this technique, we use four R151's which all have parallel inputs from the
least significant three bits. Thus, there will be four out of the 32 lines at ground for each
of the eight combinations of the least significant three bits. All that remains now is to turn
off three of the four active lines as a function of the most significant two bits. A two bit
decoder made from four two-input gates (1113 R 111) will cause one out of its four output
lines to be at ground for each of the four permutations of the two input bits. Remembering
that the R151 inhibit inputs are enabled by a ground signal, the four unique ground outputs
of the two bit decoder furnish the necessary four disabling signals for the four RI51's.
PARTS LIST
R151 4 Binary to Octal decoder
Decoder R111 2 3 two-input gates
5-Bit Buffer R203 2 3 Flip Flops
196
STEPPING MOTOR DRIVES (TRANSLATORS)
APPLICATION
NOTE
The W042 driver card was designed to handle a variety of high power driving applications,
including stepping motors. Stepping motors come in a variety of types, but the two we will
deal with here are the Superior Electric Sio-Syn (model SS50-1001l and the United Shoe
Machinery Responsyn (model HDUM-16-100-161l. The phasing of the pulses that drive the
steppers is varied. Three different phasing requirements and simple methods for deriving
them are described in this application note.
Most stepping motors have four leads plus a common lead. To make the motors step, a
sequence of pulses on the four leads is required. Three of the desired pulse timings are
shown in Figure 1. To reverse the direction of either motor, simply reverse the sequence of
pulses.
TIME 4 5
LEAD
--.J
I
4
SLO SYN
TIME 2 3 4 5
LEAD 1
2
3
4 1 .... _____ ---'
RESPONSYN - SINGLE PHASE MODE
TIME 3 4
LEAD 1
2
3 1 .... ___ ----'
4
RESPONSYN - DUAL PHASE MODE
Figure 1 Pulse Timing for Stepping Motors (Positive Level means on)
197
All of these timing sequences can be generated with two R201's in a switch tail ring counter
configuration and the diode gate on the W042 driver card (two diodes per driver). The basic
counter is shown in Figure 2.
REV
COUNT SEQUENCE
(FWDI (REVI
T A B A B
R201
O ~ o 0
B
1 ci o 1
0
3 1 1 1 1
4
O 1
1 0
5 o 0 0 0
FWD
Figure 2 Basic Counter
By using the diode gating, the three different sequences shown in Figure 1 can be decoded
from the one counter. Figures 3, 4, and 5 show the pulse sequences and logic diagram for
each configuration. The decoding requirements have been simplified to one input per
driver in the Sio-Syn and Responsyn single phase mode configurations.
B(OI
T FWD REV
A B A B
L2
B(11
" ~ ~ "
, . ~ ~
MOTOR
LEADS 1 1 L2 1 1 L2
4 L3 0 1 L4 1 0
L3 5 o 0
o 0 Ll
AtOI
L4
A( 11
Figure 3 Slo-Syn Configuration
198
T FWD REV
A B A B
IL:]]
L
t
~
L
t
Z o:::::::::QJ
LZ
~
L4
3 [C:=IJ
L3
rr:::=II
L3
4 I:2:::J]
L4
o==Q]
L
Z
5 IL:]]
L
t
~
L
t
A (0)
BIO)
Alt)
BIt) -_+-'----Dlf--.J
Figure 4 Responsyn Single Phase Mode Configuration
Lt
B 101--t)j--nrJ
T FWD
L
z
A B
A I 1l--t)j--Gtl t
" ~ ~ "
MOTOR
Z
LEADS
3
t t L3
L3
4 L4 0 t
5 o 0
B 111--t:l--Uu
A 101 --t:l--Gtl
Figure 5 Responsyn Dual Phase Mode Configuration
199
MOTOR
LEADS
REV
A B
" ~ ~ "
LZ t t
t 0 L.
o 0
The W042 driver board supplies current from -15v to the motor leads. Generally, some
external resistance will be required. The Sio-Syn motor SS50-1001 is rated at 14 v and re-
quires little to no external resistance. The time constant of the motor coils is smaller than
its minimum pulse period and does not require external resistance to decrease it (only to
insure proper motor dissipation).
The Responsyn HDUM-16-100-161, however, is a 4 v motor and has a time constant which
is larger than the minimum pulse period. It requires the external resistance and higher
supply voltage to insure maximum running rate. The values of external resistance depend
upon the motor parameter and coil ::onfiguration. Two such configurations are shown in
Figures 6 and 7. Also, the dissipation requirements of the W042 driver must be obeyed.
See the W042 data sheet for use restrictions.
MI
r
M2
2..-
R.
:042
M3
COMMON
3..-
M4
4..-
Figure 6 External Resistance for Sio-Syn and Responsyn (dual phase mode only)
R.
MI
R.
M2
ro {
R.
...---. COMMON
W042
M3
R.
M4
Figure 7 External Resistance for Responsyn (single phase or dual phase modes)
Stepping Motor Characteristics
Motor
SS50 -1010
SS50 -1009
SS50 -1008
SS50 -1001
SS15(H023
SS150-1009
SS150-1010
SS250-1006
SS250-1002
DC Volts
Sio-Syn Motors
2.0
5.5
8.0
14.0
1.4
2.5
10.0
2.5
9.0
200
Amps/Winding
3.3
1.3
0.85
0.53
8.5
4.0
1.25
5.0
l.55
..
HDUM-16-100-161
Responsyn Motor
4 7.3
Sio-Syn SS150 series:
Static Holding Torque
Resolution
310 oz.-in.
Ipart in 200 (1.8)
Speed 50 s/sec 100 s/sec 150 s/sec 200 s/sec
Torque 156 oz. in. 125 oz. in. 100 oz. in. 62 oz. in.
Responsyn HDUM-16-100-161
Static Holding Torque
Resolution
375 oz. in.
1 part in 800 (0.4)
Speed
Torque
200 s/sec 400 s/sec 800 s/sec
100 oz. in. 100 oz. in. 75 oz. in.
The Sio-Syn motor has a solid. conventional motor rotor, and hence a high inertia.
The Responsyn motor has a thin wall cup rotor with essentially negligible inertia.
Type
R201
W042
Quantity
2
1
PARTS LIST
Description
Flip Flop (Translator)
10amp Driver (Driver/decoder)
201
GENERATION OF
PSEUDO-RANDOM SEQUENCES
APPLICATION
NOTE
In transmitting information digitally, it is often desirable to code it in such a way that a single
bit of data is represented by a burst of pulses. Such techniques are useful in combatting
noise and interference of the type encountered in sonar, radar, and high-frequency com-
munications. Codes having special autocorrelation properties are widely used for this pur-
pose. Receiving systems can be constructed which will produce an analog output whose
shape is similar to the autocorrelation function of the coded signal. This analog output signal
is developed as the code train passes through a filter which is "matched" to the coded
sequence. The signal out of such a filter is usually a narrow pulse similar in shape to the
single bit of information which was encoded originally.
BARKER CODE
One of the earliest codes used for this purpose was the Barker code, which was seven bits
long. It is a bipolar signal as follows:
+1, +1, +1, -1, -1, +1, -1
A digital matched filter for receiving such a code could be implemented as shown in Figure
1. It will be observed that a signal coded in this fashion and having a peak amplitude of E
will be compressed into a single spike having a peak amplitude of 7E.
INPUT
WAVEFORM
INPUT WAVEFORM
SUMMED ANALOG OUTPUT
SHIFT REGISTER OR DELAY LINE
OUTPUT WAVEFORM
Figure 1 Barker Sequence
202
I n addition, various sidelobe signals having an amplitude of -E will be present. The width
of both the main lobe and the side lobes is equal to t, the sUb-interval used in the pulse
sequence. In order to identify the position of the received bit of information in time, it is
necessary to detect it at a threshold above the sidelobe level. The peak-to-sidelobe ratio is
therefore a characteristic of some importance in the choice of a sequence for pulse com-
pression. The Barker Code is one of a family of sequences having such properties. They are
commonly referred to as maximal-length, or merely M-sequences. This reference is to the
fact that a code of length 2n_1 can be generated by using a shift register only n bits long.
Because of the ease with which these codes can be generated and detected, they are be-
coming widely used in communications and control systems.
31-BIT SEQUENCE GENERATOR
Figure 2 shows a five-bit shift register and the necessary control electronics to generate a
sequence which is 31 bits long. It is a slight modification of a ring counter in that it uses
an exclusive OR condition of the state of the third and the fifth flip-flops to determine whether
a 1 or a 0 is to be set into the first flip-flop on each shift operation. The state of the fifth
flip-flop may be used to generate the code shown.
A filter matched to this sequence could be constructed either from a linear tapped delay
line or its digital equivalent, a shift register. In either case, a resistive summing network
could be used to produce an analog output which should be similar in shape to the auto-
correlation function of the sequence. The following module list includes the modules
necessary to construct an N-sequence generator and also a 31-bit shift register to receive
it. The control electronics will include a clock which will sample the incoming waveform
at a rate at least twice the bit rate used in sending the compressed signal.
START
FF

010000
I 0' a a a
2 00100
3 , a a , a
4 0 I a 0 t
5 , a I a a
6 I' a I a
7 0 1 I a ,
. e 0 0 , I a
9 I 0 a , 1
10 t , 0 a I
II , , , a 0
12 I I , I 0
13 , , I , I

ETC. t
Figure 2 M-Sequence Generator
203
R1I1
PARTS LIST
Shift Register
R202 3 2 Flip-Flops
R202 3 2 Flip-Flops
Burst-of-31 R302 1 2 Delays (oneshot)
R602 1 2 Pulse Amplifiers
Matched Digital R202 17 2 Flip-Flops
Filter R40l 1 Variable Clock
204
PART III: LOGIC LABORATORY
205
INTRODUCTION
The DEC Logic Laboratory is a low cost device for use in laboratory training, function
breadboarding, and testing. When used in conjunction with the Logic Laboratory
Workbook, a student who is not completely familiar with digital logic can quickly learn to
construct operating logical networks and understand their function. In addition to the
training function, the Logic Laboratory can be used to breadboard complex algorithms in
order to verify the logic design prior to its inclusion in a system.
The Laboratory is also a very effective tool for testing individual logic modules as is
evidenced by the fact that each DEC field service office uses a Logic Laboratory for
computer module testing and maintenance.
The Logic Laboratory is a completely self-contained system consisting of a power supply,
pulse generator, controls, indicators, mounting hardware, and a basic complement of logic
modules necessary to construct a workingsystem.
Education And Training
Fiexibility is the key to the excellence of the Logic Laboratory as a training device. Far from
a complicated do-it-yourself kit, the Laboratory is a well-conceived and implemented
teaching aid. The Workbook provides a step-by-step approach to building an understanding
of various digital logic functions. With on Iy three module types, the student is taught to use
the operations of NAND and NOR to perform the basic AND, OR, NOT, and EXCLUSIVE OR
(Half add) functions. In the process of constructing these, the student learns by doing, as
opposed to learning by rote.
Experiments contained in the Workbook are designed so that experiments can be short
and simple, or extended by the numerous possibilities suggested by additional problems
contained in the Workbook. The Logic Laboratory is designed to teach universal logic
principles rather than a narrow system, rigidly tied to a company's product line. The logic
symbology used in the Workbook closely resembles the generally accepted standard. These
symbols have been field-tested by leading training directors and found to be easily
understood.
206
207
The Workbook was prepared specifically for use with the Logic Laboratory. It contains 15
graded experiments,or laboratory sessions, each of over three hours duration. No prior
knowledge of electronics or digital logic is required, although the experiments should be
performed in conjunction with a course in logic design or textbook study (a number of
texts are recommended). All basic logic elements and techniques are covered, and all can
be carried out with the basic Laboratory. Each chapter or laboratory session of the
Workbook contains tutorial text and projects to be performed. A considerate amount of
latitude is allowed in the assignment of projects, so that a given chapter can stimulate well
over three hours of experimentation. The Workbook is also arranged so .that various
material may be omitted by the instructor, should he desire a less thorough coverage.
One of the most valuable features of the Logic Laboratory is its AnalogDigital Converter
section. Two chapters are allocated to the techniques of analog to digital (ADC) and digital
to analog (DAC) conversion. Since most of the measuring devices which exist today are
analog in form, an understanding of the techniques and operation of AID conversion
equipment is absolutely necessary. In addition to the data acquisition problem, an
understanding of the conversion of digital values to analog outputs for control and display
(strip charts, meters, etc.) is essential for an overall understanding of data processing.
The following list of chapter titles indicate the scope of the material covered.
l. Binary Numbers 9. Binary Coded Decimal Arithmetic
2. Binary Coded Decimal 10. Code Conversion
3. Basic Digital Circu its - Gates II. Control
4. FlipFlops and DCD Gates 12. Timing
5. From Boolean Equations to 13. Introduction to Analog-Digital
Gating Networks Conversion
6. Boolean Equations and FlipFlops 14. Advanced Studies in
7. Addition
AnalogDigital Conversion
8. Parallel Addition and Subtraction 15. Computer Design.
The Standard Logic Laboratory contains the necessary equipment to perform experiments
1 through 10. The Advanced Logic Laboratory, when added to the Standard, contains the
necessary equipment to perform experiments 11, 12, and 15. The Advanced Logic
Laboratory with AnalogDigital Conversion provides the equipment for experiments 13
and 14.
Breadboarding And Testing
The Logic Laboratory is much more than a training device when used by a competent logic
designer or technician. It is designed and manufactured to the same high quality as DEC
modules and computers as well as being electrically and physically compatible with all
Flip Chip modules. The Logic Laboratory power supply is capable of supplying drive to
about 100 modules. There is no restriction on the size of the system which can be
implemented since additional equipment can be ordered and Logic Laboratories intercon-
nected directly. Many logic problems are complex enough to require assurance beyond
the level of the Boolean equation before the final implementation. There is no substitute
for actually building the system and verifying the logic using the Laboratory. The inherent
flexibility and ease of interconnection modification results in a checkout time which is
faster than building a complete prototype. The Laboratory is also very useful in building
208
Excerpt from Experiment 1 of the Logic Laboratory Workbook
A digital computer is an assemblage of extremely simple circuits. Consider the familiar
elements in the logic laboratory - the toggle switches and push buttons. Examine
these elements in detail. Compare them.
Each has only two states. The switch may be up or down. The button may be depressed
or released. Digital circuits also have two states, a negative voltage level and a positive
level. In the logic laboratory the negative level is -3 volts; the more positive level is
ground.
Note the differences. The button makes contact only when depressed. When released,
it always returns to its original position. The switch, by contrast, always remains where
last positioned. It remembers. I n this same sense, digital circuits are divided into two
classes, those which remember and those which follow.
A BINARY COUNTER
In this experiment, you will study the binary
number system by constructing a counter,
using .the FlipFlop Type R201. The flip-flop
is the circuit equivalent of the toggle switch.
It remembers.
The flip-flop circuit is shown symbolically in
Figure 1. The two outputs are always in
opposite states. That is, if one output is at - 3
volts, the other is at ground and vice versa.
To see this, connect the two flip-flop outpu'ts
to indicator lights. One will be on while the
other is off.
To change the state of the flip-flop, connect
a push-button pulser to each of the direct
inputs. Notice that the flip-flop always re-
members which button was depressed last.
209
GATED {
INPUTS
DIRECT
INPUT
DIRECT
INPUT
GATED
INPUTS
FF
I
"one of a kind" systems for experimentation when the system is used for a single pass
operation.
Some common uses of the Logic Laboratory are listed below. Many of these are described
in detail near the back of this catalog under the heading Application Notes. The uses to
which the Laboratory can be put are without limit. Digital's highly trained field engineers
are always available to assist customers with logic design problems.
Peak Amplitude and Zero
Crossing Measurements
Computer Controlled
Communications Systems
Generation of PseudoRandom
Sequences
Digital Filtering of Analog Signals .
Highly Accurate Timing Systems
Stepping Motor Drives (Translators)
Teletype Send Receive Logic
RealTime Computer Interfaces
210
General Purpose Digital Clocks
Analog Signal Multiplexing
Decoding
Sequence Control
Data Acquisition
Hybrid Computation
Pulse Train Techniques
Typewriter Drive Logic
VoltageTime Conversion
Digital Spectrum Analyzers
Elapsed Time Measurements
Basic Equipment Lists
STANDARD LOGIC LABORATORY
The following equipment is sufficient to study the basic principles of digital logic, as set
forth in Experiments 1 through 10 in the Logic Laboratory Workbook. Total price is $886.20
1 H901 Module Mounting Panel
1 H902 Indicator-Switch Panel (complete
with W052 module)
6 R201 Flip-Flops
3 R121 Four NOR Gates
1 R122 Four NAND Gates
4 911-2" Box of 2" Patch Cords
1 7000 Power Supply and Signal Generator
Panel (with modules)
4913 Mounting Rack
ADVANCED LOGIC LABORATORY
Added to the standard equipment above, the following items permit the study of control,
timing, and computer design, as covered in Experiments 11, 12, and 15 of the Workbook.
Total price of added equipment is $321.50
1 H901 Module Mounting Panel
3 R201 Flip-Flops
1 R121 Four NOR Gates
2 R302 Two Delay One-Shots
1 R602 Two Pulse Amplifiers
ADVANCED LOGIC LABORATORY WITH ANALOGDIGITAL CONVERSION
The converter listed below opens up the large and varied field of interfacing digital logic
with external analog equipment. As described in Experiments 13 and 14 of the Wookbook,
analog-to-digital converts can convert analog outputs from thermometers, pressure gauges,
flowmeters, etc. into digital values for processing, and digital-to-analog converters can
convertthe resulting digital numbers into analog signals to control devices such as motors,
oscilloscopes, and valve actuators. Converter price $143.00
H903 AnalogDigital Conversion Panel (Provides 0 to A and A to 0 Con-
version. Complete with modules)
211
I
POWER SUPPLIES AND INPUT PANEL
TYPES 7000, 7000A
LOGIC
LABORATORY
COMPONENTS
The 700D is a combination power supply and input panel. The input devices
include a dial, three push buttons and pulsers, and a clock. The power supply
can drive approximately ten Type H901 Panels of R-series FLIP CHIP logic,
or up to 10 basic Logic Laboratories. Details on the pulser circuits are given
in the module description of the Type W501. Details on the clock are given
on the module description of the Type R401. The 7000 is a combination of
the 900 Control Panel plus an H701 Power Supply.
ELECTRICAL CHARACTERISTICS
INPUT VOLTAGE: Power Supply-700D: 115v, 60 output and within +9.2 and +11.5v for the +10v
cps, 700DA: 112.5, 123.5, 195, 220, 235v,' 50 cps. output, when load varies from minimum to maximum
See "50 cps power." and line voltage varies from 105 to 125 vdc.
OUTPUT VOLTAGE: 1Ov, -15 vdc, floating P-P RIPPLE: Less than O.6v. for +1Ov output; less
OUTPUT CURRENT: +lOv: 0 to 0.4 amp; -15v: than 0.6v for -15v output (20% more ripple on
0.5 amp to 3 amp. the 50-cps type).
LINE AND LOAD REGULATION: The output voltage LINE FREQUENCY TOLERANCE: 2% of line fre-
remains between -15.5 and -16.5v for the -15v quency
PANEL WIDTH: 19 in.
PANEL HEIGHT: 5-3/16 in.
DEPTH: 12 in.
MECHANICAL CHARACTERISTICS
Socket
FINISH: Chassis: Chromicoat; Panel: DEC Blue
POWER UNIT CONNECTION: Amphenol 160-5
212
POWER OUTPUT CONNECTION: Hayman lab Ter-
minals which fit AMP "Faston" receptacle series
250, part 41774 or Type 914 Power Jumpers.
7000 --'- $323.00
700DA - $343.00
CONTROL PANEL
TYPE 900
LOGIC
LABORATORY
COMPONENTS
The Type 900 Control Panel uses the same chassis and input controls as
the 7000 but does not contain a power supply. It is designed for multi-
student installations of DEC Logic Laboratory units powered by a single 7000
Power Supply. Connections are available on the rear of the 900 to accept
power from the master 7000 Power Supply.
The input devices include a dial, three pulsers with pushbuttons, and a
variable clock. Details of the Type W501 Pulser Circuits and the Type R401
Clock can be found in the module description section of the catalog.
PANEL WIDTH: 19 in.
DEPTH: 12 in.
FINISH: Chassis: Chromicoat; Panel: DEC Blue
AC POWER CONNECTION: Amphenol 160-5 Socket
213
is provided, but not wired
DC POWER INPUT CONNECTION: Hayman Tab Ter-
minals which fit AMP "Faston" receptacle series
250, part 41774 or Type 914 Power Jumpers.
900 - $214.25
MISCELLANEOUS ACCESSORIES
TYPES H901, 911
H901 PATCHCORD MOUNTING PANEL
LOGIC
LABORATORY
COMPONENTS
This panel provides up to ten FLIP CHIP modules with power and patch
connections. Space between patching sockets allows insertion of logic
diagrams. Logic diagrams are printed on all FLIP CHIP .module data sheets.
More permanent plastic diagrams are available for those modules required
to complete the logic laboratory workbook experiments.
PANEL WIDTH: 19 in.
DEPTH: 61f2 in. with FLIP CHIP modules inserted
FINISH: DEC ~ I u e
POWER INPUT CONNECTIONS: Tabs which fit AMP
"Faston" receptacle series 250, part 41774.
911 PATCH CORDS
DEC Type 911 Banana-Jack Patchcords are supplied
in color-coded lengths of 2 in. (brown), 4 in. (red),
8 in. (orange), 16 in. (yellow), 32 in. (green), and
64 in. (blue). Patchcords may be stacked to permit
multiple connections at any circuit point on the
graphic panels of the DEC H901 Mounting Panel.
The cords are supplied in snap-lid plastic boxes of
ten for handy storage.
H901 - $82.50
911 - $9.00/pkg.of 10
214
INDICATOR SWITCH PANEL
TYPE H902
LOGIC
LABORATORY
COMPONENTS
The H902 Panel provides facilities for control and observation of the Logic
Laboratory. It contains eight indicator lights and a lamp driver module, eight
toggle switches and four potentiometers. Connections to these devices are
made with Type 911 Stacking Banana-Jack Patchcords.
INDICATORS: Indicators inputs accepts signals of -3v and ground. A ground
input lights the indicator. If the input is returned to -3v or open circuited,
the indicator will not light. The load is 1 rna.
TOGGLE SWITCHES: The toggle switches are single pole, single throw with
a logic diagram to show the open and closed positions.
POTENTIOMETERS: The potentiometers are 20,000 ohms. They may be used
to control the frequency of delay one-shots or clock circuits" in "the H901
Mounting Panel.
PANEL WIDTH: 19 in.
PANE:L HEIGHT: 5-3/16 in.
DEPTH: 6112 in.
FINISH: DEC Blue
H902 - $112.80
215
ANALOG-DIGITAL PANEL
TYPE H903
LOGIC
LABORATORY
COMPONENTS
This panel provides facilities for experimenting with analog-digital techniques.
It contains a 4 bit variable output D-A converter and a comparator circuit.
Also includes two 8 volt panel meters and a potentiometer for producing 0
to ...,.8v test signal. Connections to these devices are made with Type 911
Stacking Banana-Jack Patchcords.
D-A CONVERTER ZERO OFFSET: :+0.4v or less
LINEARITY: 3% of full scale
ALL ONES OUTPUT (FULL SCALE): adjustable from
-7v to -8v driving 3000 ohm load
D-A CONVERTER OUTPUT IMPEDANCE: typically
less than lOOn
COMPARATOR OFFSET: 0.2v or less
COMPARATOR INPUT CURRENT: typically less than
1001'a.
INPUT VOLTAGE OPERATING RANGE: 0 to -lOv
INPUT: D-A converter inputs each require 1 rna at
ground. No load at -3v.
OUTPUT: D-A converter output may be shorted to
ground accidentally without harm. Comparator out-
put'supplies up to 8 ma at ground; 1 ma at -3v.
Because the inputs may pass through the switching
region slowly or. hesitantly in most A-D converter
applications, the comparator output transition is not
suitable for driving DCD gate pulse inputs.
POWER: +10 v(A)/8 ma; -15 v/30 mao
PANEL WIDTH: 19 in.
DEPTH: 61f2 in. with FLIP CHIP modules inserted
216
FINISH: DEC Blue
H903 - $143.00
MISCELLANEOUS ACCESSORIES
TYPES 4913, 914
4913 MOUNTING RACK'
LOGIC
LABORATORY
COMPONENTS
The 4913 Mounting Rack provides support for a 700D Power Supply and up to
four H901 Patchcord Mounting Panels, for a total of up to 40 FUP CHIP
modules ready to be patched together for experiments. It may also be used
to mount general purpose mounting panels such as the 1943 for use with the
H700 Power Supply. The power supply must be mounted at the bottom for
stability.
Height: 26V4 in.
Threads for mounting panels: 10-32
914 POWER JUMPERS
For interconnections between power supplies, mounting panels, and logic lab.
Panels these jumpers use AMP "Faston" receptacles series 250. Specify
914-7 for interconnecting adjacent mounting panels,or914-19 for other runs
of up to 19 inches. 914-7 contains 10 jumpers; 914-19 contains 5.
217
4913 - $47.00
9147 - $4.00
91419 - $4.00
218
LIJ
Z
:::;
o
LIJ
~
ORDER FORM
DIGITAL
LOGIC
LABORATORY
WORKBOOK
/
SEE PAGE 182 FOR
DETAILED DESCRIPTION
o -------------------------------------------------------------------------- FOLD HERE .------------------------------------------------------------------------
(!l
z
9

I-
::l
()
Technical Publications Dept.
146 Main Street
Maynard, Mass. 01754
GENTLEMEN: Please send me __ copy(s) of The DIGITAL LOGIC LABORATORY
WORKBOOK B250 at $5.00 per copy postpaid.
Enclosed is my check or money order for $, ______ _
Name, _______________ Title __________ _
Company _________________________ _
Street ____________________ ~ ~ ____ _
City ___________ ,State ______ ,Zip Code ____ _
Type of Business ______________________ _
o I would like a demonstration of the Logic Laboratory. Please have a salesman call.
219

EUSINESS REFLY MAIL
NO POSTAGE STAMP NECESSARY IF MAILED IN THE UNITED STATES
Postage will be paid by:
FIRST CLASS
PERMIT NO. 33
MAYNARD. MASS.

TECHNICAL PUBLICATIONS DEPT.
146 MAIN STREET
MAYNARD, MASS. 01754
220
PART IV: HARDWARE, OCTAIDSAND PANELAIDS
221
INTRODUCTION
Digital manufactures a complete line of hardware accessories in support of its module
series. Module connectors are available for as few as one module and as many as 64.
A complete line of cabinets is available to house the modules and their connector
blocks, as well as providing a convenient means for system expansion. Power supplies
for both large and small systems as well as marginal check and reference supplies are
also available.
A major new addition to the hardware line is the H20! Core Memory. The H20! gives
the logic designer the capability of designing a complete random access storage system
for off line storage, programming automatic equipment. or controlling of a general-
purpose computer.
Coupled with the recent additions to the hardware line, Digital has made every effort
to maintain or improve the high standards of reliability and performance of its present
line. Through the availability of a wide range of basic accessories, DEC feels that it is
offering the logic designer the necessary building blocks which he requires for complete
system design.
MARGINAL CHECKING
The use of variable output power supply Type 786 and marginal check switches on the
mounting panels allow preventive maintenance routines to be established on systems that
must not become inoperative without warning, even after years of use. Varying the power
supply voltages is an excellent way to determine the safety margins and can also help in
pinpointing a logic design or wiring mistake that causes baffling sporadic errors with'
normal voltage applied. -All systems built by Digital are tested under conditions of varying
+10 and -15 supply voltages before shipment, and standard practice on large computers
is to make margin checks at regular intervals, to detect any deteriorating components
before they threaten an unscheduled shutdown.
50-CYCLE POWER
Because of the demand for Digital's products in areas where l15-v, 60cps power is not
available, each of the power supplies with a -frequency-sensitive regulating transformer is
also available in a multi-voltage 50-cps version. All 50-cps supplies have the same input
connections. The line input is on pins 3 and 4. Jumpers should be connected depending
on the input voltage. _ These connections are shown below along with a schematic.
WIRING HINTS
These suggestions may help reduce mounting panel wiring time. They are flot intended to
replace any special wiring instructions given on individual module data sheets or'in appli-
cation notes. For fastest and neatest wiring, the following order is recommended.
222
HARDWARE - INTRODUCTION
(1) All power wiring (pins A, B, and C) and any horizontally bussed signal wiring. Use
Horizontal Bussing Strips Type 932.
(2) Vertical grounding wires interconnecting each chassis ground with pin C grounds.
Start these wires at the uppermost mounting panel and continue to the bottom panel.
Space the wires 2 inches apart, so each of the chassisground pins is in line with one of
them. Each vertical wire makes three connections at each mounting panel.
(3) All other ground wires. Always use the nearest pin C above the pin to be grounded,
unless a special grounding pin has been provided in the module.
(4) All signal wires in any convenient order. Pointtopoint wiring produces the shortest
wire lengths, goes in the fastest, is easiest to trace and change, and generally results
in better appearance and performance than cabled wiring. Pointto-point wiring is
strongly urged.
The recommended wire size for use with the H800 Mounting Clocks and 1943 mounting
panel is 24 for wire wrap, and 22 for soldering. The recommended size for use with H803
block is =30 wire. Larger or smaller wire may be used depending on the number of
connections to be made to each lug. Solid wire. and a heat resistant spaghetti (Teflon)
are easiest to use when soldering.
Adequate grounding is essential. In addition to the connections between mounting panels
mentioned above, there must be continuity of grounds between cabinets and between the
logic assembly and any equipment with which the logic communicates.
When soldering is done on a mounting panel containing modules, a 6v (transformer)
soldering iron should be used. A 11O-v soldering iron may damage the modules.
When wire wrapping is done on a mounting panel containing modules, steps must be taken
to avoid voltage transients that can burn out transistors. A battery- or air-operated tool is
preferred, but the filter built into some lineoperated tools affords some protection.
Even with completely isolated tools, such as those operated by batteries or compressed air,
a static charge can often build up and burn out semiconductors. In order to prevent
damage. the wire wrap tool should be grounded except when all modules are removed from
the mounting panel during wire wrapping.
AUTOMATIC WIRING
Significant cost savings can be realized in quantity production if the newest automatic
wiring techniques are utilized. Every user of FLIP CHIP modules benefits from the ex
tensive investment in high'production machinery at Digital, but some can go a step further
by taking advantage of programmed wiring for their FLI P CH I P digital systems.
While the breakeven point for hand wiring versus programmed wiring depends upon many
factors that are difficult to predict precisely, there are a few indications:
1. One-ofakind systems will probably not be economical with automatic wiring, even when
the size is fairly large; programming and administrative costs are likely to outweigh
savings due to lower costs in the wiring itself.
223
2. At the other end of the spectrum, production of 50 or 100 identical systems of almost
any size would be worth automating, not only to lower the cost of the wiring itself but also
to reduce human error. At this level of volume, machinewired costs can be expected to
be less than the cost of hand wiring.
3. For two to five systems of several thousand wires each, a decision on the basis of secon
dary factors will probably be necessary: ease of making changes, wiring lead time, reo
liability predictions, and availability of relevant skills are factors to consider.
The GardnerDenver Corporation, and Digital can supply further information to those
interested in programmed wiring techniques. At Digital, contact the Module Sales Manager,
Sales Department.
COOLING OF FLIP CHIP MODULES
The low power consumption of Rseries modules results in a total of only about 25 watts
dissipation in a typical 1943 Mounting Panel with 64 modules. This allows up to six panels
of Rseries modules to be mounted together and cooled by convection alone, if air is allowed
to circulate freely. In higherdissipation systems using modules in significant quantities
from the W, B, and A series, the number of mounting panels stacked together must be
reduced. For example, no more than three panels of Bseries modules may be mounted
together without forcedair cooling. In general, total dissipation from all modules in a
convectioncooled system should be 150 watts or less (about 9amp total current at -15v).
The regulating transformers used in most DEC power supplies have nearly constant heat
dissipation for any loading within the ratings of the supply. Power dissipated within each
supply will be roughly equal to half its maximum rated output power. If power supplies are
mounted below any of the modules in a convectioncooled system, this dissipation must be
included when checking against the 150 watt limit.
224
STANDARD MODULE SIZES
SINGLE- WIDTH FLIP CHIP MODULE
CONDUCTIVE COMPONENT LIMIT 11/32 0.056"
+ NONCONDUCTIVE COMPONENTS 3/8 max.
~ I
t ~ S'
1/16 MAXIMU! HEIGH: GOLD-PLATED CONTACTS
OF SOLDERED
COMPONENT LEADS ETCHED WIRING SURFACE
SINGLE-HEIGHT FLIP CHIP MODULE
4--.812-.j
I
=,
0
=8
=c
=0
=E
C::=::=;F
=H
=J
2 ~ 1 6
=,
2.240
=L
="
~
=N

I
=p
=.
=s
~
=,
=u
0
=v
.050-
... 687-1
~
.075
5 ~ 6
5h
225
DOUBLE-WIDTH FLIP CHIP MODULE
CONDUCTIVE COMPONENT LIMITS 13/
16
27/32 max.
:::i::=,1 =-=--
NONCONDUCTIVE
COMPONENTS
CONTACTS tl
DOUBLE-HEIGHT FLIP CHIP MODULE
r---.050
.075
.812...j
=AA
=!==1I;=1
=.D

T c:====:J Af
=AH
=AJ
=AK
=AL
=AM
=AN
.370 .140

T
BA
=BB
=oc
=BO
=B'
=B'
=BH

=BL
=BM
=BN
=BP
=BR
=BS
=Br
=BU
=BV
o
o
o
o

226
ADDRESS:
BITS:
CORE MEMORY
TYPE H201
HARDWARE
ACCESSORIES
The H201 Core Memory can store 4096, 13bit words. By combining with
other modules, a complete random access memory storage system with a
cycle time of 8 I"s can be built. This system may be used to provide off.line
data storage, program automatic equipment, or control a generalpurpose
computer.
SPECIFICATIONS
4096 RISE TIME:
0.5 I's nominal
13
u V ~ (UNDISTURBED "I"): 50 mv typical
ORGANIZATION: 4 wire, 3D
d
V
Z(DlSTURBED "0"): 8 mv typical
PEAKING TIME: 0.7 us typical
SELECTION: 2 diodes per line, 8x8
PHYSICALLY: Occupies 24 module 10
"X" matrix, 8x8 "Y"
cations, 4 vertically by
matrix
6 horizontally. Note
1/2 SELECT CURRENT: 200 ma nominal mounting panels must
be aligned to permit
SELECTION PULSE WIDTH: 2 lis nominal
core to plug in.
H201 - $2,000.00
227
I
POWER SUPPLY
TYPE H704
15. VOLTS
HARDWARE
ACCESSORIES
The Type H704 Power Supply ( 15v) is a precision unit designed to supply
the voltage requirements for up to six operational amplifier modules. It is an
all silicon modular supply capable of delivering 400 m ~ on both outputs. Re
mote sensing terminals are provided.
The H704 is supplied with all the hardware required for computer cabinet,
1943, or H900 mounting. When theH704 is installed in a 1943 or H900
Mounting Panel, it takes the place of two module connector blocks.
DIMENSIONS:34 x 3% x 5 in. height
CONNECTIONS: All input-<lutput wires must be
soldered to octal socket at the ,tJase of the power
sopply.
OPERATING TEMPERATURE: -20 to +71 C am-
bient
INPUT VOLTAGE: 105 to 125 vac; 47-420 cps.
OUTPUT VOLTAGE: floating 15v at 400 ma
OUTPUT VOLTAGE ADJiJSTMENI:1v each output
REGULATION: 0.05% line, 0.1% load for both volt-
ages
RIPPLE: 1 mv rms max for both outputs
OVERLOAD PROTECTION: The power supply is capa-
.ble of withstanding output short circuits indefinitely
without being damaged.
228
-15V SENSE
H704 - $200.00
POWER SUPPLIES
TYPES H701, H701A, 7B2, 7B2A
+10, -15 VOLTS
HARDWARE
ACCESSORIES
TYPE H701 DESIGNED FOR PLENUM
OR COMPUTER CABINET MOUNTING
TYPE 782 DESIGNED FOR 19 INCH RACK MOUNTING
The 782 and 782A power supplies are ruggedly built, low cost units that fit
into a standard 19inch rack. TheH701 and H701A are identical to these units,
except they can be mounted on a chassis or panel in applications where space
is added to an existing device. The basic supply can be mounted in various
configurations and is identical to the power supplies used in models 7000
and H900. The Types 782A and H701A are Power Supplies with 50 Hertz
transformers.
INPUT VOLTAGE: H701: 115 v 60 cps. H701A:
112.5, 123.5, 195, 220, 235 v, 50 cps. See "50
cps power"
OUTPUT VOLTAGE: +10 v, _15 vdc, floating
OUTPUT CURRENT: -15 v: lf2 to 3 amp; +10 v: 0
to 0.4 amp.
LINE AND LOAD REGULATION: The output voltage
remains between -14.5 and -16.5 v for the -15
output, and within + 9.2 and +11.5 v for the +10
output, when load varies from minimum to maximum
and line voltage varies 10%.
p.p RIPPLE: Less than 0.6 v for +10 output. Less
than 0.6 v for -15 output; 20% more ripple on the
50-cps type.
LINE FREQUENCY TOLERANCE: 2% of line fre-
quency.
LENGTH: 8"
WIDTH: 4-15/16"
HEIGHT: 5%"
FINISH: Chromicoat
229
POWER CONNECTIONS: Screw terminals are pro-
vided on transformer for input power connections.
Output power connections are made via tab termi-
nals which fit the AMP "Faston" receptacle series
250, part #41774 or Type 914 power jumpers. All
required mounting hardware is supplied with this
unit.
H701 - $116.00
H701A - $136.00
782 - $128.00
782A - $148.00
POWER SUPPLIES
TYPES 728, 728A
+10, -15 VOLTS
HARDWARE
ACCESSORIES
DESIGNED FOR COMPUTER CABINET MOUNTING
~ 8.0

~ 7.0
~ :::
1
4
.
0
u3.0
2.0
I
I
I
I
I
I
--1------
I
1.0 2.0 3.0 4.0 !l0 6.0 70 80 9.0
-15\1 CHANJlEL CURRENT 'AUPS)
, E+'OY
]
' ~ ~ : ~ O N
{YEL.lDW1
2
. _I'V
{BLUEJ
The Types 728 and 728A (+10, -15 v) Power Supplies are capable of with
standing wide line and load variations for general system use. When used
singly, the 1O-v channel can supply 0 to 7.5 amp, or the 15-v channel can
supply 1.0 to 8.5 amp. The 728 Power Supply is electrically identical to the
783 but is made on a shorter chassis specifically designed for mounting on
the plenum door of a DEC computer cabinet.
INPUT VOLTAGE, 728: 115 v, 60 cps, 728A: 112.5,
123.5, 195, 220, 225 v, 50 cps. See "50 cps power"
OUTPUT VOLTAGE: +10 Y, -15 vdc, floating.
OUTPUT CURRENT: 1) When only one output is
loaded: +10 v: 0 to 7.5 amp -15 v: 1.0 to 8.5 amp.
2) When both outputs are loaded: +lOv: Oto 7 amp"
-15 v: 1.0 to 8.0 amp." At least 1.0 amp must be
drawn from the -15Y channel to assure proper load
regulation.
remains between -14.5 to -16.5 v for the -15 v
channel and within +9.5 to +11.5 v for the +10 v
channel, when load varies from minimum to maxi
mum and line voltage varies from 105 to 125 vac.
p.p RIPPLE: Less than 0.7 v for +10 v output; less
than 0.7 v for -15 v output (20% more ripple on
the 50 cps type).
LINE FREQUENCY TOLERANCE: 2% of line fre
quency.
"The sum of the output currents is limited by the
following equation: 5 (1'0) + 6(1'5) = 53 (see Figure).
PANEL WIDTH: 16% in.
PANEL HEIGHT: 8% in.
DEPTH: 5% in.
FINISH: Chromicoat.
230
POWER INPUT CONNECTION: Screw terminals on
transformer. .
POWER OUTPUT CONNECTION: Heyman tab termi
nals to fit with AMP "Faston" receptacles series 250,
part 41774 or Type 914 power jumpers.
728 - $240.00
728A - $260.00
HARDWARE
ACCESSORIES
POWER SUPPLIES
TYPES 783,783A
+10, -15 VOLTS
DESIGNED FDR 19 INCH RACK MOUNTING
~ 7.0
!! 6.0
"
~ '.0
B 4.0
d
! 3.0
5 2.0
1.0 2.0 3.0 40 !5.0 6.0 7.0 8.0 9.0
-t5V CHANNEL CURRENT !AMPSI
The Type 783 Power Supply (+10, -15 v) is a simple, rugged supply capable
of withstanding wide line and load variation for general system use. The graph
above shows the permissible region of operation when both outputs are used.
When used singly, the 10v output can supply 0 to 7.5 amp, or the 15-v out-
put can supply 1.0 to 8.5 amp. It is designed for mounting in a standard
19-in. rack. The Type 783A is a 783 Power Supply with a 50-cps transformer.
INPUT VOLTAGE: 783: 115 v, 60 cps. 783A: 112.5,
123.5, 195, 220, or 235 v, 50. cps. See "50 cps
power"
OUTPUT VOLTAGE: +10 v, -15 vdc, floating.
OUTPUT CURRENT: 1) When only one output is
loaded: +10 v: 0 to 7.5 amp, -15 v: 1.0 to 8.5
amp; 2) When both outputs are loaded: + 10 v: 0 to
7.0 amp", -15 v: 1.0 to 8.0 amp. At least 1.0 amp
must be drawn from the -15 v channel to assure
proper load regulation.
remains between -14.5 and -16.5 v for the "':15 v
output and within +9.5 and +11.5 v for the +10 v
output, when load varies from minimum to maximum
and line voltage varies from 105 to 125 vac.
pop RIPPLE: Less than 0.7 v for +10 v output. Less
than 0.5 v for -15 v output. (20% more ripple on
the 50-cps type.)
quency.
'The sum of the output currents is limited by the fol-
lowing equation: 5(1,.)+ 6(1,,) - 53 '
PANEL WIDTH: 19 in.
PANEL HEIGHT: 8% in.
DEPTH:5,a in.
FINISH: Chromicoat
231
POWER INPUT CONNECTION: Screw terminals on
transformer.
OUTPUT POWER CONNECTION: Heyman tab termi-
nals designed to mate with AMP "Faston" recep-
tacles series 250, part #41774 or Type 9144 power
jumpers.
783 - $240.00
783A - $260.00
POWER SUPPLIES
TYPES 786, 786A
VARIABLE 0 TO 24 VOLTS
HARDWARE
ACCESSORIES
DESIGNED FOR MARGINAL CHECK APPLICATIONS
The 786 is a floating variable.power supply for mounting on a standard 19 in.
rack. A resonant input transformer isolates the output from line voltage vari-
ations, while the variable transformer and 0 to 30 v meter allow precise con-
trol of output voltage. Although designed for marginal checking of digital
systems, this rugged supply will fill many other laboratory needs.
INPUT VOLTAGE: 786: 115 v, 60 cps; 786A: 112.5,
123.5, 195, 220, or 235 v, 50 cps. See "50 cps
power!'
OUTPUT VOLTAGES: 0 to 24 vdc continuously
variable.
MAX. OUTPUT CURRENT: 2.5 amp.
LINE REGULATION: 2% for input variation 10%.
LOAD REGULATION: Maximum 3.0 v drop at 20 v
going from no load to full load.
RIPPLE: :0; 1.0 v, p-p at 20 v and 2.5 amp (20%
more on 50-cps modell.
quency.
OVERLOAD PROTECTION: 4-amp fuse accessible
from front panel.
PANEL WIDTH: 19 in ..
DEPTH BEHIND PANEL: 5 in.
FINISH: Chromicoat
POWER INPUT CONNECTION: Screw terminal strip
provided on panel.
POWER OUTPUT CONNECTION: Barrier strip with
screw terminals and tabs which fit AMP "Faston"
receptacle series 250, part 41774 or Type 914 power
jumpers.
232
786 - $215.00
786A - $235.00
CONNECTOR BLOCKS
TYPES H800-W, H800-F
HARDWARE
ACCESSORIES
This is the a-module molded socket assembly used
in FLIP CHIP mounting panels. Aside from its
function as a replacement part, there may be times
when a special mounting fixture with one or more
WIRE-WRAP TOTAL LENGTH
1",0--------- ----1

WIRE
WRAP
TERMINAL
I
1
1
I
I
I
I
I
I
1
L--?-___ .l
I" .1.

1---------- ,t" ----1
SOLDER FORK TOTAL LENGTH
H800 blocks must be made by a manufacturer who
wishes to fit a few modules into a confined or ir-
regular space. The drawings below show the perti-
nent dimensions.
CLEARANCE FOR COMPONENTS
ON THE MODULE IN THIS SLOT -.I L- 1.:
I "._",,,_, ...
Z 585 r-- COUNT(RBORE
1
/ "" DEEP
---------;.,;
-@ 0375
r----3--t

",,- J

REPLACEMENT CONTACTS TYPES H801-W, H801-F
These contacts are offered in packages of 18 for
replacement purposes. In each package, nine
straight and nine offset contacts are included,
enough to replace all cuntacts in one socket.
233
H80l-W is for wire-wrap connectors; H80l-F is
for solder-fork connectors.
H800F - $8.00
H800W - $8.00
H801F - $4.00
H80l W - $4.00
I
CONNECTOR BLOCK
TYPE H 802
HARDWARE
ACCESSORIES
Ioorf----- 2.594 R E F . - - - - - ~ ~ I
--'-f-
1.000
REF.
L-L
0125 j,!""
REF. ~ ~
-*--
0.1 25 --.J '----
REF. I I
This is a connector block for a single flipchip module like the H800, the H802
can be used to fit a single module in a confined or irregular space. Often the
H802 is used as a connector for a cable at some remote location. The H802 is
only available with wire wrap pins.
H802 - $4.00
234
MODULE SOCKETS
TYPES H803, H804, AND H805
HARDWARE
ACCESSORIES
--:I
-*--
0.125
I 2 I 2 I 2 I 2
A A A A
T
B B B B
C C C C
D D D D.
E E E E
F F F F
H H H H
J J. J J
K K K K
L L L. L.
M M M
.M.
N. N N N
p p p p
R. R R R
S S S S
T T. T T.
U U U U
V V V V
I 2 I 2 I 2 I 2
H804 PIN LAYOUT
The H803 and H804 are8 module sockets used in the
H910 and H911 mounting panels. They can also be
used separately to provide convenient sockets for up
to eight modules. The H803 is a 36pin connector
with the pins forming a 0.125inch staggered grid.
The H804 is the same connector with only half the
pins present (pins A2, B2, etc.) so that it mates with
the standard FlipChip modules. A standard module
will also plug into an H803, with 18 pins contacting
the circuit on the normal side and the additional 18
235
r--
I
I
I
I
I
L __
HB04

.070 MIN. ---t
pins on the back side (A1, B1, etc.) making no con
tact. The blocks have the same physical dimensions I
as the H800 with the exception of pin length. These
blocks are only available with wire wrap pins which
are designed to be wrapped with number 30 wire.
Pin dimensions are 0.025 inches square.
The H805 Package of 18 pins for use as spares and
replacements for those used in H803 and H804. The
package contains nine each straight and offset pins.
H803 - $13.00
H804-$ 9.00
HS05-$ 4.00
POWER CONTROL
TYPE 831
HARDWARE
ACCESSORIES
The Type 831 Power Control Panel features a 2-pole circuit breaker which
provides convenient I-step control and protection for entire systems, including
auxiliary equipment. The panel fits standard 19-in. racks and is finished with
a protective aluminum coating. Available in 4-, 10-, 20-, or 30-amp. capacity.
PANEL WIDTH: 19 in.
236
Space available for mounting other controls and
indicators: 3 in. by 8 in.
831-$51.00
MOUNTING PANELS WITH POWER
TYPES H900,H900A ,H910,H910A
HARDWARE
ACCESSORIES
TYPES H900 ANDH900A
This dual function mounting panel offers a way to build complete digital
systems of up to 32 FLIP CHIP modules into only 54' in. of rack space.
More power is available than is ever likely to be consumed in a 32 module
system; a typical mix of as many as 96 Rand W series modules (one .H900
and one 1943) can be adequately supplied. Power in excess of that required
for 32 modules can be obtained at the terminal block, which is convenient
to the input terminal block on any adjacent 1943 Mounting Panel.
INPUT VOLTAGE: H900: 115 v .. 60 cps. H900A:
112.5, 123.5, 195, 220, 235 v, 50 cps. See "50 cps
power".
. OUTPUT VOLTAGE: +10 v, -15 vdc.
OUTPUT CURRENT: -15 v: 2 to 3 amp; +10 v:
o to 0.4 amp.
remains between-14.5and-16.5 v for the-15v
237
output and within +9.2 and + 11.5 v for the + 10 v
output, when load varies from minimum to maxi-
mum and line voltage varies 10%,'
Pop RIPPLE: Less than 0.6 v for +10 v output; less
than 0.6 v for ~ 1 5 v output (20% more ripple on
the 50-cps I type).
quency.

PANEL WIDTH: 19 in.
DEPTH: 6-3/4 in.
FINISH: Chromicoat
vided on transformer.
MODULES ACCOMMODATED: 32
POWER OUTPUT CONNECTIONS: Barrier strip with
screw terminals and tabs which fit AMP "Faston"
receptacle series 250, part no. 41774 or Type 914
POWER INPUT CONNECTIONS: Screw terminals pro- power jumpers. .
H900, H900A
OPTIONS AND ORDERING
There are two kinds of sockets available for FLIP
CHIP modules: wire-wrap and solder fork. Prewiring
of pins A, B, and C for +10 v, -15 v, and ground
is also optional. The example below shows how to
specify wire-wrap connectors and prewired power:
BASIC PANEL
H900
H9DD w
CONNECTOR
W-for wire-wrap
F -for forked solder
connectors
p
T
PREWIREO POWER
(extra cost option)
Omit P if not desired.
1945-19 HOLD DOWN BAR: Secures moaules for
shipping, or other vibrational environments.
932 BUS STRIP: Simplifies wiring of register pulse
busses, power, and grounds.
TYPES H910 AND H910A
The H910 and H910A are similar to the H900 and H900A mounting panels.
The only difference being that the H910 panels are built from four H803. con .
nector blocks. The panels are then capable of housing 32, 36 pin modules
with sufficient room available for an H701 power supply. The H910 is not avail
able with the solder connectors. However, the prewired power option is avail
able at a small additional charge. Wire wrapping of the pins should be
accomplished with #30 wire.
Electrical and mechanical specifications are identical to those for the H900 .
and H900A with the exception of pin length. Pin length is 0.625" for the H910
and 0.75" for the H900.
238
H900 - $180.00
H900A - $200.00
Option P - $5.00
H910 - $200.00
H910A - $220.00
MOUNTING PANEL
TYPE 1943 I H911
TYPE 1943 MOUNTING PANEL
HARDWARE
ACCESSORIES
The 1943 Mounting Panel houses 64 modules. It is designed for mounting
in a standard 19-in. rack. The mounting panel is finished with an aluminum
conversion coating (Chromicoat). Filter capacitors are included on all power
supply lines.
Available options are solder or wire-wrap connectors, power input via terminal
strip or marginal check switches, and power wiring. The chart below shows
how the options are indicated when ordering;
239

BASIC PANEL
1943
1943 F
-
-r
CONNECTOR
F type
OR
B
p
POWER CONNECTION
B -Power input via terminal block.
80th conventional screw connec-
tions and taper tabs can be used.
W-Wirewrap connectors.
M-Marginal check switches on all
voltage inputs allow selection of
fixed or variable power*,
EXAMPLE: If you require a Type 1943 Mounting
Panel with wire-wrap connectors, marginal check
1943

BASIC PANEL
w
/'
CONNECTOR
MECHANJCAL DIMENSIONS: 19 in. wide; 5-3/16 in.
high; 6-3/4 in. deep. Tabs for power connections
fit AMP "Faston" receptacles, series 250, part
41774 or Type 914 power jumpers.
1945-19 HOLD DOWN BAR: Reduces vibration and
"Additional charge.
switches on the power connection, and prewired
power, you would order:
p

POWER WIRING
POWER CONNECTION
keeps modules securely mounted when panel or
system is moved. Adds 1/2 in. to depth of mount-
ing panel.
932 BUS STRIP: Makes wiring power and register
pulse busses easy.
H911 MOUNTING PANEL
The H911 mounting panel is similar to the 1943. It houses 64, 36 pin con
nectors. Mechanical dimensions are identical to those of the 1943 with the ex
ception of the pin length. Pins on the H803 blocks are 0.626" long while those
on the H800 are 0.75 inches long. The H911 is available with wire wrap pins
only. Marginal check ,and power wiring options are available on the H911.
240
H911B - $151.00
H911BP - $161.00
H911M - $172.00
H911MP - $182.00
932 - $ .60
1943FB - $111.00
1943WB - $111.00
1943FBP - $121.00
1943WBP - $121.00
1943FM - $132.00
1943WM - $132.00
1943FMP - $142.00
1943WMP- $142.00
1945-19 - $15.00
MOUNTING PANEL ACCESSORIES
TYPES H001, H002, 1907
HARDWARE
ACCESSORIES
HOOl PANEL COVER BRACKET
The H001 consists of a pair of Ubrackets that fit
under the mounting screws of a 1943, H900, H910,
or H911 FUP CHIP mounting panel, providing a way
to mount 1907 cover panels. The %" standoff ob
tained makes these cover plates approximately flush
with DEC cabinet edges.
H002 PANEL SETBACK BRACKET
The H002 consists of a pair of U-brackets which are
used to setback a power supply or mounting panel
from a 1907 panel cover. The H002 permits the use
241
of controls on the 1907 cover plate with protrusions.
up to 2".
DEPTH: 3 in. (bridges DEC cabinet mounting rails)
1907 MOUNTING PANEL COVER
The Type 1907 Mounting Panel Cover is designed to
cover the logic and power wiring for the 1943, H900,
H910, or H911 Mounting Panels. The 1907' should
be used with the H001 or H002 Panel set back
brackets and is finished in brown "tweed" or blue.
Dimensions are 51,4 by 19 in. Specify color when or
dering.
H001- $8.00
H002-$8.00
1907-$9.00

INDICATORS
TYPES 4908,4906,4917,4918
4908 PANEL INDICATOR ASSEMBLY
Consists of lamp, spring clip, and terminal. Facili-
tates panel mounting of individual indicator lamps.
Power required is -15 v at 30 rna. Requires a 5/16
in. hole for mounting. Wires can be any length, since
their capacitance is isolated from the logic by a
driver such as W050.
4906 INDICATOR WITH AMPLIFIER
Single indicator lamp with transistor driver. May be
panel mounted in 3A1 in. hole; bulb is replaceable
from the front. Power required is -15 v at 30 rna.
Overall dimensions are 2'h in. x % in. x Y. in. The
input to the transistor driver requires DEC standard
levels or equivalent. The input load is 1 rna at -3 v.
Minus 3 Ii lights the lamp; 0 v turns it off.
4917, 4918 Indicators with Amplifiers
The 4917 is a 9-bit indicator and the 491S is an
ISbit indicator from R series logic without the re:
quirement of additional buffering or clamped loads.
The indicators must be used in conjunction with a
WOIS connector. All 18 pins of. the cable connector
are used for signal leads. Ground and -15v are
brought to the module through quick disconnect
jumpers. Plugging the W018 into a connector slot
with +10v and -15v on pins A and B will not
damage the indicators. A -3v or open circuit turns
the indicator on. A ground input turns the indicator
off. Load at ground is 1 rna. The W018 should be
pl!lgged into the logic mounting and panel and a
W023 should be used on the end of the cable that
connects to the 4917 or 4918.
Power required is 540 rna at -15v (4918). Overall
dimensions are 9 1/2 in. x 15/16 in. Bulbs are on
7/16 centers and each requires a 5/16 in. hole.
All indicators utilize 28-v bulbs which are operated
at 15 v. This provides more than adequate iIIumin
ation and greatly extends the life of the bulbs. When
driving type 4906 from Rseries flipflops, one or
more W002 or W005 clamped loads must be added
to supply the current at -3 v demanded by these
circuits.
242
HARDWARE
. ACCESSORIES
490S-$ 3.00
4906-$ 9.50
4917-$73.00
4918 ~ $96.00
WIRING ACCESSORIES
H810, H811, H812
H8l0 PISTOL GRIP HAND WIRE WRAPPING TOOL
HARDWARE
ACCESSORIES
The type HBIO Wire Wrapping Tool is designed for wrapping #24 or #30
solid wire on Digitaltype connector pins. The HBIO Kit includes the proper
sleeves and bits. It is recommended that five turns of bare wi re be wrapped
on these pins. This tool may also be purchased from GardnerDenver Co.
(GardnerDenver part No. 14H1C) with No. 26263 bit and No. IBB40 sleeve
for wrapping #24 wire. Specify bit #504221 and sleeve #500350 for wrap
ping #30 wire. When ordering from Digital specify the sleeve and bit size
desired for #24 and #30 wire.
H8ll WRAPPING AND TYPE H8l2 UNWRAPPING TOOLS
The Type HBll Hand Wrapping tool is useful for service or repair applications.
It is designed for wrapping #24 solid wire on DEC Type HBOO-W connector
pins. This tool may also be purchased from Gardner-Denver Co. as Gardner-
Denver Part #A20557-12.
Wire wrapped connections may be removed with the Type HB12 Hand Un-
wrapping tool. This tool may also be purchased from Gardner'Denver Co. as
Gardner-Denver Part #500130.
The HBllA and HB12A are equivalent to the HBll and the HB12 except that
the A versions are designed for #30 wire. Both tools may be purchased from
Gardner-Denver directly under the following part numbers: HBI1A A20557-29;
HB12A 50 5084 - (LH).
H810(24) $99.00
HB10(30) $99.00
HBlO(24&30) $150.00
HBll(24) $21.50
HBllA(30) $43.00
H812(24) $10.50
HBI2A(30) $10.50
243

WIRING ACCESSORIES
TYPES 913,932,H820,H825
913 PATCHCORDS
HARDWARE
ACCESSORIES
Slipon patchcords for wirewrap FLIP CHIP mounting panels. Type 913 Grip
Clip Patchcords are available in colorcoded standard lengths of 2, 3, 4, 6, 8,
12, 16, 24, 32, 48, and 64 in. All cords are shipped in quantities of 100 in
handy polystyrene windowtype, snaplid boxes. These patchcords use AMP
Terminal Type 605301.
932 SOLID BUS STRIP
The 932 Bus Strip is designed for use with 1943 and H900 Mounting Panels.
The bus strip fits either wirewrap or solderfork pins and simplifies wiring of
register pulse busses, power and grounds. Length: 16 inches
H820 GRIP CLIPS FOR SLlpON PATCH CORDS
The Type H820 GRIP CLIP is identical to slipon connectors used in the Type
913 Patchcords. These connectors are shipped in packages of 1000 and permit
fabrication of patch cords to any desired length. Up to three GRIP CLIPS may
be stocked on any H800W connector pin. The GRIP CLIPS will take size 2420
awg wire and may also be purchased from Amp, Inc. as Amp part #604772.
H820 - $47.80/1000.
H825 HAND CRIMPING TOOL
The Type H825 Hand Crimping Tool may be used to crimp the Type H820 GRIP
CLIP connectors. Use of this tool insures a good electrical connection. This tool
may also be obtained from Amp, Inc. as Amp part #90084. H825 - $146.70.
244
913 - $18.00 /pkg.of 100
932- $.6b
H820- $47.80
H825 - $146.70
WIRING ACCESSORIES
POWER JUMPERS
TYPE 914
HARDWARE
ACCESSORIES
For interconnections between power supplies, mounting panels, and logic lab.
panels these jumpers use AMP "Faston" receptacles series 250. Specify 9147
for interconnecting adjacent mounting panels, or 91419 for other runs of up to
19 inches. 9147 contains 10 jumpers per package; 91419 contains 10 jump
ers per package.
245'
9147 - $4.00/pkg.
91419 - $4.00/pkg.
SOCKET ADAPTER
TYPE 4912
HARDWARE
ACCESSORIES
FLIP SYSTEM
~ ~
A---A
B
B---C
C---D
D
E---E
F---F
H---H
J---J
K---K
L---L
M---M
N---N
p---p
R---R
S---S
T---T
u---u
V---v
W
y
'------2
This accessory occupies one slot of a system module mounting panel, allow-
ing a FLIP CHIP module to be plugged in. The + 10 v for the FLIP CHIP mod-
ule is taken from pin A of the system module connector. The handle of a FLIP
CHIP module plugged into the adapter projects llh in. beyond a standard
System Module handle. Because components on the FLIP CHIP module project
above the normal limit of components on system modules, some care should
be taken to see that these components do not contact adjacent printed boards
while power is applied.
4912 - $28.00
246
Stack-On Riveting Tool
Type H830
HARDWARE
ACCESSORIES
The H830 is designed for clinching on rivets on the 1951, W992, and W993
blank modules. Fits into any vise and once the module and eyelet are posi-
tioned, a slight tap of the hammer will clinch the eyelet.
The unit has a 3
1
/ 2" height and weighs 1 lb.
H830 - $10.00
247
CABINETS
TYPES CAB-1, CAB-2, CAB-3, CAB-6, CAB-8
HARDWARE
ACCESSORIES
Digital offers a variety of cabinets which can be used
to build up special systems.
The Type CABl can be used alone or in multicabinet
systems where only one control cabinet or indicator
cabinet is desired. The standard cabinet has full
length French doors for access to logic wiring.
The Type CAB2 cabinet is used where many controls
and indicators are required. No French doors are
provided in front.
The Type CAB3 is intended as an expander cabinet
for the PDp8. The cabinet has French doors in the
front above the table top. In addition to use as an
expander cabinet, the CAB3 provides a means of
ready access to the front of the system.
The Type CAB6 can be used with other cabinets or
as a remote indicator cabinet. The French doors give
access to logic wiring. The brushed aluminum, clear
anodized panel is placed at a convenient height for
viewing indicators.
All cabinets are alike with the exception of end
panels and the French door configuration. All cabinets
come, as shown, with fan, fan housing, and filter.
A plenum door for mounting power supplies is pro
vided in the rear behind full length French doors.
Casters are provided for mobility. All cables enter
through an access cutout in the bottom of the
cabinet.
The CAB8A is a freestanding cabinet, with a winged
table with legs. The logic modules are housed be
neath this table and enclosed with short French
doors. The CAB8A is not expandable and, therefore,
DIMENSIONS
Cabinet
not recommended for systems that have multiple
bays bolted together.
The CAB8B is similar to the CAB8A with the excep
tion of the table. The table is rectangular and is
positioned so that other cabinets may be placed ad
jacent to the cabinet. The CAB 1 will normally be
bolted to the left on the CAB8B and the CAB3 will
normally be bolted to the right of the CAB8B.
Cabinets can be factory assembled into multicabinet
groups. Cabinet types can be mixed in one group
except for the CAB2, which has different end panel
and trim details.
The cooling fan built into the bottom of each cabinet
is adequate to ventilate up to 51f.! in. mounting
panels of Bseries FLIP CHIP modules mounted near
the bottom of the cabinet. If the lower dissipation
Rseries modules of W or A series modules make up
a significant portion of the system, more modules
can be installed. Fourhundred watts is the maximum
total power that should be dissipated in all of the
modules mounted in anyone cabinet. The top panel
of each cabinet must be removed when FLIP CHIP
modules are installed, and all side panels and
plenum doors should be closed except during system
checkout.
The price of the first cabinet includes end panels.
The price of each additional cabinet includes the
cabinet joining hardware. Cabinets are shipped as
sembled and on skids with the tables packed
separately.
42 irl.Wide""
27-1/16 in. deep
69a in. high
CABLACCESS
(TYf'12CA8INETS .................
RNOVABlE
END PANEL
SERVICE CLEARANCE
8% in. front
149'8 in. rear
In addition to the 728 power supply, there.are several
other items especially designed for plenum door
mounting. For example, there are the 1946 mounting
248
panel and the 734B marginal checking supply. Your
nearest DEC sales office can supply details.
CABl - $700.00 CAB6 - $800.00
$500.00 $600.00
CAB2 - $700.00 CAB.8A _ $1 10000
$500.00 ' .
CAB3 - $650.00 CAB8B - $1,000.00
CAB 1 CAB2

CAB 3
CAB6 (Inside view)
249
FREESTANDING
CABINETS
250
OCTAIDS AND PANELAIDS
Digital's new OCTAID and PANELAID kits are designed to provide the logic user with an
easy-to-assemble, time-saving group of components to achieve common logic functions,
such as up-down counting, decoding digital-to-analog and analog-to-digital conversion, and
computer interfaces_ Standard FLIP-CHIP modules and connectors are used in conjunction
with special purpose printed circuit interconnectors_
The OCTAID series has up to eight standard FLIP-CHIP modules, and the PANELAID series
has up to 64 modules_ Each kit includes the necessary modules, connectors, and specially
designed printed-circuit, back-panel wiring eliminating the necessity for hand-wi ring_ Since
hand-wiring and trouble shooting are eliminated, a significant reduction in the amount of
manufacturing time can be achieved_
Input/Output Buffer kits are designed to interface between Digital's PDP-8 or PDP-8/S
computers and other OCTAID kits or specially designed systems_ PANELAID kits, in general,
can be interfaced directly to the PDP-8 or PDP-8/S_
251

DIGITAL-ANALOG CONVERTER KITS
KITS 0001 A-F
OCTAID
SERIES
Each converter can have from 8 to 13 bits depending upon the module and input pin com
bination. Each 0001 kit includes two printedcircuit boards (F728 and F843) that are used
for all digitalanalog converter common connections, one H800W connector block, one
A704 reference supply, and the necessary converter modules.
MODULE LOCATION AND TYPES VERSUS BITS (KIT)
BITS (KID
8(A) 9(B) 10(C) 11(0) 12(E) 13(F)
A02 AGOI AGOI AGOI AGOI AGOI AGOI
A03 AGOI AGOI AGOI AGOI AGOI AGOI
A04 AGOI AGOI AGOI AGOI AG04 AG04
B02
- AG04 AG04 AG04 AG04 AG04
B03
- - - AG04 AG05 AG05
B04
- -
- - -
A605
The double height A704 fills locations A01 and 801 for all converters.
01 02 03 04
A A A

0 0 0
, , ,
A
7
0

A

A
0
A

0
oR'
0

,.

0

MODULE LOCATION DIAGRAM (WIRING SIDE)
252
The following table illustrates the digital input pins and analog output pin as specified for
each size of' converter. Bit 0 is designated the most significant bit.
INPUT/OUTPUT PINS FOR CONVERTERS
SIGNAL 8 (A) 9(B) lO(C) 11(0) 12(E) 13(F)
Analog
Output A04K B02K B02K B03K B03K B04K
Bit 0 in A04U B02U B03U B03U B03U B04U
Bit 1 in A04T B02T B02T B03T B03T B04T
Bit 2 in A04V A04U A04U B02U B02U B03U
Bit 3 in A03U A04T A04T B02T B02T B03T
Bit 4 in A03T A04V A04U A04U A04U B02U
Bit 5 in A03V A03U A03U A04T A04T B02T
Bit 6 in A02U A03T A03T A04V A03U A04U
Bit 7 in A02T A03V A03V A03U A03T A04T
Bit 8 in - A02U A02U A03T A03V A03U
Bit 9 in - - A02T A03V A02U A03T
Bit 10 in -
- - A02U A02T A03V
Bit 11 in -
- - -
A02V A02U
Bit 12 in
- - - - - A02T
INPUT: Standard Digital levels of -3 volts or ground are required. Loading for all digital
inputs is 1 ma each at ground. If all inputs on a module are not required, the most signifi
cant inputs should be used and the least significant ones should be left opencircuited.
OUTPUT: The analog output is the equivalent of the digital input. The most negative output
is -10 volts, less the value of the least significant bit. The most positive output is ground.
The offset of the least significant bit may be overcome by adjusting the -10 volt reference
to -10 volts plus the value of the LSB to obtain a full scale value. Output impedance
is 1000 ohms.
D001A ( 8 bits) $377.25
D001B ( 9 bits) _______ $439.25
DOOlC (10 bits) $439.25
00010 (11 bits) $501.25
DOOlE (12 bits) $519.25
DOOlF (13 bits) ______ $597.25
253
BCD UP-COUNTER
KIT 0002
OCTAID
SERIES
The 0002 has four binary-coded decimal (8-4-2-1) up-counters_ Each counter is operated
independently or in series to form a four-digit counter_ All inputs conform to standard
DEC pulse and level requirements_ Each kit includes one printed-circuit board type F723,
one H800 connector block, and two R202 decade modules per decade.
A
B
at 02 03 04
R R R R
2 2 2 2
0 0 0 0
2 2 2 2
R R R R
2 2 2 2
0 0 0 0
2 2 2 2
AOt - 02 DECADE'
A03- 04 DECADE 2
BOI -02 OECADE:5
803-04 DECADE 4
--- ----R2ii2---------------,
INHIBIT
CLOCK C > - ; I - - ~
I
I
L ________ . ______ _
BCD Up-Counter
254
I
I
I
I
I
I
I
_ ...J
8 (I) 8(0)
CARRY
Single Decade __ $ 62.50'
Quad Decade ___ $137.50
BIDIRECTIONAL DECADE COUNTER
KIT 0004
OCTAID
SERIES
The D004 is primarily a one-digit 8-4-2-1 up-down counter that can optionally be used
for decimal decoding and 8421-2421 conversion_ The counter is level-controlled_ All
blqcks are wired for the counting, decoding and conversion functions_ If decoding or
conversion functions are not desired, the associated modules need not be ordered_ The
basic up-down counter kit includes two printed-circuit boards (F861 and F862), four
R201 flip-flops, and one R121 gate_ To add the decimal decoding function, one R151
binary-to-octal decoder and one Rll1 gate are included_ To add the 8421-2421 conversion
function,'" one R121 gate is included_ The following figures show module location; define
the inputs, outputs, and control connections for interconnection to other devices; and
schematically illustrate the completed counter_
01 02 03 04
R R R

2 2 2 2
0 0 0 0
I I I I
R R R R
B
I I I I
2 I 5 2
I I I I
*8421 to 2421 Conversion - Wiring has been provided for 8421 to 2421 conversion for use in a
digital-to-analog converter_ The most and least significant bits (2 - - 1) are derived directly from the
o sides of the up down counter_ The remaining bits (- 4 2 -) are derived from the outputs of two R121
gates connected as NOR gates_
2 4 2 I RESET SETI sm SET3 SET4
UP
,
~
~
,
~ T
z z z z
~
~
~
,
~ ~
~
0
"
a
~
m 4
CARf!Y-
CARRY+
1(0) 1{ll2{O)2t1l4{0}4{1)8(O\B\Il 0 I 2 3 4 5 6 1 8 9
UP/DOWN DECAOE DECODER AND CONVERTER INPUTS, OUTPUTS, AND CONNECTIONS
0004 _____ $123.40
Decoding Option _____ $ 47.00
Conversion Option $ 17.00
255
I
N
U1
'"
o
RISI
IKl3
RI21
1101
1104 RI2l
\
\ 2(1) E
\
\4(1)
\ 1M
\
\
L __________ _
RESET
UP 0 , I, I I I I I I I
PULSE
1102 IN 0 I I I '1
PULSE IN
D
DOWN
-=- kJ::>-L-1 (0)

1102
L-9 Np. K 8(1)
'" L 1(1)
RI21
1101
K
Up DOWN COUNTER, DECODER, AND CONVERTER
CARRY -

SET 8
PDP-SIS INPUT BUFFER INTERFACE
KIT 0005
OCTAID
SERIES
This kit provides selected cable connection sockets, device selection level and pulses,
and data input gates for a 12-bit register. The PDP-8/S supplies a device selection code
and three 1.0_ transfer pulses (lOP I, 2 and 4), which are decoded by a WI03 device
selector module for local device usage.
A W005 clamped load module provides clamping voltage to the AC register input lines.
If, however, there are clamps on these lines in other units it may not be necessary to
use these. The complete kit includes two printed-circuit boards(types F002 and F003) ,
one H800W connector block, the WI03 and W005, and two R123 gates. An inverter
must be provided by the device to be sampled to supply an AC pulse on the Read Buffer.
I. O. command.
01 02
SOCKET
FOR
W023
CABLE
SOCKET
FOR
W023
CABLE
03 04
SOCKET
FOR
RI23
W023
CABLE
RI23 WOQ5
257
D005 __ $122.00

D E K L R 5
A ~ 2
3
t AD
'"
U1
I ~ AK
ex>
.) _ AR
BK
BL
BM
BN
BP
BR
BS
BT
'"
/ I
PDP 8/S INPUT BUFFER INTERFACE
D E K L
~ l O T Wl03
AS 1
rOT 2
4
10 SElECT
.A!::!
~
R
--I
S
I
I
I
I
PDP-SIS OUTPUT BUFFER REGISTER
KIT 0006
OCTAID
SERIES
This kit provides a convenient means of sampling, storing, and transferring to the
outside world the contents of the Central Processor's Accumulator Register during an
lOT instruction.
Four R203 triple flipflop boards are used for storage, making a total of twelve flipflops.
The kit also includes two printedcircuit boards (types FOOl and F004) , one H800W
connector block, and W103 device selector, on RIll gate, and one R202 dual flipflop.
The flip flops are gated by the corresponding AC bits (0 11) being set to the "one"
state (positive conditioning levels) and by a positive enabling pulse (supplied by lOP 4
and the chosen device select code). A strobe pulse input is required to the R202 (pin 0)
when data is strobed out of the R203's.
OUTPUT: The contents of the output buffer register may be sampled or transferred
where desired by taking the bit information directly from the output pins of the flipflops.
R202 output (pin J 804) provides a negative level to indicate new data has been deposited
in the buffer.
01 02 03 04
R R R
2 2 2
0 0 0
, , ,
W
I
0
,
*
R R R
I 2 2
I 0 0
I
,
2
* THE Wl03 IS
A. DOUBLE -
HEIGHT BOARD
259
0006 __ $215.50
I
N
'"
o
_ .......0-"
LOCATION NUMBERS FOR R203's ARE: A02, AD3, A04, B03.
1-
R203
-
A02
- - - - - --l
'EF LM STI
1R203003- - - - ,- --I"i
1 ___ . --- 'I
Jr]KDRO:O----------{)I :.,' .,' I ~
I
I
I
I
I
I
I
r---L- --
W] 03 AO] & 80]
~ H
AK
AR
so
, r
I BF
BH
.J BJ
BK
MB 3-8 BL
aM
aN
BP -<)C:i--,
I
I
I
I
I
I
1--
BR
as
BT
BU
____ !V_ _____ ---l
- - - - --- - - --V'l."".r-'--H,""'---'A /,-r--""
- - - - - - -, - - - - - - -
I
RUPT I
I
~
___ J
OUTPUT BUFFER REGISTER
R2J2 804
From
- - - - - - ~
Data available at
buffer
I
L
I
I
I
I
I
I
I
I
I
I
- __ 1
DUAL a-BIT SHIFT REGISTER
KIT 0007
OCTAID
SERIES
This kit has two independent 8-bit shift registers with the enable inputs of the first
flip-flop left floating. By appropriate connection of the floating inputs, operation as a
shift register, ring-counter, or switch-tail ring counter can be attained. In addition,
the two 8-bit circuits can be operated in series to provide a 16-bit capability. Each kit
includes two printed-circuit boards{ types F005 and F006), one H800 connector block,
and either one or two shift registers. Each shift register includes four R205 flip-flop
modules (8 flip-flops) with the "0" and "1" enable inputs connected, respectively, to
the" 1" and "0" outputs of the preceding stage.
All pulse inputs are bussed together to form a common shift line_ A common direct clear
line is also brought out for simultaneous clearing of all flip-flops in the shift register.
Loading a "1" into the shift register is accomplished by grounding any enable input at
least 400 nsec prior to initiation of a load pulse. Individual load pulses are also bussed
together to form a common load or strobe input. All pulse and level inputs must conform
to the standard DEC configuration_
01 02 03 04
R R R R
2 2 2 2
0 0 0 0
5 5 5 5
R R R R
2 2 2 2
0 0 0 0
5 5 5 5
MODULE LOCATION, DIAGRAM (WIRING SIDE)
261
Single _ $133.00
Dual __ $249_00
'"
en
'"
fAo""i------------- -- -TA02- - - - -- ---- - --TAQ3- --- - -- - - - - -- rAr)4 - - - - -- -- ---- -...,
[R20.5 IR205 IR20S ["205 I
I I I I
I I ,
T
I
'----------00---":
I
DIRECT II I FF I II LI FF 111'1 I FF 1111 FF III
ENABLEI
LEVE.LSI
LO/.l.n
I
I
I
L ___ _
,
_______ L_
I
I
I
I
I I
~ : [
_______ "- _____________ .L _____________ J
D U A ~ 8-BIT SHIFT REGISTER
Note: Loading for registers #1 and #2 are identical. For location of inputs
and outputs for register #2, substitute row B in place of row A.
DUAL a-BIT UP-COUNTER
KIT 0008
OCTAID
SERIES
This kit includes two independent eight-bit up-counters (four R202 flip-flops for each
counter), two printed,circuit boards (types FOO? and F008) , and one H800 connector
block_ Each counter can be operated independently, or should more than eight bits be
required, the counters can be connected in series to provide a 16-bit counter_
Each counter has an independent direct clear, count enable, and pulse or count input
line_ In addition, each bit of either counter can be individually set so that the counter
can start counting at some preselected number other than zero_ All pulses and levels
should conform to DEC standards.
02 03 04
R R R R
A
2 2 2 2
0 0 0 0
2 2 2 2
R R R R
B
2 2 2 2
0 0 0 0
2 2 2 2
MODULE LOCATION DIAGRAM
263
Single _$117.00
Dual __ $217.00
N
~
SET
INPUT
(LSS)
IR20i--- --- ---- --T
Alo
> - - - - - ---
I AO I I A02
I I
I I
~
- -TRZ02---- - - -- --[RZOZ- - ---------1
A03 I A04 I
I I
CABBY I_
I
I
I::: Nt. BLE it.
I
I
I
.. II I,
DC
I
L- ________ _
__ L __ _ _______ ...1.. _______ __ _ ....1._
DUAL 8BIT UP-COUNTER
,
_-l
ANALOG-DIGITAL CONVERTER KITS
KITS C001, C001 A, C002
PANELAID
SERIES
There are three 10-Bit, successive-approximation, analog-digital converters in the Panelaid
TM series_ The COOl is supplied with 60-Hertz power, the COOlA with 50-Hertz power, and
the C002 is available without AC power- All three converters accept any analog value from
o to -1O_23v (full scale) and convert this to a lO-bit absolute value_ Each unit has an
accuracy of 0_1 percent (% least significant bit) and a resolution of 1 part in 1024 (10
millivolts)_ The maximum conversion rate is 30 kh for low-impedance sources with a
maximum input current drain of 1 /.tamp_ All digital input and output signals are compatible
with DEC standard pulses and levels_ The COOl and COOlA are housed on a H900 mounting
panel, while the C002 is mounted on a 1943 panel.
I 2 3 4 5 6 7 8 9 10 II 12 13 14 15 16
AAAAAARRRR
7666652222
000000000
I I I '" 2: 2: 2: 2 2:
A RRR RRRRR
7 663 22222
a 000 00000
4 222 22222
MODULE LOCATION DIAGRAM
PARTS LIST
I-H900-W-P*
ll-R202
2-R602
l-R302
l-A502
3-A601
l-A604
l-A704
l-E724
l-E725
l-E726
l-E727
l-W980
power supply and
mounting panel
dual flip-flop
dual pulse amplifier
dual delay
comparator
D to A converter
D to A converter
reference supply
Panelaid
Panelaid
Panelaid
Panelaid
module extender
* 1943 supplied with the C002, H900A-W-P is supplied in COOIA
265
POWER SUPPLY
$1,144_00 - Kit #COOI
1,162_00 - Kit #COOlA
1,079_00 - Kit #C002

VIN
'" g:
START
+ F
M

L uJ K! "->t P,; U E K P E K P U E K f I I
YE -Tv JL v L V L V L I F 6:
K
I VP
N_..I.. I IK_..I.. 0_..1.. I IN ..I.. U ..I.. I 10_..1.. K U_..I.. I 1<1..1.. K_..I.. 11"-..1.. U
lPl ULr 1
1
0. -
JEILJP TVILJE JlQ:
I
-,/L
UP JVIUE JIUp JVIUE l
v
F-!-

64
o M

R302 Iv _ S
1.4J.1sec
5 T
R602IU
63
IOBIT ANALOGDIGITAL CONVERTER
50/60 HERTZ BCD REAL TIME CLOCK
KIT C003
PANELAID
SERIES
This kit is useful for applications in which it is desired to maintain a time of day reading, or
to measure the time lapse of events, to provide time synchronizing signals, etc. The clock
has builtin gating to permit connection of switches to preset the clock to the correct
time of day; or for elapsed time measurements to clear the clock before use.
16 15 14 13 12 II 10 09 De 07 06 05 04 03 02 0 I
xx xxx
R R R R R R R R W R R R W R R R
Z Z Z Z Z Z Z Z 0 Z Z Z 0 Z Z Z
0 0 0 0 0 0 0 0 I 0 0 0 I 0 0 0
z z z z z z z z

z z z

z z Z
R R W W W W w R R R R R W W R R

6
,

0 0 0 I I I I I 0 0 I I
0 0 0 0 z z 0 I I I I I Z 0 I I
I Z I I
, ,

I I I I I
,

I I
X X X X
X MODULES OMITTED IF SECOND SET NOT
DESIRED
XX MODULES OMITTED IF CABLE OUTPUTS
OF SECOND COUNTER NOT DESIRED
xxx MODULES OMITTED IF CABLE OUTPUTS
OF HOUR AND MINUTE COUNTERS NOT
DESIRED
The following figure illustrates the required relative positioning of the three printedcircuit
boards to each other when installp.d on a mounting panel. However, the boards need not
occupy these particular socket positions.
RELATIVE BOARD POSITIONS
It is suggested that the PANELAID boards be positioned approximately %2 inch apart and as
close to the mounting panel as possible. This facilitates attachment of testing equipment
or wiring to other devices.
267
I
The radix of the hertz counter may be set to either 50 or 60 by jumpers installed on the
outer layer board. For a 50 hertz clock a jumper is installed between each pair of lugs with
a "50" between them or for a 60hertz clock a jumper is installed between lugs which have
a "60" between them. These lugs are shown on the logic diagram.
CLOCK INPUTS: Should come from a network consisting of a small filament transformer,
and a small integrator to filter high frequency noise from the line voltage. A cable socket
input is provided. The input network must be s u p p l i ~ d by the customer.
PRESET LEVEL INPUTS: Inputs are clamped to -3 volts which cause assertions of "ones"
at the input gates. Decade switches used to generate preset levels must provide contact
closures to ground to the level inputs to negate bits not to be set; connection to these
inputs is via W023 boards in sockets provided. .
PRESET PUSH BUTTON INPUT: A momentary contact closure presented between pins S
and U of the W051 in B14 will generate the signals to read the contents of the preset
switches into the hour, minute and second counters.
OUTPUT CABLING: The "1" side negative outputs of the hour, minute, and second
counters are connected to two sockets to enable driving of indicators, connecting to
other devices etc. '
Ippd
-r
PRtSET
P .
Ipph Ippm
50/60
CV LINE INP
Ipps
REAL TIME CLOCK BLOCK DIAGRAM
50/60pps
Equipment required and prices; excluding mounting panels, power supplies, and cabling are:
1 - E007 PANELAID for C003
7 - Rll1 Diode Gates
14 - R202 Dual Flipflop
268
1 - R302 Dual Delay
1 - R601 Pulse Amplifier
2 - W005 Clamped Loads
2 - W501 Schmitt Trigger
2 - W018 Cable Connector
C003 - $639,00
" g
....
u
"'
:0
;::
....
I
i:I
a:
269
PDP-SIS 1/0 BUS INTERFACE
KIT C005
PANELAID
SERIES
This kit provides an interface for a grouping of six cables that parallel connect each periph
eral device. Each device must have associated sockets to facilitate interconnection of
peripheral gear and continuation of the I/O Bus cables. Six sockets must be provided for
input cables and six for output cables.
Device
PDP-SIS
I/O BUS CONFIGURATION
Device
2
L-__ x.;..;6 __ -?>
TO ADDITIONAL
PERIPHERALS
The kit includes two printedcircuit boards, each six sockets wide by two sockets high.
One board provides all the ground connections, the other all the wiring to parallel the two
sockets for each cable.
The 'PANELAID boards should be positioned as close to the mounting panel surface'
(and each other) as possible, for easy attachment of test equipment or wiring to
inputoutput devices.
PARTS LIST
1 - EOOI PANELAID
1 - E002 PANELAID
C005 - $12.00
270
PDP-S or PDP-SIS .
INPUT IOUTPUT BUFFER REGISTER
KIT C006
PANELAID
SERIES
This kit includes a 12-bit register that may be loaded from the computer's accumulator
register (AC); two device-selector modules to decode and generate the control or transfer
pulses; input gating to the AC input lines; a flip-flop with one output connected to I/O Skip
and Program Interrupt facilities; and a free flip-flop which may be used for enabling or
controlling functions. A C005 PANELAID TM interface is used with each C006 to provide
connections for the PDP-8 or PDP-8/S I/O Bus cables_
The following figures illustrate the required relative positioning of the C005 and C006
boards when installed on a mounting panel. Although relative board location is critical, the
exact positioning of the kit, with respect to the mounting panel, is not_
o .
,
liD 110 liD
CABLE CABLE CABLE CABl..E CABlE CABLE
110 110 110 liD 110 110
CABLE CABLE CAIl.E CABt.E CABLE CABLE
MOOULE LOCATION DIAGRAM (WIRING SIDE)
'005
I
~
RELATIVE BOARD POSITIONS
NOTE:
EOOI and E002 are board
numbers for kit # COOS
E003, E004, and E005 are
board numbers for kit #C006
The equipment required and prices, - excluding mounting panels and power supplies are:
271
1-Rll1 Diode Gate
2 - R123 Diode Gate
1 - R202 Dual Flip-flop
6 - R205 Dual Flip-flop
2 - W103 PDP-8 Device Selector
C006 - $397.00
I
N
8
8
272
PRINTED CIRCUIT BOARDS
E and F
SERIES
The E and F series are the printed circuit back panels used in the OCTAID and PANELAID
Kits. These back panels may be ordered separately by those who already have the
required FlipChip modules and connectors. Where a kit includes more than one board,
all boards must be used to perform the indicated functions.
PRICE LIST
KIT NUMBER KIT NAME
COOl
C001A
C002
C003
COO5
C006
D001A-F
D002
Analog-Digital Converter Kits E724
E725
E726
E727
TOTAL
50/60 Hertz BCD Real Time Clock E869
E007
E870
TOTAL
PDP-8/S 10 Bus Interface EOO1
E002
TOTAL
PDP-8 and PDP-8/S Input Output Buffer Register E001
E002
E003
E004
E005
TOTAL
Digital-Analog Converter F728
F843
TOTAL
BCD Up-Counter F723
*If the board fits over only one H800 connector block it will be identified by
an F and three digits. If the board fits over more than one H800 it will be
identified by an E and three digits. Any Octaid kit (D series) can contain only
F series boards. However, any Panelaid kit may contain both E and F series
boards.
273
PRICE"
$11.50
11.50
11.50
11.50
---
$46.00
16.00
16.00
16.00
---
$48.00
6.00
6.00
$12_00
6.00
6.00
10.00
10.00
10.00
$42.00
2.25
2_25
---
$ 4.50
4.50
I
PRICE LIST (continued)
KIT NUMBER KIT NAME PRICE'"
0004 Bi-Directional Decode Counter F861 $ 5.70
(Converter and Decoder Options) F862 5_70
TOTAL $11.40
0005 PDP-8/S Input Buffer Interface F002 4_50
F003 4_50
--
TOTAL $
9_00
0006 PDP-8/S Output Buffer Register FOOl 2_25
F004 2_25
TOTAL $ 4.50
0007 Single or Dual 8-Bit Shift Register F005 4.50
F006 4.50
--
TOTAL $ 9.00
0008 Single or Dual 8-Bit Up-Counter F007 4.50
F008 4.50
--
TOTAL $ 9_00
274
PART V: ANALOG - DIGITAL CONVERSION HANDBOOK I
275
PREFACE
The Analog-Digital Conversion Handbook represents the first attempt in the data process-
ing industry to assemble comprehensive information on this subject in a form that makes
it immediately useful to beginner or expert All phases of conversion are covered, from
concepts to calibration_ Many diagrams supplement the text; and tabular summaries of
terms, methods, and performance characteristics are included for comparison and refer-
ence_ Circuit modules and other equipment manufactured by Digital Equipment Corpora-
tion are mentioned specifically, so after choosing the conversion method most appropriate
to his needs, the reader can construct his system directly_
The use of circuit modules in constructing analog-digital converters yields several advan-
tages_ First, they are flexible_ Converter systems have widely varying requirements, from
pulse height analysis, where differential linearity is of utmost importance, to time-locked
averaging in biomedical work, where resolution is more critical than repeatability or even
accuracy_ Modules permit the construction of the exact type of converter needed and,
should requirements change, the same modules can be used later to build a different kind
of system_
Second, modules are economical. Aside from the interchangeability mentioned above, sav-
ings are gained in the cost of construction_, The typical cost of a digital-to-analog converter
is about $1,000; of an analog-to-digital converter, about $2,000. If several systems are
built, the cost per converter decreases since the same power supplies and mounting
panels are used for the additional units. If the speed. requirement is exceptionally high,
costs will be higher.
A third advantage of using modules is that the completed converter need never go back
to the factory for recalibration. Procedures for calibration and adjustment are included
in this handbook. Recalibration can be carried out quickly and easily by the user.
Those modules designed exclusively for use in conversion systems are specified in detail in
this handbook. The general purpose logic modules also needed are mentioned by name
and type number. Complete specifications for these and over 200 pther kinds of circuit
modules and accessories are contained in another part of this handBook.
276
~ , '
.-
CHAPTER I
BASIC ELEMENTS OF CONVERSION
Introduction
This chapter describes the general technique used to convert, to sample and hold, and
to multiplex.
For digitaltoanalog conversion, just one technique is described. Though there may be
some variations, the same technique is generally applicable for all digital-to-voltage or
digital-to-current converters.
Analogtodigital conversion is somewhat more complex and thus a variety of different
methods is commonly used. In this chapter, the four most common methods are described.
Of these, the successive approximation converter is most generally used since it provides
good performance over a wide range of applications at a reasonable cost. However, if the
converter is to be used only in a single application, various other methods may be preferred
for better performance or lower cost.
It is suggested that this chapter be read as a brief development of the principles of con-
version, rather than a delineation of specific methods. Detailed descriptions of conversion
systems will be given in Chapters 3 and 4.
Digital-to-Analog Conversion
To convert from a digital numberto an analog voltage, a resistive divider network is con-
nected to the flipflop register which holds the digital number (see Figure 1). The divider
network is weighted so that each bit of the register will contribute to the output voltage
in proportion to its value.
Figure 1. Digital-toAnalog Conversion
The digital input signal determines the analog output voltage, since the divider network
is simply a passive element. However, because digital voltage levels are not usually as
precise as required in an analog system, level amplifiers are placed between the flip-flops
and the divider network. The amplifiers switch the divider network between ground and a
277
reference voltage supplied by a precision reference supply. The output voltage range is
between these two voltage levels. In Digital systems, the range is normally 0 to -10 volts.
If the digitaltoanalog converter is to drive a long cable or a heavy load, an operational
amplifier or emitter follower is usually put on the output of the circuit to lower the output
resistance. The digitaltoanalog converter and reference supply shown in Figure 1 are
basic to a digitaltoanalog converter and are described under those headings in Chapter 5.
Analog-to-Digital Conversion
The basis of analogtodigital conversion is the comparator circuit. This circuit compares
an unknown voltage with a reference voltage and indicates which of the two is larger.
SIMULTANEOUS METHOD
Figure 2 shows how a simple simultaneous analogtodigital converter can be built using
several comparator circuits. Each comparator has a reference input signal. The other input
terminal of the comparators is driven by the unknown input analog signal, which is between
o and V volts. The comparator is called "ON" if the analog input is larger than the refer
ence input. Then, if none of the comparators are on, the analog input must be less than
+ . If C1 is on, and C2 and C3 are off, the input must be between + and -f- . Similarly,
if C1 and C2 are on, and C3 off the voltage is between -f- and ~ ; and if all the compara
tors are on, the voltage is greater than ~ .
ANALOG
INPUT
IS BETWEEN
OANO 1/ VOLTS
COMPARATOR
OUTPUTS
Figure 2.Simultaneous AnalogtoDigital Converter
C. C2
off off
on off
on on
on on
C, Input Voltage
off o to V/4
off V/4 to V/2
off V/2 to 3V/4
on 3V/4 to V
Here, the voltage range is divided into four parts, which can be coded to give two binary
bits of information. Seven comparators would give three bits of binary information. Fifteen
comparators would give four bits. In general, 2N-1 comparators will give N bits of binary
information.
278
The simultaneous method is extremely fast for small resolution systems. For large resolu
tion systems (a large number of bits), this method requires so many comparators that it
becomes unwieldly and prohibitively costly.
FEEDBACK METHODS
If the reference voltage were variable, only one comparator would be needed. Each of the
possible reference voltages could be applied in turn to determine when the reference and
the input were equal. But a digitally controlled variable reference is simply a digital to
analog converter. Thus the generalized analogtodigital converter shown in Figure 3 is
actually a closedloop feedback system. The main components are the same as a digital-to-
analog converter plus the comparator and some control logic. With a digital number in the
DAC (digital-to-analog converter) the comparator indicates whether the corresponding
voltage is larger or smaller than the input. With this information, the digital number is
modified and compared again.
ANALOG INPUT
r- ---- --------,
I I
I I
I I
I
I
I
I
I
L _ _-.J
Figure 3. Analog-toDigital Converter Incorporating Digital-toAnalog Conversion Modules
COUNTER METHOD
Numerous methods may be used for controlling the conversion. The simplest way is to
start at zero and count until the DAC output equals or exceeds the analog input.
Figure 4 shows a converter in which the DAC register is a counter, and a pulse source has
been added. The gate stops pulses from entering the counter when the comparator indio
cates that the conversion is complete.
ANALOG INPUT
Figure 4.Counter Converter
279
The tounter method is good for high resolution systems: As the number of bits is in-
creased, very little additional circuitry is needed. Multiple inputs can easily be converted
simultaneously (as described under Multiplexing later in this chapter). However, conversion
time increases rapidly with the number of bits, since an Nbit converter must allow time
for 2N counts to accumulate. The average conversion time will, of course, be half this
number.
CONTINUOUS METHOD
A slight modification of the counter method is to replace the simple counter with an up-
down counter as in Figure 5. In this case, once the proper digital representation has been
found, the converter can continuously follow the analog voltage, thus providing readout at
an extremely rapid rate. This method, called continuous conversion, is particularly useful
when a single channel of information is to be converted. The converter starts running, and
the digital equivalent of the input voltage can be sampled at any time.
ANALOG INPUT
Figure 5. Continuous Converter
The continuous method is less effective for multiple inputs or for inputs that change faster
than the converter can change. Each time the input makes a large change, the converter
may require as many as 2N steps to catch up. However, if a rapid rate of change is neces
sary, extra comparators may be added so that the up-down counter can count in units of
2,3, 4,or more (see Chapter 3).
SUCCESSIVE APPROXIMATION METHOD
For higher speed conversion of many channels, the successive approximation converter
is used. This method requires only one step per bit to convert any number. The successive
approximation analogtodigital converter operates by repeatedly dividing the voltage range
in half as follows:
<
111-=--111
~ 1 1 0 ~ 1 1 0
100/ . 1 0 1 ~ ~ g 6
~ <011-=---011
010 ~ 0 1 0
0 0 1 ~ 0 0 1
000
280
Thus, the system first tries 100, or half scale. Next it tries either quarter scale (010) or
three-fourths scale (110) depending on whether the first approximation was too large or
too smaiL After three approximations, a 3bit digital number is resolved.
Successive approximation is a little more elaborate than the previous methods since it
requires a control register to gate pulses to the first bit, then the second bit, and so on.
However, the additional cost is small and the converter hand.les all types of signals about
equally fast.
ANALOG INPUT
Figure 6. Successive Approximation Converter
The successive approximation method is good for general use. It handles many continuous
and discontinuous signals and large and small resolution conversions at a moderate speed
and moderate cost.
Sample and Hold
A sample and hold circuit is used with an analogto-digital converter when it is necessary to
convert a signal which changes too rapidly to allow an accurate conversion. A digital signal
from some timing device can signal the sample and hold to hold the analog voltage present
on its input until a time when the converter has completed its operation. The sample and
hold is basically an operational amplifier which charges a capacitor during the track or
sample mode, and retains the value of the charge of the capacitor during the hold mode.
The acquisition time of a sample and hold is the time required for the capacitor to charge
up to the value of the input signal after the switch is first shorted. The aperture time (see
definition, Chapter 2) is the time required for the switch to change state and the uncer
tainty in the time that this change of state occurs. The holding time is the iength of time
the circuit can hold a charge without dropping more than a specified percentage of its
initial value.
The sample and hold circuit can be represented as shown in Figure 7. When the switch is
closed, the capacitor is charged to the value of the input signal; then it follows the input.
When the switch is opened, the capacitor holds the same voltage that it had at the instant
the switch was opened.

Ie
T
Figure 7 Sample and Hold
281
It is possible to build a sample and hold circuit just as shown here. Often, the same circuit
is used with a high gain amplifier to increase the driving current available into the capacitor
or to isolate the capacitor from an external load on the output. In some cases, this sample
and hold is made entirely differently; but from a logical point of view, it acts as the ideal
component shown.
Figure 8 shows a sample and hold built with an A200 amplifier board and an A121 multi
plexer switch.
In Track sample, the hold capacitor is charged up by the operational amplifier; in Hold,
the capacitor is switched into the feedback loop. The input resistor and the feedback
resistor are switched to ground. Since the input to the amplifier remains within a few
microvolts of ground (except during switching), the input impedance is 10,000 ohms to
ground both in Track and Hold. The offset input allows a precise dc level to be added to the
input so that the output of the sample and hold is shifted by this value.
These principles have been incorporated into the A400 sample and hold.
INPUT
o TO + IOV
-v REF
OPTIONAL
INPUT BIASING
NETWORK
~ ~ - - - - - - - - - - - - - - - - - - '
I IOKA 9.5K A I
I% tl% I
I
I
I
t------
~ _____ T ___ R
I AI21
I STU
v P
I
3.3K.ll.
tlO%
I
I
I
~ ~ I
I
HOLD I
I I
L _______ . ____________ J
Figure 8. Sample And Hold System
282
TYPICAL SAMPLE AND HOLD PERFORMANCE
INPUT: 0 to + 10 volts
OUTPUT: 0 to -10 volts
ACCURACY: 0.05% for 10 I's
sample, 5 ms - hold
INPUT IMPEOANCE: 10KO
Higher resistances in A200 will in
crease input inpedance at the cost
of increased acquisition time, re-
duced accuracy or both.
OUTPUT IMPEDANCE: < 100
NOTE: For best noise rejection
keep amplifier negative input lead
short and add two 0.01 mfd capac
itors from supply pins to analog
ground (pin F) within A200.
Multiplexing
Often it is desirable to multiplex a number of analog channels into a single digital channel
or conversely a single digital channel into a number of analog channels. Multiplexing can
take place in the digital realm, the analog realm, or in the conversion process.
DIGITAL TOANALOG
In digitaltoanalog conversion, a common problem is to take digital information which is
arriving sequentially from one device, such as a digital computer, and to distribute this in
formation to a number of analog devices. Usually it is necessary to hold the information on
the analog channel even when it is not being addressed from the digital device. There are
two ways to multiplex. A separate digitaltoanalog converter may be used for each channel
as shown in Figure 9.
PULSES REAO INTO
SELECTED CHANNELS
\
DIGITAL INPUTS
(COMMON TO ALL CHANNELS)
Figure 9. DigitaltoAnalog Systems
In this case, the storage device is a digital buffer associated with the da. convertor. Or, a.
single digitaltoanalog convertor may be used, together with a set of analog multiplexing
switches and a sample and hold circuit on each analog channel. The cost of the first
method is slightly more than the cost of the second method, but it has the advantage that
the information can be held on the analog output for an indefinite period of time without
deteriorating. With the multiple sample and hold technique, however, it is necessary to reo
new the signal on the sample and hold at periodic intervals.
283
ANALOG-TO-DIGITAL
In analog-to-digital conversion, it is more common to multiplex the inputs in the analog
realm. Here switches, either relays or solid state, are used to connect the inputs to a com-
mon bus. This bus goes into a single analog-to-digital converter which is used for all chan-
nels (see FigurelC!l. If simultaneous time samples from all channels are required, a sample
and hold circuit can be used ahead of each multiplexer switch. In this way, all channels
would be sampled simultaneously and then switched to the converter sequentially_ The
multiplex switches and sample and holds will introduce some error into the system. How-
ever, it is usually less expensive to go to higher quality sample and hold and multiplex
circuits than add extra converters.
ANALOG
INPUTS
DIGITAL OUTPUTS
----------
Figurelo.Multiplexed Analog-to-Digital Conversion System
In a simple analog-to-digital converter with a single comparator circuit, it is also possible
to multiplex by using a separate comparator for each analog channel. One input of each
comparator is tied to the voltage generating device in the converter. The other inputs are
tied to the separate analog channels_ The comparator to be used can be selected digitally_
This method is particularly good when a small number of channels is to be multiplexed
since it is quite simple and requires little additional control. For a large number of chan-
nels, separate multiplexer switches are usually less expensive and more accurate as they
do not put any load on the voltage generating device of the converter.
OAe
COUNTER
Figure 11.Counter Type Analog-to-Digital Converter with Multiplexed Input
The comparator multiplexing technique is particularly useful with the counter type analog-
to-digital converter. This technique is shown in Figure 11. Several comparators are at-
tached to one converter_ The counter is cleared; then count pulses are applied_ When one
of the comparators signals that the digital-to-analog output is greater than the input volt-
age on that channel, the contents of the counter are read out. Counting is then resumed
until the next signal is received.
284
CHAPTER 2
MEASURES OF CONVERTER PERFORMANCE
Accuracy
Since the end result of conversion is the representation of a given value in different
terms, it is important to know how accurate the representation is. In systems where
accuracy requirements are not too stringent, say in the order of 1 percent, an overall
accuracy specification is usually sufficient. In cases where the desired accuracy is 0.1
percent or greater, it is necessary to isolate the various sources of error; and since a
converter is a hybrid device, both digital and analog sources must be taken into account.
In high accuracy systems particularly, accuracy figures given in the general specifica
tions may not include isolated sources of error, e.g., noise. Thus, it is important to know
the various types of errors, their causes, how they are measured and specified, and
when they are important. Figure 12 shows a breakdown of various types of errors.
Figure 12. Measures of Cumulative Error
DIGITAL ERROR SOURCES
When a continuous signal is quantized, there is an error which is equal in magnitude to
the smallest quantum. For a linear converter, the smallest quantum is the least significant
bit. In most converters the quantization error is centered so that it is equal to 1/2 the
least significant bit, written as 1/2 LSB.
In a continuous converter or digital voltmeter, accuracy may not be as important as
avoiding chatter. That is, if the input is right on the dividing line between two quantiza
tion states, the output should not oscillate. If hysteresis is introduced so that the quantiza
tion error is just under 1 LSB, then oscillations will normally be avoided and the
accuracy will not be greatly impaired.
285
The digitaltoanalog converter reproduces exactly all the digital input information which
it accepts. Hence digital error is not included in its accuracy specifications. However,
if the input has more bits than the converter, there will be a quantization error in the
readin process which should be taken into account. If desired, a 1f2 LSB offset can be
built into the converter so that the readin will round off, rather than truncate, more
precise digital information.
ANALOG ERROR SOURCES
The dc accuracy of the converter (or switching point accuracy) depends on the offset,
the gain calibration, and the linearity. Nonlinearities are due to the variation in gain (or
common mode effect) in going from the smallest input to the largest input. Some of
these will be longterm, because of the common mode effect of the comparator circuit in
the analogtodigital converter, for example. Some will be shorter term, because of dis
continuities in the divider network or insufficient settling time of the comparator. The
offset and the gain can be adjusted in the calibration until their effects are essentially
negligible.
The measurement of analog error in a digitaltoanalog converter is easily made by put
ting in a digital number (the same word length as the converter) and observing the output.
In an analogtodigital converter, the analog error is difficult to locate since the quantiza
tion error is always present. However, the point where the output oscillates approximately
equally between two neighboring digital numbers is fairly welldefined. This point, called
the switching point, can be measured and compared with the theoretical value.
The ripple on the reference supply and other sources of noise are often measured sepa
rately since one or the other can sometimes be neglected in the final result. The two can be
separated by measuring the ripple in the reference supply and subtracting it from the
measured noise, or by running the input source in the converter from the same reference,
thereby giving a direct measurement of all noise sources except the reference supply rip
pie. In a digitaltoanalog converter the noise and ripple can be measured by observing the
output with a scope. In an analogtodigital converter they can be measured by observing
the input range which causes the output to oscillate between two states.
DIFFERENTIAL LINEARITY
Differential linearity is the variation in the size of the r'equired voltage change that causes
an analogtodigital converter to go from one switching point to another. That is, it is
the variation in the size of the states and is generally quoted as a percent of the size
of the states. It is a part of the overall linearity discussed above, but deserves special
mention because of its importance when an analogtodigital converter is being used in
histogram applications. For example, when plotting the number of inputs versus the
digital state, if one of the states is twice as big as its neighbor, it will tend to accumulate
twice as many counts. Naturally, a very misleading output results.
Differential linearity is one of the few accuracy characteristics which is affected by the
conversion technique. The differential linearity tends to be best when the converter goes
286
through all the states sequentially as in the counter type converter described in Chapter 1
or the ramp variation described in Chapter 3. In an approximation converter, such as the
successive approximation type, the large transients which result in going from, say, half
scale to quarter scale require a long time to settle down, and any hysteresis in the com
parator circuit causes relatively large variations in the state size. However, the differential
linearity of an approximation converter can be improved by running it at very low speed.
Differential linearity is also affected by variations in the divider networks (although they
are relatively small). It can be avoided by using a ramp converter.
The shorter the converter word length, of course, the better the differential linearity
will tend to be. However, this gain may well be compromised, since small resolution
could result in the loss or the smoothing of very sharp peaks in the histogram.
Techniques commonly used to overcome difficulties with differential linearity are: chang
ing the offset on the converter (or equivalently the bias on the input signal) and changing
the word length of the converter. Switches can be mounted on the converter for this
purpose, or the change can be made programmable so that the controlling device can
make the change automatically.
DISTRIBUTION OF ERROR
How much of the total error should be in the digital circuitry and how much in the analog
portion? For converters in the range of up to 10 or 11 bits, the digital error generally
accounts for about V3 to 1f2 of the total. Thus, a typical 10bit system would have a quanti
zation error of 1/2LSB and an analog error of O. 1 %.
If the accuracy requirement is low, the word length may be the major source of error.
The total error may then be treated simply as roundoff. If the accuracy requirements
are stringent, it is desirable fo minimize all sources of error, analog and digital. The
digital error is quite simple to minimize by extending the number of bits within practical
limits. A converter with an overall error of 0.1 % and a word length of 20 bits would be
unjustified.
Requiring monotonicity is one way to assure that all the bits are meaningful. This means
that all states must exist and they must be in the correct order. In terms of converter
operation, as the number going into the digitaltoanalog converter is increased, the output
voltage must also increase; it should never dip back down at any point. Similarly, if the input
voltage to an analogtodigital converter is increased, the digital output should stay at the
same value or increase and should not skip over any states.
The converter is most likely to lose monotonicity when switching between digital states
such as 0111 and 1000. If the weighting of the bits is not quite correct, in a digitalto
analog converter the higher state might correspond to a lower voltage, and in the analog
todigital converter the output might jump directly from 0110 to 1000.
287
Measures of Speed
DIGITAL TOANALOGCONVERSION
The maximum conversion rate is theoretically limited only by the minimum time between
read ins to the converter flip-flops, and can easily be as high as 10 megacycles. However,
such a figure may be misleading. The desired ratio of settling time to non-settling time
usually determines the maximum usable conversion rate.
The settling time of a converter is measured from the time the digital. read in is performed
to the time when the analog output has settled to within the specified limits of accuracy.
How the output approaches its final value depends on the output circuit, as discussed
below:
The divider output will have high frequency transients before it begins to settle. If the
o!Jtput is going to a low frequency device, the transients can be ignored. In some applica-
tions, it is more desirable to smooth the transition between states than to minimize the
total time, in which case the oscillations can be damped with the capacitor or a low pass
.
If the output is from an amplifier circuit,the settling time will be determined by the maxi-
mum rate of change of the amplifier. Thus, the first readin may take longer to settle than
subsequent read ins, which usually do not change the converter by sLich a percentage of
the full scale. .
ANALOG TODIGITAL CONVERSION
Conversion time is measured from when a request is given to when a digital output is
available_ In converters like the successive approximation type, where all conversions are
completely independent, time must be allowed for completion of entire steps in the con-
version process. In converter, the conversion time is usually just that time
required to synchronize the request and get the number. .
The conversion rate is usually the inverse of the conversion time. In some systems; an
amplifier or comparator recovery time is required between conversions; thus the rate is
lower. However, comparators manufactured by Digital Equipment Corporation do not
have a recovery time. The conversion rate will also be slower if logical operations must be
carried out between conversions. In some cases, such as the counter converter performing
a number of simultaneous conversions or the synchronous sequential converter, the con-
version rate is actually faster than the inverse of the conversion time.
If the input signal is changing with respect to time, it is very important to know when the
signal had the value given by the output. The uncertainty in this time measure is called the
aperture time (sometimes also called window or sample time). The size of the aperture and
the time when the aperture occurs vary depending on the conversion method.
Figures 13 through 16 illustrate how the aperture varies with different conversion tech-
niques. In each case, the upper portion of the figure shows how the converter arrives at
an output. The lower portion of each figure shows how the input might be reconstructed
from the digital data.
288
IJOLTAGE
VOLTAGE
ANALOG /' - ...... ,
INPUT / "
~ /
I
I
I
I
I
READ
OUT
2
RECONSTRUCTED
SIGNAL
I = CONVERSION RESULTS
I
REAO
OUT
,
READ
OUT
4
- ~ - - , - , - = - = - = - - - - ~ .. --
TIME
Figure 13. Counter Converter
In the counter converter (Figure 13) the aperture occurs at the end of the conversion.
This is not constant with respect to the beginning of the conversion, but it may be calcu
lated from the digital output.
For the continuous converter (Figure 14) the aperture is the time for the last step. Here
the assumption is made that the input signal does not change more than l LSB between
conversion steps. To meet this requirement, the maximum rate of change of the input
voltage must not exceed the maximum rate of change of the converter. This is
V, ./2"lIT where V reference is the full scale voltage, N is the number of bits, and liT is the
time per step. The maximum rate of change of the sine wave is 2"V,f. or "V"f. Thus, if the
converter is to follow the input, the maximum frequency components in the input must
satisfy the following equation:
"Verf =V",/2
N
lI T
and if the peaktopeak voltage is assumed equal to the converter reference, then the maxi
mum frequency is:
f=_l_
2
N
lIT"
289

VOLTAGE
VOLTAGE
VOLTAGE
TIME
...... - ....... ,
ANALOG / ,
INPUT / ""
'y '."
/ ..... , .. _ . ~ _ ~ ~ . - : - : . " 7 " : . ..,...
/ ...... ~ . . . . . . ~ . ~ . ~
" ' '\RECONSTRUCTED '"';": -:-:-.-;.
I SIGNAL
I EACH DOT INDICATES A READ OUT
I ...
TIME
Figure 14. Continuous Converter
READ REAO READ REAO READ READ READ READ READ READ READ READ
OUT OUT OUT OUT OUT OUT OUT OUT OUT OUT OUT OUT
1 2 :3 4 5 6 7 8 9 10 11 12
EACH HORIZONTAL LINE INDICATES
THE OIGITAL READOUT AND APERTURE
Figure 15. Successive Approximation Converter
290
TIME
For a successive approximation converter (Figure 15) the digital output corresponds to
some value the analog input had.during the conversion. Thus, aperture is equal to the
total conversion time. Aperture time of the successive approximation converter can be
reduced by using the redundancy techniques outlined in Chapter 3 or by using a sample
and hold circuit. The sample and hold is illustrated in Figure 16.
VOLTAGE
VOLTAGE
RECONSTRUCTED
SIGNAL.
READOUT
Figure 16. Successive Approximation Converter with Sample and Hold
SELECTING A CONVERSION METHOD
Chapters 1 and 2. have summarized several methods of conversion and the performance
characteristics that may be expected from them. These criteria for choosing a specific
conversion method are condensed in Table 1. The table is organized like the handbook
with applicable chapters called out for quick reference to detailed descriptions of the
methods.
The decision to choose one converter over another is principally a matter of speed, aper-
ture, .cost, and whether multiplexing or a single continuous input is to be used. Exact
conversion times, aperture times, and cost depend on the number of bits, type of circuitry,
291

I\)
lS
Method
BASIC METHODS (Chap. 1 & 4)
Simultaneous
Counter
Continuous
Continuous Input
Discontinuous Input
Successive Approximation

c.-=

"0

.Eg
-:nu

COo
Both
M
C
C
M
VARIATIONS ON BASIC METHODS (Chap. 3)
Ramp M
Section Counter M
Continuous with add. compo
Continuous Input C
Discontinuous Input C
Successive Approximation M
with Redundancy
ADVANCED (Chap. 3)
Subranging M
Subranging with Redundancy M
Seq. Approx. (NonSynchronous) M
Seq. Approx. (Synchronous) C
Quantizing
Continuous Input Both
Discontinuous Input Both
""See text
TABLE 1 ANALOG TO DIGITAL CONVERSION TECHNIQUES
.
" E
f="U
,,0>

'e"':;-
"",

gco
u",
.
.-
O-U
0>
"'"
:?'"
ell;::
1: CO
00
u-
*
EU'
.- "
0-",

"'" 1:'::::
1l.
co
<'"

"u
E"
t=S

t:co
1l.0
<-
____ Not Applicable ____
24 avo 1792 avo 1.5 3.S.
1.5 2 I.S 2
24 avo 1024 avo
7.S 36 7.S 36
16 avo S12 avo 1 1
18 112224 I.S 3.S
1.5 2 I.S 2
612 avo 32512 avo I.S 2
9 27 1.5 3
_.
34 1020 34 1020
2.5 69 I.S 3
7.5 25
1.5i9t I.S
2 3 2 3
6 18 2 3
'0
" E
i=

"'''
-:nt:
""
00.
u<
Ves
No
Ves
Ves
Ves
No
No
Ves
Ves
Ves
Ves
Ves
Ves
Ves
Ves
Ves

;-:;;
"0
a::u
Deperids on
resolution
Low
Low to
Medium
Medium
on
resolution
Low to
medium
Medium to
High
Medium to
High
High
High
High
High
High
High
Remarks
Excellent for low resolution systems
in about 100 nanoseconds
operates
Allows many conversions simultaneously
Extremely high speed for continuous input but falls
behind on sharp rate of change
General purpose good speed dolla_r ____
Good differential linearity - low cost for low
resolution systems
Used with digital voltmeter
Similar to continuous but has faster responses to
discontinuous or high speed signals
Good speed per dollar in high resolution systems.
Small aperture, good differential linearity
-- --
Excellent for S to 8 bits
Excellent for 7 bits or more
May make errors. Requires sample and hold
tTime between Ct?nversions total time
Excellent for both multiplexed and cO[ltinuous in
puts. Automatically follows fast input with low
resolution and slow input with high resolution
and variations in system design. The speeds given in the table were derived assuming that
the system was designed for maximum speed per dollar. Actual speeds will usually be
within a factor of 2 for basic conversion methods and within a factor of 5 for the others.
The basic conversion methods, as described in Chapters 1 and 4, will satisfy most require-
ments. If Table 1 confirms the choice of one of the basic methods, the reader can go
directly to Chapter 4 for specific information on the equipment required. The other methods
are variations of basic methods and advanced techniques primarily for increased speed.
They are described in general terms in Chapter 3.
293
Signed Digital Numbers
In the following chapters, the most significant bit represents -5 volts, the next most sig-
nificant bit, -2.5 volts, and. so on. Thus, the all ZERO state corresponds to 0 volts and
the all ONE state corresponds to 1 LSB less -10 volts. This conversion can be reversed
simply by using the opposite side of the flip-flop to drive the divider.
If only the most significant bit is reversed, the numbers are signed, 2's complement, as
shown below. Since more numbers are negative than positive, the negative numbers are
used for 0 to -5 volts and the positive numbers go from -5 to slightly less than -10
volts.
Voltage Unsigned Signed
-0-
000000 100000
000001 100001
011110 111110
011111 111111
-5 100000 000000
100001 000001
<-10 111110 011110
111111 011111
In l's complement, the weighting of the sign bit is reduced so that +0 equals -0. (Thus,
it cannot be used in a system with redundancy, a variable word length or any other feature
which does not give a constant one to one correspondence between voltage and number.)
Weighting is done by increasing the resistance of the MSB (most significant bit), and for
a system of 9 or 10 bits the ladder potentiometer is sufficient. For low resolution systems,
add a small resistor (about 16 ohms for an 8-bit system, 32 ohms for 7 bits, etc.) in series
with the MSB input. It need not be precise (since the potentiometer will adjust for it) nor
have a low temperature coefficient (since a large change in this resistor will have a small
effect on the output voltage). Use the standard DAC adjustment procedure and look for
a straight line instead of a square wave on the most significant bit.
294
Bipolar Voltages
BIPOLAR 0 TO A OUTPUTS
Binary weighted DAC modules made by Digital have an output impedance of 1000 ohms.
The external load can reach 2000 ohms without any noticeable effect on the linearity of the
system. Thus the output can be made bipolar with a large resistor to a positive bias voltage.
The equivalent circuit is shown below.
OUTPUT
"'BIAS
With a '--10 volt reference driving the digitaltoanalog converter, the output voltage swing
is reduced to R volts. The output voltage swing is centered when the bias voltage
equals +R/200. The bias voltage, of course, should be stable and noise free.
BIPOLAR A TO 0 INPUTS
If multiplexing is being done or if the input signal cannot drive a heavy load, an amplifier
should be used for signal conversion as described in Chapter 5. In other cases, a simple
divider can be used in front of the comparator. The basic circuit is shown below.
R,
INPUT
1 I COMP
NEGATIVE REF
The impedance level should be kept low since the comparator will draw some current even
at balance (a fraction of a microampere). The input, as seen by the comparator, should
be as near as possible to the full 0 to -10 volt range.
295
An alternative to the above scheme is to buck the input against the digitaltoanalog can
verter in the feedback loop, as follows:
.,
r - ~ . / W ' - -10 VOLTS
The range seen by the comparator is reduced, but common mode effects are avoided
since balance is always at the same point. R, should be less than 1000 ohms, and R, plus
R. should be about 500 ohms or less.
Table of Voltages
Octal Numbers Voltage
Signed
Unsigned (Negative)
2's Compo
4000 0000 O.
4001 0001 0.00244140625
4002 0002 0.0048828125
4004 0004 0.009765625
4010 0010 0.01953125
4020 0020 0.0390625
4040 0040 0.078125
4100 0100 0.15625
4200 0200 0.3125
4400 0400 0.625
5000 1000 1.25
6000 2000 2.5
0000 4000 5.
2000 6000 7.5
3000 7000 8.75
3400 7400 9.375
3600 7600 9.6875
3700 7700 9.84375
3740 7740 9.921875
3760 7760 9.9609375
3770 7770 9.98046875
3774 7774 9.990234375
3776 7776 9.9951171875
3777 7777 9.99755859375
- 10000 10.
296
CHAPTER 3
SPECIAL ANALOG-TO-DIGITAL
CONVERSION TECHNIQUES
The analogtodigital conversion techniques described in Chapter 1 are the most com
monly used methods but not necessarily the only ones. There is an extremely large
variety of techniques, not all of which have been investigated. Some of the other methods
are described in the following section.
Variations In Basic Techniques
SECTION COUNTER
The counter converter is a simple technique for performing conversions. However, if the
digital word becomes long, the 2" steps required to complete the conversion may be
too many.
One way to decrease the time at a minimum of cost is to divide the counter into sections.
For example, a lObit converter could be divided into 2 sections of 5 bits each. At the
beginning of the conversion the least significant counter is set to all ones and counts are
inserted into the most significant counter until the comparator indicates that the input has
been exceeded. The least significant counter is cleared and counted up until the correct
value is reached. The maximum number of steps required to complete a conversion is
2' for the most significant counter and 2' for the least significant counter, giving a total
of 2' steps. This is a maximum of 64 counts versus 1024 counts for the standard coun
ter converter.
Other types of section counters might use more parts and operate by counting one
counter up and the next down. The total conversion time, of course, depends on the
number of sections.
The section counter technique is frequently used in digital voltmeters where the output
is to be in decimal: Each section of the section counter thus represents one decimal digit.
RAMP METHOD
In the counter converter, each count input is increasing the voltage out of the DAC by
one step, effectively generating a ramp out of the DAC. Thus the level amplifiers, refer
ence supply, and divider network could be replaced by an external ramp generator
circuit. If accuracy is not too important, the ramp can be made by charging a capacitor
with a current source and using the linear part of the exponential. In higher resolution
converters, the ramp might be made by using the operational amplifier as an integrator.
The ramp technique is somewhat faster than the counter technique because carry and
DAC set up time is not required before gating the next count pulse. The differential linear
297
ity, over a short span of a ramp converter, is bound to be fairly good, since the ramp is a
continuous signal. Although there may be some noise, the general slope will not change
significantly over a 'short span.
Both the ramp method and the counter technique approach the final value in small steps
and from one direction only. This puts considerably less strain on the comparator circuit
than a technique such as subranging or successive approximation where the comparator
is receiving large input voltage changes in different directions and being asked to resolve
small differences. In general, all smooth conversion techniques (the counter, ramp, and
continuous converters) generally operate at a considerably faster time-per-step and
produce better .differential linearity than the approximation methods (subranging, succes-
sive ,!pproximation, sequential approximation, etc.).
CONTINUOUS CONVERTERS WITH ADDITIONAL COMPARATORS
A continuous converter is an extremely fast and relatively inexpensive device for follow-
ing a continuous signal. However sometimes the input rate of change exceeds that of
the converter. To close this gap determine if the differential error exceeds a specified
amount and add or subtract a correction count in a more significant bit. For example, a
small amount of logic added to a lO-bit continuous converter could measure large differ-
ences between the input and the contents of the converter. If the difference is more than
+8 counts, it adds a count in the third flip-flop from the least significant end. If the
Figure 17. High Speed Continuous Converter
298
difference is more negative than - 8 counts, it subtracts a count from this stage. Thus
the converter operates on high frequency signals with reduced accuracy and on low
frequency signals with the full accuracy.
A continuous converter with additional comparators is shown in Figure 17. Two com
parators and additional gating and synchronizing logic have been added to a basic con
tinuous converter.
SUCCESSIVE APPROXIMATION CONVERTER WITH REDUNDANCY
Redundancy is useful where high resolution and high speed are both required. It can
also be used to improve differential linearity and aperture.
The successive approximation converter is extremely efficient; but, since the results of
each step are irrevocable, each step must be allowed to settle to within the total system
accuracy. For high resolution systems, the settling time can be quite long. With redun
dancy, the first steps are done with a limited accuracy; then a correction step is inserted
to improve the accuracy. Only the correction step and the following steps need to settle to
final accuracy. Steps before correction need only settle within '12 of the correction amount.
The correction can be implemented by adding or subtracting one bit, as in a continuous
converter. If the steps preceding the correction are offset, only add circuitry is necessary.
For fastest operation, a special divider with redundant inputs can be used so that the
addition can be done without generating carries. The digital summing can be done in an
output buffer where the carries will not interfere with the analogtodigital feedback loop.
The correction step can also be used to compensate for changes in the input analog
signal during earlier steps, thereby reducing the aperture. It also improves the differential
linearity of the converter since a large part of the variation in state size is due to the
large transients during the early conversion steps.
Advanced Techniques
SUBRANGING
This method is very good for converting a large number of input channels since the I
conversion begins without assuming anything regarding the previous state of the input.
It also converts quite rapidly and allows a tradeoff between cost and complexity and
speed. Thus, if extremely high speed is required, numerous comparators are used and
not many steps are required. In the case where less speed is required, perhaps only one
comparator circuit would be used.
The subranging method operates by dividing the total input signal range by the number
of subranges, selecting the appropriate subrange and then dividing this into subranges
as before, repeating until the desired resolution is achieved.
Figure 18 shows how subranging works. At the start of the conversion, the only informa
tion available about the input signal is that it lies somewhere in the range of zero to the
full scale voltage, VFS ' The first step of the conversion divides the full scale voltage into
299
subranges, in this case four. Simultaneously, comparisons are made between the input
voltage and the three .subrange boundaries, V" V" and V . It can be determined whether
the input voltage is higher than or lower than each of these boundaries. If the input
signal is lower than all of the boundaries, it must fall in the lowest range. If it is higher
than V, but lower than V, and V., it must fall in the next to the lowest range, and so forth.
Once this information is determined, the selected subrange can be divided into four
more subranges and the process repeated .
. ",.---.,------.,------,
.,1-------1------1
.,1-------1------1
.,1-------1---'----1
START STEP I STEP 2
Figure 18 Subranges for a Converter with Four Subranges Per Step
If there are M ranges per step and S steps, the total resolution of this conversion will
be (* )'. For example, a 12bit system requiring a total resolution of ~ could be
implemented in 12 steps [ (+)" = ~ ] , in six steps [ (+)' = .m ], in four
steps [ (+)4 = .-on ], or in three steps [ ( -to)' = ~ ]. The step resolution does
not have to be an integer power of two. However, except in a binary coded decimal system
where it is useful to make M equal to 10, the saving in control circuitry is usually suffi
cient to justify increasing the step resolution to the next power of 2.
Figure 19 shows a subranging converter. Here two digitaltoanalog converters and a
number of comparators are referenced at equally spaced intervals in the range between
the value of the two converters. The technique is similar to the simultaneous method.
The system starts with the lower DAC (digitaltoanalog converter) at zero, the upper
one at the maximum voltage. The output of the comparators indicates which range con
tains the input, say between the reference applied at CK and the reference applied at
CK+,. Then the reference voltage from C. is applied to the lower DAC, and the reference
voltage that was at CK+, is applied to the upper DAC. A new, smaller set of ranges is pro
duced. The process is then repeated.
300
ANALOG INPUT
OAC
TIMING
a
CONTROl.
Figure 19. Subranging Converter
If the number of subranges obtained in a single step is equal to the total system resolution,
this method becomes' the simultaneous method described in the introduction. If the number
of subranges per step is reduced to two, this method becomes the successive approxima-
tion conversion.
SUBRANGING WITH REDUNDANCY
Redundancy, as described for the successive approximation converter, can be applied in
the same manner to a subranging converter. It is particularly useful here as the capacitance
of many comparators in parallel causes the settling time to be quite long.
SEQUENTIAL APPROXIMATION
Sequential approximation, also called parallel approximation, uses a separate analog-to-
digital converter for each binary bit of information to be obtained. There are two methods
ef operation, synchronous and non-synchronous.
Figure 20 shows how the non-synchronous type operates. In one example shown at the
top of the figure, the analog input comes into a comparator which compares the input
with half scale. If the input is larger, the comparator applies a voltage to the most signifi-
cant bit of each of theDAC circuits down the line. As soon as the first comparator has
301

settled, the second comparator can start to make its decision. Speed is gained because
there is no flip flop delay in this system. But more importan't, most comparators . will
make decisions relatively quickly, since the analog input cannot be very close to the
boundaries of more than two subranges (the last one and one other), Thus, the average
amount of time required per decision is considerably less than the maximum. In a clocked
system, the rnaximum required time must be allowed for each step. Here, only the
average time is allowed, '
ANALOG
INPUT
ANALOG
INPUT
FLAG
!--------.DIGITAL
OUTPUT
Figure 20. Nonsynchronous Sequential Approximation
302
The converter shown at the lower half of Figure 20 is quite similar except that the
individual DAC networks are replaced by operational amplifiers. The analog input goes
to the first comparator. If the input is above half. scale, the comparator produces a
voltage corresponding to half scale. This is subtracted from the input signal, and the
result is multiplied by two and passed on to the next comparator.
In both examples, speed is gained by the fact that full settling time is not needed by
those comparators which are not making a critical decision. On the other hand, some
difficulties are encountered if the input signal should change slightly before the digital
readout has occurred. One of the comparators may change value, but the results may
not carry to the end of the chain before readout. Thus, the comparator should have
builtin hysteresis so that small noise spikes will not cause an error, and the digital output
should always be read into a buffer and double checked with the input.
In synchronous sequential approximation, the time required to perform a complete con
version is essentially the same as in a successive approximation converter; however, the
conversion rate is much faster. Erroneous readout is eliminated, since the converter is
buffered and synchronous. This type of converter is particularly useful for systems with
a single input.
The synchronous or clocked type sequential approximation converter also uses one
converter per bit. It differs from the nonsynchronous type because there is a delay line
between each converter (see Figure 21). The analog information arrives at the first
comparator, which makes a decision and stores the information in a shift register for
use by later converters. By the time the second converter is set up and ready to make
a decision, the same analog information is just arriving at the second comparator. This
converter decides on the second bit of the output word, based on exactly the same analog
voltage as was at the first converter when the first bit decision was made. This process
is continued for however many bits are necessary.
DIGITAL OUTPUT
Figure 21. Sequential Approximation (Synchronous)
303

QUANTIZING ENCODER
The quantizing encoder was developed by Dr. Jerome Cox and Donald Glaser at the
Central Institute for the Deaf. It is the most general purpose, high speed converter
because it is fast for both continuous and multiplexed inputs.
The quantizing encoder uses one digitaltoanalog converter, a number of amplifiers,
and a number of comparators. It examines the difference between the input voltage and
th.e DAC output, quantizes the difference to the nearest power of 2, and adds or sub
tracts this from the DAC. Thus, the quantizing encoder can follow a continuous signal,
staying within one bit of the correct value for low frequencies. For high frequencies,
it will always be within the nearest pqssible power of 2 of the correct answer. (Thus,
if the input signal should suddenly change, the quantizing encoder will make a similar
jump corresponding to the nearest power of 2 change, while the standard continuous
converter could increase by only one count.)
For multiplexed input, the quantizing encoder will operate similar to a successive approxi
mation converter but twice as fast. It requires only one step for-each two binary bits
(or fraction thereof). In addition, since it includes a selfcorrecting ability, the time
per step can be quite fast.
Figure 22 shows how the quantizing encoder could arrive at the result when used as a
4bit encoder. In the left-hand example, it is used with a multiplex input. At the start
.... 1111
1110
1101
... 1100
1011
.,<:--.... 1010
._---1-.1001
___
------1-.0111
...... 0110
0101
"<E--.... 0100
0011
0010
0001
0000
START STEP 1 STEP 2
FOR MULTIPLEXED INPUT

----____ 1110
.... 1101
1100
.... 1011
----.... 1010
.... 1001
..:.---. 1000
.... ---. 0111
-----...... 0110
...... "..--.... 0101
. 0100
--:::.....-... 0011
__ --.... 0010
Jr---+OOOI
.... --..... 0000
START JUMP JUMP JUMP
FOR CONTINUOUS INPUT
Figure 22. Quantizing Encoder Method
304
of the conversion the converter is set to mid-scale_ At the end of the first step it will
go to any of the points shown, and at the end of the sp.cond step it always has arrived
at the correct answer_
The right-hand example, Figure 22, shows how the converter would react if it were hold-
ing its maximum value and the input suddenly dropped to a much lower value. If the
new value were within 1 or 2 counts, it would immediately arrive at the exact answer.
Otherwise, it would make a power of 2 jump to the nearest correct value.
305

CHAPTER 4
TYPICAL CONVERTER LOGIC
Digital-to-Analog Conversion
Figure 23 shows a typical digital-to-analog converter_ The basic components of this circuit
are a flip-flop register, DAC modules, and a reference supply. The digital signals are brought
in with a pair of complementary levels for each bit. This information is jammed simultan-
eously into all the flip-flops and is automatically converted to the appropriate voltage by the
divider network.
READ IN
PULSE
DIGITAL INPUT DATA
2 LINES PERBtT
Figure 23.Digital:to.Analog Converter
'"
SUPPLY
The settling time of the digital-toanalog converter depends .on the number of flipflops
that change, as well as the voltage difference between the two states involved. For example,
in switching from a number such as 0111 to a number such as 1000, all of the flip-flops
change state. Even though the two final values of the analog voltage are very close, tran
sients occur on the divider output for the following reasons: variation in ,transition times
from flipflop to flip-flop and from level amplifier to level amplifier; transient current drawn
from the reference supply; the fact that the flip-flops have a slower fall than rise time;
and the fact that signals must propagate through the divider network. The worst case is
switching from mid-scale (1000 ... ) to one count less (0111 ... ). Here the transients
are as much as 1 volt.
However, the transients are quite short in duration and return to within 1/2 LSB of their
final value within at most 2.5 microseconds for the medium speed combination's listed in
the table. In most cases, these transients will be faster than the load can respond and
hence can be ignored. These units will settle to within 0.05 per cent within 2.5 micro .
seconds.
306
RECOMMENDED MODULES FOR MEDIUM SPEED DIGITALTOANALOG CONVERTER
No. of DigitaltoAnalog Reference
Bits FlipFlops Conversion Module Supply
Up to 4 R200 R202 R203 A601 A704
5 R200 R202 R203 A601 A704
6 R200 R202 R203 A601 A704
B R200 R202 R203 A601 A704
10 R200 R202 R203 A604 A704
12 R200 R202 R203 A604,A605 A704
A small choke can be usedbetween the flipflop output and the DAC .input on the more
significant bits to equalize the switching times. This will reduce the transients to about a
0.5 microsecond duration on high accuracy systems. If further smoothing is desired, a
low pass filter should be used on the output.
Analog-to-Digital Conversion
SIMULTANEOUS
The simultaneous conversion technique is simple, inexpensive, and extremely fast for a
small resolution system. Figure 24 illustrates a simultaneous converter with a resolution
of 3 bits. It uses Type W520 as comparator circuits for the input. These units have a res
olution of 0.1 volts and are therefore suitable for a simultaneous convertor of up to 4 bits.
The reference voltages for the level standardizers are made by dividing a +10 volt refer
ence with a series of identical resistors. Although the tolerance on these resistors is not
wide, in systems of 3 or 4 bits small trimming potentiometers should be put in series with
the resistors so that the reference voltages can be adjusted to offset the common mode
effects and the zero offset of the comparators. Since the comparators also draw a current
through the resistors, the potentiometers can trim the value of the resistors to compensate
for this current.
The outputs of comparators are coded in a Gray code and jammed into a simple flipflop
register, made by crosscoupling inverters and diode gates. A Graytobinary decoder on
the output produces standard binary notation.
307
ANALOG
INPUT
COMPARATORS
1018 $INPUT:!;5\10t81
Figure 24.Simultaneous Converter
GRAY TO
BINARY
DECODER
COOING
HETWOfIK
For medium speed systems, Type RIll, R1l3, RI21 and RI22 gates can be used with
DEC Standard 100 nanosecond pulses. With these gates, pulses can occur 0.5 microsecond
apart, thus giving a conversion every 1 microseconds.
There are two factors which limit extending this system to large resolution systems. One
factor is cost of so many comparators. The other factor is the current drawn and the input
capacitance, which become extremely large if too many comparators are tied in parallel.
For this reason it is recommended that the Type W520 be used for systems up to four bits
only. Detailed information on the Type W520 is included in the W Series Section of
this handbook.
308
COUNTER METHOD
Figure 25 illustrates a typical circuit for the counter type analogtodigital converter. The
start signal clears the counter and inserts a single pulse in the delay chain. Each time the
pulse goes around the chain, one count is added to the flipflop register. When the divider
output is equal to the analog input, the comparator will switch. The next pulse sets the can
trol flip flop, indicating the end of conversion and inhibiting the pulse from circulating. The
circuit shown in Figure 25 used Type R201, R202, R205 flipflops for the counter. The can
trol flip flop may be any unbuffered flipflop such as the Type R200, R202, R203, R204. At
the end of conversion, this flip flop will be set by grounding the ZERO output terminal.
The Pulse Amplifiers Type R602 perform pulse standardization as well as amplification.
The clear pulse should be 400 nanoseconds in duration.
The type of delay unit depends upon the number of bits in the counter, since this deter
mines the maximum time required between counts. If an .R302 Delay is used, at least two
units must be in the loop to provide the required recovery time. The Type R303 does not
have a recovery time requirement. -
A complete conversion requires 2" steps, where N is the number of bits in a counter. The
average number of steps is 2"-'. Calculations of the time per step must take into account
the following:
Carry propagate time of the flip flops
Total transition time of the flip flops
Delay of the level amplifiers
Delay through the ladder network
Transition time of the comparator and
settling time (0.15 +0.05N microseconds)
Gating time
Synchronization time (if required)
309
310
If buffered flipflops are used, the control flipflop should also be buffered and would be
set through the input terminal. Also the comparator signal should not gate the counter
input directly. Any noise on either the analog input or the ladder output could cause the
comparator input to be in a transient state at the time the out pulse occurs. This could
result in a split or partial pulse which might not propagate fully. To avoid this possibility
'the gating inverter on the counter may be' either synchronized or eliminated, since the
pulse will be inhibited from continuing through the loop.
CONTINUOUS CONVERSION
Continuous analogtodigital conversion can be performed using an updown counter. At
each step the counter output is compared with the analog input, and a pulse is added to or
subtracted from the counter, as necessary. By proper adjustment of the comparator, it is
also possible to inhibit counts when the analog signal is appr'oximately equal to the digital
number. The inhibit signal is formed by adjusting the comparator outputs so that they
do not switch simultaneously. For ideal operation, the two outputs should be in the same
state whenever the digital feedback signal is within =,='12 part in 2N of the appropriate
input signal.
Synchronization (the use of control flipflops) is required in all continuous converters. Any
noise on the inputs to a comparator could cause the outputs to be in a transient state at
the time they were sampled. Thus, if the signals are not synchronized, add and subtract
pulses could enter the counter at the same time.
A continuous converter is illustrated in Figure 26. This system uses the R series unbuffered
flipflops with level change carry propagate. The illustration consist of four basic parts: the
updown counter, the DAC, the comparator, and the synchronizer and control logic. Two
control pulses are formed by a clock and a delay unit. The synchronizer pulse sets the up
sync flip flop if the enable level from the comparator indicates that the feedback signal is
smaller than the input signal. Similarly, the downsync flipflop is set if the analog input is
larger than the feedback input. Two sets of diode gates are used to inhibit counting which
would cause the counter to overflow.
The outputs of the set flipflops are exclusive ORed together to assure that no count sig
nals will be generated if both flipflops are set. After these signals have had time to set up,
the clock generates a count pulse which samples the levels and produces a count up pulse
or a count down pulse. At the same time the up and downsync flipflops are reset so that
the enable signals can be read in the nexttime. All of the synchronizer and control logic
should be from the same speed line. The flipflops illustrated here are Type R202s, the
clock is a Type R401, the delay a Type R302, and the inverters and diode gates are 2
megacycle logic.
The counter can be. a lower speed logic than the synchronizer if desired. In this case, the
up and down count pulses should be stretched with Type R602 Pulse Amplifiers to pro
duce pulses of appropriate duration. In the illustration using Type R202 FlipFlops, the
outputs are buffered since the flipflop outputs drive a capacitor diode gate level input,
a capacitor diode gate pulse input, and a diode gate input, as well as providing the signals
to the DAC. The inverters used for buffering are 10 megacycle units which have a minimum
of capacitance. Readout from the counter register should take place from the output of
the inverter buffers so as not to exceed the loading on the flipflops.
311
The continuous conversion method is applicable when the maximum rate of change of the
analog voltage is less than the fastest possible rate of change in the converter. That is:
where
( ~ ; ) input ~ ~ : : ~
( ~ ; ) input
is the rate of change of the input analog voltage, V"' is the full scale voltage of the con
verter, N is the number of bits, and ~ T is the time per step. L'lT is the sum of:
Carry propagate time for the flipflops
Total transition time of the flipflops
Delay through .the divider network
Transition time of the comparator and settling time
(0.15 0.05N microseconds)
Total transition time for the synchronizer
15 microseconds
Delay through gates and pulse amplifiers if in feedback loop
312
ANALOG
INPUT
'" SUPPLY
F F O ( O ) ~ ~ - r - ........
FFI(O)
FF2 (0)
FF4(Q)
Figure 26.Continuous Converter with Unbuffered FlipFlop
313
Carry propagate time may be reduced splitting the carry chain. For example, if flip flops 2,
3, and 4 are in the ONE state, the countup signal can be gated to complement flipflop 1 as
well as flipflop 4. Of course, the normal carry input to flipflop 1 is not used.
SUCCESSIVE APPROXIMATION CONVERTER
This method repeatedly approximates the input voltage. At each step, the possible range
of the input signal is divided in half. The converter uses a digital register with gatable
ONE and ZERO inputs, a digitaltoanalog converter, a comparison circuit, a control timing
loop, and a flipflop distributor register that determines which step' is taking place (see
Figure 27). The distributor register is like a ring counter with a single ONE circulating to
indicate which step is taking place. At the beginning of the conversion, both the digital
register and the distribution register are set with a ONE in the most significant bit and a
ZERO in all bits of lesser significance.
At the same time, a pulse enters the delay chain. When this pulse has had sufficient time
to make one complete loop through the chain, the digitaltoanalog converter and the com
parator have settled and the comparator output determines whether the next digital
approximation should be larger or smaller. At this time, the next most significant bit of
the digital register is set to a ONE, and the most significant bit either remains in the ONE
state or is reset to a ZERO, depending on the comparator output. The single ONE in the
distribution register is shifted to the next position. This procedure is repeated until the
final approximation has been corrected, making a total of N steps, plus settling time for
the last flipflop. .
The total time
o
required is N ~ T, where ~ T, the time per step, is at least as large as the
sum of:
Total transition time for the flip flops
Delay through the divider network
Transition time of the comparator and settling time
1 microsecond for N<5,
0.2 N microseconds for 5<N<8,
2.4 microseconds for N=9,
3.0 microseconds for N= 10
Delay through pulse amplifier and gates
No synchronization time is required for this method since the comparator never controls
the action of more than one flip flop.
The digital register and the distribution register use the Type R202 flip-flops. Due to the
set up time of the internal gates, the time per step must be at least 1 microsecond.
The control delay chain uses Type R302 delays and R602 pulse amplifiers.
314
w

01
START
Y,N
E

+ F
M M
EOG
9
EO
f
N K IN La NaN a N a
I P L 0 J, P U J E K P U E K P U E K P U E K ..10 1
N_...L D_-L
R602 K I K
H 83

J'E Tv L V L V L V L I F 82
L J
M
Ip
S T
o R20Z 1
R 610
K...1.. D . ..L. I IN
UL
9v
[U
E
JlQ:
1
0 R202 \
F 89
R
u' D . .!.. K
vi [fE LIUP VIUE
Figure 27 logic Diagram of Successive Approximation

LIUP

316
CHAPTER 5
BASIC CIRCUITS
This chapter includes general information on the use and importance of various charac-
teristics of converter circuits. Detailed performance characteristics are given for specific
Digital modules. Definitions of symbols and terminology are included in the appendix.
Any of the appropriate modules described in the Digital Module Catalog can be used
for the flip-flop registers and the control and gating logic in a converter. The choice is
governed by speed requirements in the system.
DIGITAL - ANALOG CONVERSION MODULES
The A60l (Figure 28) is a three-binary-bit digitaltoanalog converter utilizing a star-type
divider network and three precision germanium-transistor level amplifiers. It may be con-
nected in series with other converters to form higher resolution converters. The accuracy of
the A60l is suitable for up to eight bits of conversion. For higher resolution, it should be
combined with the Types A604 and A605.
A -3vinput signal at all digital inputs produces ground out. The input load is 1 ma at
ground. If all inputs are not required, the most significant inputs should be used, and the
least significant ones should be left open circuitEid .. The converter input may be driven from
the converter output of another module in order to' provide higher resolution. If not driven
from another unit, it shGluld be terminated with 1000 ohms to ground. A termination
resistor is included in the module. The reference input requires a -15 ma DEC A704
supply. The supply should be adjusted to approximately-lO.Olv to overcome the satura-
tion resistance in the level amplifiers. High Quality Ground is the ground return for the
reference supply and should be connected to the supply terminal and eventually to chassis
ground at a noise-free location.
The output is the analog equivalent of the digital input. The most positive output is Ov. The
most negative output is -lOv less the value of the least significant bit. The output imped-
ance is 1000 ohms. If a bipolar or reduced output swing is required, the output may be
loaded with 1000 ohms or more without affecting the accuracy.
The A604 and A605 (Figure 29) are twobinarybit digital-to analog converters for use with
the A60l in forming high resolution, high accuracy converters. Inputs and outputs are
identical to the A60l except that a terminating resistor is not included. Germanium tran-
sistors are used.
317
ACCURACY:'
TEMPERATURE
COEFFICIENT:
OUTPUT
IMPEDANCE:
SETTLING
TIME:
DAC OUTPUT
GROUND r-
I
DIGITAL GROUND 1.._
(MUST BE GROUNDED
EXTERNALLY)
TERMINATING
RESISTOR
DIGITAL INPUTS
DAC INPUT
REFERENCE
HIGH QUALITY
GROUND
Figure 28. A60! DIGITALANALOG CONVERTER
-DAC OUTPUT
GROUND r-
I
DIGITAL GROUND 1.._
(MUST BE GROUNDED
EXTERNALLY)
-DIGITAL INPUTS
DAC INPUT
REFERENCE
HIGH QUALITY
GROUND
Figure 29. A604, A605 DIGITALANALOG CONVERTERS
A60! A604 A605
0.25% of expected 0.025 % of expected 0.005% of expected
value or value or value or
O.5 mv, whichever O.25 mv, whichever O.05 mv, whichever
is greater is greater is greater
100 ppm/DC max 25 ppm/DC (from 10 ppm/DC (from
(from + IODC to
+ 10C to +45C) + 10C to + 45C)
+45C)
1000 ohms 0.1 % 1000 ohms 0.1 % 1000 ohms 0.1 %
.-'
300 nsec 300 nsec 1.5 II.sec
"At 25C Includes tolerance of 1.5 v on the 10 v and -15 v supplies.
The following combinations of modules are recommended.
Resolutions % of Ana log Accu racy Units
(bits) Full Scale (% of Full Scale) (quantity-type)
upto 8 down to 0.39 % 0.25% 3A601
910 0.195% to 0.098% 0.082% l-A604, 3-A601
11 0.049% 0.038% 2A604, 3'A601
12 0.024% 0.014% l-A605, 2-A604, 2-A601
13 0.012% 0.01% 2A605, 2A604, 2A601
318
Reference Supplies
The reference supply determines the voltage range of the converter. It is important that
the supply be of good quality, since any error in the reference voltage will translate directly
into error in the converter. That is, a 0.1 percent ripple in the reference produces a 0.1
percent ripple in the analog output.
Digital manufactures a reference supply: The Type A704, which may be used for systems
of 13 bits or less. The supply is mounted on a Digital module and is driven by Digital
standard power voltages. The characteristics are shown in Figure 30.
AM AN
AT + SENSE
AV - SENSE
AE -OUTPUT
AS POWER INPUT
Type A704. Precision Power Supply
Module
Type
Output Current
A704 -10v -90 to +40 ma
Module Adjustment Input
Type Resolution Power
REF
SUPPlY
Stability
1mv/Bhrs
AT
AC
AE
1 mv/15 to 35C
4 mv/O to ooC
Use
Ripple
Regulation Peak to
Peak
0.1 my. no
load to lull . 0.1 mv
load
Output
Impedance
-15 volts/lOa ma Load with 5000
10 ma pi at load.
-152 volts/
See below lor
A704 0.01 mv sensing and 0.0025 ohms
250 ma
preloading
TYPE A704 PRECISION POWER SUPPLY
REMOTE SENSING
The input to the regular circuits of the
A704 is connected at sense terminals at
(+) and AV (-). Connection from these
points to the load voltage at the most
critical location provides maximum reg
ulation at a selected point in a distrib
uted or remote load. When the sense
terminals are connected to the load at a
relatively distant location, a capacitor of
approximately 100 microfarads should
be connected across the load at the
sensing point.
PRELOADING
The supplies may be preloaded to ground
or -15 volts to increase the current
available in either direction. -125 ma
maximum can be obtained by connecting
a 270[2 5%, 1 watt, resistor from the
-10 v pin AE reference output to pin AC
ground.
Figure 30. Reference Supply Specifications
319
I
Comparators
The comparator is an unusual circuit because it is a hybrid, partially analog and par-
tially digital. Basically, it is a very high gain difference amplifier_ The outputs "saturate"
quickly so that they do not exceed standard levels_ (For Digital equipment these are
o and -3 volts_) The comparator, Type A502, has an input range of 0 to -10 volts_
When the input differential is large, the dual outputs are complementary Digital levels_
The time required for the comparator to switch states depends on the desired
resolution and the conversion method. That is, the comparator takes longer to respond
to a 10-millivolt differential iriput than for a 100-millivolt differential input. A 7-bit system,
for example, seldom requires information about lO-millivolt differentials. Similarly, the
comparator switches faster in a counter or continuous converter system, where the
differential input is being reduced gradually, than in a successive approximation con-
verter, where the differential voltage may go from 5 volts to 0 in one step.
FACTORS AFFECTING COMPARATOR ACCURACY
As the input to the comparator circuit varies from 0 to -10 volts, the switching point
may vary also. That is, one of the inputs may have to go somewhat more negative than
the before the outputs switch. A similar effect occurs with changes in temperature.
In the Type A502, the switching point does not move more than 5 millivolts as the input
voltage changes from 0 to -10 volts and the temperature changes over a range of
20C (around room temperature). Power supply fluctuations of the Type 728 or an
equivalent supply change the switching point by less than 0.6 millivolts.
Both outputs will not switch simultaneously unless the amplifier is in perfect balance.
If the switching point changes mentioned above affect both outputs equally, the effect
is referred to as a shift in the common switching level. If one of the outputs is affected
more than the other, there is an offset between the two sides of the amplifier.
USE OF THE COMPARATOR
Illustrations of the use of the comparator are included in Chapter 4. In other uses, it
must be remembered that the comparator is a hybrid circuit. If the differential input
is small, but balance is not quite right, the outputs may not be complementary. If there
is a small amount of ripple on the input, the outputs may oscillate. Normally this effect
is of no concern because the errors have already been taken into account in the common
mode and resolution specifications. However, if the results are to be read into more than
one flip-flop, output oscillations can cause different information to be read into the
different flip-flops. Thus, the outputs must be synchronized before being read into more
than one flipflop. Similarly, the comparator level changes should not be used as infor-
mation unless it is certain that transient signals (such as those that occur when a DAC
switches) will not cause false outputs.
320
:
+F -1/
Type A502 Comparator
Specifications:
Input Range: 0 to -10 volts
Input Impedance: 1 microampere, 125 picofarads (The input current depends on the
relative polarity of the two inputs. The more positive input may draw up to 1 micro
ampere and the more negative input may supply up to 1 microampere. The maximum
current difference between states is 1 microampere.)
Outputs: Two outputs, 0 and -3 volt levels
Output Loading: 7 units base load at dc, 1 unit for maximum speed
Resolution: 1 millivolt at dc
Common Mode & Temperature: 5-millivolt maximum equivalent input offset for 10-volt
common mode change and 20C change
Speed:
Depends on application, principally on the ratio of the voltage difference before passing
through the switching pOint (VB) to the voltage afterward (VA)' Speed is affected to a
lesser degree by the length of time the input difference is at VB, by the f11agnitude of VA>
by the source impedance, and by the load. Typical speeds in an analogtodigital con-
verter system where the source is a ladder network and level amplifiers, and the load is
a 1 unit base load, are listed below. (These speeds include allowances for extra divider
settling times at high accuracies.) For more information on specific applications, see
Chapter 4.
V./VA VA in mv Time in fLsec
-512 10 3.0.
-128 40 1.6
-32 160 1.2
-2 20 0.6
-2 80 0.5
-1/512 10 0.15
Adjustment:
Two potentiometers control zero set and common balance. See Chapter 6 for adjustment.
Power: -15 volts/55 ma; +10 volts (A)/O; +10 volts (8)/21 mao
Figure 31. Comparator Specifications
321
I
When the comparator is used in a digital voltmeter or a continuous converter, it is
usually desirable to have builtin hysteresis which is just slightly less than % LSB.
The hysteresis avoids converter chatter (switch back and forth between two states)
when the input voltage lies on a boundary between the two states. It is possible to
introduce some hysteresis into the A502 by adjusting the common switching level and
offset control so that the two outputs do not switch simultaneously. In a continuous con
verter, counting would then take place only when the two outputs were of opposite
polarity. No action would take place when both outputs had the same polarity. The same
type of logic would be applied in a digital voltmeter, the exact action depending on the
conversion method used to arrive at the results.
Multiplexer Switches
ANALOG MULTIPLEX SWITCHES
An analog multiplex switch is like a relay, in that two points are opened or shorted on
command from an external source. Digital manufactures a relay switch, Type AlII, for low
speed operations, and a solid state switch, Type A12l, for high speed operations. The con
trol inputs to these switches are 2input AND gates, each with a separate control input and
all with one input in common. If the control inputs are driven from binary to octal decoders,
up to 512 switches can be placed in parallel. The accuracy and speed limitations are the
switch capacitance and the amount of leakage current back through the switches.
The switches can be tied in parallel, cascaded to give double level multiplexing for large
systems, or used in other applications such as the sample and hold circuitry.
USE OF THE TYPE A121 SWITCH
In the off position, most types of A12l switches can have upto 15 volts across them, ranging
from +10 volts to -5 volts referenced to ground.
322
180
AI21 Solid State Switch
M R P N S
(GUARD) 9 9 (GUARD)
1 I
r---t- --1---,
o-f- '""'"-+ __ --+---+---+1...-,
J 1 I

LO--+- :-1 -+------+------4 .......
1

L
1 L ______ I J
-v
,-
I
1
1
I
L_
-1---- -----;-
I
1
1
1
--,--,
I 1---,-
I I I
1 1 I 180
_"!_J 6H
(COIL SHIELD) L-__ ,.
I
I
1
1
--,-""
I I
1 1
I I
1 1
___ ...J
E F I
D I
I I
______________
AliI Relay Switch
Figure 32. Multiplexer Switchs
323
I
Multiplexer Switch Specifications
TYPE NUMBER AI21 Alii
Type of Switch Solid state Relay
Number of Circuits 4, single pole, independent 2, double pole, outputs bussed
Control
Signals Digital levels Digital levels
Enable -3v -3v
Load 1.3 ma load shared 3 ma load
among grounded inputs shared among ground inputs
Signal
Max. voltage +10 -5 lOv
Max. cu rrent +1 ma I ma
"On" offset (max.) 0
"On" resistance 4501] 021]
(max.)
"Off" resistance, 10 na 5 x 10'1]
leakage
Speed
50% input to .01 % Delay + sync + charging time Delay + bounce setting
of output
Turn on delay 100 nsec . 0.9 msec
Turn off delay 50 nsec 0.06 msec
Bounce setting 0.3 msec '
Life 109 operations at low loads
TABLE 8 USEFUL LOGS
tIT
1 - e tIT e tIT
3 0.95021 0.04979
4 0.98168 0.01832
5 0.99326 0.00674
6 0.99752 0.00248
7 0.99909 0.00091
8 0.99966 0.00034
9 0.99988 0.00012
10 0.99995 0.00005
The switch is turned on when the two control level inputs are negative voltage (or open).
When changing the state of the switches, care should be taken that two switches tied to a
common node are never turned on simultaneously. If the control levels come from 10-
megahertz flip-flops which are all changed simultaneously and which are decoded by 5-
megahertz binary octal decoders, the switching is fast enough so that there is no danger of
shorting. If low speed circuitry is used to drive the switch, or if the controlling flip-flops are
not all changed simultaneously, one of the enabling inputs should be grounded before the
state of the switches is changed. This will put all of the switches in the off position and
assure that there will never be a make before break situation. If the switches are turned on
324
simultaneously, no damage will occur if the voltage ratings are observed, but large signal
and transients will occur. It should be noted that the fall time of the output waveform of a
single multiplexer switch is almost entirely determined by the load impedance and may be
quite long for high load impedances.
RELAY MULTIPLEXER TYPE A111
The Type AlII contains two double pole switches which can be used for differential multi
plexing. The switch outputs are connected to an output bus. The speed of the relay is
determined by the delay in turnon plus the bounce settling time. For the Type All1 this
delay is 1.2 milliseconds. There are two control !evel inputs for each relay; one of these is
common to both relays on the module, and one is independent. The relay is turned on when
both of its control inputs are negative.
Analog Amplifiers
Amplifiers are sometimes used at the input of an analogtodigital converter to shift the
input range, scale the input range, provide a differential input, or isolate the input signal
from the converter. Amplifiers are used on the output of digitaltoanalog circuits to shift
or scale the output range, to reference the output signal to the external ground, and to
lower the output impedance. The last two features are important when the two pieces of
equipment are separated by a distance that makes noise pickup likely. In this case, it is
best to put the amplifiers at the driving source end, that is, at the output of the divider
network in a digitaltoanalog conversion or at the signal source for analogtodigital.
The most usefui amplifier for these applications is the operational amplifier. Its high input
impedance and high gain make it a building block just as a flipflop or nor gate is a digital
building block.
Usually an operational amplifier is designed to roll off at 6 db per octave. This ensures less
than 180 degrees additional phase shift so that the amplifier will not oscillate. This also
produces a simplification in calculating the dynamic aspects of putting an amplifier in the
system, although for a rigorous analysis, much more detail about the characteristics of the
amplifier must be known, and the calculations are far more complex.
The approximation of infinite input impedance and infinite gain can be used in designing
with good quality operational amplifiers with negligible error.
R,
Figure 33a shows the amplifier operating as a simple inverter. The gain iSR: The imput
impedance is R; and is returned to virtual ground (within microvolts of true ground). The
gain accuracy and stability is that of R, and Ri The output impedance depends on the dif
ference between open loop gain and closed loop gain. Typically, the closed loop gain is not
very high, less than 100, and the open loop gain is in the order of 10'. Under these circum
stances, the output impedance is less than one ohm.
325
-v
(A)
E out
R,
(B)
(e)
E out
.Figure 33. Typical Amplifier Configuration for Scaling and Biasing
Digital-to-Analog or Analog-to-Digital Inputs
326
E out
Figure 33b shows how a signal may be offset with a negative or positive reference voltage.
In b, the positive reference is summed with the input voltage, E in. The amplifier will keep
the turning point at ground so that a sample calculation of resistances and currents is all
that is required to determine the offset. Since the summing point is at ground, each input is
independent of the other. If a half scale offset is required,
+V 2R,
~ = T
the input signal current applied thru Ri will be superimposed on the above dc level. Figure
a utilizes a gain of two and a times '/2 alternator. If the input voltage swings from -5 volts
to +5 volts, the alternator output will vary from zero to -10 volts.
Figure 34c shows a noninverting "potentiometric" amplifier configuration. The output will
equal the input and the gain will approach + 1. The current drawn by the non inverting
input will be nearly zero. The input impedances achieved can approach 500 megohms.
Scaling, offsetting and differential input to an analogtodigital circuit follow exactly the
same method. The input resistor would be R1 minus the appropriate output impedance
of the driving signal.
When amplifiers are required on converters, it is generally best to use the same type of
amplifier as is being used in the analog portion of the circuitry to keep the performance
characteristics the same. When an amplifier is being used internally in the converter, such
as between a group of multiplexer switches and an analogtodigital converter, a higher
performance amplifier is required, since it will be asked to take full scale changes and
settle within a very short time.
The A200 consists of a DEC amplifier (part #1505379) mounted on an A990 amplifier
board and includes a gain trim and balance potentiometer. Mounting holes are provided for
input and feedback networks and roll off capacitor. The amplifier is supplied to Digital by
Analog Devices and is identical with the Analog Devices 102 A.
327
Open Loop Gain:
Rated output voltage (a20 mal
Frequency response
Unity gain, small signal
Full output voltage
Slewing rate
Overload recovery
Input voltage offset
Average vs temp.
Vs supply voltage
Vs time
Input current offset
Average vs temp.
Vs supply voltage
Input inpedanee
Between inputs
Common mode
Input voltage
Input
Max common mode
Common mode rejection
Pawer
Voltage
Current at rated load
-INPUT
+INPUT
-15V
2 x 10'
l1v
10 me
300 ke
30v/"see
200
1
,see
Adjustable to Zero
20 uv;CC.
15 uv/%
10uv/day
2 na
0.4 na;CC.
0.15 na/%
6 megohm
500 megohm
15 volts
10 volts
20,000
8 "volts
15 to 16 volts
35 ma
+15V
V OUTPUT
F COMMON
Figure 34. A200 Operational Amplifier
328
CHAPTER 6
INTERCONNECTION AND CALIBRATION
Grounding and Shielding
If the converter system operates with eight bits or more, careshould be taken with the
system wiring to avoid noise pickup and ground potential differences between the analog
equipment and the converter. Since the digital voltages are low level, the major noise
source within the converter is fastswitching transients, particularly pulses. Their effects
can be minimized by isolating the analog portions (the divider network, level amplifier,
reference supply and comparator) from the digital portions of the converter. Pulsegen
erating devices, such as clocks and pulse amplifiers, should be farthest away from the
major analog components.
Single wires can be used within a mounting panel or between two panels if no noise sources
are nearby. Coaxial cable is best, of course, for long leads. However, a twisted pair is
usually sufficient, even in high accuracy systems, if pulse lines and other fast transients
are avoided. The shield conductors of the coaxial or twisted pair should be tied down at
one end only, and this end should go to a good ground, not near a pulse transformer or
other high frequency device.
On the precision level amplifiers, there is a separate input for the high quality ground. These
inputs can be tied together and fastened to chassis ground at a good solid point. In general,
excessive analog ground loops should be avoided.
Similarly, in large systems sense wires from the reference voltage supply should be
brought to a point near the load so that the supply will regulate the voltage as seen by the
load, not the voltage as generated at the pOwer supply. If the load is distributed, capacitors
at the main load points will reduce transients caused by the rapid switching of the DAC's. A
separate groundshield is brought out on the Type AlII Multiplexer Switch. It isolates the
analog signal from noise transients generated by the multiplexer control signals. Any solid
ground source can be used here, but this analog signal is not a ground reference for
the system.
Signals from a high impedance output are more sensitive to noise pickup than those from
a low impedance output. Thus, if a digitaltoanalog converter is to drive long leads where
noise could be picked up, the output should be buffered with an amplifier having a low
output impedance.
The size of the ground potential differences which can occur between the converter and
the analog input or output signal should also be minimized. If it is not possible to place
the two pieces of equipment close to each other with the grounds tied tightly together, a
329
heavy ground strap can be run between them. Alternatively, the ground potential differ
ences can be subtracted out. In an analogtodigital converter this is done by using a
differential amplifier at the input, or by using two standard operational amplifiers. In the
latter case, one of the amplifiers is used to invert the ground; then the signal and inverted
ground are summed. In digitaltoanalog conversion, the ground from the signal destina
tion is brought back to the converter, inverted with an operational amplifier, and summed
with the signal.
CAUTION
The multiplexer switches are low impedance switching
circuits. Precautions should be taken against possible
shorting of the analog inputs and outputs of these circuits
to any other low impedance source, including ground.
such shorting could damage either the circuits or the
signal sources.
CALIBRATION
EQUIPMENT NEEDED
The adjustment and calibration procedures outlined here are designed to be as simple as
possible. Three pieces of equipment are needed, as follows:
1. A digital "module extender.
2. An oscilloscope with a high gain accoupled vertical amplifier and a dual trace
amplifier.
3. A reference for determining proper gain setting. Can be a standard voltage or a refer
ence from the analog equipment.
GENERAL PROCEDURE
There are six kinds of calibration needed for basic conversion systems. They apply to
digitaltoanalog converters and to three types of analogtodigital converters: the counter,
continuous, and successive approximation types. Other conversion systems require
. basically the same kind of adjustments, with certain steps added or omitted depending
upon the circuits used. In this chapter, the six general procedures for calibration are pre
sented as follows:
Steady State Calibration
DAC Networks
Comparators (analogtodigital conversion only)
Offset and Gain
Speed
Noise and Ripple
DigitaltoAnalog Adjustment
AnalogtoDigital Adjustment
330
Calibration should follow in the same order as the procedure given. In steady state calibra
tion particularly, the DAC, offset, and gain adjustments carried out in that order make it
unnecessary to repeat previous adjustments for fine trimming. The effects of these adjust
ments can be seen easily if output is ploUed against input. Normally, with the digital
number 0 in, the output should be 0 volts out, and vice versa. Similarly, maximum input
should yield full scale output. Intermediate points should fall on a straight line between
these two points.
An uncalibrated converter, however, produces the non linear plot shown below.
OUTPUT
INPUT
By adjusting the DAC networks, the irregularities are removed from the curve and it
becomes a straight line. Next a small amount of offset is added to the network; shifting
the curve up and down, as shown below, until the zero input gives a zero output.
OUTPUT
INPUT
Finally the gain is adjusted until a relatively large input produces the correct output. The
slope of the curve will change as follows:
OUTPUT
INPUT
331
I
DIVIDER ALIGNMENT
The DAC is aligned to compensate for variations in resistors in the divider network and for
variations in the output impedance ot"the level amplifiers. The output voltage from the bit to
be calibrated is compared with the output voltage resulting from all of the bits of lesser
significance. The difference is trimmed so that it is equal to one least significant bit. A
simple setup for making this 'adjustment is shown in Figure 35. The clock, delay oneshot,
and inverter simulate a digital input to the converter. Here they are shown switching the
inputs between 00100000 and 00011111. Thus, the bit under test is the third bit,and the
adjustment is made with the trimpot on that bit.
Figure 35. DAC Adiustment
REF
SUPPLY
The output should be monitored by an oscilloscope with a high gain, ac coupled vertical
amplifier. The amplitude of the output should be one least significant bit voltage contribu
tion. Since this alignment also adjusts for variation in output impedance of the DAC's, the
level DAC's should be aligned in the same relative order as in the final system. Any unused
bits of lesser significance should be connected to -3v. in the test setup just as in the
system. Inputs to DAC's in more significant positions may be either grounded or connected
to -3 volts, as long as they remain constant throughout the alignment.
In starting the alignment, it is advisable to check a nonadjustable bit first to make sure
that the setup is correct. Using this method, the adjustment of bits of lesser significance
is independent of the adjustment of bits of more significance. Therefore, the least sig'
nificant adjustable bit should be checked first, then the next in order, and so on toward
the most significant bit. The alignment is then completed in one step, without the need
of going back to realign any portion.
332
The sensitivity to the trim pot motion depends on the number of bits being used. In a lObit
system, where the voltage differential being observed is approximately 9.8 millivolts, the
motion of the least significant potentiometer is barely seen on the scope. Working toward
the most significant bit, the adjustment range will become larger and larger until it may
be possible to invert the relative values of the outputs. To avoid such inversion, note
whether the longer portion of the rectangular wave corresponds to the more negative part
of the signal or to the more positive part of the signal. Be sure that this relationship con-
tinues the same for all of the bits.
Since the DAC'suse fine resolution wirewound trim pots, ascertain that they are in a stable
position and that the slider arm is not resting on a single wire where it could jump away,
possibly in the wrong direction. After trimming, tap the pot once or twice with the end of a
screwdriver and, if the output changes, retrim to the stable position closest to the
ideal value.
There are two advantages of this particular meihod of alignment. As the trimpot is changed,
the dc level will move up and down. However, since the levels are being observed simul
taneously, and only the difference is being monitored, the scope can be put on ac and the
picture will stay in the center of the scope face. Also, since the adjustment is made on the
differential between two states, a 10 per cent error in the adjustment will give an error
in the system of only 10 per cent divided by 2".
An ohmmeter should not be used to adjust the ladder network, since it will not take into
account the output impedance of the level amplifiers. Do not try to adjust the output with
a voltmeter since the dc level shift would require switching between the two states, and
the measurements would be extremely confusing and timeconsuming.
NEED FOR REALIGNMENT
Realignment should not be necessary under normal conditions. The system should be
checked if the modules have been subjected to a drastic change of temperature or to a
mechanical shock sufficient to change the trimpot settings. Realignment should be done
if one of the DAC modules is changed.
THE COMPARATOR TYPE A502
The comparator can be adjusted easily with a dual trace oscilloscope and a clean-SN dc
source applied simultaneously to the inputs.
The two outputs of the A502 are viewed simultaneously on a dual trace oscilloscope. With
both traces synchronized to a single point, the two signals will appear as roughly comple
mentary square waves. For most applications the comparator should be balanced; that is,
both outputs change .simultaneously when the relative .polarity of the inputs changes.
For continuous converters or digital voltmeters, however, such adjustment would cause
the converter to oscillate around a dc level. In this application it is desirable to delay
switching of the outputs until the input analog signal is almost V2 LSB away from the
divider input. The small amount of hysteresis introduced prevents converter chatter.
333
ADJUSTMENT FOR BALANCE
The comparator should be adjusted so that the two outputs are perfect complementary
square waves. Adjust the upper potentiometer to make the two waveforms complemen:
tary. Adjust the lower potentiometer for equal positive and negative portions of the square
wave. These controls are somewhat interdependent, so it is necessary to repeat the
adjustments until the optimum symmetry is obs.erved. The resolution can be increased
by reducing the size of the input sine wave and repeating the adjustment.
COMPARATOR WITH HYSTERESIS
Begin the adjustment with a difference voltage, applied to the input, that is equal to the
amount of hysteresis desired. Proceed to adjust the comparator as above, under Adjust-
ment For Balance_
The comparator adjustment. can change with time, temperature, or a mechanical shock
severe enough to jar the potentiometers .. The need for readjustment depends on the
accuracy required and the environment. Usually a monthly check is more than sufficient.
Testing can be done by taking the comparator out of the converter and employing the
above method or by testing the overall system, as described in the ne,xt chapter,
OFFSET AND GAIN
Offsetting and gain adjustments should be made on the assembled system. In a digital-to-
analog converter, a digital number is put in and the output is observed with'a voltmeter. In
an analog-to-digital system, a voltage is put in and the switching points are observed. Offset
and gain adjustment are necessary to compensate for the open-circuit voltage drop of the
DAC's, which can be as high as.10 millivolts in precision converters. In an analog-to-digital
converter offsetting is also necessary to center the. quantization error; that is, if the state
zero corresponds to 0 volts and the s t ~ t e one corresponds to 10 millivolts, the converter
should switch between states zero and one at an input of 5 millivolts.
OFFSET
The calibration should begin with the offset. A- positive voltage can be applied through a
large resistor (usually on the order of 1 to 10 megohms) to the digital-to-analog converter
output. The size of the bias resistor, or the amplitude of the bias voltage can be varied until
the offset is correct. For digital-to-analog conversion, zero in gives zero out. For analog-to.
digital, an input of % LSB, produces the first switching point.
The offset voltage source can be the standard +10 volts (for systems of up to 10 bits).
The percentage variations in the voltage supply for the offset signal are scaled -according.
to the amount of. bias obtained. That is, a 10 percent ripple on the bias supply produces
a 10 percent ripple in the offset; so if the offset is 10 millivolts, a 10 percent ripple
would be 1 millivolt.
GAIN
The gain adjustment on a digital-to-analog converter is made by setting the digital number
to half-scale or full'scale and adjusting the reference voltage supply until the'output has
334
the correct value. The Type 1562 has one trimming potentiometer for this adjustment.
The Type A204 has both a fine and a coarse adjustment. In checking the output voltage,
remember that the output impedance of the divider network is about 1000 ohms. Thus,
loading the output with one megohm would reduce the output voltage by 0.1 percent.
This adjustment should be done with the same load as in the final system.
The gain adjustment for an analogtodigital converter differs only in that the common
mode effect of the comparator must be taken into account. The comparator has been
balanced with a 5 volt common mode. The offset has been applied so that the lowest
switching point is correct. Therefore, the reference adjustment which gives the correct
halfscale switching point is different from that which gives the correct fullscale switching
point. Generally, maximum accuracy is desired in the lower part of the scale, so the
mid scale point should be used. Alternatively, the reference may be i1diusted for th1!
best fit between halfscale and full scale points. In calculating the switching points for
these measurements, be sure to remember the quantization offset that has been intro-
duced. This adjustment should be made at low speed (10 to 20 microseconds' per step).
SPEED ADJUSTMENTS
NOISE AND RIPPLE
When the DAC's switch, a transient current is drawn from the reference supply. Normally
this supply is loaded with a capacitor to reduce noise (see power supplies, Figure 30), but
it may also be desirable to place small capacitors at the reference inputs of the individual
DAC's and possibly also between the high quality ground and the chassis ground. The
voltage at the load can be monitored on a scope with a highgain dc-coupled plug-in unit.
Care should be taken that the noise being observed is actually there and not introduced
through the scope or by a ground lead attached to the wrong point. The reference and
the ground tend to move together. If the input signal source is referenced to this ground,
it also moves. Thus, the scope should generally be disconnected from ground at the power
and connected to the converter ground at a good solid point, and the cable running to this
scope should be prevented from introducing additional noise.
ANALOG -TO- DIGITAL ADJUSTMENTS
As mentioned previously, the speed and accuracy of an analog-todigital converter are inter-
related. That is, if the converter is run too fast, the DAC's and the comparators do not have
enough time to settle to final value. In the range of 6 to 10 bits, even a tenth of a micro-
second per step can make a considerable difference in the system accuracy. Thus, the
speed may be adjusted for the maximum allowable time and hence the maximum accuracy,
or it may be adjusted for the minimum time required to give the required accuracy.
335
I
Test for speed should be made by checking the major switching points. For most con
verters' (including the counter, continuous, and successive approximation types) these
are around one-fourth, one-half, and of full scale, as shown below.
Area Being Checked From To
'/4
0011 ... 110 0011 ... 111
0011 ... 111 0100 ... 000
0100 ... 000 0100 ... DOl
'/2
01l1 ... 1l0 0111 ... 111
0111 ... 111 1000 ... 000
1000 ... 000 1000 ... DOl
%
1011 ... 110 1011 ... 111
1011 ... 111 1100 ... 000
1100 ... 000 1100 ... 001
For a more detailed check, the switching points around 1/8, 7/8, 1/16, etc., might also
be included.
DIGITAL-TO-ANALOG ADJUSTMENTS
The digital-to-analog converter output contains transients when many bits are changed
simultaneously, such as in going from 01111 to 10000. These transients are caused by
variation in flip-flop transient times and propagation time through the divider.
Transients can be reduced by adding a small choke in series between the flip-flop and
level amplifier, or by loading the flip-flop with a resistor to -15 volts. Further reduction
can be made with a low pass filter on the output (remember that the digital-to-analog
output impedance is 1000 ohms.)
336
CHAPTER 7
TESTING AN ANALOG-TO-DIGITAL CONVERTER
The adjustment and calibration procedures detailed in the previous chapter should result
in a converter that operates correctly over the whole range. To make sure the converter
meets specific accuracy requirements, testing may be desired, and simple operating
checks should be repeated at regular intervals to assure continued correct operation.
If the converter is part of a general purpose computing facility, complete testing can be
performed easily under program control. If, on the other hand, the converter is part of a
specialized system or is to be tested before installation in such a system, manual or semi
automatic testing IS necessary and will probably cover only the worst cases. For most
converters (counter, continuous, and successive approximation) the worst case's are seen
at the major switching points; namely, onefourth, one half, and threefourths full scale
(see Chapter 6).
The following sections describe tests that can be performed to measure the various con-
verter characteristics either manually or by computer. The equipment required depends
on the tests to be performed. Some of the tests require very specialized equipment, while
others can be p.,rformed with quite simple equipment.
MONOTONICITY
This simple test requires a minimum of precision equipment. It does not guarantee il
specific accuracy but gives a good indication. If a converter with a star / type divider
passes a monotonicity check, the relative error in the DAC will be small, probably less
than -+-1 LSB.
DIGITALTOANALOG - Monotonicity can be checked by driving the converter from a
counter and observing the output on a high gain scope. The output should be a staircase
pattern. '.
ANALOG-TO-DIGITAL - The input can be any noise free power supply (such as a battery) I
and a potentiometer of less than 2000 ohms. In testing high resolution systems, poten-
tiometers should have a coarse and fine control with overlapping ranges. Starting at zero,
increase the input voltage and check that each state exists and that these states are in
the correct order. A similar computer-controlled test can be done using a saw tooth gen-
erator as the input signal.
STEADY STATE ACCURACY
DIGITAL-TO-ANALOG - Set an input to a known digital number and observe the output
with a high accuracy meter. Compare with the theoretical value.
ANALOG TO-DIGITAL - The input can be a high accuracy voltage reference or a stable,
ripple free, variable power supply with a high accuracy meter. When the converter is run
at a rapid rate, the indicator lights.will show quite clearly where the switching points are_
337
The input voltage at the switching point is measured and compared with the theoretical
value. Computer controlled checking can be done in a similar manner using a precision
programmable reference as the input signal.
NOISE
DIGITALTOANALOG - Noise can be measured on a scope with a high gain, accoupled
plugin unit.
ANALOGTODIGITAL - The noise appears as a band around the switching point, where
the converter output is oscillating between two neighboring states.
INTERMITTENT ERRORS
Intermittent errors can be caused by pickup or loss of a bit in the digital section or by
noise picked up in the analog section. The test for intermittent errors should be done with
automatic or semiautomatic equipment where the converter is run at full speed for an
extended period of time. The equipment should be installed in its final configuration so
that the transmission of the information is included in the test. It is important to check
the states where there is only a single a or a single 1 for possible pickup or loss of infor
mation in the digital transfer.
In a general purpose system which includes two way conversion, an intermittent error
check can be run in a closed loop. The computer can generate a pattern wave of digital
numbers which are converted to analog, then reconverted to digital. The results are
checked to see that the two numbers agree within their specified tolerance.
DIGITALTOANALOG - In a generalpurpose system, limits for two specific numbers can
be set up with two comparators, and the state of the comparators can be sampled by
the computer after the corresponding number has been brought in. Where other numbers
are read in, the comparator outputs would not be sampled. of course. For semiautomatic
testing, a similar system might be set up with a counter driving some of the bits of the
converter and toggle switches driving other bits.
ANALOG TO DIGITAL - In a general purpose system, a dc voltage input would be applied
and the computer would monitor it to make sure that all the readouts produced the same
number or two adjacent numbers.
To test semiautomatically, set a dc voltage input that is as far as possible from any
switching point and insert the equivalent number into a bank of toggle switches. The Type
R121 AND/NOR Gate can be used to compare the output with the toggle switches. A clock
and a few gates can be set up so the converter runs at its maximum rate and stops if the
toggle switches and the analogtodigital converter do not agree.
SETTLING TIME (OIGITAL-TO-ANALOG)
In most applications, the digitaltoanalog converter is asked to go through small changes
at a time. The worst case transients occur when all the flip flops change, that is, when the
states change from 0111 to 1000.
338
The settling time with respect to large transients is most important when the converter
output is being multiplexed. It can be observed by looking at the signal on a single channel
with a high gain scope.
In a system where the multiplexing is done digitally, or where there is only a single chan
nel, the response to large transients is only important when a group of conversions is
started; after which the converter will be changing in relatively small steps. If. an analog
todigital converter had been constructed with the same modules, then the response to
large transients can be inferred from previous operation. For example, in a successive
approximation converter, the settling time for a quarter-scale step must be less than the
time per step of the converter.
To observe the settling time more directly, a comparator can be used with one input set
to the desired threshold of the dc value of the digitaltoanalog converter .. The Digitalto
Analog can be switched back and forth, and the comparator output can be monitored on
the scope.
RESPONSE TO TRANSIENTS (ANALOG-TO-DIGITAL)
Transient response is extremely important in a converter with multiplexed inputs. The
response can be tested in the same way that switching point accuracy is tested. Alternate
the input. between a test channel and an offset Ihannel. Vary the voltage on the test
channel until a switching point is found, and compare this with the switching point that was
observed in the steady state test. If the output is observed visually on indicator lights,
the voltage on the offset channel should be one which gives all zeros or all ones, so that
the alternate voltage can be read clearly.
If the output is being monitored by a computer, the steady state arid transient switching
points can be measured simultaneously by performing several conversions before chang
ing the channel. The first conversion will give the transient results, the last conversion will
give the steady state results. A check should be made with the offset and test channels at
nearly opposite ends of the voltage range. Do not use end points,as the converter saturates
and overshoot would not be deteoted. The first decision point should be tested. In a suc
cessive approximation converter, for example, the first decision is whether the input is
above or below halfscale.
In a single channel system, the transient response is only important for the first conver
sion. It can be checked manually, running the converter from a pushbutton and changing
the input voltage manually. The general approach would be the same as for a multiplexed
system.
Operating Checks
Operating checks are made to assure that the equipment has not been damaged, wires
have. not been pulled off, or other catastrophic failures have not occurred. If properly
set up, the check also detects drift, so that the converter will never actually reach a point
where it needs realignment. Generally the test should be simple and should be a part of
the overall preventative maintenance routine for the equipment. In a general purpose com
puting facility with both types of conversion systems, a closed loop test can be run very
simply by plugging the digitaltoanalog converter into the analogtodigital converter and
comparing the results that come back with the original number.
If a converter is being tested separately, a simple test can be made on the worst case
points. If precision equipment is not readily available for the test, the converter can be
checked against a divideddown value of its own internal reference.
339
GENERAL-PURPOSE ANALOG-TO-DIGITAL
CONVERTER AND MULTIPLEXER CONTROL
Digital is now offering its generalpurpose analogtodigital converter (ADC1)
and multiplexer control (AMX1) as separate units or as a combined converter
multiplexer (CMX 1). Optional equipment includes input amplifiers to obtain
high impedance or "standardize" the input signal, sample and hold circuitry,
and interfacing for the PDPS, PDPS/S, or PDP9 computers.
ADCl CONVERTER SPECIFICATIONS
The ADC1 converts an analog voltage to a binary number. Three convenient
switches are mounted on the INDICATOR/CONTROL PANEL; a POWER ON/
OFF switch is a 117volt input power disconnect. The ADC switch is a normally
open pushbutton that initiates an A/D conversion whenever the switch is
activated. The WORD LENGTH control is a rotary switch used to select the word
length, the conversion accuracy, and the conversion time. The WORD LENGTH
switch selects the following characteristics:
TABLE L CONVERSION ACCURACY AND
TIME AT SELECTED WORD LENGTHS
A completely wired back panel includes an A/D
Converter, interfacing for the PDP computers, an
optional multiplexer control with up to 64 input
channels and amplifier output, and provisions for a
sample and hold amplifier (MOO).
Word Length Max Switching
(No. of bits) Point Error*
6 1.6%
7 O.B%
B 0,4%
9 0,2%
10 0.1%
11 0.05%
12 0.025%
1f2 LSB for quantizing error.
Conversion Time
(I'sec)
9.0
10.5
12.0
13.5
1B.0
25.0
35.0
340
If the converter is used with a PDP computer. can
versions are initiated by an IN/OUT Transfer instruc
tion. If the converter is used separately, a CONVERT
AD PULSE is necessary to initiate conversion.
ACCURACY: See Table 1.
CONVERSION TIME: See Table 1.
APERTURE TIME: Same as conversion time .
CONVERTER RECOVERY TIME: None.
INPUT: 0 to -10v standard. Input scaling may be
specified using the amplifier option.
INPUT LOADING: 11,2 amp and 125 pf for 0 to
-10v input signal.
OUTPUT: Binary number of 6 to 12 bits, with
negative numbers represented in 2's complement
notation. A Ov input gives a 4000,; a -5v input a
0000, and a -10v (minus 1 LSB) input gives 3777,
number.
CONTROLS: Power ON/OFF switch, ADC switch,
binary readout indicators and a seven position rotary
switch which selects word length and conversion rate
are provided.
The convert A/D pulse input requires a negative
pulse (0 to -3v) of at least 150nsec duration. The
pulse loading is 1 ma at ground.
At the completion of the conversion process, two
complementary A/D DONE levels initiate external
reading of the converted data. This level remains
in the A/D DONE state until an external clear flag
pulse is generated or another convert A/D pulse is
given. If used with a PDP computer, the flag is
cleared when the read buffer command is given.
The clear flag pulse requires a negative pulse (0 to
-3v) of at least 100nsec duration. The pulse load
ing is 1 ma at ground.
POWER: Module power is supplied through one H701
power supply and one H704 regulated power supply.
Input power; 117 volts at less than % amp.
OPERATING TEMPERATURE RANGE: OOC to 50C
MECHANICAL: Panel Width: 19 inches
Panel Height: 8
11
/" inches
Depth: 19% inches
AMXl MULTIPLEXER CONTROL SPECIFICATIONS
The AMX1 includes from 1 to 16 A121 multiplexer
switch modules, depending upon the number of
channels required by the user. The user may select
any multiple' of four channels to a maximum of 64,
In the random address mode, the control routes the
analog signal from any selected channel to the A/D
Converter input. In the sequential address mode,
the multiplexer control advances its channel address
by one each time an indexing command is received.
After indexing through a predetermined number of
channels, the address is returned to 2ero. When
using the sequential operation, the conditioning
levels for random addressing are ignored.
Three convenience switches are mounted on the
INDICATOR/CONTROL PANEL; a POWER ON/OFF
switch, a CLR switch, and an INDEX switch. The
POWER ON/OFF switch is a 117volt input power
disconnect. The CLR switch is a normally open push
button that clears the multiplexer address register
and selects channel zero whenever the switch is
activated, The INDEX switch is also a normally open
pushbutton that increments the channel address by
one each time the sWitch is activated.
A completely wired back panel includes the multi
plexer control with up to 64 input channels and
amplifier output, interfacing for the. PDP computers,
an optional 6 to 12 bit generalpurpose A/D Can
verter, and provisions for a sample and hold ampli
fier (MOO),
If the multiplexer is used with a PDP computer, can
trol is carried on by In/Out Transfer instructions. If
the multiplexer is used separately see specifications.
Multiplexer Address Input-six lines accept DEC
standard levels of 0 and --3 volt, with Ov for asser-
tion. Load is 1 ma at ground.
341
Random Address Control-the channel select pulse
input requires a negative pulse 0 to 3v of at least
100nsec duration. The pulse loading is 1 ma at
ground.
The channel address levels must be brought to their
final values at least 400 nsec before the channel
select pulse occurs.
Sequential Address Control-the index _ pulse has
the same characteristics as the random address
control.
Operate Time-the time required to switch from one
channel to another is 10 I,sec to within 1 millivolt
of the final voltage. This time is preset within the
control and starts when a set-or index command is
received,
Indicators-binary readout indicators; six are pro
vided.
A121 Multiplexer Switch
POWER: Module power is supplied through one
H701 power supply and one H704 regulated power
supply. Input power: 117 volts at less than % amp.
CONTROL:
Signals Digital levels, -3 volts for assertion.
Load 1.3 milliampere load shared between its
grounded inputs.
SIGNAL:
Input Operating Signal voltages
Output voltage
.Output current
Input Impedance
Output Impedance
Input Current Offset
Input Voltage Offset
"Off leakage"
Capacitance
SPEED:
10% input to .01 % output
MECHANICAL:
Panel Width:
Panel Height:
Depth:
19 inches
8"/ .. inches
19
1
/2 inches
+10v to -5v
O'to -lOv
5 ma
6 Mohm (min)
1 ohm (max)
2 ohm (max)
1 mv. (max)
10 na. (max)
10 pf (max)
10 usee
The CMX-l combination with computer interface is
available as the AFOIA (PDP-8, 8/S) AFO 1 B (PDP-9)
interface. As such, it is considered a computer periph-
eral, and check .. out and installation is included in
the price.
The computer interface options also may be pur-
chased as modules and cables from this catalog and
installed by the customer who takes responsibility
for check-out.
Modules required for PDP8 and 8/S interface:
1 RIll
2 R123
2 WI03
Bus Cables
Modules for PDP-9 Interface:
3 W103
1 W500
1 R202
1 RI0l
4 R123
2 W640
Bus Cables
A/D Converter (ADC 1) - $2,000.00
64 Channel Multiplexer (AMX 1) - $2,250.00
Plus $16.25 per channel
Converter-Multiplexer (CMX-l) - $3,300.00
Plus $16.25 per channel
Options
Input Amplifier
Sample and Hold
AFOIA, AFOIB
342
-$ 300.00
-$ 500.00
-$4,500.00
A
SERIES
. ~ ......
343
MULTIPLEXER SWITCHES
TYPES A100, A103 , and A121
1-----------------1
I I
I I
I I
N
I
I I
I I
I M I
L _________________
AIDD, AID3 MULTIPLEXER SWITCHES
Control
Signals
Enable
Load
Signal
Max voltage
Max current
"On" offset (max,)
liOn" resistance (max.)
IIOff" leakage, capacitance
Carrier cross talk (with light Iiltering)
Speed
AI21
MULTIPLEXER SWITCH
AlOO AI03
Digital levels and 5-mc square wave
- 3 v (5-mc square wave pin E)
11/4 ma shared among grounded inputs
12v 30v
I rna I rna
200l'v 300l'v
SOn SOn
2 na, 10 pi 2 na, 10 pi
10 mv p'p 10 mv p.p
I
50% input to tolerance output Delay + sync + charging time (RC)
Turn on delay 400 nsec
I
400 nsec
I
600 nsec
I
Turn off delay 200 nsec
i
400 nsec
I
1000 nsec
I
Synchronization 100 nsec I 100 nsec 100 nsec
AI21
-3v
10v
I rna
0
480n
2 na, 10 pi
0
200nsec
2090 nsec
100 nsec
The AlOO and Al03 multiplexer modules contain
two, single-pole, high-speed, solid-state switches.
The switch drive is transformer-coupled so that the
switch may be completely isolated from ground. The
switch is turned on when the three control inputs are
at -3v (or open-circuited) and the carrier is receiv-
ing a 5-mc square wave. The square wave can be
made using a lO-mc clock and a 10-mc flipflop.
Since the switches are low impedance, care should
be taken to avoid shorting signal terminals to ground
or to each other, or simultaneously turning on two
switches which have a common connection. There is
a shield on Pin N that should be grounded. In newer
modules, this connection is made internally. Better
performance results if Pin N is also grounded ex-
ternally.
344
The A12l multiplexer module contains four single
pole, high-speed, insulated-gate FET switches. The
switch is turned on when its two inputs are at -3
volts.
AlOO - $100.00
Al03 - $ 78.00
A12l - $ 65.00
GUARDED RELAY MULTIPLEXER SWITCH QJ
TYPE A111
(Standard height, double width) SERIES
tGUARD) {OUTPUTI {GUARD} {OUTPUT} (GUARD)
M R P N S

+
INPUTS _ INPUTS
L o----j-- --j-o v
L-i- ----------i--'
'--1--- -- ------1-.,
1-l2v ->-' -.2V ->-' I
I I
I .oon lOon I
I Y Y I
I I
* = I
I
C{SHIELDIO---+- - - - - _ - - _ - _____ -.J
Al11 GUARDED RELAY MULTIPLEXER
At low levels, multiplexing of analog signals must
usually include guarding and shielding provisions
to control noise pickup. Fortunately, transducers
having lowlevel outputs are often slow speed
devices like thermocouples, so that the limited
speed of a relay multiplexer is not a serious
problem, and the superiority of relay contacts
for ultralow-offset sWitching can be fully utilized.
The two James Microscan 3-pole relays in the
Alll are specially designed for this purpose,
and are mounted on a double-clad circuit board
which shields the analog from the digital circuitry
and provides guarded contact wiring. The large
size of the special relays used requires that two
module slots be allowed for each Alll module.
shield on the component side of the board covers
the driver circuitry and is connected to pin C.
Another shield covering the area under the relay
coils is connected independently to pin C. Contacts
close when inputs are at -3v.
INPUTS: Each relay driver requires 3 rna drive
at ground, shared among grounded inputs. A
345
OUTPUT: Signal Contacts - IOv and I rna,
max. Contacts switch within I msec. Life expec-
tancy - 10
9
operations. Limits can be extended
to 30v and 10 rna below 25 cps at short duty
cycles. Guard Contacts- Designed for high
voltage, high current transients. Guard contacts
close before signal contacts close, open after
signal contacts open.
POWER: +10 v(A}JO rna; -15v(B}J85 rna.
All1-$93.00
OPERATIONAL AMPLIFIER*
TYPE A200
The A200 is an. operational amplifier mounted on an A990 amplifier board.
Provisions are made on the board for the mounting of potentiometers for gain
trim and balance. Mounting holes are also provided for input and feedback
networks, and rolloff capacitor.
OPEN LOOP GAIN:
RATED OUTPUT
Voltage:
Current:
FREQUENCY RESPONSE
Unity gain, small signal:
Full output voltage:
Slewing rate:
Overload recovery:
INPUT VOLTAGE OFFSET (Adjustable to Zero)
Average vs. Temperature:
Average vs. Supply voltage:
Average vs. Time:
INPUT CURRENT OFFSET:
Average vs. Temperature:
Average vs. Supply voltage:
INPUT IMPEDANCE
Between inputs:
Common mode:
INPUT VOLTAGE
Maximum:
Maximum common mode:
Common mode rejeclion:
POWER
Voltage:
Cu rrent at rated load:
"REFER TO A990 FOR CONNECTIONS
346
2x10'
l1v
20 rna
10 MHz
300 kHz
30vltAsec
200l'sec
20l'vrC
15I'v/%
10 I'v./day
2 na
0.4 narC
0.15 na/%
6 megohm
500 megohm
15 volts
10 volts
20,000
15 volts
35 ma
A 200 - $130.00
SAMPLE AND HOLD AMPLIFIER
TYPE A400
L-_________ <D_O_U_B_L_E_H_E_IG_H_T_,_D_O_U_BL_E_W __ ID_T_H_> ________ SERIES

INPUTS
AS o-----"'vv---..
A r o-------<l
(AI
OUTPUT
AV
(SI
1---1>0------.---- (AI
DIGITAL
CONTROL
SO
INPUT
(SI
The A400 is an accurate sample and hold amplifier
capable of tracking a full scale excursion in 12
micro-seconds to 0.025% accuracy. In the hold
mode, the droop (a decay) is less than 1 millivolt
per millisecond. Two analog inputs are provided ..
Pin AS is connected to a 10Kn resistor which pro'
vides for unity gain. Pin AT is connected to' a point
which allows for the insertion of different. resistors
to effect a gain change. The resistor connected to
this point must., be a preciSion 1 % resistor. with a
temperature -coefficient of 25 ppm.
An optional internal offset network which uses the'
15 volt supply can be included. Connections are
made according to the following table:
+15
I
AD
347
ANALOG DIGITAL
-15 GROUND GROUND
I Jl f1
AE AF AC
TO OFFSET OUTPUT
PIN NEGATIVE POSITIVE
BK NO CONNECTION -15 VOLT GROUND
BL +15 VOLT SUPPLY NO CONNECTION
BM NO CONNECTION -15 VOLT SUPPLY
BN +15 VOLT GROUND - NO CONNECTION
Offsets of up to 6 volts can be achieved in this
manner. The digital control input (8D) requires "the
standard -3 volt level to sample (track) and ground
to hold.
The A400 can be used to sample fast time varying
TRACK TIME TO 0.025%:
APERTURE:
DROOP:
GAIN:
INPUT IMPEDANCE:
FULL SCALE INPUT:
OUTPUT CURRENT:
wave forms and produce a time invariant output
sufficient for analog to digital conversion. Several
sample and holds may be used to simultaneously
sample a number of inputs and be multiplexed into
an A to D converter. The A400 is mounted on a
double height double width board. Therefore, the
unit requires 4 card slots (2 x 2).
12 I'see
Less than 150 nanasee
Less than 1 volt/sec
1.000 (Adjustable to 0.025 %)
lOKn 0.1% (AT)
1O Volts
10 MA
TEMPERATURE COEFFICIENT (IN SAMPLE): 20 I'valtrC Offset
TEMPERATURE COEFFICIENT (IN HOLD): 0.10 Voll/Sec/"C
POWER REQUIREMENTS: 15 Volts/50 MA
A400 $330.00
Optional Offset 50.00
348
/1 ..... _____ ____ __I11 SE!ES I
OUTPUTS
+F -v
8
-N +pl
INPUTS
A502 COMPARATOR
The A502 Comparator is a high speed difference
amplifier which compares two input voltages and in-
dicates which of the two is the more negative. The
comparator has a resolution of 1 mv, and an input
range of 0 to -lOv. The maximum combined error
due to a change in the common input voltage from
o to -lOv and a 20C temperature change is 5 mv
equivalent input offset. Two potentiometers allow
adjustment of the zero set and common balance.
As seen in the module diagram, .when the input
polarity of pins Nand Pare - and +, respectively,
then the output polarity of pins F and V are + and
-, respectively.
The comparator switching time is less than 250 nsec
for a 10 mv square wave. The switching time is
also less than 250 nsec when one input is at -5.0Ov
and the other. is switched from ground to -5.02v.
For finer resolution, the switching time is increased.
When the comparator is driven from a high imped-
ance, fast switching source, such as a digital-to-
analog converter, time should also be allowed fOT
349
transients to settle. The analog-digital conversion
application notes show illustrations of various
combinations of divider networks and comparators
in typical converter applications.
INPUT: 0 to -lOv. The input draws up to 1 ,.a, de-
pending on the relative polarity of the two voltage
inputs. The maximum current difference ,between.
positive and negative input voltages is 1.,.a. The
difference input capacitance is 75 pf.
OUTPUT: The outputs produce standard levels of
ground and -3 v. Each output will supply 5 rna (2 rna
at maximum speed) at ground, and 14 rna (2 rna at
maximum speed) at -3 v.
POWER: +10 v(A)f21 rna; -15 v(B)/55 rna.
NOTE: See "WIRING" sectionof reference supplies
data sheet.
A502 - $110.00

CONVERSION MODULE
TYPE A601
OAe OUTPUT
GROUND i
DIGITAL GROUND L..
(MUST BE GROUNDED
EXTERNALLY) ,
TERMINATING
RESISTOR
REFERENCE
DIGITAL INPUTS
AGOl DIGITALANALOG CONVERTER
The A601 is a threebit digitaltoanalog conversion
module utilizing a startype divider network and three
precision germaniumtransistor level amplifiers. It
may be connected in series with other converters to
form higher resolution converters. The accuracy of
the A601 is suitable for up to eight bits of conver
sion. For higher resolution, it should be combined
with the Types A604 and A605.
ACCURACY': 0.25% of expected value or 0.5
mv, whichever is greater
TEMPERATURE COEFFICIENT: 100 ppml'C max
from +lO'C to +45'C
OUTPUT IMPEDANCE: 1000 ohms 0.1%
SWITCHING TIME: 300 nsec
SETTLING TIME: The settling time is determined by
the capacitive loading at the output. Approximately
10 nsec/pf should be allowed in addition to the
switching time.
DIGITAL INPUT: DEC standard levels. A -3v input
signal at all digital inputs produces ground out. The
input load is 1 ma at ground. If all inputs are not re-
quired, the most significant inputs should be used,
and the least significant ones,should be left open'
circuited. Converter Input - The converter input
may be driven from the converter output of another
module in order to provide higher resolution. If' not
driven from another unit, it should be terminated
with 1000 ohms to ground. A termination resistor is
included in the module. Reference Input-The
reference input requires a -15 ma DEC Type A702
or A704 Supply. The supply should be adjusted to
approximately -10.01 v to overcome the saturation
resistance in the level amplifiers. High Quality
Ground - This is the ground return for the refer-
ence supply and should be connected to the supply
terminal and eventually to chassis ground at a noise-
free location.
OUTPUT: The output is the analog equivalent of the
digital input. The most positive output is Ov. The
most negative output is -lOv less the value of the
least significant bit. The output impedance is 1000
ohms. If a bipolar or reduced output swing is re-
quired, the output may be loaded with 1000 ohms
or more without affecting the accuracy.
POWER: +10 vl1 ma; -15 v/40 ma; -10 v refl
-9ma.
At 25'C includes tolerance -of 1.5: v on the + 10 v
and -15 v power supplies.
See CAUTION on AGOG specifications.
A60l - $60.00
350
DIGITAL-ANALOG
CONVERSION MODULES
TYPE A604,605
DAC OUTPUT
GROUNI:! r
DIGITAL GROUND ...
(MUST BE GROUNDED
EXTERNALLY)
DIGITAL INPUTS
AB04, AB05 DIGITAL-ANALOG CONVERTERS
The AG04 and A605 are twobit digitaltoanalog
conversion modules for use with the AG01 in forming
high resolution, high accuracy converters. Inputs
A604
and outputs are identical to the AG01 except that a
terminating resistor is not included. Germanium
transistors are used.
A605
ACCURACY:' :!:0.025% of expected value or :!:0.005% of expected value or
:!:0.05 mv. whichever is' greater :!:0.25 mv whichever is greater
EMPERATURE :!:25 ppm/"C (from + lO"C to :!:10 ppm/"C (from + 10"C to
COEFFICIENT: +45"C) +45"C)

1000 ohms :!:0.1 % 1000 ohms :!:0.01 %
SWITCHING TIME: 300 nsec 1.5 fLsec
-
At 25"C includes tolerances of :!:1.5 v on the +10 v and -15 v supplies.
SETTLING TIME: The settling time is determined by
the capacitive loading at the output. Approximately
10 nsec/pf should be allowed in addition to the
switching time.
The following combinations of modules are recommended.
Resolutions % of Analog Accuracy Units
(bits) Full Scale (% of Full Scale) (quantitytype)
up to 8 down to 0.39%
910 0.195% to 0.098%
11 0.049%
12 0.024%
13 0.012%
These modules have been factory aligned; however,
for maximum accuracy, the assembled system
should be calibrated as a whole. Offset compensa
tion has been made for standard digitaltoanalog
conversion. Additional offset may be added for
analogtodigital conversion.
0.25% 3-A601
0.082% 1A604, 3-A60 1
0.038% 2-A604, 3A601
0.014% 1A605, 2A604, 2A601
0.01% 2A605, 2A604, 2A601
Note: See "WIRING" section of reference supplies
data sheet.
POWER: viI rna; +15 v/30 rna; -10 v
ref/-lO rna. AGOS: +10 v All rna; -15 v/30 rna;
-10 v ref/- 9 rna.
See CAUTION on A606 specifications.
351
A604 - $62.00
AG05 - $78.00
I
DIGITAL-ANALOG
CONVERSION MODULE
TYPE A606
'--r-,r-...
L
KLESS SIG:,[h ~ ~ ~ N ~ ~ N G
DECADE ( ~ F 1 RESISlOR
OUTPUT
A606 2BIT DAC
TRUTH TABLE:
Decimal Number 8 4
0 0 0
1 0 0
2 0 0
3 0 0
4 0 ,1
5 0 1
6 0 1
7 0 1
8 1 0
9 1 0
This module is similar to the A604 but with differ
ent values of ladder resistors. It is designed to be
used in conjunction with an A604 to form one
decade of BCD Digital to Analog conversion. The
digital inputs of the decade must be 2421
_ weighing (a conversion scheme from an 8421
flipflop register to a satisfactory 2421 code is
shown). Overall accuracy and other characteristics
are the same as for A604, except as shown in the
logic diagram.
POWER: +lOv(A)/1.0 ma; -15v/30 ma; +'lOv
ref./-9 mao
CAUTION
Care should be taken when using power supplies
with separate + 10v and -15v onoff controls. If
this is the case, the -15v must be turned off first
and on last; otherwise, damage to the DACs may
result.
2 1
0 0
0 1
1 0
1 1
0 0
0 1
1 0
1 1
0 0
0 1
A B c D
2 4 2 1
0 0 0 0
0 0 0 1
0 0 1 0
0 0 1 1
0 1 0 0
0 1 0 1
0 1 1 0
0 1 1
' 1
1 1 1 0
1 1 1 1
0.----' --00
8.4-21 ,FLIP FLOP REGISTER
2-4-2-1
OUTPUT
A606 - $62.00
352
REFERENCE SUPPLIES
TYPES A702,A704
AT + SENSE
AV -SENSE
AE -OUTPUT
A702 REFERENCE SUPPLY A704 PRECISION REFERENCE SUPPLY
(Double height module)
Ripple
Module
Type
Output Current Temperature Coefficient Regulation Peak to
Peak
A702 10 v 60ma 1mv/'C 30 mY, no 10 mv
load to full
load
A704 -10v -90 to +40"ma 1 mv/B hrs 0.1 mv, no 0.1 mv
1 mv/15' to 35'C load to full
4 mv/O' to 50'C load
Module Adjustment Input
Use
Output
Type Resolution Power Impedance
A702 5 mv -15 vi 100 ma Load with 500 I'f 0.5 ohms
+10 v (8)/10 ma at load. May also
be preloaded if desired
A704 0.01 mv -15 2 v/250 ma See below for sensing 0.0025 ohms
REMOTE SENSING: The input to the regulating cir-
cuits of the A704 is connected at sense terminals
AT (+J and AV (-J. Connection from these points to
the load voltage at the most critical location pro-
vides maximum regulation at a selected point in a
distributed or remote load. When the sense termi-
nals are connected to the load at a relatively distant
location, a,capacitor of, approximately 100 I'f should
be connected across the load at the sensing point.
The supplies may be preloaded to
ground or -15v to change the amount of current
available in either direction. For driving DEC
Digital-Analog Converter modules, -125 rna
maximum can be obtained by connecting a
270!l 5% 1 watt resistor from the -lOv pin AE
reference output to pin AC ground (A704 only}.
PIN CONNECTIONS: The A704 is a double-sized
module. The top pin letters are prefixed A.
353
and preloading
Wiring:: Digital-analog and analog-digital converters.
perform best when module locations and wiring are
optimized. All Digital-Analog Converter modules
should be side-by-side, with Type 932 bus strip
used to bus pins E and pins F together on all
converter modules. In an analog-digital converter,
the comparator should be mounted next to the
converter module for the bits of most significance.
The reference supply module should be mounted
nearby, and if the A704 is used, its sense ter-
minals should be wired to pins E and F ofthe most-
significant-bits converter module. The high qual-
ity ground must be connected to the common
ground only at pin AC of the reference supply
module, and this point should also be the com-
mon ground for analog inputs to analog-digital
converters. Do not mount A-series modules
closer than necessary to power supply trans-
formers or other sources of fluctuating electric
or magnetic fields.
A702 - $ 58.00
A704 - $184.00
AMPLIFIER BOARDS
TYPES A990, A992
Many types of commercially available operational
amplifiers can be mounted in the holes provided
on these predrilled etched boards. Mounting holes
and printed wires provide for balance trim, gain
trim, and feedback networks required to build such
common operational devices as voltage followers,
inv.erting or non-inverting amplifiers, integrators,
differentiators, summers and subtractors. Most am-
plifiers listed in the table below require 15v regu-
lated supplies which are readily available from the
amplifier manufacturers. Notable exceptions are
Analog Devices' Models 101, 103, and 104 which
may be used with standard DEC +lOv, -15v sup-
plies at some sacrifice in voltage range (+5, -lOv)
and noise.
POWER: Positive at pin D, negative at pin E, com-
mon at pin F for all types. Space is provided for
mounting bypass capacitors used with some high
frequency amplifiers.
TRIMMING: Mounting holes on I" centers at the
handle end accept wirewound potentiometers for
balance and feedback (gain) trimming. Gain rheostat
may be connected in series with feedback com-
ponents to allow precise adjustment of gain using
inexpensive 1 % feedback resistors. Board is etched
to allow for use without gain trimming, and one
printed conductor must be cut at caret marks to
put a rheostat in the circuit. Gain rheostat stray
capacitance to ground is driven by amplifier output.
Amplifier Supplier
Types accepted Types accepted by
by A990 A992 (boosters too)
Analog Devices 101, 102, 104, etc. 103, 106, 107, etc.
Burr-Brown* 1500-46, 1500-68 -
Data Device Corp.
-
most types, except boosters
Nexus Case K or Case L Case Q
Philbrick -
Case PP
Union Carbide
-
most types
Zeltex
-
Case A
'Except Burr-Brown differential output and chopper stabilized types. Perforated
boarg W994 or other blank module may be used to mount non-standard con-
figurations.
354
F 00---------0()0 F
eo 5>l.
VOLTAGE FOLLOWER (NON-INVERTING)
>--'--'--.() V
Fo----L--------------.()F
INVERTING AMPLIFIER
:: 3J
NON-INVERTING AMPLIFIER
::
INTEGRATOR
DIFFERENTIATING
CAPACITOR
GAIN \

T U
IGH -FREQUENCY
ROLL -OFF NETWORK
DIFFERENTIATING
RESISTOR
V
FG-----------L----------o
DIFFERENTIATOR WITH DOUBLE BAND STOPS
R
S
T
U
>--'---.() V
Fo----L------------OF
FOUR-INPUT SUMMER
MATCHING TRIMMER CAN BE
RI
U G-----.JW....--J'--f'
RI
R2 PLUS
112 AT
__ 'ft.':ON-
EFFECTS GAIN.
>--'--'---.() V

SUBTRACTOR AND COMMON-MODE REJECTOR
355
A990-$4.00
A992-$4.00
I
356
PART VI: COMPUTER CATALOG
357
PDP COMPUTERS
PDP generalpurpose digital computers are used for a wide variety of data processing and
control functions. PDP's are constructed of highly reliable FLlpCHIP digital circuit modules,.
and include builtin provisions for marginal checking. The resulting overall reliability has
earned PDP's a reputation for troublefree performance. An exceptionally varied line of
inputoutput devices are available, and versatile facilities are provided in the computers to
handle these and other devices.
A complete, welldocumented package of programming aids accompanies each PDP com
puter. The package includes a FORTRAN compiler, a symbolic assembler, on line debugging
routines, an editor, and utility, arithmetic, and maintenance routines. Editing and online
debugging programs use the same symbolic language as the assembly systems. This means
that debugging is carried out in the same language as the program being debugged,
eliminating the creation and reassembly of new symbolic tapes each time an error is found.
The arithmetic subroutines include a floating point package. Inputoutput subroutines are
prepared for most of Digital's standard optional devices. Extensive maintenance routines
are provided. Supporting these programming aids are free training courses at Digital and
membership in DECUS, the Digital Equipment Computer Users Society. DECUS povides a
means for users to exchange ideas and programs through regularly scheduled symposia. A
library of fully documeted programs is maintained.
358
359
PDP-8
The PDP-8 is a general-purpose, stored-program computer, featuring a 1.5 microsecond
random access core memory, a fast arithmetic processor, and a buffered input-output
control. These features combine to make the PDP-8 one of the most popular on-line
computers for physics and biomedical analysis and process control. The PDP-8 is also used
in large systems as a control element and as a training computer.
The PDP-8 is easy to install, maintain, and use, with comprehensive software, customer-
tested in over 500 installations. The basic system-includes 4096 words of 12-bit ferrite
core memory, keyboard-printer and tape reader-punch, eight auto-index registers, wired-in
analog-to-digital converter, program interrupt, data interrupt, and indirect addressing.
A partial list of central processor options includes the Extended Arithmetic Element for
high speed, double precision arithmetic; Memory Modules and Control for increasing
memory size in increments of 4096 words to 32,768 words; a Data Channel Multiplexer
providing direct memory access for seven external devices; and a Serial Drum for storage
of 65,536 to 262,144 words.
The applications success of the PDP-8 has led Digital to develop a series of computers
based on the PDP-8 to meet a number of special needs, resulting in a unique family of small
computer products. These include the DISPLAY 8, the LlNC-8, the TYPESETTING-8, the
MUL TIANAL YZER-8, and the new.PDP-8/S.
SPECIFICATIONS
Word Length: 12 bits
Memory: 4096 to 32,768 words; cycle time 1.5 microseconds
Add Time: 3.0 microseconds
In-Out Transfer Rates: 7,992,000 bits per second
Standard I/O Devices: Printer-keyboard with paper tape punch and reader
Instructions: 49 with standard equipment, expandable to over 100 as optional equipment
is added
360
361
PDP-SIS
The PDPB/S is the first fullscale, generalpurpose, core memory digital computer selling
for under $10,000; it is designed for data handling and for controlling complex process
systems.
The PDPB/S has the same size memory, the same input/output capabilities, the same
extensive set of standard options as the PDPB. Both use the same software. The difference
between the two machines is in speed and physical size. The PDPB/S adds in 36 micro
seconds compared with an add time of 3.0 microseconds for the PDpB. The basic 12bit
word PDPB/S features an Bmicrosecond, 4096word, expandable core memory; a
comprehensive software package, including FORTRAN; and an ASR33 Teletype. Although
the PDpB/S combines a fully parallel core memory and input/output facility with a serial
arithmetic unit, the machine appears to be fully parallel to the user. Flexible, high capacity,
input/output capabilities of the computer operate a variety of peripheral equipment. In
addition to the standard teletype and perforated tape equipment, the system can operate in
conjunction with most of the optional devices offered in the PDpB family line. Equipment of
special design is easily adapted for connection into the PDPS/S system. The computer
need not be modified to add peripheral devices.
SPECIFICATIONS
Word Length: 12 bits
Memory: 4096 to 32,76S: cycle time S.O microseconds
Add Time: 36 microseconds
InOut Transfer Rate: 1,500,000 bits per second
Standard I/O Devices: Printerkeyboard with paper tape punch and reader
362
363
LINC-8
The LlNC-S is a computer-based system designed to ,control experiments and collect and
analyze data in the laboratory. The system combines the features of the PDP-S and the
LlNC computers, and allows the researcher to choose between the two programming
systems available. The researcher simply uses one of the two consoles in the system. Typical
biomedical applications for the new system are: arterial shock wave measurements in-phase
triggering of stimuli from EEG alpha waves, processing of single-unit data from the nervous
system. EKG processing, and operative conditioning applications.
Other applications for the LlNCS include research in physics, chemistry, meteorology,
oceanography, psychology .. radiation, seismology, and acoustics.
The original LlNC hardware and software were developed for on-line, real-time laboratory
research under grants from the National Institutes of Health and the National Aeronautics
and Space Administration. Development began at Massachusetts Institute of Technology
and continued at Washington University in St. Louis.
The LlNC-S system includes:a built-in multiplexed analog-to-digital input facility, a relay
register, dual digital LlNCtape transports, an alphanumeric oscilloscope display and an
ASR-33 teletypewriter. The LlNC-S takes advantage of the PDP-S's input/output bus for
additional convenience in interfacing other laboratory instrumentation to the LlNC-S system.
With the LlNC-S, the researcher has the option of using the LlNC software which has been
designed to allow the researcher to write his own programs after minimum instruction or he
may use the more advanced PDP-S programming system which includes FORTRAN. The
LlNC-S system "talks" with researchers by displaying instructions and results on the
oscilloscope display. Displays combine English language with data displays. To familiarize
customers with the new system, Digital offers four courses in programming and maintenance
of the LlNC-S. These are included in the basic system purchase price.
364

365
DISPLAY8
The DISPLAY-8 (Type 338 Programmed Buffered Display) is an integrated cathode-ray-tube
system containing its own 'general-purpose computer. It is capable of precisely displaying
points, lines, and characters, and of performing extensive computation using the computer
order code and a complete software package.
The computer is a PDp8: It is fast enough to perform 200,000 additions per second while
displaying 300,000 points, 600 inches of vector, or 700 characters flicker free at the same
time. The highly flexible character generator produces alphabetic characters or special
symbols, similar to those used on electronic circuit schematic, with equal ease.
The 338 can be used as a self-contained display system or as a buffered display station in a
large computer system. The 338 can control interfaces to external data sources, such as
the central computer in a large system, and can' handle real time requests, such as data
phone interrupts. The 338 can be programmed to view selected small areas of a large
stored drawing: 10 by 10 inch window can be moved randomly about a 6 by 6 foot drawing
for detailed examination and modification.
The system contains the following features for general purpose computations: An extensive
software package that includes FORTRAN, symbolic assembler, debugging programs, float-
ing point arithmetic, and display maintenance programs; 4096 words of core memory;
program interrupt; and keyboard-printer and.lO-hertz paper-reader punch. The 338 may
be expanded using any standard PDP-8 plug-in units.
366
I
367
PDP-9
The PDP-9 is a stored-program, general-purpose digital computer, designed to handle a
variety of on-line and real-time scientific applications calling for more computation power
than offered by the PDP-8. The basic PDP-9 features a 2-microsecond add time; 8,192 words
of 18 bit (plus optional parity bit) core memory; a real-time clock; a 300-character-per-
second paper tape reader; a 50-character-per-second tape punch; and input-output tele-
printer (Teletype Model KSR-33), Input/Output can be via programmed transfers, data
channel transfers, or direct memory access_ The maximum I/O transfer rate is 18,000,000
bits-per-second_ .
Single address instructions are used, with auto-indexing and one level of indirect addressing
permitted_ A single memory reference instruction can directly address any location in a
block of 8,192 words of memory. PDP-9 has a Direct Memory Access channel plus four
built-in Data Channels.
The memory can be expanded in 8,192-word increments to a total of 32,768 words. Mass
storage devices, such as DECtape, IBM compatible magnetic tape, disks and drums are
available as options for the PDP-9, as are a wide variety of other input-output devices and
central-processor additions.
A comprehensive software package including FORTRAN IV, a MACRO Symbolic Assembler,
a monitor system, and diagnostic routines is provided with the basic machine. With the
modular software package, PDP-9 users can program in a device-independent environment
to take full advantage of configurations with mass storage devices and central processor
options.
Applications for the PDP-9 include its use in biomedicine, process control, chemical instru-
mentation, display processing, hybrid systems and data communications. A special
configuration, the PDP-9 MULTI ANAL YZER, has been designed for physics applications.
SPECIFICATIONS
Word length: 18 bits
Memory: 8,192 to 32,768 words in 8,192 word increments.
Cycle Time: 1.0 microseconds
Add Time: 2 microseconds
In-Out Transfer Rate: Up to 18,000,000 Bits per second
Standard I/O Devices: A 300-character-per-second paper tape reader, a 50-character-per-
second paper tape punch and a lO-character-per-second KSR-33 teletype
Options: DECtape, IBM Compatible magnetic tape, drums, CTRS, AID converters, line
printers, card readers, plotters, etc.
368

369
PDP-10
PDP-10 is an expandable, 36-bit computer system available in five configurations
(PDP-10/1O, 10/20, 10/30, 10/40, and 10/50) and offering optimum power and versatility
in the medium price range.
The PDP-10 includes an extremely powerful processor with 15 index registers, 16 accumu-
lators, and 8,192 words of 36-bit core memory, a 300-character-per-second paper tape
reader, a 50-character-per-second paper tape punch, a console teleprinter, and a two-level
priority interrupt subsystem. PDP-10/20 adds two DECtapes. PDP-IO/30 includes 16,
384 words of memory and additional I/O devices. PDP-10/40 adds an extended order code
and a memory protection and relocation feature. And PDP-IO/50 permits swapping
between 32,768 words or more of memory and fast access desk file via the multiplexer/
selector channel, and includes multiprogramming time-sharing software.
The PDP-1O is designed for on-line and real-time applications such as physics and bio-
medical research, process control, as a departmental computation facility, in simulation and
aerospace, chemical instrumentation, display processing and as a science teaching aid.
The software package includes real-time FORTRAN IV, a control monitor, a macro
assembler, a context editor, a symbolic debugging program, an I/O controller, a peripheral
interchange program, a desk calculator and library programs. All software systems assure
upward compatibility from the standard 8,192 words of memory through the multi-program-
ming and swapping systems at both the symbolic and relocatable binary level.
PDP-10 features a I-microsecond cycle time, -a 2.1-microsecond add time, I/O transfer
rates. up to. 7,200,000 bits per second and a modular, proven software package that
expands to make full use of all hardware configurations. Memory can be expanded in 8,192
word increments to the maximum directly addressable 262,144 words.
370
----- ~
---- ~
- ~ --------~
- - - - - - - - ~
~
I
371
DIGITAL TEST SYSTEMS
DEC designs and manufactures' a variety of devices for testing computer components and
similar products. The Company uses these products in its own operations and also markets
them to a regularly growing list of customers.
DEC memory test products are used by nearly every major manufacturer of core memories
(memories such as the ones DEC uses in its computers). This equipment is used to check
each stage of the computer memory assembly from the single core to the completed unit.
DEC has recently announced a memory test system, the PMA8 (Programmable Memory
Analyzer8), which incorporates the PDp8 as its control element. The speed and versatility
of the PMA8 is a major contribution to the memory testing field.
Automatic Module Testers are another Digital Test System product; Equipment such as the
tester shown in the photo, is used by DEC and other companies who manufacture large
quantities of their own modules to perform functional tests on completed modules. The
tester controlled by a PDP computer, can perform from 10 to 100 different static and
dynamic tests on a module in one second.
372
373
INPUT-OUTPUT OPTIONS
MAGNETIC TAPE EQUIPMENT
DECtape, a unique fixed address magnetic tape system, allows on-line program debugging
or high speed loading and readout Density is 375 60 bpi; tape speed is 80 ips with a 15
kc character rate_ Reads and writes in both directions: redundant tracks allow less than one
transient error in 10'0 characters. Total storage, the equivalent of 4000 feet of perforated
tape, is three million bits per reel.
Other magnetic tape systems include automatic and programmed controls and high or low
density transports. Formats are IBM compatible at recording densities of 200, 556, and
800 bpi. Transfer rates range from 15 to 90 thousand characters per second. Transports
include an electro-pneumatic design of high performance and low tape stress and wear.
MAGNETIC DRUM SYSTEMS
Drums provide auxiliary mass storage with direct access to memory. Sizes range from a
32,768 word drum to 262,144 words.
DISPLAY AND PLOnlNG EQUIPMENT
Precision and incremental cathode ray tube displays convert digital data into graphic and
tabular form. Light Pen detects plotted points to initiate computer action; Symbol Generator
plots alphanumeric or special symbols in four sizes on scope face. Incremental Plotters give
hard-copy graphs and histograms.
PRINTERS
Automatic line printers produce hard-copy output data from 300 to 1000 lines per minute
with 120 or 132 column lines and any of 64 characters per column. Teleprinters permit on-
line inputs and outputs from the computer console or remote stations at 10 characters
per second. Character sets are ASCII.
ANALOG-DIGITAL CONVERTERS
General purpose analog to digital converters offer seven front-panel selections of speed
and word length. Maximum speed: 6 bits 1.6%. 9 microseconds. Maximum accuracy: 12
bits 0.025% 35 microseconds. Digital-to-analog equipment has maximum conversion time
to an accuracy of one least significant bit of 2 p,sec. Speeds may be limited by the repetition
rate of the associated equipment
PERFORATED TAPE AND CARD EQUIPMENT
Paper tape punches operate at 10 to 63 characters per second; readers at 10,300, and 400.
Card punch controls permit operation at 100 or 300 cards per minute; card readers at
100, 200, or 800.
374
375
376
APPENDIX 1
MIL-STD-S06B AND DEC SYMBOL COMPARISON
This comparison of MIL-STD and DEC Symbology relates MIL-STD symbols to the DEC
logical equivalents or combination of equivalents which perform the same function_
Designations of high (H) and low (L) are used instead of "1 's" and "O's" to avoid the
problem of positive and negative assertion_
DEC R-series FLIP-CHIP modules also are related to the appropriate MIL-STD equivalent
In cases where many input gates are provided (such as flip-flops), only a few of the many
methods of input connections are illustrated_ Output triggering of fiip-flops is not shown,
but is one of the possible methods of clearing and setting_
MIL-STD-8068 gating symbols use a small circles(s) at the input(s) of the logic element
to indicate that a relatively low (L) input signal activates the function_ The absence of a
circle indicates that a relatively high (H) signal activates the function_ The presence or
absence of a circle at the output of an element indicates that the output is low (L) or
high (H) respectively in the activated state_
DEC's high and low level symbology uses this same principle except for a difference in
signal indication_ A solid diamond ( .) indicates DEC's low (L) signal level (-3
volts), whereas open diamond (---< indicates the high (H) signal level (0 volts or
ground)_ Correspondingly, for pulse signals, a solid arrowhead ( ~ ) signifies a
negative (low) pulse, and open arrowhead (-----t- a positive (high) pulse_
The MIL-STD-8068 logic symbols with their DEC equivalents are listed in Table 1. Logically
equivalent AND and OR elements for each system with the appropriate Table of Com-
binations are included_
MIL-STD-8068 logic symbols, with their DEC counterpart, for flip-flops, one-shot multi-
vibrators, inverters, etc_, are illustrated in Table 2. It should be noted that the input
connections of the DEC elements correspond to the MIL-STD operation, symbolically
described.
In most cases, the DEC element is capable of greater flexibility of operation if maximum
use of input/output gating is utilized_
MIL-STD nomenclature for flip-flop inputs is, set (S), clear (reset) (C), and toggle
(trigger) (T)_ DEC terminology is identical except for the "toggle" input which is termed
"complement". .
R-series module elements are shown in Table 3 with their MIL-STD symbol. For gating
functions, two logically equivalent MIL-STD symbols are compared with the appropriate
DEC symbol (shown without the polarity indicators). A table of combinations is included
in instances where clarification may be necessary.
377
DEC R series flip flops, pulse amplifiers, and one shots use diodecapacitordiode gates
to generate and steer pulses to these elements. Operation of the DCD gate is explained
in the R-Series section of the handbook. The DCD gate can be represented by symbols
listed in MILSTD8068 as shown below:
The 400 nanosecond level input represents the set up time of the gate and the + level
change delay of 100 nanosecond represents. the minimum duration cif the positive'levei.
The DEC symbol is:
+ LEVEL
CHANGE
OUTPUT
LEVEL INPUT
In order that the unwieldy MIL representation need not be shown in the following
representations, a new symbol is defined as being the equivalent of the DCD gate.
+LEVEL=-
CHANGE
LEVEL
INPUT
OUTPUT
Jl
The inputs are defined in the positions shown above, regardless of the orientation of
the gate.
In cases where a large number of connections may cause confusion, terminal points are
letter referenced for both symbologies.
378
AND OR AND OR A B X
::pH-'
H H H
-:pH-'
H L L
:=o-x : ~ X
L H L
L L L
B ~
::P}' ::P}'
H H L
H L L
A:::[)- A:::[>-
B X B X
L H H
L L L
::P}' ::P}'
H H L
:::[)---x
H L H
::L>-X
L H L
L L L
:::1}' ::J}'
H H .L
:=rJ--
X
:=I>-X
H L L
L H L
L L H
: ~ ' : ~ '
H H H
:=lJ-
X
H L H
:=L>-X
L H H
L L L
:$' :$'
H H H
:==Lr-X
H L L
: :::[)---x
L H H
L L H
::HJ}.
::HJ}'
H H H
H L H
A=:[)-
:::D-
X
L H L
B X
L L H
:::p}-
::P}'
H H L
:=o-x ::::L>--X
H L H
L H H
L L H
EXCLUSIVE OR EXCLUSIVE OR
:=tr>-X
: ~ X
L L L
L H H
H L H
H H L
TABLE 1. GATING SYMBOL EQUIVALENTS
379
FLIP-FLOPS
~ t l
l
SFFe FF
I 0 e S
e -D"-----r-'
S
~ - - - T
MONOSTABLE MULTIVIBRATOR
W
e
S . FF e
I 0
S
T---....
INVERTER
PULSE AMPLIFIER
OR
TABLE 2. LOGIC SYMBOL EQUIVALENTS
380
TYPE DEC MIL - STD - 8068 A 8 C D X
RI07
.--},
H L
INVERTER
A-{>-X A-i>-X
L H
RI21
:4,
H H L
NEGATIVE
:=O-X :=1>-X
H L H
NAND L H L
POSITIVE NOR L L H
R III WITH ROOI
~ ~ ~ ~
H H H H L
NEGATIVE
~
L H H H L
NAND
8 . X
L L H H L
L L L H L
POSITIVE NOR
C .. NODE_
L L L L H
D -
ETC
A
RI41 A H H H H L
8
NEGATIVE 8 L H L H L
AND/NOR
C
C H L H L L
POSITIVE
D
OR/NAND
D H H L L H
L L H H H
L L L L H
X X
RI22
:+4-'
H H L
NEGATIVE NOR H L H
POSITIVE NAND
:I:)--X
:=I>--X
L H H
L L H
TABLE 3. RSERIES MODULE SYMBOL EQUIVALENTS
381
R2
n
OO
o I
FF
C S
R202

C-h----r'
L..----+-T
C
T ...... ----'
FLIP - FLOPS
R203

LJ-----fJ
C
R205
C
382
FLIP - FLOPS
R201
MONOSTABLE MULTIVIBRATOR
R302
383
PULSE AMPLIFIERS
PA ~
R603
BUS DRIVER
R650
31
1
BUS I
DRIVER t--.C--
384
APPENDIX 2
POWERS OF TWO
n -n
2 n 2
8
1.
32
64
128
25.
512
1 024
2 048
4 09.
1.0
0.5
0.25
0.125
0.062 5
0.031 25
0.015 625
0.007 812 5
B 0.003 906 25
0.001 953 125
10 0.000 976 562 5
11 0.000 488 281 25
12 0.000 244 140 625
0.000 122 070 312 5 8 192 13
16 384 14 0.000 061 035 156 25
32 768 15 0.000 030 517 578 125
65 536 16 O.DOO 015 258 789 062 5
131 072 17 0.000 007 629 394 531 25
262 144 18 0.000 003 814 697 265 625
524 288 19 0.000 001 907 348 632 812 5
1 048 576 20 0.000 000 953 674 316 406 25
2 097 152 21 0.000 000 476 837 158 203 125
4 194 304 22 0.000 000 238 418 579 101 562 5
8 388 608 23 0.000 000 119 209 289 550 781 25
16 777 216 24 0.000 000 059 604 644 775 390 625
33 554 432 25 0.000 000 029 802 322 387 695 312 5
67 loa 864 26 0.000 000 014 901 161 193 847 656 25
134 217 728 27 0.000 000 007 450 580 596 923 828 125
268 435 456 28 0.000 000 003 725 290 298 461 914 062 5
536 870 912 29 0.000 000 001 862 645 149 230 957 031 25
1 073 741 824 30 0.000 000 000 931 322 574 615 478 515 625
2 147 483 648 31 0.000 000 000 465 661 287 307 739 257 812 5
4 294 967 296 32 0.000 000 000 232 830 643 653 869 628 906 25
8 589 934 592 33 0.000 000 000 116 415 321 826 934 814 453 125
17 179 869 184 34 0.000 000 000 058 207 660 913 467 407 226 562 5
34 359 738 368 35 0.000 000 000 029 103 830 456 733 703 613 281 25
68 719 476 736 36 0.000 000 000 014 551 915 228 366 851 806 640 625
137 438 953 472 37 0.000 000 000 007 275 957 614 183 425 903 320 312 5
274 877 906 944 38 0.000 000 000 003 637 978 807 091 712 951 660 156 25
549 755 813 888 39 0.000 000 000 001 818 989 403 545 856 475 830 078 125
1 099 511 627 776 40 0.000 000 000 000 909 494 701 772 928 237 915 039 062 5
2 199 023 255 552 41 0.000 000 000 000 454 747 350 886 464 118 957 519 531 25
4 398 046 511 104 42 0.000 000 000 000 227 373 675 443 232 059 478 759 765 625
8 796 093 022 208 43 0.000 000 000 000 113 686 837 721 616 029 739 379 882 812 5
17 592 186 044 416 44 0.000 000 000 000 056 843 418 860 808 014 869 689 941 406 25
35 184 372 088 832 45 0.000 000 000 000 028 421 709 430 404 007 434 844 970 703 125
70 368 744 177 664 46 0.000 000 000 000 014 210 854 715 202 003 717 422 485 351 562 5
140 737 488 355 328 47 0.000 000 000 000 007 105 427 357 601 001 858 711 242 675 781 25
281 474 976 710 656 48 0.000 000 000 000 003 552 713 678 800 500 929 355 621 337 890 625
562 949 953 421 312 49 0.000 000 000 000 001 776 356 839 400 250 464 677 810 668 945 312 5
1 125 899 906 842 624 50 0.000 000 000 000 000 as8 178 419 700 125 232 338 905 334 472 656 25
2 251 799 813 685 248 51 0.000 000 000 000 000 444 089 209 850 062 616 169 452 667 236 328 125
4 503 599 627 370 496 52 0.000 000 000 000 000 222 044 604 925 031 308 084 726 333 618 164 062 5
9 007 199 254 740 992 53 0.000 000 000 000 000 III 022 302 462 515 654 042 363 166 809 082 031 25
18 014 398 509 481 984 54 0.000 000 000 000 000 055 511 151 231 257 827 021 181 583 404 541 015 625
36 028 797 018 963 968 55 0.000 000 000 000 000 027 755 575 615 628 913 510 590 791 702 270 5u7 812 5
72 057 594 037 927 936
144 115 188 075 855 872
288 230 376 151 711 744
576 460 752 303 423 488
1 152 921 504 606 846 976
~ ~ __ ~ = ~ __ ~ _ ~ ___ ~ __ m
57 0.000 000 000 000 000 006 938 893 903 907 228 377 647 697 925 567 626 953 125
58 0.000 000 000 000 000 003 469 446 951 953 614 188 823 848 962 783 813 476 562 5
59 0.000 000 000 000 000 001 734 723 475 976 807 094 411 924 481 391 906 738 281 25
60 0.000 000 000 000 000 000 867 361 737 988 403 547 205 962 240 695 953 369 140. 625
2 305 843 009 213 693 952 61 0.000 000 000 000 000 000 433 680 868 994 201 773 602 981 120 347 976 684 570 312 5
4 611 686 018 427 387 904 62 0.000 000 000 000 000 000 216 840' 434 497 100 886 801 490 560 173 988 342 285 156 25
9 223 372 036 854 775 808 63 0.000 000 000 000 000 000 108 420 217 248 550 443 400 745 280 086 994 171 142 578 125
18 446 744 073 709 551 616 64
36 893 488 147 419 103 232 65
0.000 000 000 000 000 000 054 210 108 624 275 221 700 372 640 043 497 085 571 289 062 5
0.000 000 000 000 000 000 027 105 054 312 137 610 850 186 320 021 748 542 785 644 531 25
73 786 976 294 838 206 464 66 0.000 000 000 000 000 000 013 552 527 156 e68 805 425 093 160 010 874 271 392 822 265 625
147 573 952 589 676 412 928 67 0.000 000 000 000 000 000 006 776 263 578 034 402 712 546 580 005 437 135 696 411 132 812 5
295 147 905 179 352 825 856 68 0.000 000 000 000 000 000 003 388 131 789 017 201 356 273 290 002 718 567 848 205 566 406 25
590 295 810 358 705 651 712 69 0.000 000 000 000 000 000 001 694 065 894 508 600 678 136 645 001 359 283 924 102 783 203 125
1 180 591 620 717 411 303 424 70 0.000000 000 000 000 000 000 847 032 947 254 300 339 068 322 500 679 641 962 051 391 601 562 5
2 361 183 241 434 822 606 848 71 0.000 000 000 000 000 000 000 423 516 473 627 150 169 534 161 250 339 820 981 025 695 800 781 25
4 722 366 482 869 645 213 696 72 0.000 000 000 000 000 000 000 211 758 236 813 575 084 767 080 625 169 910 490 512 847 900 390 6 2 ~
385
APPENDIX 3
ABBREVIATIONS
Frequency
Hz ~ . Hertz C ~ _ cycles per second
KHz = 10' Hertz
MHz -= 10' Hertz
Time .
msec = millisecond = 10-
3
second
. fASec ::::; microsecond = 10-
6
second
nsec::::; nanosecond::::; 10- 9 second
Current
amp = ampere
ma = milliampere::::; 10-
2
amper.e
"a::::; microampere::::; 10-
6
ampere
na::::; nanoampere = 10-
9
ampere
Voltage
v = volt
mv = millivolt = 10-
2
volt
/,-v = microvolt = 10-
6
volt
386
Resistance
K::::; kilohm = lOs ohms
meg = megohm::::; 10
6
ohm
Capacitance
I,J = microfarad::::; 10-
G
farad
nf =nanofarad = 10-
9
farad
pf = picofarad = 10- 12 farad
Other
ac ::::; alternating current
BD::::; bus driver
CD = capacitor-diode
D::::; delay
dc = direct current
DCD = diode-capacitor-diode
DEC =-= Digital Equipment Corporation
FF= flip-flop
LA::::; level amplifier
MP = mounting panel
PA::::; pulse amplifier
PG = pulse generator
pop = peak to peak
PS ::::; power supply
APPENDIX 4
DEFINITIONS
TIMING DEFINITIONS
Level Delay Time is the time delay between the point of 10% input change and the point of
10% output change in a given circuit. The input is assumed to be the output of a flip-flop of
the same frequency series as the circuit under discussion, or a level change with similar
characteristics. This is also referred to as delay for output fall or delay for output rise.
Pulse Delay Time is the time delay between the point of 10% input change and the point of
10% output change in a given circuit, when the input is a standard pulse of the same fre-
quency line as the circuit under discussion.
Propagation Delay is the average signal delay per stage for many similar circuits connected
in cascade.
Rise Time and Fall Time are the time delay between the 10% and 90% points of a voltage
change.
Total Transition Time is the time delay between the point of 10% input change and the point
of 90% output change in a given circuit. It is the sum of delay time and. rise (or fall) time.
Rise TTl is total transition time for rising output. Fall TTl is total transition time for falling
output.
Set-up Time is the time required for a diodecapacitor-diode gate to open or close after a
change of input level. This time is measured from the point of 10% input change.
387
. '
APPENDIX 5
BIBLIOGRAPHY OF DIGITAL LOGIC
INTRODUCTORY BOOKS
A large number of books are currently becoming available on the principles of digital
computers and digital logic. The following four are particularly well written and
concise. They are on a level which would be suitable for an undergraduate course, an
industrial training course, or for self-study. Only a knowledge of algebra, trigonometry
and basic electronics is assumed .
Siegel, Paul" Understanding Digital Computers
New York: John Wiley & Sons, Inc., 1961.
Discusses logic and arithmetic, components and circuits lIsed in logical building
blocks, and tile functional units of the digital computer.
Bartee, Thomas C., Digital Computer Fundamentals
New York: McGraw-Hili, 1960,
Includes a discussion of computer operations, programming; number systems, basic
logical circuits and logical design, and the functional elements of a general purpose
computer.
Irwin, Wayne C., Digital Computer Principles
Princeton, New Jersey: D. Van Nostrand Company, 1960,
Puts slightly more emphasis on logic and less on circuitry. It includes a discussion
of number systems, a brief discussion of the circuitry, timing, and digital arithmetic,
as well as Venn diagrams, Karnaugh maps. and Harvard minimizing chart.
Smith, Charles V. L" Electronic Digital Computers
New York: McGraw-Hili, 1959,
Discusses digital computer aritlimetic, instruction codes, basic logic circuits, and
functional elements of computers. It includes a variety of specific examples, prin-
cipally from the parallel direct-coupled asynchronous machine developed at the
Institute for Advanced Study at Aberdeen Proving Ground.
LOGICAL DESIGN
The following texts provide a more theoretical treatment of switching theory and machine
design. They are suitable for a fourth year undergraduate course, a first year graduate
course, or home study by a practicing design engineer or research scientist. The reader
should be familiar with college mathematics and basic electronics.
Bartee, Thomas C., Lebow, Irwin L., and Reed, Irving S., Theory and Design of Digital
MaChines New York: McGraw-Hili, 1962
Combines switching theory and machin'e design, Including the desigll of general
purpose, special purpose, and sequential machines.
388
Chu, Yaohan, Digital Computer Design Fundamentals New York: McGraw-Hili, 1962
Discusses arithmetic operations in binary, BCD, floating point, and residue numbers.
Also describes circuitry, using different types of modern circuit elements. The text
leads to the design of a simple computer.
Ledley, Robert S., Digital Computer and Control Engineering New York: McGraw-Hili,
1960
A thick book that includes programming, systems design, logic design, and circuit
design. A simple computer, Pedagac, is designed from start to finish.
Phister, Montgomery, Jr., Logical Design of Digital Computers New York: John Wiley &
Sons, Inc., 1958
Develops the design of a computer from boolean equations. It covers all subjects,
including circuits, memory, and input output, from a strictly mathematical approach.
Scott, Norman R., Analog and Digital Computer Technology New York: McGraw-Hili,
1960
Excellent basic book on principles and applications of analog and digital computers.
Analog topics include general approach to problem solving, representation of non-
linear functions, and amplifier design. Digital topics covered are problem solving
approach (very little on programming), number systems, switching and logic cir-
cuits and their design, arithmetic and control circuits, and memory elements. Written
for graduate and advanced undergraduate electrical engineers.
REFERENCE WORKS
These books are intended primarily as an aid to the practicing designer. Each section is
prepared by a specialist in the field, and contains detailed, concise information.
Grabbe, Ramo and Wooldridge (editors)
Handbook of Automation, Computation and Control
New York: John Wiley & Sons, 1959.
Volume 1 - Control Fundamentals. Emphasizes mathematics including sets and
relations, Boolean algebra, prc.bability, and statistics, as well as numerical analysis,
operations serearch, and information theory.
Volume 2 - Computers and Data Processing. Discusses computer terminology,
digital computer programming, the design and use of digital computers, data
processors, analog computers, and unusual computer systems. Included in this is
a discussion of digital computer circuits, logical design, and teChniques for reli-
ability.
Volume. 3 - Systems and Components. Includes systems engineering, manufactur-
ing process control, chemical process control, and industrial control. The component
section treats selection, mathematical description, and integration of components
into systems.
Huskey, Harry D., & Korn, Granino A .. Computer Handbook New York: McGraw-Hitl,
1962
Section 1 discusses analog computers including terminology, basic building blocks,
design of computer systems, and computer applications. Section 2 deals with digi-
tal computers. including definitions, components, circuits, logic design, program-
ming. system design, and applications.
389
390
NUMERICAL INDEX
A100 Multiplexer Switch. . ........... 344 H704 Power Supply. . ............ 228
A103 Multiplexer Switch. . ....................... 344 H800 Connector Block. . ............ 233
A111 Multiplexer Guarded Relay ... 345 H801 Replacements Contacts. . ... 233
A121 Multiplexer Switches ........ 344 H802 Single Connector Block. . .. 234
A200 Operational Amplifier .. 346 H803 36Pin Connector Block. .235
A400 Sample and Hold Amplifier. 348 H804 Pin Connector Block ....... 235
A502 Comparator. . .. 349 H805 Replacement Pins. . ....... 235
A601 DigitalAnalog Converter. . .... 350 H810 Pistol Grip Wire Wrapping Tool. ..243
A604 DigitalAnalog Converter. 351 H811 Hand Wrapping Tool. 243
A605 DigitalAnalog Converter 351 H812 Hand Unwrapping Tool. 243
A606 DigitalAnalog Converter. 352 H820 Grip Clips. 244
A702 Reference Supply ......................... 353 H825 Hand Crimping Tool. . ... 244
A704 Reference Supply ... 353 H830 Stackon Riveting Tool. . ................ 247
A990 Amplifier Board ... ..: ................. 354 H900 Mounting Panel w/power .......................... 237
A992 Amplifier Board. . ..... 354 H900A Mounting Panel w/power .................... 237
B104 Inverter ............. 108 H901 Mounting Panel ....................... 214
B105 Inverter ................................................ 108 H902 Switch and Indicator Panel 215
B113 NAND/NOR Gate. . ............................ 110 H903 AnalogDigital Panel. . ...... 216
B115 NAND/NOR Gate ..110 H910 Pin EquivalentjH900 ................................ 237
B117 NAND/Nor Gate .................110 H910A Pin EquivalentjH900A ............................ 237
B123 Inverter ..... 108 H911B Pin Equivalentj1943 Series. . ....... 239
B124 Inverter ...... 108 H911 BP Pin Equivalent/H911 B . . .. 239
B130 Threebit Parity Circuit .. . ... 112 H911M Pin Equivalent/1943FM ... . ... 239
B155 Half Binary to Octal Decoder. . ............. 113 H911MP Pin EquivalentjH911M . . ......... 239
Bl71 NAND/NOR Gate. . ................ 110 R001 Diode Network. 57
8200 FlipFlop . . ......................... 114 R002 Diode Network. 57
B201 FlipFlop . . ...... : ..... 116 R107 Inverter 58
B204 Quadruple FlipFlop . . .......................... 118 R111 Expandable NAND/NOR Gate. 59
B301 Delay (One Shot) . ..119 R113 NAND/NOR Gate. 60
B310 Delay ......... 121 R121 NAND/NOR Gate. . 61
B360 Delay with Pulse Amplifier ............................ 122 R122 NOR/NAND Gate 62
B401 Clock ......................... 123 R123 Input Bus. 63
B405 Clock .................. 123 R131 Exclusive OR. 64
B602 Pulse Amplifier. . ......................... 124 R141 AND/NOR Gate.. 65
B620 Carry Pulse Amplifier .......... 125 R151 BinarytoOctal Decoder. 66
B681 Power Inverter ......................................... 126 R181 DC Carry Chain. 67
B684 Bus Driver ....................... 127 R200 FlipFlop . 69
CABl Cabinet .............................. 247 R201 FlipFlop ..... 70
CAB2 Cabinet. . . . ........... 248 R202 Dual FlipFlop . 71
CAB3 Cabinet . .. 248 R203 Triple FlipFlop . 72
CAB6 Cabinet . . . ..................................... 248 R204 Quadruple FlipFlop .. 73
CAB8A Cabinet. . ..... 248 R205 Dual FlipFlop . 74
CAB8B Cabinet. . ............................... 248 R302 Dual Delay Multivibrator . 75
H001 Brackets ................................................ 241 R303 Integrating One Shot. 77
H002 Brackets ........................................ 241 R401 Variable Clock. 78
H201 Core Memory .............................................. 227 R405 Crystal Clock 79
H701 Power Supply. . ................................. 229 R601 Pulse Amplifier. 80
H701A Power Supply. . .. 229 R602 Pulse Amplifier. 81
391
R603 Pulse Amplifier 81 W991 Blank Module/Double Height/36'pins 172
R650 Bus Driver 82 W992 Blank Module/Copper Clad/18pins 172
W002 Clamped Load .132 W993 Blank Module/Double Height/36pins 172
W005 Clamped Load 132 W994 Blank Module/Perforated Board 172
W018 Cable Connectors for Indicator Amplifiers 133 W995 Blank Module/Perforated Board/
W021 Cable Connector for Levels and Pulses 134 Double Size 172
W022 Cable Connector for Levels and Pulses 134 700D Power Supply 212
W023 Cable Connector 133 700DA Power Supply 212
W028R Cable Connector 134 728 Cabinet/Plenumdoor mounting 230
W040 Solenoid Driver 136 728A Cabinet/Plenumdoor mounting 230
W042 Solenoid Driver 137 782 Power Supply /Rackmounled 229
W043 Solenoid Driver 136 782A Power Supply /Rackmounted 229
W050 30 ma Indicator Driver 138 783 Power Supply 231
W051 100 ma Indicator and Relay Driver 139 783A Power Supply 231
W061 Relay Driver 140 786 Power Supply 232
W080 Isolated ACDC Switch 141 786A Power Supply 232
WI03 Device Selector 142 831 Power Control 236
WI08 Decoding Driver 143 900 Control Panels 213
W500 High Impedance Follower 144 911 Patchcords 214
W501 Schmitt Trigger 145 913 Patchcords 244
W502 PhotonCoupled Trigger 147 914 Power Jumper 245
W510 Positive Input Converter 148 932 Bus Strip 244
W511 Negative Input Converter 149 1907 Mounting Panel Cover 241
W512 Positive Level Converter 150 1943 Mounting Panel 238
W520 Comparator 151 1945 Hold Down Bar 237
W532 Initial Transient Detector 152 4906 Indicator with Amplifier 242
W533 Rectifying Slicer 153 4908 Panel Indicator Assembly 242
W590 IBM N Line/DEC Converter 154 4912 Socket Adapter 246
W600 Negative Ouput Converter 155 4913 Mounting Rack 217
W601 Positive Output Converter 156 4917 Indicators with Amplifiers 242
W602 Bipolar Output Converter 157 4918 Indicators with Amplifiers .242
W603 Positive Level Amplifier 158 DOOIA DA Converter 252
W607 Pulse Output Converter 159 DOOIB DA Converter 252
W640 Pulse Output Converter 159 DOOIC DA Converter 252
W690 DEC/IBM N Line Converter 160 DOOID DA Converter 252
W700 Switch Filier 161 DOO I E DA Converter 252
W705 +3.6 Volt Power Supply 162 DOOIF DA Converter 252
W706 Teletype Receiver 163 D002 BCD UpCounter 254
W707 Teletype Transmitter 166 D004 BiDirectional Decade Counter 255
W708 Communications Adapter 169 D005 Input Buffer Interface 257
W800 Relay 170 D006 Output Buffer Interface 259
W802 Relay Multiplexer 171 D007 Dual Shift Register 261
W970 Blank Module/36'pins 172 D008 Dual UpCounter 263
W971 Blank Module/Double Height/72'pins 172 COOl AD Converter 265
W972 Blank Module/CopperClad/36pins 172 COOIA AD Converter 265
W973 Blank Modules/Double Height/72pins 172 C002 AD Converter 265
W980 Module Extender 173 C003 BCD Real Time Clock 267
W985 System Module Adapter 173 C005 I/O Bus Interface 270
W990 Blank 172 C006 I/O Buffer Register 271
392
ROOI
*R002
';'Rl07
*Rlll
Rll3
"R12l
"R122
RI23
RI3I
*R14l
'RlSl
RI8l
*R200
*R20I
*R202
*R203
*R204
*R205
*R302
R303
*R40l
'*R405
R60I
*R602
DIODE NETWORK
DIODE NETWORK
INVERTER
EXPANDABLE
NAND/NOR GATE
NAND/NOR GATE
NAND/NOR GATE
NOR/NAND GATE
INPUT BUS GATE
EXCLUSIVE OR
AND/NOR GATE
OCTAL DECODER
DC CARRY CHAIN
FlIPFLOP
FlIpFLOP
DUAL FlIPFLOP
TRIPLE FlIpFLOP
QUADRUPLE
FlIpFLOP
DUAL FlIpFLOP
DELAY
INTEGRATING
ONE SHOT
VARIABLE CLOCK
CRYSTAL CLOCK
PULSE AMPLIFIER
PULSE AMPLIFIER
'Normal Delivery offtheshelf..
"2mc only offtheshelf delivery.
PRICE LIST
EFFECTIVE APRIL 1, 1967
R SERIES
Seven diodes for adding inputs to gates and flipflops.
Five 2input diode networks; OR for ground AND for nega
tive signals, when connected to diodegate node points.
Seven oneinput diode gates. One inverter also has node
input.
Three 2input diode transistor gates. NOR's for ground
and NAND's for negative inputs. Node terminal provided
for additional inputs.
Five 2input diode transistor gates. NOR's for ground and
NAND's for negative inputs.
Four Rllltype gates with loads internally connected and
without nodes.
Logical complement of R12l, but somewhat slower. One
ma input and 18 ma outputs like R121.
Six twoinput Rllltype gates without loads or nodes.
Each gate shares one input with another.
Four gates which perform exclusive OR function.
Seven 2input gates. OR's NAND'ed together for ground
inputs; AND's NOR'ed together for negative inputs.
One of eight outputs is grounded for each combination of
states in three pairs of complementary inputs.
Six cascaded AND gates for counting without carry delay.
Twotransistor flipflop is set or cleared from its own out
puts or direct set and clear inputs. Operates at up to 2
megacycles.
FlipFlop with direct set and clear inputs and five diode
capacitor diode gates. Can be set or cleared from its
outputs.
Two flipflops, each with clear input and two diode
capacitordiode input gates. Can be set or cleared from
its outputs.
Three flipflops, each with direct clear input and diode
capacitordiode gate. Can be set or cleared from its
outputs.
Four flipflops, each with direct clear and set inputs. Two
flipflops share each directclear input. Can be set or
cleared from its outputs.
Two flip flops with common clear and each having three
diodecapacitordiode gates. Can be set or cleared from
its outputs.
Two one shot multivibrators triggered by capacitordiode
gates. Independent delays controlled internally or
externally.
Zero recovery time multivibrator with complementary out
put buffers. Delays controlled externally or internally.
Pulse amplifier and six capacitordiode input gates. Driven
by 40 to 100nanosecond, positivegoing pulses at up to 2
megacycles or 400nanosecond pulses at up to I mega
cycle.
Two pulse amplifiers, each with two diodecapacitordiode
gates and one ungated input Input frequency up to 2
megacycles.
393
$ 4.00
$ 5.00
$ 24.00.
$ 14.00
$ 20.00
$ 17.00
$ 26.00
$ 19.00
$ 35.00
$ 13.00
$ 33.00
$ 35.00
$ 9.50
$ 22.00
$ 25.00
$ 28.00
$ 28.00
$ 29.00
$ 44.00
$ 45.00
$ 4S.00
$100.00
$ 2S.00
$ 22.00
R SERIES (Cont.)
. Three pulse amplifiers. each with one diodecapacitor
"R603 PULSE AMPLIFIER diode gate and one ungated input. 'Input frequency up to
$ 2S.00
2 megacycles.
Two inverting bus drivers, each with 2-input NOR or NAND
':'R650 BUS DRIVER
diodetransistor gate. Node terminal provided for addi
$ 23.00
tiona I inputs.
W SERIES (Cont.)
"W002 CLAM PED LOAD Fifteen 2milliampere clamped loads.
$ 13.00
"W005 CLAMPED LOAD Fifteen 5milliampere clamped loads.
$ 15.00
W01S CONNECTOR MODULE
For 1Sline, ribboncable connections.
$ 18.00
For each foot of cable or fraction.
$ .60
W018U CONNECTOR MODULE Without cable.
$ 9.00
W021R
CONNECTOR For 9 direct ribboncable connections:
$ 13.00
MODULE For each foot of cable, or fraction:
$ .60
W021-RU
CONNECTOR
MODULE
Without cable.
$ 4.00
W021-C
CONNECTOR For 9 direct coaxial connections:
$ 31.00
$ 1.50
W021CU
CONNECTOR
MODULE
Without cable.
$ 4.50
W022-R
CONNECTOR For 9 terminated ribbon connections:
$ 13.50
$ .60
W022RU
CONNECTOR
Without cable. $ 4.50
MODULE
W022-C
CONNECTOR For 9 terminated coax connections:
$ 33.00
MODULE For each foot of cable, or fraction: $ 1.50
W022CU
CONNECTOR
Without cable.
$ 6.50
MODULE
CONNECTOR MODULE
For 1Sline, ribboncable connections. $ 13.00
W023
For each foot of cable or fraction. $ .60
W023U CONNECTOR MODULE Without cable. $ 4.00
CONNECTOR For 9 jumperable ribbon connections: $ 13.00
W028-R
MODULE For each foot of cable, or fraction: $ .60
W028-RU
CONNECTOR
MODULE
CONNECTOR For 9 jumperable coax connections: $ 31.00
W028-C
MODULE For each foot of cable, or fraction: $ 1.50
W028-CU
CONNECTOR
MODULE
SOLENOID
Two highcurrentdrive amplifiers, each with a diode NOR
$
':W040
or NAND gate. Node terminal provided for additional 36.00
DRIVER
inputs.
W042
10 AMP Four germanium transistor drivers which can provide up
$ 80.00
DRIVER to 10 amps of DC drive.
Two highcurrentdrive amplifiers, each with a diode NOR
W043 SOLENOID DRIVER or NAND gate. Node terminal provided for additional $ 35.00
inputs.
"W050
INDICATOR
Seven amplifiers for miniature, incandescent indicators. $ 13.00
DRIVER
W051 100ma DR IVER Seven highercurrent circuits. $ 22.00
W061 RELAY DRIVER
Four quarteramp, 55 volt, allsilicon drivers for loads
$ 35.00
returned to positive voltage.
ISOLATED Two floating switches controlled by light beams. Allows
W080 ACDC isolated control of lineoperated devices. Onequarter amp $ 60.00
SWITCH at 125v AC or DC.
"Normal Delivery olttheshelf.
394
W SERIES (Cant.)
WI03
DEVICE
For PDp8.
$ 52.00
SELECTOR
WI08
DECODING
300ma, bipolar drive with 8 drivers. $ 75.00
DRIVER
HIGH
Seven fault protected circuits, each comprising two cas
W500 IMPEDANCE
$ 25.00
FOLLOWER
caded emitterfollower amplifiers. Up to ,1:30v in.
LEVEL
Produces standard pulses from contact closures or non
CONVERTER
W501
AND SCHMID
standard negative logic levels. Switching thresholds can $ 13.00
TRIGGER
be 0 to -2.5 volts. Contactbounce integrator included.
PHOTON Two isolated trigger circuits; each responds to light from
W502 COUPLED tungsten filament falling on a photocell. Inputs 14v or $ 38.00
TRIGGER 48v.
W51O
POSITIVE LEVEL Three inverters with input thresholds which can be set at
$ 17.00
CONVERTER 0, + 1 v, or + 2v.
NEGATIVE LEVEL
Two circuits for converting inputs between - 50 volts and
'W511
CONVERTER
+ 25 volts to 3 volt levels. Switching point can be set to $ 17.00
0, -1, -2,or -3 volts.
W512
POSITIVE LEVEL
Wand A Series Interface for Positive Logic Systems. $ 25.00
CONVERTER
An inexpensive comparator for A/D work, or a general pur
W520 COMPARATOR
pose input level converter. Three circuits, each is a 4
$ 43.00
transistor difference amplifier, DEC Standard levels at
the output.
W532 DUAL AMPLIFIERS Two ACcoupled differential amplifiers. $ 30.00
W533
DUAL RECTIFYING Two slicers that amplifies signals from a W532 and con
$ 30.00
SLICER verts them to DEC signal.
IBM N LINE
Each of five inverting amplifiers provides inputs compati
W590 TO DEC
$ 26.00
CONVERTER
ble with three types of IBM N lines.
NEGATIVE LEVEL
Three diodetransistor inverting amplifiers convert stan
W600 dard levels to outputs of ground and -1 to -15 volts. $ 12.00
AMPLIFIER
Node terminals provide for additional inputs.
POSITIVE LEVEL
Three diodetransistot inverting amplifiers convert stan
W601
dard levels to ground and 1 to 20volt outputs. Additional $ 13.00
AMPLIFIER
gates added through node terminals.
BIPOLAR
Three inverting amplifiers for output levels at +6, +3,0,
W602 LEVEL
$ 40.00
AMPLIFIER
-3, or -6 volts.
W603
POSITIVE LEVEL Amplifier drives positive logic systems from FLIP CHIP
$ 23.00
AMPLIFIER systems.
W607
PULSE Three pulse amplifiers. Input: 40nanosecond (or wider)
$ 42.00
AMPLIFIER pulses at up to 2.5 megacyc!es.
PULSE
Three standardizing amplifiers with transformercoupled
W640
outputs of 400 nsec at up to 500 kc, or 1 I,sec at up to $ 42.00
AMPLIFIER
200 kc.
DEC TO IBM N
~ a c h of four inverting drivers provides outputs compatible
W690 LINE
$ 36.00
CONVERTER
with three types of IBM N lines. .
/ W700 SWITCH FILTER Six switch filters reduce the effects of contact bounce. $ 20.00
W705
POWER SUPPLY
Supplies up to 1 amp at +3.6 volts. $ 15.00
(+3.6v)
W706
TELETYPE Integratedcircuit, serialtoparallel Teletype code
$150.00
RECEIVER converter.
W707
TELETYPE Integratedcircuit, paralleltoserial Teletype code
$150.00
TRANSMIDER converter.
TELETYPE
Provides special gating controls and clock synchronization
W708
for Teletype and data communications systems, when used $ 55.00
INTERFACER
with the W706 and W707.
*Normal Delivery offtheshelf.
395
W800
W802
'W970
*W971
*W972
W973
W980
W985
W990
.*W991
'W992
*W993
W994
*W995
B104
*B105
*B113
B115
B117
B123
B124
B130
B155
Bl71
B200
*B201
B204
*B301
B310
B360
*B401
B405
B602
B620
RELAY
RELAY
MULTIPLEXER
BLANK BOARD
BLANK BOARD
CLAD BOARD
CLAD BOARD
MODULE EXTENDER
ADAPTER
BLANK BOARD
BLANK BOARD
CLAD BOARD
CLAD BOARD
PERFORATED
BOARD
PERFORATED
BOARD
INVERTER
INVERTER
NAND/NOR GATE
NAND/NOR GATE
NAND/NOR GATE
INVERTER
INVERTER
3BIT PARITY
CIRCUIT
HALF BINARYTO
OCTAL DECODER
NAND/NOR GATE
FLlpFLOP
FLlPFLOP
QUADRUPLE
FLlPFLOP
DELAY
DELAY
DELAY (WITH
PULSE
AMPLIFIER)
VARIABLE CLOCK
CRYSTAL CLOCK
PULSE
AMPLIFIER
CARRY PULSE
AMPLIFIER
*Normal Delivery offtheshelf.
W SERIES (Cont.)
Two normally open reed relays for up to ~ amp, 200 volts
with contact protection provided.
Eight doublepole, normally open relays, their drivers and
a gating circuit.
36pins
Doubleheight W970, 72 pins
36pins
Doubleheight W972, 72 pins
Allows access to circuits' without breaking connections.
Puts a System Module in four FLIP CHIP slots.
18Pins
Doubleheight W990, 36'pins
18pins
Doubleheight W992, 36'pins
A doubleheight W994.
B SERIES (Cont.)
Four lOmegacycle inverters and three clamped load
resistors.
Five 10megacycle inverters and five clamped load
resistors.
Four 2input diodetransistor NOR or NAND gates with
three clamped load resistors.
Three 3input diode transistor NOR or NAND gates and
three clamped load resistors.
Two 6input diodetransistor NOR or NAND gates.
Three 10ma clamped loads, 4 pairs transistor inverters.
Three groups of three inverters, and three clamped load
resistors. Logically complementary to Bl15.
Two levels of gating at high speed, with complementary
outputs.
One output is grounded for each of four input combina
tions, if each tlf three auxiliary inputs is arso enabled.
One 12input diode gate with two transistors. NOR and OR
for ground signals; AND and NAND for negative signals.
For building simple 10 mc registers at low cost. Short
delay from pulse input to flipflop output.
General purpose flipflop and nine input transistors.
Four flipflops with an inverter for common clear. Can be
set or cleared from their outputs.
One shot multivibrator with both pulse and level outputs.
Four delay lines, each with maximum delay of 50 nano
seconds in 12.5 nanosecond steps.
Adjustable delay line (250 nanoseconds, maximum) and a
pulse amplifier.
Two pulse amplifiers for negative or positive pulses.
Two stages for long B201 counters.
396
$ 45.00
$160.00
$ 4.00
$ 8.00
$ 3.00
$ 6.00
$ 14.00
$ 34.00
$ 2.50
$ 5.00
$ . 2.00
$ 4.00
$ 4.40
$ 8.80
$ 17.00
$ 21.00
$ 23.00
$ 21.00
$ 14.00
$ 31.00
$ 31.00
$ 50.00
$ 25.00
$ 18.00
$ 25.00
$ 56.00
$ 29.00
$ 73.00
$ 66.00
$ 84.00
$ 57.00
$100.00
$ 36.00
$ 47.00
B SERIES (Cont.)
Four inverters each equivalent to two paralleled B105
B681
POWER type inverters. Includes a 20 ma clamped load internally
$ 25.00
INVERTER connected to each collector, and four separate 10 ma
loads.
B684 BUS DRIVER Two 40 ma drivers for direct or terminated connection. $ 52.00
A SERIES
AlOO
Solidstate analog switches. Transition time: less than one
$100.00
MULTIPLIER
microsecond; offset; 100 to 300 microvolts. Controlled by
Al03
SWITCHES
standard levels and high frequency carrier. Transformer
$ 78.00
A12l
coupled for isolation.
$ 65.00
A11l
MULTIPLEXER Low level relay multiplexer with guarded contact wiring.
$ 93.00
SWITCH Two threepole relays.
OPERATIONAL
15 volt amplifier with a 2xlO' openloop voltage gain and a
tA200
l5megahertz unity gain. Differential inputs accept up to $130.00
AMPLIFIER
10volt common mode signals.
Highspeed, 10volt amplifier with a 10,000 openloop
A20l
OPERATIONAL voltage gain and a l5megacycle gainbandwidth product.
$180.00
AMPLIFIER Differential inputs accept up to 10volt commonmode
signals ..
A400
SAMPLE AND 10 volt sample and hold able to track a full scale ex
$330.00
HOLD AMPLIFIER cursion in 12 microseconds to 0.25% accuracy.
*A502 COMPARATOR
Difference amplifier for signals as small as one millivolt.
$110.00
Standard output drives R or Bseries modules.
DIGITALTO
Threebit star networks with drivers and terminating
*A60l ANALOG
$ 60.00
CONVERTER
resistor.
DIGITALTO
*A604 ANALOG Twobit network with drivers for more accuracy. $ 62.00
CONVERTER
DIGITALTO
Twobit wirewound network with drivers for maximum
A605 ANALOG
$ 78.00
CONVERTER
accuracy.
DIGITALTO
Twobit network used with A604 to form one decade of
A606 ANALOG
$ 62.00
CONVERTER
BCD D to A conversion.
A702 REFERENCE
Both -lOvolt references. A702: 10millivolt ripple, 30
$ 58.00
millivolt load regulation. A704: Olmillivolt ripple, 0.1
*A704 SUPPLIES
millivolt load regulation.
$184.00
AMPLIFIER
Predrilled etched board on which can be mounted one of
A990
several dozen types of operational amplifiers manufac $ 4.00
BOARD
tured by three different companies.
AMPLIFIER
Similar to A990, and also receives booster amplifiers.
A992
Accepts even more different types, from many $ 4.00
BOARD
manufacturers.
HARDWARE ACCESSORIES
CABl CABINET
Includes End Panels $ 700.00t
Additional Cabinets $ 500.00t
CAB2 CABINET
CAB3 CABINETS Expander cabinet for PDp8 $ 650.00
CAB6 CABINET
CAB8A CABINETS Free standing with winged table. $1,100.00
'Normal Delivery offtheshelf.
tNondiscountable
397
HARDWARE ACCESSORIES (Cont.)
CAB-SB CABINETS Free standing with rectangular table. $1,000.00
':'H001
BRACKETS
One pair brackets to mount a 1907 cover plate on a 19"
mounting panel. % inch standoff, cover plate
flush with cabinet.
$ S.OO
"H002 BRACKETS
One pair brackets to mount a 1907 cover plate on a 19
inch mounting panel. Provides 2 inch setback to leave $ S.OO
room for lights, controls.
H201 CORE MEMORY 4096 x 13 bit core memory (90 day guarantee) $2,000.00
"H701 POWER Small, chassis mounted supplies electrically similar to $ 116.00
*H701A SUPPLIES 7S2A.
$ 136.00
Will drive six operational amplifier modules. All silicon,
H704 POWER SUPPLY regulated, floating supply providing 15 volts at up to $ 200.00t
400 ma on both outputs.
"HSOOF CONNECTOR
Unmounted 144 pin connectors for S FLIP CHIP modules.
$ 8.00
"HSOOW BLOCKS $ 8.00
*HS01F CONNECTOR
Set of IS spare pins to fit FLIP CHIP sockets.
$ 4.00
':'HS01W PINS $ 4.00
HS02 CONNECTOR BLOCK For single FLIP CHIP modules. $ 4.00
HS03 MODULE SOCKETS Unmounted 2SS pin connectors for S-36 pin modules. $ 13.00
HS04 MODULE SOCKETS Unmounted 144 pin connector for SlS pin modules. $ 9.00
HS05 CONNECTOR PINS Set of IS spare pins to fit HS03 and HS04 connectors. $ 4.00
HS10
PISTOL GRIP
For wrapping #24 solid wire on DEC Type HSOO-W Can
HAND WIRE
$
(24)
WRAPPING TOOL
nector Pins.
99.00t
HS10
PISTOL GRIP
For wrapping #30 solid wire on DEC Type HSOOW Can
HAND WIRE
$ 99.00
(30)
WRAPPING TOOL
nectar Pins.
HSlO
PISTOL GRIP
For wrapping #24 and 30 wire on DEC type HSOOW and
HAND WIRE
$ 150.00t
(24 & 30)
WRAPPING TOOL
HS03 Connector Pins.
HS11 HAND WRAPPING For wrapping #24 wire on Connector Pins in service or
$ 21.50t
(24) TOOL repai r applications.
HS11A HAND WRAPPING For wrapping #30 wire on Connector Pins in service or
$ 21.50t
(30) TOOL repair applications.
HS12 HAND UNWRAPPING
For removing #24 wire wrapped connections.
$ 10.50t
(24) TOOL
HS12A HAND UNWRAPPING
For removing #30 wire wrapped connections.
$ 10.50t
(30) TOOL
Identical to slip-on connectors used on Type 913 patch
HS20 GRIP CLIPS cords. Fit sizes 2420 AWG wire. Shipred in packages of
1000.
$ 47.S0t
HS25
HAND For crimping Type HS20 Grip Clips. Insures a good elec
$ 146.70t
CRIMPING TOOL trical connection.
HS30
STACKON RIVETING
$ 10.00
TOOL
*H900
MOUNTING
For up to 32 modules. Power supplies electrically similar
$ lS0.00
PANELS WITH
H900A
POWER
to 7S2, 7S2A.
$ 200.00
"H901
MOUNTING
For patching up to 10 modules.
$ S2.50
PANEL
SWITCH AND
5in. panel for experimental use. Eight switches, eight
indicator lamps with. groundenabled drivers, .four- rhea
*H902 INDICATOR
stats for controlling clocks and delays. Complements H901
$ 112.S0
PANEL
and 700D. Complete with indicator driver module.
*H903
ANALOG-DIGITAL
D-A converter and comparator (including module).
$ 143.00
PANEL
H910 36pin equivalent of H900.
$ 200.00
H910A 36-pin equivalent of H900A.
"Normal Delivery offthe-shelf.
tNon-discountable
39S
H911B
H911BP
H911M
H911MP
''700D
700DA
':'728
728A
"782
782A
"783
783A
786
786A
831
"900
MOUNTING PANEL
MOUNTING PANEL
MOUNTING PANEL
MOUNTING PANEL
POWER SUPPLY
WITH DIAL
POWER
SUPPLIES
POWER
SUPPLIES
POWER
SUPPLIES
VARIABLE
POWER
SUPPLIES
POWER
CONTROL
CONTROL PANEL
HARDWARE ACCESSORIES (Cont.)
Similar to the 1943 but uses H803 and H804 connectors.
Following options available as indicated by the suffix
Marginal Check Switches (M) or terminal block
(B), prewired power (P) or not.
5in. panel for powering and pulsing experimental logic
systems patched together on H901 panels. Contains a
variable clock, three pulsers, with a telephone dial and
three pushbuttons for driving the pulsers. Designed es-
pecially for educational use.
50cycle version of 700D.
+ 10 v @ 0.7 amp, 15 v @ 1-8 amp, for 17" DEC
cabinet plenumdoor mounting.
Rack-mounted power supplies with regulated +10 and
-15 volt output at up to 0.4 and 3 amperes, respectively.
Input: 782: 115 volts, 60 cycles. 782A: 112.5 to 235
volts, 50 cycles.
Rackmounted power supplies with electrical character-
istics similar to 728 and 728A.
Standard rack mounting 0-24 volt variable supplies, 0-2.5
amp.
Specify 4, 10,20, or 30 amp circuitbreaker.
A front panel and chassis exactly duplicating 700D but
without power. For multi-student installations on a Logic
Laboratory.
$ 151.00
$ 161.00
$ 172.00
$ 182.00
$ 323.00
$ 343.00
$ 240.00
$ 260.00
$ 128.00
$ 148.00
$ 240.00
$ 260.00
$ 215.00
$ 235.00
$ 51.00
$ 214.25
_'_'9_1_1 ______ __ bo_x_. __________ $ ____ 9_.0_0_
WIRE WRAP
"913 PATCHCORDS Hundred each 2, 4, 8, 16,32 or 64 inches. One pkg. $ 18.00
"914-7
POWER
JUMPERS Ten each 7 inches long. One pkg. $ 4.00
"914-19 POWER T h 19' h 10k $ 400
JUMPERS en eac inC es ong. ne p g. .

"1907
MOUNTING
PANEL COVER
51;4 by 19 inch cover plate to protect power and logic
wiring and permit installation of lights and control $
switches on a 19 inch mounting panel. Use with one pair
of brackets, HOOI or H002 (order separately).
9.00
" 1943-F-B $ 111.00
':' 1943W-B $ 111.00
"1943-F-B-P Houses 64 modules in a standard 19-inch rack. Following $ 121.00
"1943-W-B-P MOUNTING options available as indicated by the suffix lelters: Forked- $ 121.00
"1943.F.M PANEL pin solder (F) or wirewrap (W) connectors; marginal- $ 132_00
':'1943-W-M check switches (M) or terminal block (B); prewired power $ 132.00
':' 1943-F-M.P (P) or not. $ 142.00
*1943-WMP $ 142.00
194519 HOLD ________________ -=$ __
SINGLE
INDICATOR With amplifier. $ 9.50
4906
SINGLE
INDICATOR Lamp and mounting hardware. $ 3.00
..
':'4908,
"4913
MOUNTING F I d f 51 V: , I $ 47.00
RACK or a power supp y an our 2 ' pane s.
4917 INDICATORS 9 bit indicators with amplifiers. $ 73.00
4918 INDICATORS 18 bit indicator with amplifiers. $ 96.00
"'Normal Delivery offtheshelf.
399
OCTAIDS and PANELAIDSA
DOOIA DA CONVERTER 8 Bits $ 377.25
D001B DA CONVERTER 9 Bits $ 439.25
D001C DA CONVERTER 10 Bits $ 439.25
00010 DA CONVERTER 11 Bits $ 501.25
DOOlE DA CONVERTER 12 Bits $ 519.25
D001F DA CONVERTER 13 Bits $ 597.25
0002 BCD UPCOUNTER
Single Decade $ 62.50
Quad Decade $ 137.20
BIDIRECTIONAL Decoding Option
$ 123.40
0004 $ 47.00
DECADE COUNTER Conversion Option
$ 17.00
0005
INPUT BUFFER
For PDP8/S $ 122.00
INTERFACE
0006
OUTPUT BUFFER
For PDP8/S $ 215.50
INTERFACE
0007
DUAL SHIFT 8,Bits $ 249.00
REGISTER Single Register $ 133.00
0008 DUAL UPCOUNTER
8 Bits $ 217,00
Single UpCounter $ 117.00
COOl AD CONVERTER 10 Bits, 60 Hertz $1,144.00
C001A AD CONVERTER 10 Bits, 50 Hertz $1,162,00
C002 AD CONVERTER 10 Bits, without AC power, $1,079.00
C003
BCD REAL
50/60 Hertz $ 639,00
TIME CLOCK
C005
I/O BUS
For PDP8/S $ 12.00
INTERFACE
C006
I/O BUFFER
For PDp8 or PDP8/S $ 397.00
REGISTER
ASee Printed Circuit Board Price List on Page 273.
All shipments are F.O.B. Maynard, Massachusetts, and prices do not include state or local taxes.
Prices and specifications are subject to change-without notice.
QUANTITY DISCOUNTS
$5,000-3%; $10,000-5%; $20,000-10%; $40,000-15%;
$70,000 - 18%; $100,000 - 20%; $250,000 - 22%; $500,000 - 25%
Discounts apply to any combination of FLIP CHIP, Cabinets are not discountable.
400
FOR MORE MODULE INFORMATION
INFORMATION REQUEST
o Please add my name to your mailing list to receive future technical bulletins and
module application notes as they are issued ..
o Please forward any available information on the following application(s):
Please have a Digital engineer contact me for an appointment.
o I would like a demonstration of the Logic Laboratory.
o I would like to discuss the following application(s):
Please send technical literature on the following Digital computer products:
o PDp7 0 PDp8 0 LlNCEIGHT 0 CRT Displays
o Other __________________________________________________ _
Name ___________________________________ ,Position ______________ _
Company ____________________________________________________ __
Dept. or Div. __________________________________________________ __
Business ____________________________________________________ __
Street ______________________________________________________ __
City ______________ -:-____________ State ____________________ _
ZI P Code No. ______________________ Telephone ____________________ __
401
........................................................................ FOLD HERE ......................................................................... .
BUSINESS REFLY MAIL
NO POSTAGE STAMP NECESSARY IF MAILED. IN THE UNITED STATES
FIRST CLASS
PERMIT NO. 33
MAYNARD, MASS.
._
146 MAIN STREET
402
PERHAPS YOU HAVE A FRIEND WHO WOULD LIKE TO
RECEIVE A FREE COpy OF THE 1967
~ D m D D m D LOGIC HANDBOOK, THE NEW SMALL
,COMPUTER HANDBOOK, OR THE NEW
,
,
,
: INDUSTRIAL CONTROL HANDBOOK.
,
,
:
,
,
I IF SO, PLEASE FILL OUT THE CARD BELOW.
r
r
:
r
1
z
~
o
IIJ
-------------------------------------'FOLD HERE -------------------------------------
CJ
Z
9
4(
I-
::J
~ ..
r
I
,
I
r
r
I
r
:
I
Tech n i.ca I Publications Dept.
146 Main Street
Maynard, Mass. 01754
GENTLEMEN:
Please send a free copy of the Digital Logic Handbook ___ , the In-
dustrial . Control Handbook __ , Small Computer Handbook ___ , or all
three __ to:
Name ______________________ __
Position _____________ ~ _______ _
Company _____________________ __
Business _____________________ __
~ ~ & - - - - - - - - - - - - - - - - - - - - - - - - -
Telephone ____________________ _
City ________________________ _
State _____________ Zip Code ______ _
BUSINESS REPLY MAIL
NO POSTAGE STAMP NECESSARY IF MAILED IN THE UNITED STATES
FIRST CLASS
PERMIT NO. 33
MAYNARD, MASS.

146 MAIN STREET
404
SALES AND S E ~ V U ~ E
MAUNl OFFDCE AND PLANT
146 Main Street, Maynard, Massachusetts 01754
Telephone: From Metropolitan Boston: 6468600
Elsewhere: (617)-897-8821
TWX: 710-347-0212 Cable: Di9ital Mayn. Telex: 9-4819
DCMESTDC
NORTHEASTERN
NORTHEAST OFFICE:
146 Main Street, Maynard, Massachusetts 01754
Telephone: (617)-646-8600 TWX: 710-347-0212
BOSTON OFFICE:
a9g Main Street, Cambridge, Massachusetts 02142
Telephone: (617)-491-6130 TWX: 710-320-0468
ROCHESTER OFFICE:
455 Empire Boulevard, Rochester, New York 14609
Telephone: (716)-482-2310 TWX: 510-253-3078
CONNECTICUT OFFICE:
129 College Street, New Haven, Connecticut 06510
Telephone: (203)-777-5797 TWX: 710-465-0692
MIO-ATLANTIC
NEW YORK OFFICE:
1259 Route 46, Parsippany, New Jersey 07054
Telephone: (201)-335-3300 or (212)-279-4735
TWX: 510-235-8319
PHILADELPHIA OFFICE:
Southampton Medical & Professional Bldg.
Second Street Pike, Southampton, Pa. 18966
Telephone: (215)-357-9450 (9451) TWX: 510-667-1724
WASHINGTON OFFICE:
Executive Building
71 00 Baltimore Ave., College Park, Maryland 20740
Tolephone: (301)-779-1100 TWX: 710-826-9662
SOUTHERN
HUNTSVILLE OFFICE:
Suite 41 - Holiday Office Center
3322 Memorial Parkway S.W., Huntsville, Ala. 35801
Telephone: (205)-881-7730 TWX: 510-579-2122
CANAOA
Digital Equipment of Canada, Ltd.
150 Rosamond Street, Carleton Place, Ontario, Canada
Telephone: (613)-257-2615 Telex: 013-442
Digital Equipment of Canada, Ltd.
39 Dundas Road East, Cooksville, Ontario, Canada
Telephone: (416)-279-1690 Telex: 022-9665
GERMANY
Digital Equipment GmbH
5 Koeln, Neue Weyerstr. 10, West Germany
Telephone: 235501 Telex: 8882269
Telegram: Flip Chip Koeln
Digital Equipment GmbH
8 Muenchen, Theresienstr. 29, West Germany
Telephone: 299407 and 292566 Telex: 841 24226
Telegram: Digital Muenchen
ENGLAND
Digital Equipment Corporation (U.K.) Ltd.
3 Arkwright Road, Reading, Berkshire, England
Telephone: Reading 83366 Telex: 851-84327
Digital Equipment Corporation (U.K.) Ltd.
The Precinct
46 Bolton Road, Walkden, Manchester, England
Telephone: Walkden 5660
HOUSTON OFFICE:
3417 Milam Street, Suite A, Houston, Texas 77002
Telephone: (713)-523-2529 TWX: 910-881-1651
MIOWEST
PITTSBURGH OFFICE:
300 Seco Road, Monroeville, Pennsylvania 15146
Telephone: (412)-351-0700 TWX: 710-797-3657
CHICAGO OFFICE:
910 North Busse Highway, Park Ridge, Illinois 60068
Telephone: (312)-825-6626 TWX: 910-253-0342
ANN ARBOR OFFICE:
3853 Research Park Drive, Ann Arbor, Michigan 48104
Telephone: (313)-761-1150 TWX: 810-223-6053
WESTERN
LOS ANGELES OFFICE:
801 E. Ball Road, Anaheim, California 92805
Telephone: (714)-776-6932 or (213)-625-7669
TWX: 910-591-1189
SAN FRANCISCO OFFICE:
560 San Antonio Road, Palo Alto, California 94306
Telephone: (415)-326-5640 TWX: 910-373-1266
SEATTLE OFFICE:
McAusland Building, 10210 N.E. 8th Street
Bellevue, Washington 98004
Telephone: (206)-454-4058 TWX: 910-440-2306
DENVER OFFICE:
Suite 205-
5200 South Quebec Way, Englewood. Col. 80110
Telephone: (303)-771-1180 TWX: 910-935-0711
FRANCE
Equipement Digital
22, rue du Champ de l'Alouette, 75-Paris 8", France
Telephone: 336-0384 Telex: 842-26705
AUSTRALIA
Digital Equipment Australia Pty. Ltd.
89 Berry Street
North Sydney, New South Wales, AUstralia
Telephone: 92-0919 Telex: 790AA-20740
Cable: Digital, Sydney
Digital Equipment Australia Pty. ltd.
36 Outram Street
West Perth, Western Australia
Telephone: 214-993
.JAPAN
Rikei Trading Co., Ltd.
Kozato-Kaikan Bldg.
No. 18-14, Nishishimbashi 1-chome
Minto-ku, Tokyo, Japan
Telephone: 5915246 Telex: 7814208
SWEOEN
Digital Equipment Corporation
Birger Jarlsgatan 110 Stockholm, Sweden
Telephone: 32 25 98
DIGITAL announces price reductions for its FLIP CHlpTM modules in
this issue of the Logic Handbook. New integratedcircuit modules, noise
immune industrial modules, functional Teletype receiver-transmitter
modules, and OCTAIDTM and PANELAIDTM Kits are introduced. Included
i s a complete Analog-to-Digital Conversion Handbook.
PRINTED IN U.S.A. 17503/67

Digital Logic Handbook 1967

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Digital Logic Handbook 1967

Uploaded by

Copyright:

Available Formats

mamaomo

You might also like