I
BM
ELECTRIC PUNCHED CARD
ACCOUNTING MACHINES
CUSTOMER ENGINEERING MANUAL OF INSTRUCTION
ELECTRONIC MULTIPLIER
TYPE
603
INTERNATIONAL BUSINESS MACHINES CORPORATION
NEW
YORK.
NEW
YORK
Published by
Department of Education
International Business Machines Corporation
Endicott, N. Y.
Copyright
1948
International Business Machines Corporation
590
Madison Avenue, New York
22,
N.
Y.
Printed in
U.
S.
A.
Form 22-3838-0
CONTENTS
INTRODUCTION
OPERATING FEATURES
................................................
Main Line Switch
......................................................
Power Indicating Light (Green)
......................................
Start Key
........................................................................
Stop Key
..........................................................................
Error Indicating Light (Red)
.....................................
................................................
Error Reset Push Button
.................................................
Factor Reversal Switch
Card Hopper
................................................................
Card Stacker
...................................................................
Speed
.............................................................................
............................................................................
Current
..............................................................
Control Panel
FUNCTIONAL PRINCIPLES
Multiplication
.....................................................................
Decimals
..............................................................................
One-Half Correction
...........................................................
Double Punch and Blank Column Detection
................
Multiplication Check
.......................................................
Group Multiplication
..........................................................
Product Overflow
................................................................
Product Summary
...............................................................
Column Splits
......................................................................
Punch Suppression
..............................................................
Class Selectors (Optional)
................................................
Distributor (Optional)
......................................................
Sign Control (Optional)
...............................................
Sign Corltrol Checking
.....................................................
MECHANICAL PRINCIPLES
Location of Parts
........................................................
Drive Mechanism
............................................................
Punch Clutch
..............................................................
Feeding Mechanisms
....................................................
Index and Cycles
...............................................................
Geneva Mechanism
..................................................
Single Revolution Timing Cam
..................................
Principle of Punching
................................................
Magnet Unit
.....................................................................
.........................................................................
Oil Pump
Cam Contacts
....................................................................
ADJUSTMENTS
PUNCH CLUTCH
......................................................................
GENEVA MECHANISM
..........................................................
Single Revolution Timing Cam
..........................................
Single Revolution Timing Cam Bracket
.........................
Geneva Clutch
.................................................................
PUNCH UNIT
.............................................................................
Belt Tension
.......................................................................
Feed Roll Tension
................................................................
..........................................................
Feed Knife Guides
Feed Knife Projection
..................................................
...............................................................
Feed Knife Block
........................................................
Hopper Guide Posts
...........................................................................
Throat
......................................................................................
Die
............................................................
Punch Bail Tongue
Interposer Pawl Lock Bar and Spring Bail
....................
Punch Magnet Armatures
................................................
......................................
Punch Bail Connecting Links
Punch Hopper Side Plates
.............................................
Punch Brush Lateral Alignment
..................................
Anchor Slide Adjustment
..............................................
Brush Timing
................................................................
Vertical Registration
.......................................
.....................................................................
Stacker Plate
..........................................................
Stacker Timing
.....................................................................
CONTACTS
Adjustment of Cam Contacts
........................................
...........
Adiustinent of Card Lever Contacts
ELECTRICAL PRINCIPLES
PUNCH UNIT CIRCUITS
........................................................
............................................................
Time Delay Circuits
........................................
Start Circuit and Bias Interlock
Start Interlock Circuits (Gang Punching Only;
............
Start and Running Circuits (No Cards in Machine)
-~----
Start and Running Circuits (Cards in Machine)
...........
......................................................
Reading Brush Circuits
I,
lectronic Computing Control Circuits
..........................
..............................................................
Punch Suppression
...............................................................
Product Summary
...............................................................
Product Overflow
Factor Reversal and Check Circuits
..................................
..............................................................
Group Multiplying
Column Split,
0
and
X,
Hot
1
.....................................
Sign Control (Optional Feature)
.....................................
Class Selectors (Optional)
.................................................
Distributor (Optional)
.....................................................
POWER SUPPLY CIRCUITS
....................................................
Main Transformer and Selenium Rectifiers
........................
Tube Power Supply Chassis
...........................................
Constant Ratio Voltage Regulator
...................................
Screen Grid Supply
.............................................................
Voltage Adjustments
......................................................
PRINCIPLE OF MULTIPLYING
Computing
Circuits-General
BASIC CIRCUITS
TRIGGER CIRCUIT
.......................................................
77
Theory of Operation
....................................................
77
Coupling of Trigger Circuits
........................................
85
Trigger Circuits Used in Electronic Computing
..
86
Control of Other Tubes by Triggers
................................
89
ELECTRONIC COUNTER
........................................................
91
Indicator Blocks
..................................................................
97
Counter Read-In from Card
...........................................
98
Read-In Pulse Circuit
........................................................
100
Counter Read-In at High Speed
..................................
101
COMPUTING CIRCUITS
....................................................
103
Multivibrator and Clippers
........................................
103
Electronic Timers
...................................................
110
Compute Start Control and Primary Timer
...................
112
Secondary and Tertiary Timers
....................................
113
Multiplier Advancing Pulses
....................................
114
Ten-Pulse Control
................................................
119
Multiplicand Counter Rolling Control
.........................
121
Multiplicand Read-Out
...............................................
124
Column Shift Switches
............................................
127
Tertiary Timer and Colun~n Shift Control
.................
13 1
Carry Control and Carry Circuits
.............................
134
Half-Entry
..............................................
139
Compute Stop
...................................................
140
Read-Out Circuits
.................................................
142
TESTING PROCEDURE AND TROUBLE ANALYSIS
Use of Neon Indicator Bulb. Voltmeter and Oscilloscope
149
PURPOSE OF TUBES IN THE COMPUTING SECTION
Feed Hoppar Start-Stop Keys
\
Joggle Plate
/
/
Read
and
Punch
Unit
Electronic Computing
Unit
ELECTRONIC MULTIPLIER
Type
603
IBM ELECTRONIC MULTIPLIER
TYPE
603
INTRODUCTION
THE
CONVENTIONAL
mulriplying machine using
mechanical counters for the computation of products is considerably limited in its speed of oper-
arion because of the inertia of moving parts. By
the use of electrical computation circuits, calculat-
ing speeds can be increased considerably. The Electronic Multiplier makes use of recently-developed
electronic circuits which calculations at
extremely high speeds.
Thus the burdensome and
usually slow-speed process of computing products
is reduced to an automatic high-speed process in
keeping with the other high-speed functions of the
IBM Accounting Machine Method. Calculations
involving earnings, material costs, discounts, inventories, and many other computations can be
effected automatically to
speed up the accounting
routines which normally require much time and
ond, between the reading and punching of each
card.
OPERATING FEATURES
THE
operating controls and features of this ma-
chine, which can be seen in Figure
1,
are all located
on the punch unit.
Main Line Switch
This switch must be
ON
for the machine to be
operative.
It must not be turned
OFF
while cards
are feeding through the machine.
Power Indicating Light (Green)
When this light is
ON,
the machine is ready for
operation.
It will not turn on until
suficient time
has been allowed for the electronic tubes to warm
effort.
UP.
The Type 603 Electronic Multiplier consists es-
Start Key
sentially of a unit for reading and punching and
an electronic computing unit connected
by a cable
This key is depressed to start the feeding of
as shown by the
general view of the machine on the
cards at the beginning of a run.
It must be held
frontispiece. The factors
punched in an IBM Card
down through three machine cycles, when first
are read by the reading unit, computations are
starting, before
automaric operation begins.
automatically made by the electronic computing
unit, and
+the result is then punched in the same
card by the punching unit. No time is lost waiting for the completion of the computing operations; all computations are
performed between
the time a card leaves the reading brushes and the
time it reaches the punching position. The machine is equipped with a control
panel which makes
it entirely flexible as to the reading and
punching
of information.
The IBM Electronic Multiplier, Type 603, represents the first commercial use of electronics for
multiplication.
The use of electronic circuits for
computing permits operation of this multiplier at
maximusr punching speed of 6000 cards per hour.
The multiplication itself is
performed in .027 sec-
Stop Key
This key is depressed for manual control of stopping the feeding of cards while the machine is
through running.
Error Indicating Light (Red)
This light glows when an error is detected by the
Double Punch and Blank Column Detection Device, or when a
product exceeds the card field
capacity as indicated by the Product Overflow
feature.
Error Reset Push Button
This button is depressed to extinguish the error
light and restore the machine to normal operation,
after the machine has stopped because of an error.
2
TYPB,
603
'E.LECTR.ON.lt 'MULTlPLF,ER
;
Card
Face
Edge
Figure
I.
Operating Features
Indicating
(Red)
#
Factor Reversal Switch Speed
When set
ON,
this switch autamatically reverses
The operating speed of rhis machine is
100
cards
the multiplier and multiplicand entry hubs.
It is
per minute regardless of the number of columns
used in checking operations.
punched and the size of the
multiplier or multi-
Card Hopper
plicand fields.
Cards are placed in the card hopper face down,
current
9
edge first. The capacity of the hopper is approxi-
This machine is supplied to operate only on
11
5
mately
800
cards.
volts or
230
volts
A.C.,
SO
or
60
cycle current.
Card Stacker
Control Panel
After leaving the last set of brushes, cards enter
the stacker which has a capacity of approximately
The automatic control panel provides a means
1000
cards.
If the stacker fills
to
capacity, the
for flexible setup of rhe machine for all operations.
machine will be stopped
automatically by the
Figure 2 shows a control panel with the function
stacker stop switch.
of each hub described.
Figure
2.
Control Panel-Explanation of Hubs
FUNCTIONAL PRINCIPLES
Multiplication
The Electronic Multiplier can multiply two
6-
digit factors to produce a 12-digit product. Figure
3
shows control panel wiring for an individual
mulriplication problem. The multiplier and multiplicand are read from the card at the first reading station, represented on the control panel by
the 80 exits labelled "Read for Entry and Control." They are wired to rhe multiplicand and
multiplier counter entries.
The
product is available for punching at Product Exit. The product
exit positions from which the product is read must
be wired
TO
the Punch Entry hubs representing the
columns in which the product is to be punched.
order. These positions are represenred on the control panel by the hubs labelled "Double Punch and
Blank
Columh Entry."
These hubs are supple-
mented by
10 control panel switches labelled
"Blank
Column Switches" which correspond to the
10 DPBC entry hubs. The blank column switches
must be set
ON
for blank column checking.
Multiplication Check
In order thar the double punch and blank column feature may be used for checking multiplication, it should be used on each original multiplying run to prove that only one hole has been
punched in each product column and that no coI-
Decimals
umns are unpunched. For this purpose, the product
When either of the two factors contains deci-
field columns in Read for Checking should be
mals, the decimal
in the product will
wired ro Double Punch and
lank
column ~ntr~,
equal the sum of the decimal ~ositions of .the fac-
as shown in Figure
3.
tors.
Only those decimal positions which are to
When the
punching has been checked in the
be retained in the product are wired from Product
original run, the double punch and blank column
Exit to Punch Entry.
detection feature may then be used in a separate
One-Half Correction
When some decimal places are dropped from a
product, rhe product can be corrected to the nearest whole number or decimal position by
adding
5
to the first position following the retained pro-
duct. In the Electronic
Multiplier, this
'/2
correction can be made in any of rhe six right-hand
positions of the product counter. On the control
panel,
52
entries are located directly above the
product exits, and the
'/2
entry common is ad-
jacent. The
5
for
'/2
correction will be entered
inro the product counter during multiplication,
once for each card.
run to prove the calculation.
The cards are re-run
through the machine with the wiring of
.the mul-
tiplicand and multiplier counters reversed by
plat-
ing the factor reversal switch
ON.
During this
second operation the machine again multiplies and
punches the resulr in the product field.
With the
product field wired from Read for Checking to
the DPBC Entry, any product punched in this
re-run which differs from the original product will
cause a
double-punched column and will therefore
be sensed
by double punch detection. If an error
is detected, the machine will
srop and turn on the
red
light at the front of the machine. The reset
button must be depressed to turn the
lighr out.
Double Punch and Blank Column Detection
The start key must then be depressed for one card
The Electronic Multiplier is
equipped with 10
cycle, at the end of which time the card in error
double punch and blank column detection posi-
will be in the top position in the stacker and may
tions.
Ten additional positions are available on
be removed for review.
FUNCTIONAL PRINCIPLES
5
Figure
3.
Multiplication
The multiplier field (two decimal places) is entered in
the multiplier counter.
The multiplicand field (four decimal places) is entered
in the multiplicand counter.
Only two decimal positions are retained in the product;
5 is added to the third decimal position to correct the
product to the nearest whole cent.
The product (two decimal places) is punched in col-
umns
76
-
80.
The punched product field is checked for double punching.
The punched product field is checked for blank columns.
For group multiplication, this dotted wiring should be
added.
Only
5
positions are allowed on the card for punching
the product.
If a product carries to six positions,
a 12
is punched in card column
70.
Wiring for checking is the same as for the orig-
inal run; the
only difference is the reversal of the
factors
by the factor reversal switch.
Group Multiplication
The Electronic Multiplier can be used for group
multiplication, in which one factor remains constant for all cards in a group. The common factor
for group multiplication in the Electronic Multiplier must be wired to the multiplier counter. (The
multiplicand counter may be used to carry the
group multiplier if the factor reversal switch is
set
ON.)
The multiplier must be punched in a
card
designated by a 9 punch and placed at the
front of each group. The multiplier counter will
not reset until a group is finished and the special
multiplier card for the next group is about to be
read.
For group multiplication, the group multiplier
control panel switch should be wired
ON.
The
column punched
9
in the group multiplier card
must also be wired from Read for Entry and Control to one of the group multiplier pickups labelled
"9".
The reading of a 9 by the pickup causes the
machine to reset the multiplier counter, read in the
new multiplier, and eliminate punching and checking of the rate card.
With the group multiplier switch wired
ON,
the
reading of the
9
by the pickup hub causes only the
multiplier field to be read from the group multiplier card and only the multiplicand field to be
read from the detail cards. The basic wiring for
group multiplication is the same as for individual
multiplication. The additional wiring required is
shown dotted in Figure
3.
Product Overflow
Often the number of columns set aside on a
card to punch the product of a multiplication is
not
large enough to permit punching the largest
products. However, if
only a very small percentage of the total number of cards exceeds the capacity of the field, it may not be worth increasing
6
TYPE
603
ELECTRONIC MULTIPLIER
the size of the card field, thereby limiting the
number of columns available for other punching.
To take care of such cases, the
product overflow
feature is furnished on this machine.
To place the product overflow feature in
opera-
tion, the product counver position
next
to
the
highest order wired to punch is wired to the Product Overflow Entry hub. As long as the product
counter position wired to Product Overflow con-
tains
0, nothing happens. Any digit from 1
through 9 in this position causes the error light to
glow and the machine to stop when the card in
error is just
ready to enter the stacker.
Since
athe error light is used for other error signals, it is desirable to distinguish between errors
and overflow products. This is accomplished by
punching
a
12 hole in a card in which the product
exceeds the card field capacity. To punch the
12,
the Product Overflow 12 hub is wired to any
punch magnet, and
a
12 hole will be punched in
the overflow card.
With this arrangement,
a
card is examined after
an error light, and if
a
12 hole is punched, it is
apparent that the error was due to an overflow
product. This card can then be processed manually.
The necessary control
panel wiring is shown by
wiring
8
in Figure
3.
A 12 is punched in column
70 of the card with an overflow
product (Figure
3).
Product Summary
Often it is desired
TO
accumulate several products before punching. This permits special operations, such as
crossfooting on two cards, punching
the sum of several products, etc. To accomplish
this it is necessary to prevent reading out and clearing of the
product counter.
The product summary feature is
placed in opera-
tion by wiring the
PROD
SUM
control panel switch
ON.
This feature may be either No
X
or
X
controlled. Digit control is also provided. If
it is desired to punch the
product only in
X
Figure
4.
Use of Product Summary and Column Splits
I.
The multiplier field is entered in the multiplier counter.
2.
The multiplicand field is entered in the multiplicand
counter.
3.
The product is half corrected by adding
5
to the 3rd
position.
4.
The product is punched in card columns
54
-
60 sub-
ject to the Product Summary Wiring.
5.
The Product Summary feature is placed
in
operation.
6.
X
punchings in the cards to be punched with the
products are sensed.
7.
When an X-punched card passes the die, the product
is punched in the card and the counter is cleared.
8.
An X is punched in column 80 and an
0
in
column
I,
in all cards by means of the 0 and X and Column Split
features.
FUNCTIONAL PRINCIPLES
7
punched cards, the control panel wiring would be
as shown in Figure
4.
The standard wiring of the
multiplicand, multiplier, half-entry, etc. remains
the same as before. Of course, blank column check-
ing cannot be done in this case because many of
the cards are blank. However, double punch
checking may be used if desired.
If
ir
is
desired to punch in No X (or digit)
cards, the
N
hub is wired to the
PROD
SUM
hub,
below it. In this case X
punched cards will not
be punched, and the product will not clear.
A
blank card with the proper control punching
must precede a
product summary run to insure rhe
clearing of the
product counter provided a pro-
duct summary run is made immediately after turn-
ing the machine on. This card will insure the clearing of any random figures from the product
counter, resulting from turning the power on. On
normal runs this is not necessary because the product counter is cleared before starting the first
computation.
The
1 hub, which emits a 1 impulse during each
card cycle, may be used as a unit multiplier for
special
crossfooring operations from card to card
in connection with the product summary feature.
Column Splits
Two positions of column split are available as
standard on this machine.
The 0-9 hubs of the
column split are connected
wirh the C hubs from
9 through O of the card, and the 11-12 hubs are
connected with the
C
hubs from 11-12. This de-
vice permits an X or
12 punching over a 0-9 digit
to be recognized independently or to be ignored.
Wiring
8
on Figure 4 shows a typical use of the
column
splir device in connection with the O and
X
hubs for automatic punching of X's or 0's.
Punch Suppression
If it is desired to suppress punching in a card,
the card is either X punched or
punched with a
control digit in a specified column.
The control
punching
$hen causes the suppression of punching
on that card with proper wiring of rhe control
panel. No X (or digit) control is also furnished,
so that the control punching can appear on the
cards to be punched. This feature permits standard group (or interspersed) gang punching
oper-
a'tions on this machine.
The use of the punch suppression device in connection with two special features, the distributor
and a class selector, is shown in Figure
5.
In this
example, an offset interspersed gang
punching op-
eration is being performed.
If
No X (or digit) control is desired, rhe N hub
is wired to the
PCH
SUP
hub below. This setup
causes only X (or digit) punched cards to be
punched. Master cards would not be punched
with a control
punching.
Class Selectors (Optional)
Class selectors are optional features on rhis ma-
chine.
Two class selectors may be installed on
order.
The selectors are arranged for either X or
D
pickup and for normal or delayed operation. If it
is desired to transfer the selector during the cycle
following the reading of the X (or digit), the
GR
PLG
is wired to the Exit 1 hub. Wiring the
GR
PLG
to the Exit 2 hub causes the selector to transfer
during the second cycle following the reading of
the X or
D
hole.
Figure
5
shows an example of class selector 1
picking up from a 3 punch in column 80 and
transferring during the second cycle following the
reading of the
3.
Distributor (Optional)
A
conventional 12-segment distributor can be
installed as an optional feature on this machine.
The distributor can be used as a
digir emitter by
'
wiring the CB hub to the distributor C hub. The
individual hubs of the distributor then emit timed
impulses corresponding to rhe hub label.
The distributor can also be used as a digit selec-
tor by wiring from the brushes to the
C
hub of the
8
TYPE
603
ELECTRONIC MULTIPLIER
Figure 5. Use of Class Selector for Offset Gang Punching
1. A 3 in column 80 is sensed for punch suppression control and class selector pickup, using a digit selector.
2. Punching is suppressed as the 3-punched card passes
the die and stripper.
3. The class selector
I
transfers while the 3-punched card
passes the second set of brushes.
4.
The No X card following the 3-punched card is punched
in card columns 56
-
60 from columns 21 -25.
5.
All No
X
cards gang punch in columns 56 - 60.
distributor. Then only the desired digit in any
card column can be recognized by proper wiring
of the
9-12
hubs of the distributor. Figure 5 shows
the distributor used as a digit selector.
Sign Control (Optional)
Sign control permits determination of the sign
(plus or minus) of a product by analysis of the
sign of its factors. Two factors having the same
sign produce a positive product, but if one factor
is positive and the other negative, the product will
be negative.
For multiplication, factors should
always be
punched as true figures whether their sign is plus
or minus. In an IBM card the minus sign may be
indicated by an X punch.
Any column may be
used for the X punch, indicating the sign of a factor, but preferably it should be the unit column of
the factor field.
Wiring for sign control is shown in Figure
6.
To
place the sign control feature in operation, the
SIGN
CTRL
control panel switch must be wired
ON.
There are two sign control pickups on the control
panel, labelled
MCX
(multiplicand X) and
MPX
(multiplier X). One of the two common hubs,
MCX
should be wired from the column in Read for
Entry and Control containing the minus X for
,the multiplicand. One of the two common hubs
MPX
should then be wired similarly for the multi-
plier minus
X.
If only one of the sign control pickups reads an
X, the product should be negative. The machine
will punch the product in true figures, and to
designate the product as negative, will punch an
X in the units column of the product field. To
punch the minus
X
in the units column, the units
position of the product field must be wired from
the
PRX
PCH
hub to the Punch Entry hub (Fig-
ure
6).
The wiring is taken through the column
split to permit
X
punching over the units digit. If
it is desired, the negative product X may be
punched in any column of the card.
Sign Control Checking
When sign control is used in the original calcu-
lation, with a negative product indicated by an X
FUNCTIONAL PRINCIPLES
9
(All X's identifying negative amounts punched over
units position of corresponding field.)
The multiplier factor is entered in the multiplier counter; the units position is brought through the MPX
hubs to recognize negative multipliers.
The multiplicand factor is entered in the multiplicand
counter; the units position is brought through the MCX
hubs to recognize negative multiplicands.
The product is punched in card columns
74
-
80,
the
units position is taken through the column split to permit both
a
digit and the sign control X to be punched
in the units position.
An
X
identifying a negative product is punched in the
units position of the product field through the column
split.
An
X
identifying a negative product is read when sign
control checking only.
The product is checked for double punchings and blank
columns; the units position of the product is taken
through the column split hub because of the
X
punched
over the units position in negative products.
Figure
6.
Multiplication
and
Checking with Sign Control
punch, the punching of this minus X for the product may also be checked during the re-run. In
the re-run, the wiring of the two sign control pickups, as well as wiring of rhe entries to multiplicand
and multiplier counters, is reversed by the factor
reversal switch.
To check the minus
X
punch in
the product, the units column of the product field
from Read for Entry and Control is wired to one
of the two common hubs labelled
PRX
CHK
(Fig-
ure
6).
These
PRX
CHK
entry hubs should read an
X
when only one of the two sign control pickups
has read an
X.
If one of these conditions occurs
without the other, an error condition will be indi-
cated.
When checking
multiplication with sign con-
trol for double punching, the
X
punched for nega-
tive products must be
eliminared from the double
punch check, if it is punched over the product
field. This is done by means of the column split
device as shown in Figure
6.
MEC'H'ANICAL PRINCIPLES
A
STUDY
of the mechanical principles of this ma-
chine is limited to the read and punch
uni.t, because
the only mechanical units on the electronic unit
are the blowers. Only the
locarion of parts on the
electronic unit will be given in
,this section. The
read and
punch unit is essentially the same as a
gang summary punch, and covers are removed in
exactly the same manner.
Location
of
Parts
The five general views of the read and punch
unit in Figures 7 through
11
show the location of
all parts and units which are visible at a glance.
Certain other features not readily visible must be
illustrated schematically.
The front view (Figure
7)
shows the card lever
contacts which are mounted on a
plate at the front
of the machine. The contacts have been
placed
ourside for convenient access, although the card
levers remain in the same relative location as in the
gang summary punch. Also visible from the front
is the cam contact unit located directly under the
hopper and a portion of the rube power supply
chassis located on the lower base. The tube power
Check Brush
CLC
Entry
Brush
CLC
::nE:"
Die
CLC
/
Unit
Relay
Figure
7.
Read and Punch Unit-Front View
10
'
MECH.ANICAL
PRINCIPLES
'
11
Terminal Punch Magnet
80
Terminal Punch Magnet
10
Selt
(Full
Common Terminal
Motor Starting
/
Capacitor
Connector Latch
miurn Rectifier
Figure
8.
Read and Punch Unit-Right Side View
supply chassis extends across the entire depth of
the lower left section of the punch unit. The rest
of the power supply chassis-can be seen in Figures
10
and
11.
The right side view (Figure
8)
shows the punch
magnet terminal connections.
It will be observed
that these connections are the reverse of standard
gang punch connections. This is because the cards
are fed into
rhis machine face down, 9 edge first.
connector used on summary
punches.
To permit
access to the rear of the cable connector, the frame
on which the cable connector and the selenium
rectifier are mounted can be swung
down.if the
latch holding
%he frame in place is 'released. The
selenium rectifier shown below the cable connector
is a full-wave rectifier which supplies
40
volts'
D.
C.
in conjunction with the main transformer for the
operation of the relays and punch magnets
.in the
Relays are mounted on the right side only if the
punch unit.
The filrer capacitors for this rectifier
class selectors and sign control features are in- are mounted on the left side.
stalled. The cable connector which provides a
con- The left side view (Figure
9)
shows the cam
venient means of electrically connecting the read contact unit which is mounted under the card hopand punch unit ro the electronic unit is a standard
per.
There is space for
52
cam contacts in this
12
TYPE
603
ELECTRONIC MULTIPLIER
,Eccentric
Shah
P1
Cam Contact
P27
Cam Contact
200
mld
Capachrs
I140V
Supply)
Figure
9.
Reed and Punch Unit-Left Side View
unit, numbered from front
to
rear, top to bottom. bleeder resistor shown below the rectifier, supplies
However, no cams beyond P41 are used, and although cams
9,
13, and 15 are not used, they retain
their numbers. The
12 amp fuses and 20 amp
fusetrons shown at the top of the fuse
panel are in
the main transformer circuit. The glass fuses are
in the punch circuits and in the tube power supply
circuits. The conventional arc-suppressing capacitors are mounted between the relay brackets and
above the half-wave selenium rectifier.
This selen-
ium rectifier, together with its filter capacitors and
140 volts
D.
C.
for the read-out power tubes in
the electronic unit. The four
2000 mfd. capacitors
shown below the 140 volt
D.
C.
selenium rectifier
are the filter capacitors for the 40 volt
D.
C.
sup-
ply. The double
punch and blank column detection
relays
37
through 57 are mounted on the left rear
gate. If
10 additional positions of DPBC detection
are installed, relays
58
through 77 are mounted
just to the left of
R37-R57.
a
MECHANICAL PRINCIPLES
13
Oil
Level
Crank
Stud Indicator
Gear Housing
\
Oil
Cup
Figure 10. Read and Punch Unit-Rear View
The rear view in Figure 10 shows the mechan-
ical features visible from the rear as well as the
main transformer and the tube power
supply chas-
sis. The tube power
supply furnishes
D.
C.
volt-
ages of
100
volts, 150 volrs, and 250 volts for the
operation of tubes in the electronic unit. The main
transformer supplies
A.
C.
of proper voltage to the
40 volt and 140 volt selenium rectifiers; it also
supplies -the filaments of all tubes except the gasfilled rectifier tubes.
In the close-up view of Figure
11,
a better pic-
ture of
-the main transformer and tube power sup-
ply chassis is shown. Note
particularly the system
for numbering terminals on both the transformer
and the power
supply chassis. The
EL-3C
and
EL-1C
tubes are gas-filled full-wave rectifier tubes,
14
TYPE
603
ELECTR0,NIC
MULTIPLIER
Figure
11.
Closeup
of
Main Transformer and Tube Power Supply Chassis
while
all
other tubes on the chassis are vacuum
tubes of the type indicated. All other components
in the power
supply circuit are mounted under-
neath the chassis.
Figure
12
shows schematically the location of
the brush assemblies, die and stripper assembly, and
the card levers. The hopper card lever is located
directly under the hopper. The punch brush
1
and die card levers are mounted
oh
the front side
frame. The former is located directly under the
punch brush
I
con,tact roll while the latter is
located under the second set of feed rolls. The
punch brush
2 card lever is mounted directly
above the punch brush
2
contact roll.
However,
all
card lever contacts are mounted on a plate at
the front of the machine as shown in Figure
7.
Note from Figures 13, 14 and
15
that there are
no
relays in the electronic unit. Figure
15
shows
the gates open with all connections accessible. The
general function of each tube chassis is given, but
no effort will be made to discuss these further un-
til the section on
Electrical Principles.
The switch
and push buttons shown in Figure
15
are not ac-
cessible unless the large gate is open; rhey are in-
tended solely as an aid in servicing the unit. The
blowers shown are provided to cool the tubes. Over
1200
watts of heat from the filaments alone must
be dissipated. One blower is provided for each
side of the electronic unit.
MECHANICAL PRINCIPLES
15
Hopper
i
Reading Station
I
Punching Station
I
I
(Entry and Control)
I
I
Reading Station
(Checking and
I
Gang Punching)
I
Figure 12. Schematic of Read and Punch Unit
Stacker
H
Chassis
M
Chassis
N
Chassis
Product
Input lnverlers
Figure
13.
Electronic Computing Unit-Front View
16
TYPE
603
ELECTRONIC MULTIPLIER
Figure
14.
Electronic Computing Unit-Rear
View
Drive
Mechanism
Power to drive the read and punch unit is furnished by the drive motor which can be seen in
Figure
8.
The drive motor transmits power to the
gear housing through a
V
belt and pulley. Practically all mechanisms are under control of the
punch clutch. All feeding operations are under
further control of the intermittent feed clutch
(geneva clutch). Figure
16
shows schmatically
the various units under the control of the two
clutches. When only the motor operates (with
neither clutch engaged), the drive pulley rotates
and drives the drive
pulley shaft to which the pul-
power with this cover off because oil will be
thrown out of the housing.) The eccentric shaft
drive gear operates the eccentric shaft which in
turn transmits motion to the punch bail. (The
operation of this bail is discussed in connection
with the principle of punching.) The geneva drive
gear operates the geneva and geneva
pawl and also
the
punch clutch idler gear and shaft. On the outside of the idler gear shaft is pinned a small gear
which drives the index gear (Figure
10).
The
punch clutch one-tooth ratchet is a part of the index gear assembly and rotates continuously as long
as the motor is in operation. The index gear and
ratchet rotate on the
~unch clutch shaft but are
ley is keyed.
Attached to this shaft inside the
not pinned to it.
gear housing are two gears, the geneva drive gear
In order to place the rest of -the machine units
and the eccentric shaft drive gear.
The mechan-
in operation it is necessary to unlatch the punch
isms and gear trains inside the gear housing may clutch pawl from its armature and allow it to en-
be
seen by removing the top cover from the hous-
gage in the continuously running one-~00th ratch-
ing. (Caution: Do not operate the machine under et, which is a part of the index gear assembly.
MECHANICAL PRINCIPLES
17
-
I
I
\
X
Chassis
Gate
Stop
Rod
Cover
&i&
Block
All
Chassis
Terminals are
marked
Figure
15.
Electronic Computing Unit with Gates Open
Punch
Brushes
2
Eccentric Shaft
Figure
16.
Schematic of Drive Mechanism
When the clutch pawl engages the one-tooth
ratchet, the punch clutch shaft
.turns with the
ratchet. The gear mounted on the outside end of
the punch clutch shaft in turn drives the P-cam
shaft, on which are mounted the P-cams. Within
the gear housing there are two sets of complementary cams pinned to the punch clutch shaft.
One set of cams operates the feed knives and the
other set
conrrols the engaging of the geneva clutch
pawl with its ratchet. The geneva ratchet is nor-
mally stationary; but when rhe geneva pawl engages with it, the ratchet is driven
by the geneva,
which imparts an intermittent motion
TO
this
ratchet. Riveted to the geneva ratchet is the
ratchet gear which serves as the drive gear for all
feed rolls, contact rolls, and the stacker roll. Since
all these rolls are driven from the geneva, they all
turn intermittently. The
intermirtent movement
is
nccessary to have the card in a stationary posi-
tion while
punching.
(This is discussed in more
detail in the section on the
Gerzeva Mechanis-in.)
Only the upper feed rolls and the punch brush
2
contact roll are driven from the gear train
in the housing.
The lower feed rolls are driven
by their corresponding upper rolls through gears
at the front of ,the machine. Also, the punch brush
1
contact roll is driven from the first upper feed
roll. The stacker roll is
driven by a gear train from
the last feed roll.
Punch Clutch
The punch clutch shown in Figure
17
is of the
one-tooth rarchet type commonly used on EAM
equipment, and its operation should be thoroughly
understood.
The principal parts of the clutch are
a
continuously running one-tooth ratchet, a clurch
pawl, a latching mechanism, and a magnet.
The
magnet provides a means of electrically controlling
the operation of the clutch.
The clutch magnet
armature serves as the latching mechanism to latch
the pawl and keep it from engaging in the ratchet.
When rhe magnet is energized, the armature is
attracted and the pawl is released or unlatched. The
One
Tooth Ratchet
Figure
17.
Clutch
pawl spring causes the pawl to pivot in a clockwise
direction and engage the one-tooth ratchet when
the
r2tchet tooth reaches the pawl. The pawl pivots
on a stud riveted to the clutch pawl arm which is
pinned to the punch clutch shaft. Thus, when the
pawl turns with the ratchet, the shaft must also
turn. Once the pawl is unlatched, it must make
one complete revolution before it can be relatched
since there is but one
la~chin~ point. For this reason it is necessary to keep the armature attracted
only
!ong enough to allow the pawl to engage the
ratchet. When rhe pawl reaches the end of its
cycle, the armature has been returned to its normal
position by the return spring and the tail of the
strikes the armature, causing the pawl to be
cammed out of mesh with the one-tooth ratchet.
When the pawl has been cammed out of mesh, the
keeper drops behind the clutch pawl arm and prevents ,the clutch shaft from turning backward.
Without the keeper, the shaft might turn back-
ward because of the rebound; then the pawl would
drop against the ratchet and
ca,tch on the tooth
once each cycle. This nipping action has a tendency to round off the edge of the one-tooth
ratchet. This is objectionable because a rounded
edge on the
ra~chet tooth may cause the pawl to
pull out of mesh under load.
MECHANICAL PRINCIPLES 19
of four bearing shoes held against the feed roll
shaft
by compression springs.
The card also passes two sets of brushes and contact rolls. The brush assemblies are identical except for minor constructional differences; each
consists of 80 individual brushes mounted in a
brush holder so that they are insulated from each
other. The contact rolls are made of beryllium
copper and are geared to turn at a higher speed
than the feed rolls to provide a
wiping action by
the card.
Index and Cycles
In explaining machine operations it is necessary
to make reference to one operation in terms of
Figure
18.
Punch Feed Knife Drive
another.
By using an index for a common refer-
ence point this becomes possible.
The index gear
Feeding Mechanisms
serves as the common reference for all machine
The purpose of the feed knives is to feed one
operations.
One complete revolution of the index
card through the throat into the first set of feed
gear is called one cycle. If the punch clutch is en-
rolls for each revolution of the punch clutch. The
gaged,
a card would move from the first set of
feed rolls rhen carry the card past the brush sta-
brushes to a corresponding posirion at the die (or
tions and punching station to the stacker.
The
from the die to the second set of brushes) during
knives are driven back and forth by gear sectors
one revolution
of
the index.
For convenience in
which mesh with the feed knife racks. The gear
measurement one cycle is divided into units called
sectors are pinned to a shaft which oscillates under
cycle
poifzts.
The most logical unit of division is
the
control of a complementary cam and follower
the distance between successive
punching positions
mounted on the punch clutch shaft (Figure 18).
on the card.
Therefore the distance from the
9
When a card is fed from the magazine, it is fed
punching position to the 8 punching position in
between the first pair of feed rolls. The feed rolls
one card represents one cycle point, while the dis-
operate intermittently, hence
rhey will be station-
tance from the
9
punching posirion in one card to
ary during a portion of the time a card is being fed
the
9
punching position in the following card is
between them.
The feed knife carries a card up
one cycle.
to the first feed rolls while the feed rolls are
sta-
There are 12 punching positions on the card.
rionary.
To insure that the first feed rolls will
Each
punching position is
j/4
inch from the next,
pick up the card, the knife buckles the card slight-
therefore, for each cycle point the card moves
'/4
ly just before the feed rolls start turning.
inch on its
path rhrough the machine.
Since the
As indicated in Figure 16, the card passes card is 3
'/4
inches wide, it requires 13 cycle points
through four sets of feed rolls on its way to
he
to advance a card past any given point.
In this
stacker. The upper feed rolls are mounted in fixed
machine there is
'/4
inch between cards, therefore,
bearings while the lower feed rolls are provided
the cycle consists of 14 cycle points. The teeth on
with
pivoted bearings to allow separation of the
the index gear are used for further subdivisions.
rolls when a card is fed between them.
Feed roll
Thus a timing given as 14.1 indicates one tooth
tension is provided by a pressure bracket consisting
past the
14
index mark.
20
TYPE
603
ELECTRONIC MULTIPLIER
Geneva Mechanism
As indicated previously, the feed rolls in this
machine
operare intermittently to allow punching
of the card. The card must not be in motion while
the are being driven through the card and
withdrawn.
If the card is moving, the holes will
not be clean cut, but ragged and torn.
Since the
card
musr be standing still while it is punched,
then moved to
a
new punching position fourteen
times each cycle, the motion is necessarily
inter-
Gene
mittent. This intermittent mo.tion is obtained by
means of a geneva mechanism.
The geneva drive gear is located just inside the
gear housing and pinned to the pulley shaft.
A
stud and roller fastened ro this gear operate in the
slots of the driven member of the geneva gear
(Figure
19).
The hub of the geneva drive gear is a cam sur-
face for approximately two-thirds of its periphery.
This cam surface holds the feed rolls in a station-
ary position during punching time by locking rhe
geneva in position.
The geneva disc has seven deep slots and seven
shallow cuts in it. The roller of the drive gear
operates in the deep cuts in the geneva disc and
the cam surface rides in the shallow cuts. As the
drive roller leaves the deep cut of the geneva disc,
Figure
19.
Geneva Mechanism
Figure
20.
Geneva
Pawl
and Ratchet
the cam surface turns into the low cut and stops
the geneva disc from turning and
holding it until
the drive gear has rotated to a point where
the
drive roller enters the next deep slot of the geneva
disc and starts driving.
Then the cam surface has
turned to
a
poinr where it releases the disc and al-
lows it to turn freely.
The geneva disc turns con-
tinuously as long as the drive motor runs. However, no motion is transmitted to the feed rolls until the geneva pawl is engaged with its one-tooth
ratchet. The geneva
pawl is pinned
to
the same
shaft
as
the geneva disc.
This shaft runs through
the hub of the one-tooth ratchet and gear. The
one-toorh ratchet is free on the shaft and does not
turn unless the geneva pawl is engaged. The onetooth ratchet gear (Figure
20)
is meshed with the
feed roll drive gears.
Pawl Disengaging Roller
Geneva Pawl
Figure
21.
Pawl
Disengaging
Roller
MECHANICAL PRINCIPLES
2
1
When the punch clutch is not engaged, the gen-
eva
awl
rides on the surface of the one-tooth
ratchet during the greater part of the
cycle. When
the
pawl reaches a point opposite the single tooth,
the tail of the
pawl strikes the
awl
disengaging
roller (Figure
2 1
)
and is cammed away from the
ratchet until it has moved past the point where
it may engage in the single tooth of the ratcher.
By cranking the machine by hand, it can be noted
how the pawl disengaging roller prevents the
ge-
neva pawl from engaging.
From the above, it is
evident that the operation of the geneva
pawl is
controlled by the pawl disengaging roller. The
pawl disengaging roller is mounted on a
triangular
plate (Figure
21)
which is free to pivot on the
latch cam roller arm. The latch cam roller arm
(Figure
22)
is operated by the latch cam which
turns only when the punch clutch is engaged.
When the punch clutch is engaged, the latch cam
turns, causing the latch cam arm to rotate in a
counterclockwise direction. As the latch cam arm
rotates, the upper end moves to the left and down
allowing the
pawl disengaging roller to move past
the single revolution timing cam and the geneva
pawl to engage in the one-tooth ratchet.
As the
cycle is completed, the larch cam causes the latch
cam arm to rotate in a clockwise direction carrying
the pawl disengaging roller to the right. The roller strikes the tail of the geneva
pawl and disen-
gages the pawl from
the one-tooth ratchet when
the roller is backed
by the single revolution timing
cam.
Single Revolution Timing Cam
The geneva disc has 7 cuts in it and moves the
card one
cycle point for each cut. The machine is
a fourteen point cycle machine; therefore, the
geneva disc must make two revolutions per machine
cycle, which means that the geneva pawl will pass
the
pawl disengaging roller twice during each
cycle. The purpose of the single revolution timing
cam is to prevent the feed rolls from stopping in
a position half way through a cycle.
At the end
of the first revolution of the geneva pawl and disc,
the flat side of the single revolution timing cam
should be down (Figure
22).
The pawl disengaging roller is free to swing away from the tail of
the pawl. Therefore, in case the geneva
pawl had
become disengaged from the one-tooth ratchet, it
would be free to drop into the one-tooth ratchet
on the next revolution to complete the
cycle. This
assures that the
pawl will not be disengaged by the
pawl disengaging roller until the punch unit mechansim has reached its proper latching position and
that the geneva makes two revolutions for each
cycle.
Principle of Punching
The mechanism for punching holes consists of
80 individual punches, each controlled by an in-
/
Cam Assembly
terposer, 80 punch magnets together with arma-
Geneva Pawl
Disengaging Roller
(Pinned to this
shaft which carries
tures and pull wires to control the 80 interposers,
punch clutch pawl)
.
a punch bail to drive the punches through the card,
and the eccentric drive shaft and links to operate
L\-<j
the punch bail. There is a separate punch for each
card column; any one punch may be required to
punch any hole from
9
to
12.
The cards feed in
,
I
',
'/
.___.',
9
edge first, and every 9 to be punched is punched
:
0
at the same time. The card then moves to the
8
Figure
22.
Single Revolution Timing Cam
position and every 8 is punched, and so on until
22
TYPE
603
ELECTRONIC MULTIPLIER
Eccentric Shaft
R
I
Magnet
Pull
Wire
Stripper
Die
A
Figure
23.
PI
all positions have been punched. Thus, all possible
~unchin~ is done in 12 cycle points.
As
previously mentioned, the eccentric shaft
operates continuously as
long as the drive motor is
in operation. The purpose of the eccentric shaft
is to convert the rotary motion of the shaft to the
reciprocating motion necessary to operate the
punch bail. The punch bail operates up and down
once for each cycle point. This up and down
morion is imparted to the punch bail through the
punch bail connecting links (Figure
23A). When
the magnet is de-energized, the punch bail may
move up and down without contacting the interposer; therefore no
punching takes place. When
the
punch magnet is energized, its armature is at-
tracted, and through the
pull wire the correspond-
ing punch interposer
is
pulled into engagement
with the punch bail tongue. Since the punch bail
tongue operates up and down, it carries the punch
interposer and the
punch connected to it down
through the cards. On the return stroke, the punch
is positively withdrawn from the card by the action of the
punch bail. The purpose of the knockoff bar is to disengage the interposer from the
punch bail tongue. As the interposer is returned
Magnet
lnte
B
.inciple
of
Punching
to its normal position, the upper rounded edge
strikes the knockoff bar and the interposer is
cam-
med away from the punch bail tongue.
The in-
terposer spring then holds the interposer in normal
.
position.
Figure 23A shows the bail in its upward position before the interposer is moved under it, while
Figure 23B shows the bail driving the
punch down
through the die.
Magnet Unit
The punch magnet unit consists of 80 punch
magnets along with their armatures and pull wires,
80 interposers and
punches, the die and stripper
assembly, and the
punch bail assembly. Figure 24
shows a magnet unit with the punch bail and interposer knockoff bar removed. The insert shows
a
closeup of the interposers.
There are three types of interposer and
punch
assemblies.
One type is used in the first column,
another type in the 80th column, and yet another
is used in all columns from
2 to
79
inclusive. The
interposer used in the
1st column is provided with
a
long stud for the eye of the magnet ~ull wire to
prevent it from slipping off the stud. The
inter-
MECHANICAL PRINCIPLES
23
Figure
24.
Magnet Unit
poser used in column
80
is attached to the punch
Column
2-79
Column
1
to prevent the interposer from slipping off.
The
orher interposers, being protected on both sides by
other interposers, do not require any such precau-
tionary design. The three types are shown in Figure
25.
The interposers are being referred to as
they are located in
rhe Type
603
punch unit and
not according to the column of the card as placed
in the gang punches. It can be seen that the interposers have been relieved to
provide a space be-
tween them. This space
prevefits the interposers
from sticking together and causing extra holes to
be punched. The interposers should be kept free
of foreign particles and gumming oil.
The magnet unit is held in place by four mount-
ing screws and located
by two aligning screws
(Figure
26).
The adjusting screws at the left end
locate the unit in a vertical position at the left
end to permit proper fit of the die assembly. The
Column
\
Figure
25.
Three Types
of
Interposers
24
TYPE
603
ELECTRONIC MULTIPLIER
C
Figure
26.
Magnet
Unit-Top
View
aligning screws locate the magnet unir laterally to
Cam
Contacts
permit adjustment of the vertical punching regis-
The cam contact used on this machine makes it
tration.
possible to obtain any desired duration of contact
Oil
Pump
ranging from a fraction of a cycle point to a com-
The oil pump is a simple rotary-vane type pump.
It is located inside the gear housing on the shaft
of the small gear which drives the index. It pumps
the oil from the bottom of the gear housing to the
top where it is free to run down over the geneva
and gears.
The rotor is
pivoted off-center in the housing
as shown in Figure
27.
The expansion chamber at
the inlet provides
a
vacuum and causes the oil to
enter the pump from the well below. The compression chamber at rhe outlet causes oil to be
forced out at the top.
plete cycle.
This contact is available in two styles, latching
and non-latching
The latching style (Figure
28)
is used for all contacts that operate bur once during a
cycle; the non-latching style is used for contacts which must make more than once during a
cycle, such as the circuit breakers which feed impulses to rhe brushes for reading the card.
All contacts are closed by a lobe on a bronze
cam which operates against the contact plunger
and carries it beyond the latching point so that the
conract latch lever may support the contact
MECHANICAL PRINCIPLES
25
Outlet
-
=-%
Rotor Pump
Compression
Chamber
Pump Vane
Latching
pump housing
\\
Vanes
Contact
Plunger
hQ
Figure
27.
Oil Pump Dismantled
Figure
28.
Contact Cam
Rotor
Pump Luring
\-
Expansion
Chamber
li
Inlet
plunger.
The unlatching cam may be adjusted to
any position with respect to the
periphery of the
bronze cam. This cam strikes the contact latch
lever and unlatches rhe contact plunger. In this
manner the contact duration may be adjusted.
There is a maximum of
41
cams and contacts
mounted in the P-cam unit numbered from front
to rear,
rop to bottom. Cams
9,
13,
and
1F
are
not used but they retain their number. The P-cam
unit is arranged so that it can be swung to the left
to permit access ro the underside of the cams and
contacts. Care must be exercised when remeshing
the unit to see that the unit is installed in time with
the index.
ADJUSTMENTS
PUNCH CLUTCH
1.
Mount clutch assembly with rhe three
mounting screws.
2.
Set clutch latch stop screw for
.OIj"
clearance between the latch point and the tail of the
clurch pawl when the latch is against the latch stop
screw. Move magnets if necessary (Figure
29).
3.
Set the clutch latch backstop screw for
&"
overlap of the latch over the tail of the pawl (Figure
30).
4.
Move the magnet coil mounting plate in
elongated holes to provide for
.008"
to
.OIO"
clear-
ance between the armature and cores when the
latch is against the stop screw (Figure
3 I
)
.
.
There should be
.003"
clearance between
the keeper and the clutch
pawl arm when the pawl
is latched. This is obtained
by s,toning or peening
the keeper (Figure
32).
If
the clutch is removed
and replaced, the clutch plate should be mounted
with the mounting screws in the center of the
oversize holes. The other adjustments should
then be checked.
Figure
29.
Clutch Adjustment
w
Figure
30.
Clutch Adjustment
Clutch
Pawl
Arm
Magnet Coil
Mounting
Plate
Figure
31.
Clutch Adjustment
ADJUSTMENTS
27
Clutch
Figure
32.
Clutch Adjustment
GENEVA MECHANISM
Single Revolution Timing Cam
This cam should be timed so that the flat side
of the cam is up and in a horizontal position at
one
,tooth past 14 on the index.
Move the cam
and gear out on the shaft far enough to unmesh
the teeth and
remesh for above condition (Figure
33).
The geneva pawl should be engaged when
checking this adjustment.
Single Revolution Timing Cam Bracket
Loosen the holding screw.
Move the bracket
up or down until the cam holds the geneva
pawl
Single Revolution
Timing
Cam
/
~eneva Pawl
Figure
33.
Single Revolution Timing Cam and Bracket
Adjustment
disengaging roller in position to hold the geneva
pawl away frmi the one-tooth ratchet on the geneva gear when the punch clutch is latched up at
D
and the machine is operated by hand. Only the
raised portion of the tail of the pawl should operate
against the roller.
Set bracket so that geneva clutch pawl just nips
the one-tooth ratchet. Then move the left end of
the bracket
up until the pawl just clears the one-
tooth ratchet.
An approximate adjustment may be obtained by
setting the bracket so that the locking screw is
about one-third of rhe way up from the bottom of
the elongated hole (Figure
3
3
) .
Geneva Clutch
A two-pawl geneva clutch is used with the 603
punch unit.
This assembly is more positive in its
action than the single-pawl type, and is interchangeable with the geneva clutch assembly on
all high-speed
punches.
Two pawls, one a driving pawl and the other a
detenting
pawl, are used to eliminate the critical
knockout timing of the clutch. The
pawl disengaging roller is not adjustable because it is pinned
to the bracket.
PUNCH UNIT
Belt Tension
The belt tension is adjustable by moving the
motor up or down on its
pedestal. The belt should
be adjusted for enough tension to prevent
slippage;
excessive tension, however, will cause the motor
bearings to overheat and should be avoided.
Feed Roll Tension
Feed roll tension is determined by compression
springs in the feed roll pressure bracket. If for
any reason any pressure spring in any one bracket
is replaced, all springs in the bracket should be re-
placed to provide for even tension. The pressure
bracket is equipped with holes tapped for 5-40
screws which may be used to lock the pressure shoe
under spring tension before removing the bracket.
28
TYPE
603
ELECTR
ONlC MULTIPLIER
Figure
34.
Feed
Hopper
Feed Knife Guides
The adjustable guides for the feed knife racks
should be positioned for a minimum of play of the
racks between the guides without causing any
binds. Check this over the entire length of the
stroke (Figure
34).
Feed Knife Projection
The card feed knives are adjustable and should
be set evenly on each side for a projection of
.004"
to
.0045".
To adjust a feed knife, remove the
feed knife block from the feed knife rack. This
may be done without changing the left-to-right
position of the knife block if the knife holder adjusting screw is held with
a
i"
open end wrench
while the knife block holding screw is removed.
A Go-No Go gauge is
~rovided with two accur-
ately ground surfaces at opposite ends.
The sur-
face marked "Go" is cut
.0045"
deep and the
opposite surface marked "No Go" is cut
.004"
deep. When the feed knives are adjusted, the locking screws should be loosened until they are just
snug enough to prevent the knives from moving
with a slight pressure. Turn either in or out on
the adjusting screws, as required, to raise or lower
the feed knife. If the knife is
properly adjusted,
the "Go" end of the gauge will pass over the feed
knife on either side and the "No Go" end will not
pass over the projection.
The
ltnife must project evenly all the way across,
and both knives must be adjusted for the same
condition.
After a period of use, the knife blocks may
wear,
particularly at a point near the feed knife,
in which case accurate adjustment is not obtain-
able. If such a condition exists, it should be remedied in the following manner before the above
adjusrment is
performed.
Set feed knife for the slightest possible projec-
tion above the knife block. Stone knife and block
assembly until the top surface of the knifen and
block are even. Then proceed as above.
Feed Knife Block
The knife block pivot screws should be adjusted
so that at the extreme forward stroke
(2%
teeth
Cj
before
1
) the feed knife edge travels
.OI
5"
past
the edge of the card with the card against the first
feed rolls. This provides the proper buckle of the
card to insure good feeding (Figure
3
$A).
Hopper Guide Posts
I
The hopper guide posts are positioned by means
I
of shims so that there is at least
.OIOf'
clearance
over the width of the card. Also check to see that
at their extreme left position, the feed knives travel at least
.03OU
beyond the edge of a card held
against the guide posts.
Throat
With the throat knife and block only snug
under their holding screws,
place a
.O1OW
feeler
gauge into the throat opening. This gauge must
be separated from all other gauges.
Adjust the
throat block and throat knife until the
.OIO"
gauge
between the knife and the roller is
parallel with
the card feed knife slides. This can be determined
ti@
ADJUSTMENTS
29
Vert~cal center
line
of
A B
Figure
35.
Feed Adjustments
by laying a straight edge across the feed knife die assembly used on high-speed punches, are for
slides.
The
.OlO" gauge should just touch the
the purpose of keeping the die level when placing
under edge of the straight edge, which can be a
it in position in the machine. The following m'eth-
IBR4 card on edge or a scale (Figure 3 5B).
This will insure that the face of the throat knife
is in the correct relationship with the roller, which
is important if the throat adjustment is to be realized. Secure the throat block in
,this position, keep-
ing it to one side in its locating channel. This will
keep the axis of the roller
parallel with the face
of the throat knife.
Adiust the throat knife so
od should be used to adjust these guides correctly.
If these guides are properly adjusted, it will be
easier
,to remove and replace the die in the machine.
Install the die assembly in the machine with the
angle guide loose, and lock the die in place. Each
angle should then be
pressed lightly toward the
side frame and tightened in this position. An ex-
that the
.O1O,,
gauge
en;ers
the throat opening
cessive amounr of pressure must not be placed on
freely when held parallel to the card line. Tighten
these guides, or they will bind and make it difficult
the rhroat knife and check to see that a .OlO"
to remove the die.
gauge will not enter the throat from any angle
Punch Bail Tongue
other than the horizontal card line. Under no con-
dition should the
.O11" gauge enter the throat.
The punch bail tongue should be adjusted so
that it is 2-17/32" from the front edge of the
Die
tongue to the back of the punch bail pivot shaft
The stop studs in the stripper must be main-
(Figure 36). Loosen the four holding screws.
tained snugly against the die. These studs provide
Then position the tongue in relation to the bail by
minimum clearance between rhe die and the strip-
means of the two adjusting screws.
This
should
per.
Adjust this
by positioning the left end of
not require adjustment unless a punch bail or
the magnet unit assembly up or down by means of
punch bail tongue is replaced.
the adjusting screws, one beside each of the two
Interposer Pawl Lock Bar and Spring Bail
clamping screws at the left end of the unit. Remove and replace the die several times to be sure
The
inrerposer pawl lock bar is positioned and
that the latching bars have a
slight drag as they
pinned at the factory so there is a minimum
clear-
enter and leave the castings.
ance between the interposers and the punch bail
The two angle guides, one on each end of
rhe
tongue when the interposers are engaged with the
TYPE
603
ELECTRONIC MULTIPLIER
Bend
here
Fi&re
36.
Punch Bail Tonkue Adjustment
punch bail tongue and are driven
TO
their extreme
downward limit
by -the punch bail.
The interposer spring bail should be
positioned
so that it does not touch the interposers when the
interposers are engaged with the punch bail
'tongue and are driven to their extreme downward
limit. Check several interposers at each end.
Punch Magnet Armatures
The magnet armatures should be adjusted so that
the interposers will move
li8"
toward the magnets
when the armatures are
a~tracted. The
1/8"
tra-
vel is obtained by increasing or decreasing the
armature air gap
by bending the armature just
above the point where the
pull rod connects (Fig-
ure 37). The
inrerposers should line up when in
a normal position and should move freely.
Punch Bail Connecting Links
There must be a perceptible movement between
the punch bail tongue and the interposers when
the bail is in its uppermost position. This condi-
tion
prevenn binds and also provides for a mini-
mum travel of the
punches into the die. Proceed
as follows:
1.
Remove the front punch bail connecting
link pin.
2.
Turn the machine until the punch bail is in
the extreme upward position (eccentric up).
Figure
37.
Punch Magnet Armature Adjustment
3.
Adjust the rear punch bail connecting link
adjusting screw for
a
slight clearance (.003") be-
tween the punch bail tongue and the interposers.
(Figure 36). If there is any variation in the clear-
ance from one end to the other, the
.003" clearance should apply to the closest end. This may be
checked with a leaf gauge or
by moving the inter-
posers.
4.
Adjust the front connecting link adjusting
screw so that the front punch bail connecting pin
'
will slide freely into posi.tion in the punch bail and
punch bail connecting link. This assures an even
adjustment on both connecting links and eliminates strain on the punch bail.
After adjusting the connecting link adjusting
screws, check to see rhat the punches are not
jammed down against the
punch stop bar. Check
in the following manner:
1.
Engage the interposer at each end of the
punch bail and turn the machine until the punch
bail is at its exrreme downward limit of travel.
2.
Press on the top of the interposer with a
screwdriver and check for a slight movement.
The
punch stop bar should be set as near the
punch as possible without inrerfering with the
movement of the punch.
ADJUSTMENTS
31
Punch Hopper Side Plates
The punch hopper front and rear side plates are
adjustable forward or backward so that cards
are punched in proper horizontal alignment on the
card registration gauge.
Adju~ment is provided
by means of an elongation of the holes provided
for the mounting screws. Shifting the side plates
should result in the
punched holes lining up with
the grooves in the first upper feed roll. The hopper
side plates should then be located for a minimum
clearance over the
length of the cards.
Punch Brush Lateral Alignment
The punch brushes should be positioned so that
the brush strands will track
.through the center of
the holes in the card. This may be visually checked
if a deposit of some soft substance, such as wax
crayon or carpenter's chalk, is
placed across the
O
1
m
and 9 edges of the printed surface of a card
punched 9,
0, 9, 0, ac., and the card is run
through the machine. The brush tracks will be
plainly visible on the chalked surface of the card.
The brush holders and separators can be moved to
the front or back if the three locking screws in
the slide assembly are loosened (Figure 38).
Scribed
Lin
Ad-,--'
Holding Screws
Figure
38.
Brush
Slide Assembly
Common
Brush
Anchor
Slide Adjusting
Screw
Figure
39.
Anchor
Slide
Adjustment
Anchor Slide Adjustment
The clearance between the conract roll and the
brush separators must be
.O12" to .018".
Adjust
the brush slide unit up or down by means of the
anchor slide adjusting screws in the front and rear
support castings (Figure 39) for gang
punch and
check brushes.
The entry and
conrrol brushes are not provided
with an anchor slide adjustment.
In order to ob-
tain the proper clearance between contact roll and
brush separators of the first set of brushes, it is
necessary
.to shim the card guide plate or to grind
away the card guide support. Because of the design
of the original punch, the location of the read for
entry and control brushes require a special brush
block assembly.
Brush Timing
Both sets of brushes are set in their holder so that
the heels of the brushes are aligned wirh the scribed
line (Figure
3
8).
The brush projection should be
sf'
above the separators (Figure 40) which re-
quires a measurement of
$$"
from the brush block
32
TYPE
603
ELECTRONIC MULTIPLIER
I"
Holding
Screw
Figure
40.
Brush Projection
to the toe of the brush. Adjust the brush holders
in their assembly by loosening the holding screws
shown in Figure
40
so that the brushes make
through the holes in the card
3/4
to
j/2
tooth before
the line of index (check with a test light).
Vertical Registration
Cards must be punched in proper vertical align-
ment.
Registra~ion should be checked with a card
gauge. To
change the registration, loosen the
4
magnet unit mounting screws and adjust the
2
magnet unit &ping screws to position The magnet
unit assembly toward the right or left, for proper
vertical registration of the holes
punched in the
card. (Move the
vertical registration aligning
screws evenly and only when the magnet unit
holding screws are loosened; otherwise, the unit
may be strained and incorrect horizontal registration may result.) Be sure the
aligning screws and
holding screws are tight after making this adjustment.
After repositioning the magnet unit for the pro-
per vertical alignment as in the above adjustment,
recheck for the
slight clearance between the punch
bail tongue and the interposers when the punch
bail
is
at the top of its stroke, because repositioning
the magnet unit will
aflect tha~ adjustment.
Stacker Plate
If the stacker has a flat plate (no indentations
under rubber rollers), adjust the stop
nu.t on the
rod in the bottom of the stacker tube so that there
is
.006"
to
.OlO"
clearance from the sracker plate
to the face of the rubber rollers when the rubber
rollers are in
The extreme downward position. The
felt washer serves as a brake to prevent the stacker
from returning to its upper position too fast when
cards are removed. The braking action is adjusted
by compressing the felt washer
by means of an
adjus'ting nut.
,,
Stacker Timing
To time the stacker roll:
I.
Remove the blue steel clip from the stud of
the idler gear.
2.
Disengage the idler gear.
3.
Engage the punch clutch and turn the
machine to
3
on the index.
4.
Turn 'the stacker roll so that the high side
is down and
remesh the idler gear. This timing
should cause the card to be carried to within
A''
of the right side of the stacker.
Form
here
-
A
Figure
41.
Circuit Breaker Adjustment
ADJUSTMENTS
33
CONTACTS
Adjustment of Cam Contacts
1.
The lower contact strap should be formed
at point
A
(Figure 41) to provide proper tension.
At the factory these straps are adjusted so that a
force of 160
plus or minus 10 grams (approxi-
mately 6
02.) applied at the tip of the lower strap,
point
B,
will just close the points. This tension
must be maintained accurately to avoid
a
bouncing
condition.
2.
Place shims beneath the plunger stop plate
as required to obtain .040" to
.O5O"
travel of the
plunger before latching up occurs.
If the contact
plunger is overlapped by the latch by an amount
equal to the thickness of the latch metal, this
should
provide the .040" to
.O5OU
travel (Figure
42).
Place
5
:
shims
here
,040".
,050
Latch
Figure
42.
Circuit Breaker Adjustment
3.
Place shims between the lower contact ter-
minal block and the contact strap to provide
.O15"
to .018" air gap between the contact points (Figure 43).
4. Check to be sure that the
plunger does not
bind.
The design of the split bushing is such that
the coil spring spreads the bushing to create a drag
between the bushing and frame which increases
the pressure required to close the contact from 160
Place
shims
here
A-
u
Figure
43.
Circuit Breaker Adjustment
grams (pressure required to compress the spring)
to
22 5 grams (approximately 8 02.). This friction
is used to dampen the rebound when the contact
closes. Check to be
slure that a maximum of 240
grams applied to
the plunger will close the con-
tact (Figure 44).
5.
A
pressure of 600 ~lus or minus 20 grams
(approximately
21 02.) on the contact plunger
should
be
required to compress the plunger spring
to the latching point (Figure
4 5 ) . These values
have been tested and found to
~rovide a good oper-
ating condition.
225
f
15
Grams pressure
lo
close contact
Figure 44. Circuit Breaker Adjustment
34
TYPE
603
ELECTRONIC MULTIPLIER
600
f
2OGrams
Figure
45.
Circuit Breaker Adjustment
6.
Locate the cam contact unir on the mount-
ing bar
at
its extreme limit of travel away from
the cam, and with the
plunger on the highes,t point
of the cam lobe, advance the adjusting screw until
the
plunger latches; then advance the screw one-
half turn additional to obtain
.O
10"
to
.O1IU
movement of the plunger beyond the latch point (Figure
46).
This will provide clearance between the
low dwell of the cam and the plunger. On the
non-latch type rhere should be a
.003"
minimum
clearance
between the low dwell of the cam and
the contact
plunger when the plunger is against
its stop (Figure
47).
If a latching cam contact does
not latch, it may be recognized
by
the fact that
the contact points will close for
11
",
or approxi-
mately
4
teeth on the index.
7.
To adjust the make time of the contact,
loosen the screws holding the cam to rhe shaft until
the cam is just snug on the shaft. Turn the machine to the
index point corresponding with the
Unlatching Cam Screws
-
\
,010"-,015"
Latching Clearance
Figure
46.
~ikcuit Breaker Adjustment
.003~'-.005"
Minimum Clearance
Cam
Figure
47.
Circuit Breaker Adjustment
ADJUSTMENTS
35
make ,time of the cam. Move the cam on the shaft
in the direction of rotation until the contact just
closes. The machine may now be turned to a point
where the cam
holding screws can be ,tightened.
An accurate adjustment may be obtained by inserting a screwdriver in the slots provided on the
periphery of the cam for moving it on the shaft.
8.
To adjust the break time of the contact,
loosen the contact unlatching cam screws (Figure
46).
Turn the machine to the proper index point
and move the
unlatching cam in its slot until the
contact opens. Tighten the holding screws. There
are six
possible positions for holding screws, only
two of which will be used at any one time.
Adjustment of Card Lever Contacts
All card lever con.tacts should have at least
A"
air gap when open, and at least
&"
rise off the
support strap when closed. Adjust by bending the
brass support, or by shifting the entire contact as-
sembly. The
opera,ting lever must be positioned
on the shaft so that the lever is just against the operating strap when the contact is open.
ELECTRICAL PRINCIPLES
THE
circuits for this machine will be studied in
three sections; namely, punch unit circuits, D. C.
power supply circuits, and electronic computing
unit circuits. In all circuit explanations, the location of the various circuit components both on the
machine and on the wiring diagram will be
stressed. Owing to the newness of electronic computing circuits to most Customer Engineers, the
electronic circuits will be analyzed in as much detail as
possible.
PUNCH UNIT ClRClllTS
WIRING diagram 213639A shows the power supply, punch unit circuits and the electronic circuits.
The timing chart for the punch unit is a part of
this wiring diagram. An electronic timing chart
will be found on the last page of the manual. The
diagram is indexed to facilitate location of circuit
components. Before proceeding with the discussion of these circuits, the various terminals used
on this diagram will be listed together with an indication of their location on the machine. All
terminals are counted from left to right, top to
bottom, front to rear, facing the machine.
On machines equipped for
1 15 volt A.C. opera-
tion only, there are twenty terminals for the
11
5
volt A.C. and 40 volt D.C. connections in the
punch unit. These terminals are located on top of
the main transformer. Post
1 and post 3 are the
40 volt D.C.
supply terminals while posts 7 to 20
are 11 5 volt A.C. and transformer terminals. Post
2 and 4 are not used. The terminals on the tube
power supply chassis are designated by a number
preceded by CH. There are ten of these terminals
located at the rear of the power
supply chassis.
Post CH3 represents power
supply chassis terminal
3.
On machines
equipped for 11 5-230 volt A.C.
operation, all transformers have two primary, coils
and more terminals are necessary. In these ma-
chines there are 24 terminals on the main trans-
former and
15 on ,the tube power supply chassis.
The terminals designated as CN, for example
CN9, represent the connections on the cable connector. CN9 is the
9 connection on the cable connector on the punch unit, counted as indicated on
the wiring diagram, Section
22B. The cable is detachable at both ends, hence, a sub-letter is used
to indicate which connector is meant. Every other
terminal is used to lessen the
possibility of a short
across ,terminals.
Relays and cam contacts can be located on the
circuit diagram
by means of the location chart
shown in Section
17- 1 8 of the wiring diagram. In
tracing circuits, normally closed points will be abbreviated
by
N/C,
normally open points by
N/O,
and the operating strap
O/P.
All power is furnished by the commercial supply
through the attachment cord which is connected
across terminals
7 and 8 on top of the main trans-
former. The main line sentinel switch turns power
to the machine
ON
or
OFF.
Most machines made
require a
I 15 volt A.C. supply. However, trans-
formers are available which permit either
1 15 volts
or 230 volts A.C. operation.
The power supply
circuits for 115-230 volts
A.C. operation appear
in Section
3 of the wiring diagram, while Section
4 shows the circuits for 11 5 volts A.C. operation
only.
HD3 and HD4
relays are A.C. relays wired di-
rectly across the A.C. line from post
9
to post 10.
When the main line switch is
ON,
HD4 is picked
up and its points complete a circuit to the primary
of the main transformer.
The main transformer
is protected by two
20 ampere fusetrons.
The reason for this arrangement is that no sentinel switch
was available to handle the total current demand
when this circuit was designed. With this arrangement, HD4 relay points act as a switch, and the
sentinel switch protects
only the punch drive
motor and the tube power
supply transformers;
s
ELECTRICAL
PRINCIPLES
37
the main transformer is protected by fusetrons.
HD3 points open the
ground connection when the
switch is
OFF.
The tube power supply transformer primary
connections are made across posts
9 and 10.
The
drive motor is directly across the main line switch
and hence operates on the commercial A.C. supply
under the control of
HDI relay. It is protected
by the 10.9 ampere element in the sentinel switch.
A split
plug allows the motor to be disconnected
from rhe machine.
Note that the test lamp outlet is "hot" as long
as the attachment cord is
plugged in, regardless of
the
seting of the sentinel switch. The test lamp
outlet is protected by 3 ampere glass fuses 12 and
13.
As soon as the main line switch is turned
ON,
40
volt
,D.C. supplied by the selenium rectifier is
available across posts
I and
3.
Post 1 is the posi-
tive side of the line and post
3
is the negative side.
For sake of consistency, wherever
possible, all cir-
cuits will
be
traced fronz ~zegativc ternzi?znls to
positive.
Time Delay Circuit
Owing to the necessity of heating the cathodes
of the rectifier tubes and the vacuum tubes to op-
erating temperature before putting them in opera-
tion, it is necessary to allow an interval of time
before starting rhe machine.
To insure full oper-
ating temperature, approximately
50 seconds is allowed for heating the tube cathodes. The rectifier
tubes in the power
supply chassis are gas-filled
tubes and will be damaged if the load is applied
before their filaments have reached operating temperature. If vacuum tubes are placed in operation
before they have been fully heated, the tubes will
not be damaged, but
erraric operations will result.
The time delay is effected by a thermal delay re-
lay
(RF in machine). One contact strap of this
relay is
a
bi-metal strip, i.e., it consists of two
strips of different metal placed one on top of the
other and bonded together.
These two metals ex-
pand at a different ratio when heated and hence the
strap will bend. A coil of
high resistance wire is
wound around this strap to serve as the heating
element. The lower strip of metal has the greater
expansion rate; therefore, the strap will bend upward when heated, and the contact points will
touch to close a circuit to the coil and complete
circuits for the operation of the machine. The
upper contact point is in the form of a screw
tipped with contact metal. This screw provides
a means of varying the time interval required for
the relay points to close by varying the distance
between the contact points. The adjusting screw
is normally set to
provide 50 seconds delay. An
air gap of approximately
A"
will provide a good
starting point in setting this adjustment. This time
is applicable when starting with a cold machine.
Naturally, the time delay will vary if the machine
is turned
ON
and
OFF
several times in a few min-
utes because the heating unit will not have time
to cool.
The time delay relay operates on the regular 40
volt supply of the
p~nch unit furnished by the
selenium rectifier. With the main line switch
ON,
a circuit is complete to .the time delay heating ele-
ment as follows: post
3,
terminal CNp5, through
ground cable in the electronic unit and back to
CNpl3, through heater element,
N/C
points of
R5BL, to R2A N/C, P16, PI, post 1. Also, as soon
as the main line switch is closed,
R2 coil will be
energized in
parallel with the heating element. The
purpose of
R2 and the cable jumper will be ex-
plained shortly.
When the heating element has
su&ciently heated
the bimetal
conract strap, which is one strap of
R5A points, the R5A points will close, and a cir-
cuit will be completed to the coil of
RF as follows:
post 3, terminal
CNp5, through ground cable to
CNpl3, through R5 coil, RFA contact, and out
to post
1.
Relay F will then pick up and its BL and BU
points will transfer.
When RFBL transfers, the
38
TYPE
603
ELECTRONIC MULTIPLIER
circuit to the heating element is opened and a holding circuit for
R5 is completed.
When
R5BU points close, the HD2 relay will
be energized since
R2 is already energized and the
R2A contacts transferred. The circuit is as follows: post 3, terminal
CNp5, through ground
cable, power switch in electronic unit (Section
22A), back through cable to post CNpl5, *through
HD2 coil, R5BU, R2A
N/O
to P16, PI, and post 1.
The
HD2 contacts are in the ~rimarv of the tube
power supply transformer and are shown on the
power supply circuit diagram. Power for the
,tubes
in the electronic unit becomes available as soon as
HD2 is energized.
Start Circuit and Bias Interlock
In order to insure that the tube power supply
is functioning
properly, the starting of the machine is under the control of start circuit interlock
relay R6 and bias interlock relay R4. R4 must
pick up to complete the anode power supply circuit and also to pick up R6. The bias interlock is
necessary to insure that the grid bias
voltage is
applied before the anode voltage is applied; otherwise, excessive anode current will flow through the
tubes and overload them before the control grid
comes into operation. Eventually this will damage
the tubes. When R4 is picked up, the
R4 contacts
complete a circuit to R6, which is connected across
-the grid and anode voltage supplies; hence R6
cannot pick up unless the anode voltage is applied
to the vacuum tubes.
The voltage output terminal
CH3 (Section 4A)
on the power
supply chassis is the ground terminal,
or point of reference for all voltages, and is considered the zero point for all voltage references.
Note that CH3 is connected to post
3 which means
that post
3 may be considered as a zero voltage
reference point. Terminal CH4 supplies
plus 150
volts,
i.e., 1 50 volts above ground, while CH2 sup-
plies minus
100 volts, or 100 volts belotv grozl~~d,
making the voltage between CH2 and CH4, 250
volts. Terminal
CHI supplies minus 250 volts.
Each of the power lines is protected
by a 2 amp
glass fuse. Fuse
9
protects the -2 50 volt line, fuse
10, the -100 volt line, and fuse 11 the +1 50 volt
line.
Note that the coil of R4 is connected directly
across the
-1 00 volt supply (bias supply), between
fuse
10 and the ground connection at the relay
common. The circuit for picking up R4 is as follows: post 3, fuse 6, to
R1, R3, R7, through R4
coil, to fuse
10. This relay will then remain ener-
gized as long as the bias voltage is applied.
Bias
voltage is furnished to the electronic unit through
the cable connector
CNp3.
When the R4A points close, a circuit is com-
pleted to
supply +I 50 volts to the electronic unit
through
2 amp fuse 11, R4A, and the cable con-
nector
CNp7. The R4A contacts also complete
,the following circuit: fuse 11, R4A, R6 coil, three
25,000 ohm resistors in parallel, to fuse
10. Thus,
R6 remains energized as long as R4 is picked up
and as long as the
+I 50 volt supply is available.
The purpose of the three
25,000 ohm resistors
(net resistance approximately
8000 ohms) is to
create a
sufficient voltage drop across the 250 volts
between fuses
11 and 10 so that the proper voltage
is applied to the coil of R6. The purpose of the
capacitor across the R4A contacts is to absorb the
arc across R4A in case a fuse blows, etc.
Once R6 is picked up and the R6A contacts
close, the machine is ready to start.
The green
pilot light indicates that the punch is ready to start.
The circuit to light the
pilot light is as follows:
post 3, through pilot light, post
5,
die contact, post
6,
knockoff bar contact, R6A contacts, to center
strap of
RZA, P16, PI and post I. The knockoff
bar contact and die contact are both safety contacts. The die contact is closed only if the die is
properly in
place with its latching bars in position
in the grooves in the side frame. The knockoff
bar contact is closed only if the knockoff bar in
the magnet unit is
properly in place.
ELECTRICAL PRINCIPLES
39
Start Interlock Circuits (Gang Punching Only)
In case it is desired to use the punch unit for
gang punching only, it is desirable to disable the
start interlock circuits and open the tube power
supply circuit, although it is not necessary.
This
can be done by disconnecting the cable at either
end.
If the cable is to be disconnected at eirher end,
it is recommended that the main line switch first
be turned
OFF.
With the cable disconnected, the
circuir to R2 and Rj coils is opened.
If
R2 and
R5 fail to pick up, the HD2 relay cannot pick up,
and no power will be furnished by the power sup-
ply chassis to pick up R6. Note, however, that
the
N/C
side of the R2A contacts provides a path
around R6A to post 1.
Start and Running Circuits (No Cards in Machine)
It is most important to bear in mind, when
tracing any circuit or set of circuits, the objective
of the circuit. The objective of rhe start and running circuits is to control the feeding of cards
through the punch. This is done
by starting the
motor, causing the punch clutch ro engage, and
then keeping the machine running once cards have
reached the second set of brushes. The motor is
under the control of
HDI relay which in turn is
controlled by
RIO. The punch clutch is also under
the
conrrol of RIO; hence it is obvious that R1O
must be energized to cause the motor to run and to
cause the punch clutch to engage.
Depressing the
start key and closing the contact will cause
R1O
to pick up (assuming R6 is picked up). R10 will
remain energized as long as the start key is held
depressed through the following circuit: post 3,
through fuse 6 to
RIP, RlH, HD1, through R10
coil, rhrough start key contact, stop key contact,
R57B contact, post
5,
die contact, post 6,
KO
bar
contact, R6A contact, to center strap
R2A, to P16,
PI, post 1.
The closing of rhe
R1O contacts simultaneously
completes a circuit to the P-clutch and HD1
motor relay. Circuit for the P-clutch is as follows:
post 3, fuse
5, punch clutch coils, RlOBU, P16
cam contact, to
PI and post 1.
This circuir ener-
gizes the P-clutch magnets, which in turn unlatch
the clutch
pawl so that it may engage the rarchet
at D on rhe index.
The circuit to the motor relay is as follows: post
3, through fuse 6, 40 volt
HD1 coil, RlOBL con-
tacts, and on to post
1 as in the start key circuit.
The
HDI contacts complete a circuit to the drive
motor and keep the drive motor in operation as
long as
R1O remains energized.
The
2mfd capaci-
tor across the
HD1 contacts suppresses arcs which
mighr cause trouble in the electronic circuits. This
capacitor is
nor shown on the wiring diagram,
213639A.
The motor drives the continuously running
mechanisms; at D, the P-clutch engages, and all
units go into operation.
Shortly after the P-clutch
engages, rhe circuit to the P-clutch magnet is
broken by P16 at 14 on the index.
Remember,
from the mechanical operation of the clutch, that
once the clutch engages it must make a complete
revolution before
relatching.
If rhe start key is
held closed for more than one
cycle, P16 makes
again at 12.3 on the index to energize the P-clutch
magnet and prevent
relatching of the pawl.
In order to prevent the machine
sropping part
way through a cycle when the start key is released,
a P-cam provides a hold circuit for
R1O to insure
the completion of
cycle during which the start
key may be released (or when the machine stops
owing to any other reason).
The start relay hold
circuit is completed through P19 as follows: post
ie con-
3, through fuse 6,
R10 coil, RlOAL, P19, d'
tact, and on to post 1 as in the srart key circuit.
Start and Running Circuits
(Cards in Machine)
The running circuits are designed to keep the
machine running automatically once cards have
reached the second set of brushes and to stop the
machine for any of
,the following reasons ( 1 ) when
the last card runs out of the magazine, (2) when
the stacker is filled to capacity, (3) when an error
TYPE
603
ELECTRONIC MULTIPLIER
is made in checking, or (4) when any card fails to
advance to the die or to the
punch brushes 2 (read
for gang punching and checking).
Card lever contacts are provided to signal the
presence of cards in the magazine or at any of the
stations through the machine. A stacker stop
switch signals a full hopper, while a relay is provided to signal an error when checking.
As soon as cards are
placed in the hopper, R3
will pick up as follows: post 3, through fuse 6,
through
R3 coil, hopper card lever contact, to post
1. The R3A contacts are in the
R1O running hold
circuit.
The other card lever contact relays
R1, R7, and
R8 are energized through their corresponding card
lever contacts when the contacts are closed by the
cards.
After the cards are
placed in the hopper, it is
necessary to hold the start
key depressed through
three complete cycles before the machine runs
continuously.
After the third cycle, the machine runs automatically. The card levers are designed so that
their contacts do not open between cards; therefore all card lever contacts remain closed once cards
start through the machine. As
a
safety precaution,
R1 is held by P33 between cards, and R8 is held
by P16 between cards. This insures the card lever
relays remaining energized even if the card lever
contacts open for an instant when corner cut cards
are used.
All that is necessary to keep the machine run-
ning continuously is to keep
R10 energized. This
is done through the following circuits once cards
reach the second set of punch brushes: post
3,
through R10 coil, RlOAU, R7BL, R3A, RIAL,
stacker stop switch, stop key contact, RF7B, die
contact,
KO
bar contact, R6A, to R2A
O/P
and
post
1. This circuit is complete until any one of
the contacts is opened.
Reading
Brush Circuits
R8B contact is used in the read for entry and
control brush reading circuit, while R7A contacts
are similarly used in read for gang
punching and
checking brush reading circuit. The purpose of
these contacts is to prevent readings from the
brushes when no cards are in position at the
brushes.
PI through P4 (Section 7A) serve as circuit
breakers for both sets of brushes.
P1 and P2 con-
trol the
making time at the line of index while P3
and P4 control the breaking time at three teeth
past the line and
provide CB pulses of 3-tooth
duration.
The purpose of the first set of brushes is to read
the card for entering the factors in the various
counters and for
X
control, group multiplier con-
trol, sign control, while the purpose of the second
set of brushes is to read the card for checking or
for gang punching operations. The complete cir-
cuit for reading into a counter will be deferred until the electronic circuits are discussed but the portion shown on the punch unit diagram will be
traced. To simplify the tracing of this circuit it
will be necessary to depart from the rule established and start at the positive terminal. The circuit for reading into any one position of the multiplier
counrer is as follows: post l, through PI,
P2, P3, P4, R8B, common brush, contact roll,
through hole in card, individual brush, brush hub,
entry hub for multiplier counter (Section 1 FA),
through factor reversal relays, cable connector and
cable to electronic counter.
In gang punching, the
punch hubs are con-
nected to the second set of brushes, and a punch
magnet is energized when a hole is read in the column to which the punch magnet is wired. The
circuit is as follows: post 3, through
R12B (Sec-
tion
lFB), punch magnet common connection,
punch magnet coil, punch hub, wire to hub, read
for gang
punching brushes (Section 7A), indi-
vidual brush, through hole in card, contact roll,
common brush, R7A contacts, P3, P4,
PI, P2, to
post 1.
1
ELECTRICAL PRINCIPLES
014
9
8
7
6
5
4
3
2
1
0
11
I2 13014
9
8
7
6
5
4 3
2
1
0
11
12 130
11111111111111
IIIIIIIOIII
ad Card
2-j]
~c~~yrp
MP MC Cancel
Ion Grp. Mult.) Compute
Card
I
-
MP MC Cancel
(Group
Mult.) Punch Card I
-
PrOduci
I
P- I0 Main Cancel
I
-
13.D 14-3
P-
10
I
1
I
P-l l Product Cancel 0.6
11.0
P- ll 0.5
1
,
P-12 MP-MC Cancel (Group)
Figure 48.
Timing Chart for Computing Operations
14.
9.4
P-12
9.) 8;s
LS
8-5
~3
4-5
3.5
e-s 1-3
Reod- in CB's
P -24
Compute Start
Electronic Computing Control Circuits
The circuits in the punch unit for controlling
the electronic computing unit are
primarily timing
circuits to
*time the computing operations to the
movement of the cards through the
punch unit.
No attempt will be made to trace complete circuits
for these
conrrols; only the general function of
each timing cam will be given. Figure 48 is a
timing chart showing the relationship between the
computing functions and the card movement
through the punch unit.
Electronic counters are restored to zero by open-
ing the
-100 volt grid bias circuit by means of a
P-cam contact. Other circuits are similarly restored to normal by opening a contact.
PIO (Sec-
tion
!A) is the main cancel contact and conrrols
the restoration of all circuits except the product
counter to normal. This restoration occurs
a.t the
very beginning of a cycle before any figures are
read from a card. When group multiplying, R26
is picked up,
R26A points transfer, and either the
multiplier or the multiplicand counter, depending
upon the setting of the factor reversal switch, is
cleared by
PI2 under further control of R27-3
points. P12, which opens slightly after PI0 closes,
is effective only during cycles that R27 is picked
UP-
I
I
9;s 8-5
-
1~5
Read-inb';
-
OslLariZlO
Read-out CB's
P-24
11.5- 11.9
The product counter cannot be cleared until
after
punching is complete, consequently the clear-
ing is separately controlled
by the product cancel
cam contact
PII, which is under further control
of
R19A. PI1 opens at 0.5, and the product
counter is cancelled or cleared, provided R19, the
product summary relay, is
n0.t picked.
When it
is desired to store and accumulate
products, R19
is picked up and
PI 1 is no longer effective.
The reading of factors into the electronic
counters occurs at mid-index points under control
of
PI through P8 (Section 8A) ; PI and P6 pro-
vide read-in
pulses of three-teeth duration at mid-
index points. P17 permits only
9 of these pulses
to be used. P7 and P8 provide pulses of three-
teeth duration at index poinrs to control the elec-
tronic read-out circuits.
PI8 allows 10 of these
pulses to be used. The reason for the two sets of
CB's will be apparent in the discussion of the electronic read-in and read-out circuits.
Once factors are read into the counters from a
card, it is necessary to start the computing section.
This is timed by P24 cam contact which
provides
a pulse to start computing at 5 teeth past II on
the index if
RlBU points are closed.
RlBU prevents computing operations in case the punch is
running without cards.
1
42
TYPE
603
ELECTRONIC MULTIPLIER
Read-out is also under rhe control of P5 through
P8 except that read-out pulses are index line pulses.
Also, read-out is under further control if
RlBL
(die card lever relay), R19B (product summary
relay), and
R29AL ( group multiplier relay). If
any of these three relay points are open, the product cannot be read from the electronic unit to
the punch.
Read-out is further controlled by the P14 cam
contact (Section 4B), which applies anode potential to the read-out power tubes only during
,the
9
through O punching time of the cycle. The electronic portion of the read-out circuit is reserved
for the section on
,the electronic computing cir-
cuits; only the
portion on the punch unit will be
covered. Read-out pulses from the electronic unit
come to the punch under the control of
P7 and
P8 through the cable and to the hubs labelled
PRODUCT
(Section 15B). Normal wiring for
punching the result is from the
PRODUCT
hubs
to the
PUNCH
hubs. The circuit for energizing a
punch magnet from the product counter is as follows: post
3 (Section YB), through R12B (Section
1 FB), punch magnet common, through punch
magnet, hub, wire, to hub marked
PRODUCT,
to
cable connector, through cable to electronic unit,
back through cable to
CNp28 (Section 4A), P14,
to +I40 volt terminal.
Punch Suppression
If it is desired to suppress punching in a card,
the card is either X-punched or punched with a
control digit in a specified column. The control
punching then causes
he
pickup of R12 to sup-
press punching on the cycle following.
R15 (Section 9B) is the X relay which picks up
from 0.3 through
1 1.3 of each cycle under control
of P37. Points of R1T are used in the X pickup
circuits of various relays. For example, RIYAU
is in the X pickup circuit of R11 (Section 8B), so
that
R11 can pickup only on X's if the X pickup
is wired from rhe first set of brushes (read for entry and control). If it is desired to pick
R11 on
other digits, the
D
pickup hub is wired from the
first set of brushes.
The sequence of operations for punch suppres-
sion is shown in Figure 49.
When
RI 1 picks, the
R
1 1
A points close to hold R11 energized through
P21 until 14.5 of the next cycle. The RI 1B points
establish a circuit to pick
R12 under control of
P22. If the punch suppression device is wired to
suppress on X's, R12 will pick up at 13.6, provided
R11 is already energized. R12A points close
to hold
R12 through the R12H coil until P23
breaks at 13.4 of the same cycle.
R12B is in the
punch magnet common circuit so that no punching can occur as long as R12 is energized.
X
Cord Pamsrm
X
Card Possrs Dir
Entry
Brushas
k
--I+--
(Punching Suppressed)
Figure
49.
Sequence of Operations on Punch Suppression
ELECTRICAL PRINCIPLES
43
It is obvious from the circuit that if the punch
suppression feature is wired for
N
suppression, R12
will pick up every cycle that
R11 is
not
picked up.
Therefore,
punching will be suppressed when there
is no X (or digit) punched in a card.
The punch suppression feature can be used to
perform group gang punching operations, that is,
information in interspersed master cards can be
punched into the detail cards following the master.
Either the master cards or
sthe detail cards can be
X
(or digit) punched.
Product Summary
Often it is desired to accumulate several pro-
ducts before punching.
This permits special oper-
ations, such
as
crossfooting on two cards, punching
the sum of several products, etc. To accomplish
this, it is necessary to prevent reading out and
clearing of the product counter. The product summary feature has the same type of controls as punch
suppression,
i.e., X or digit pickup and X or
N
control. To place the product summary in operation, a jack must be
in the
ON
position of
this switch. This causes R19 to be
picked up as
long as
R18B points are closed.
R18 picks under
the control of
R17, which in turn picks under
control of an
X
or digit punching. The X or
D
hubs of the product summary device are wired to
the first set of brushes (read for entry and control). The
X
hub is under control of R15BL points
so that
only X's can energize R17.
Once R17 is picked, it
is
held by the R17H coil
through
R17A and P21 until 14.5 of the next
cycle. Then with
R17B transferred, R18 picks
when P22 makes, provided
X
control is being used.
R18 holds through
R18H coil, R18A, and P23
until
13.4 of the cycle following the reading of the
X
(or digit). During the cycle that R18 is ener-
gized and
R18B points are open, R19 cannot be
energized. When R19 is in a normal position,
R19B
points (Section 8A) are closed to permit read-out
from the product counter, and
R19A points (Sec-
tion 5A) are open to permit cancelling the
product
counter when PI1 opens.
Thus, an
X
or digit
punching will cause normal operation when set for
X
control.
If the
product summary is set for N control,
R!8 picks every cycle that no X or digit punching
appears and normal operation continues until an
X
or digit punching appears. An X or digit will
cause progressive addition of
products and sup-
pression of
punching when set for N control.
A
blank card with the proper control punching
must
precede a product summary run if the ma-
chine has just been
,turned
ON.
Otherwise, the first
group may be over because of the random figures
which
might set up in the product counter when
the switch
is
turned
ON.
Remember that on normal multiplication the product counter is cleared
before the first computation, whereas on a product
summary run, it
is
not.
Product Overflow
Often the number of columns set aside on a
card for punching the product of a
multiplication
is not large enough to permit ~unching the largest
products. However, if only a very small percent-
age of the total number of cards exceeds the capacity of the field, it may not be worth increasing the
size of the card field and thereby limiting the num-
ber of columns available for other punching.
TO
take care of such cases the product overflow feature is furnished on this machine.
To
place the product overflow feature in opera-
tion, the product counter position
next
to
the
highest order wired to
punch is wired to the product overflow entry hub. Observe that P38 cam
contact (Section 9B) is in this circuit and that
P38 will permit any impulse from
9
through
I
to pass and pick up R16. As long as the product
counter position wired to
product overflow con-
tains O's, nothing happens. Any
digit from
1
through 9 in this position causes R16 to pick. R16
then holds through its
R16H coil, R16AL, and P21
cam contact. R16AU points complete a circuit to
the error light and error
relay R57 (Section IlB),
TYPE
603
ELECTRONIC MULTIPLIER
so that the machine will stop and indicate wi~h a
red
light that the ~roduct just computed and
~unched exceeds the capacity of the card field.
Since
.the error light is used for other error signals, it is desirable to distinguish between errors
and overflow products. This is accomplished by
punching a
12 in a card in which the product ex-
ceeded the card field capacity. To punch the
12
the product overflow 12 hub is wired to any punch
magnet, and a 12 hole will punch under control of
P3 6 (Secrion 9A) provided R16B points are closed.
With this arrangement, a card is examined after an
error light and if a
12 hole is punched, it is appar-
ent that the error was due to an overflow
product.
This card can then be handled manually.
Factor Reversal and Check Circuits
Multiplication is checked by re-mulriplying with
the multiplier and multiplicand factors reversed,
then checking for double punchings or blank columns by means of the double punch and blank
column detection circuit. This checks the multiplication, because if a different answer is
ob~ained
on a check run than was obtained on a multiplying
run, the card will be double punched in some column. The check circuit also checks for failure
to punch during the first multiplication run and
again checks for blank columns during the check
run. An error will cause the machine to stop and
the error light to glow. Before restarting the machine, the double punch reset push button must be
depressed to extinguish
-the error light.
Factor reversal is accomplished by switch control to permit rapid change-over from one setup
to another. When the factor reversal switch is
ON,
relays R2 1 through R2 5 are picked up and re-
main energized as long as the
swi~ch is
ON.
R21,
R22, and R23 are used to reverse the entry to multiplier and multiplicand counters so that the factors normally entering the multiplier
enrer the
multiplicand and vice versa. R24 reverses the
multiplier and multiplicand read-in control and
cancel circuits on a group multiplying setup, while
R25 is used in connection with sign control.
One
con~rol panel switch is provided for each
position of blank column checking. Double
punch
checking may be used alone or in combination
with blank column checking. Blank column checking cannot be done alone, but must be used in combination with double punch checking. The product
field in the card is wired from the second set of
brushes (read for gang punching and checking) to
the double punch entry hubs for double punch
de-
rection.
If
blank column checking is also desired,
the corresponding blank column
swi~ches are set
ON.
When checking multiplication, it is necessary
to use both double punch and blank column
checking.
As the card passes the first set of brushes, the
factors are read in to re-multiply. The result is
again
punched, and the card is checked as it passes
the second set of brushes.
As the card passes
rhe
second set of brushes, a circuit will be completed
through the hole in the corresponding column of
the card to pick up R37. Only the first position
will be considered in tracing
circui~s.
The R37B
points indicate the presence of a hole in the column
being checked. R38 then
picks up and sets up a
circuit through its
R38A points to the error relay,
R57. If another hole is present in .the same column
after R38 is picked up, the error relay will be energized and the machine will be stopped.
The circuit for energizing R37 is as follows:
pat
3, through fuse 7, relay common, through R37 coil,
R38A
N/C,
DPBC entry hub, wire, read for gang
~unch and check brushes, hole in card, contact roll,
common brush,
R7A, P3, P4, P2, P1 to post 1.
This circuit picks up R37, closes the R37A
points, and opens the R37B points. The opening
of the
R37B points prevents the energization of
R57H coil when P3 5 makes.
R37 is held through
R37A and P34 until 13.3
of the same cycle.
R38 will not pick up as soon
as
R37A points close, because there is a shunt cir-
cui~ around the R38 coil through the hole in the
ELECTRICAL
PRINCIPLES
45
card and the CB's. When the CB's open, R3 8 will
be energized through the R37 hold circuit.
R38A points then transfer and open the pickup
circuit to R37 and set up a circuit to the
R57P
coil. If no o,ther hole is punched in the same column, nothing further happens and no error is in-
dicated. However, if another hole is sensed in the
same column, after the foregoing circuits have
been established,
R57P will be energized as follows:
post 3, fuse 7,
R57P coil, R31BL, R38A
N/O,
hub,
wire to read for gang
~unchin~ and checking
brushes, and out through hole in card to post 1.
R57 will then be held
by its holding coil and
R57A points until the reset push button is depressed. As the error light is in
parallel with the
R57H coil, it will light when the hold circuit is
completed and remain
lighred until the push but-
ton is depressed.
When R57
picks up, its R57B points open, and
as soon as P19 opens,
R10 will drop. This will open
the circuit to the P-clutch magnets and to the
HDI motor relay, thus stopping ,the machine. The
error light will also glow, indicating an error.
In case of a blank column in the units position,
R37 will not
pick up and the following circuit will
be completed to pick up R57 when
P35 makes at
12.2
:
post 3, through fuse 7, R57H coil (and error
light in parallel), R37B, hub, jackplug, R31AL,
R7BU, P3 5, to post 1.
R57 and the error light then remain energized
through the
R57A points until the reset push but-
ton is depressed.
The circuirs are the same for any position other
than the first. The circuits are designed so that
an error in any one column will stop the machine.
As an optional feature, ten additional positions
of double punch and blank column checking are
available. When the additional ten positions are
installed, relays
58 through 77 are used.
Group Multiplying
Group multiplying signifies multiplication of a
group of cards by rhe same multiplier. The multi-
plier factor is
~unched in the first card of each
group and is known as the
rate card
or
?lzaster
card.
The rate card is identified by a 9 punching
in any column. As the rate cards are interspersed
in front of each group, the machine senses the ap-
roach
of a new rate card by the 9 ~unchin~ and
clears the multiplier counter
TO
allow a new rate
to enter. Also the machine suppresses
~unchin~ of
the rate card during the cycle the rate card is pass-
ing the die. When checking, the machine must
also suppress the checking
operarions when the rate
card passes through the machine.
Relays 26 through
3 1 are the group multiplying
relays. Relay 27 is a wire contact relay. This relay is used in this circuit because of its fast
pickup.
This relay must pick up between 9 time, when a
hole is read, and
5
*teeth past 9.
To allow a suffi-
cient safety factor the wire contact relay is used.
The contacts on the wire contact relays are in line,
and therefore the identification of the contacts is
by contact number. Thus R27-3 indicates the
number
3
contact on wire contact relay 27.
VlJhen it is desired to group multiply, ,the con-
trol panel switch labelled
GRP
h4P
is wired
ON.
This causes R26 to pick up and remain energized
as
long as the
GRP
MP
switch is wired
ON
(Sec-
tion
10B).
The R26BU points set up the circuit to the
9
pickup hub ro control the pickup of R27. R26BL
points transfer to place the screens of the multi-
plier or multiplicand read-in switch tubes (depending on the setting of the factor reversal
switch) under the control of
R27-2 points. R26A
points transfer the control of the multiplier or
multiplicand cancelling circuit (depending on setting of factor reversal switch) from
PI0 to P12
and the R27-3 poinrs. A pentode can be blocked
by opening the screen potential circuit. Use is
made of this fact in controlling read-in to the
counter containing the rate. When it is desired
ro block all entries to a counter, the screen poten-
tial circuit for the read-in switch tubes of that
counter is opened.
46.
TYPE
603
ELECTRONIC MULTIPLIER
R
27~U'0'e-e'oP26
("9"
Ma in rota card)
I
I
I
C
Rota Card
Passas P Brushes
I
Multiplier
Cancellad
Naw
Rde Entars
-
--
Rate Cord Rota Card
Pos3es Dia-- Possas P Brushes 2--
Punching Supprassad Chacking Suppressed
Figure
50.
Sequence of Operation for Group Multiplication
To signal a rate card, a 9 is punched in a predetermined column and this card column is wired
to the
9 pickup hub. Then when a 9 is read as
the rate card
en,ters the first set of brushes, a cir-
cuit is completed to
pick R27 as follows: post 3,
through fuse 7,
relay common, R27P coil, R26BU,
P26, 9 hub, control wire, brush hub, individual
brush in first set of brushes,
9 hole in card, contact
roll, common brush,
R8B, CB's, to post 1. P26
allows
only 9 holes to be recognized. R27 is then
held through its R27H coil, R27-1 points, and
P27 until 13.4 of the same cycle. The sequence
of operations of the above and succeeding circuits
for group multiplying is shown in Figure
50. This
in con-
sequence chart should be studied carefully
'
nection with the group multiplying circuits.
In the following circuits it is assumed that the
factor reversal switch is
OFF
and that the multiplier
counrer contains the rate (or group multiplier).
When R27 picks up, R27-3 points open and permit
PI2 to cancel the previous multiplier as P12
is open at this time. P12 is required in addition
to
PI0 because the cancelling of the multiplier
counter must be delayed until it has been determined whether or not the card entering the first
punch brushes
contains a 9. R27-2 points close
to complete a circuit which provides screen potential to the tubes controlling multiplier read-in,
and thus permit reading of a new factor into the
multiplier counter.
R27-4 sets up the circuit to
R28P coil so that when P28 makes at 12, R28 is
picked, and it holds through its
R28H coil, R28A,
and P29 until 8 of the next cycle. The R28B points
set up a circuit to cause
R29P coil to be energized
when
P3O makes at 14.5 of the cycle following the
pickup of R28. R29 is held through its
R29H
coil, R29AU points, and P31 until 13.4 of the
same cycle.
ELECTRICAL PRlNCl PLES
47
While
R29
is energized, the
R29AL
points (Sec-
zion
8B)
are open to suppress the product read-out
circuits so that no
punching of the rate card from
the product counter can take place.
R29BU
points set up a circuit to pick
R30
when
P28
makes at
12-. R30
then holds through its
R30H
coil,
R30A
points, and
P29
until 8 of the
following
cycle.
R30B
points set up a circuit to
R31
so that
R31P
is energized when
P3O
makes at
14.5
at the beginning of the cycle during which
the rate card passes the second set of brushes.
R3
1
then remains energized through this entire cycle
through its
R3
1H
coil,
R3 1AU
points, and
P3 1.
R31A1,
points open to suppress the blank col-
umn check circuits as the rate card passes the sec-
ond set of
punch brushes.
R3 1B
points open the
circuit to
R57P
coil, thereby suppressing the double
punch detection circuit.
As
previously stated, the sequence of operations
from the reading of the
9
hole in the rate card
through the holding of
R3 1 is shown in Figure
50.
A
careful study of this chart should provide a
complete picture of the group multiplying opera-
tions.
Column Split, 0 and
X,
Hot
1
Two positions of column split are available as
standard on this machine.
R33
picks up each cycle
when
P33
makes from
0.2
through
12.2.
Thus the
0-9
hubs of the column split are connected with
the
C
hubs from 9 through
0,
and the
11-12
hubs
are connected with the
C
hubs for
11
and
12.
This
device permits an
X
or
12
punching over a
0-9
digit to be recognized independently or to be ignored.
The
O
and X hub provides a source for impulses
at
11
and O time to punch X's or 0's under the con-
trol of
P32
cam contact (Section
7A).
If only an
X
or 0 is desired, the O and X impulse
is
wired
through a column split to separate the
O
from the
X.
The Hot 1 hub provides a source for a 1 impulse
under control of
P20
cam contact (Section
7A).
This may be used to punch
1's
or to enter a 1 into
the electronic counters for unit multiplication.
The hub labelled
CB
is a source for impulses at
all index points. It is used
primarily in connection
with the distributor which is an optional feature.
Sign Control (Optional Feature)
As an optional feature, this machine is designed
to handle either positive or negative factors. Negative factors are
identified by an X punch. The
X
may be punched in any column to identify a negative factor.
If
either
factor is negative, the result
is negative. If
neither
factor, or
both
factors are
negative, then the result is positive.
Hence, it is
obvious that an
X
must be punched in the card
if either the multiplier or the multiplicand is negative.
The
MCX
and
MPX
hubs are wired to the first
set of brushes. The
MCX
hub is wired to the brush
in the column containing the
X
to identify a nega-
tive multiplicand, while the
MPX
hub is wired to
the brush in the column containing the
X
to iden-
tify a negative multiplier.
The
PR
x
PCH
hub is
wired to the punch magnet corresponding to the
card column in which the
product
X
is to be
punched. To make the sign control wiring effective the
SIGN
CTRL
switch must be wired
ON
(Sec-
tion
13B).
The objective of the sign control circuit is to
sense a negative result and punch an
X
over rhe
proper column to identify a negative product.
The
general sequence of events for punching
an
X
to identify a negative product is as follows:
1. RIOO
picks up at X time.
2.
An X is sensed in either the
MC
or the
MP
field, and the corresponding sign control relay picks up.
3.
The sign control relay is held until after
X
time and the sign delay relay picks up
and is held through the next
cycle when
the card is in punching position.
4.
At
X
time of the second cycle an X is
punched to identify a negative product.
48
'TYPE
603
ELECTRONIC MUL'TIPLIER
@
PUNCHING
n
loo pu0.'~'.' P 37
Punch
Magnet-CB(X
iim
RIOOPU-~37
R
25 ( Factor
Reversal)
R 95 PU-CB (X
Hole
-MC)
R ~~Pu"'-~.'P 40
R97H ".'P41
R
100
PU- P 37
RlOOPU-P37
1
I
@
CHECKING (Error orrumad)
1
Figure
51.
Sequence
of
Operations for Sign Control
The actual contacts and relays involved in this
operation together with the timings are shown in
sequence chart in Figure
5
1A.
The circuits traced in the order in which they
operate are given below.
RlOO picks up at X time
every cycle when P37 makes provided the
SIGN
CTRL
switch is wired
ON.
The points of RlOO con-
trol the pickup of all other sign control circuits.
Assuming an
X
in the multiplicand field, R95
will then pick up when the card passes the first
set of brushes and the X hole is read. R96 would
be energized if there were an
X
identifying a negative multiplier. R95 holds through its holding coil
R95AU, and P39 until 5 of the following cycle,
and R95B points s& up a circuit to pick R97 when
P40 makes.
The sign delay relay R97 is energized when P40
makes at 6.7 of the cycle following the reading of
the
X.
Note that if
lzeither
or
both
R95 and R96
are energized, R97 cannot
pick up. R97 is then
held through its
R97H coil, R97A points, and P41
un-til 11.5 of the same cycle. At X time of this
cycle the punch magnet wired to
PR x PCH
is en-
ergized as follows: post 3, through
R12B, punch
magnet common, punch magnet coil,
PUNCH
hub,
control panel wire,
PR x PCH
hub, R25AL
N/C,
R97B, RlOOBU, through CB's, to post 1.
For standardization it is recommended that the
X
identifying negative amounts always be punched
in the units column of the field. In this manner
a negative amount can always be spotted
a-t a
glance.
ELECTRICAL PRINCIPLES
49
Sign control checking is accomplished by comparing X-punchings in the multiplier and multiplicand fields against X's in the product field. If
the factors indicate a negative product and the
product is not so identified (or vice versa), the machine will stop and indicate an error.
The only additional control panel wiring necessary for a checking operation is the wiring of the
card column
con,taining the product X to the
PR
x
CHK
hub from the first set of brushes. The X's
for the multiplier or multiplicand are sensed
through the
MPX
or
MCX
hubs. However, when
checking, the factor reversal switch is
ON
and R25
is picked up; consequently the
MPX
and
MCX
hubs are reversed so that a mul'tiplier X picks R95
and
a
inultiplicand X picks R96.
In order to set up a given problem, assume an
X punch in the multiplicand field only. The single
X means that the product is negative and that an
X should be punched. If for some reason rhe pro-
duct X is not punched, the machine must detect
this
2s
an error.
The sequence of events involved in a sign check
.Ire shown on the chart in Figure 5 1B.
Note that
when an error is sensed, the machine is not stopped
until the end of the cycle during which the card
in
error passes the second set of brushes.
The circuits are completed in the sequence in
which they are traced below. Remember that
R25
is energized during a check run and rhat RlOO
picks up at X time of each cycle under control of
P3 7.
When the card passes the first set of brushes,
the multiplicand X is read and R96 picks as follows: post 3, through fuse 8,
R96P coil, Rl OOAL,
R25BL
N/O
MCX
hub, control wire to hub of
read for entry and control brushes, individual
brush, X hole, contact roll, common brush,
R8B,
CB's, to post 1. R99 would be energized if the pro-
duct X were punched. Since the assumption is
that no product X is punched, R99 is not picked
up. R96 holds through its
R96H coil, R96AU
points, and P39 until 5 of the nexr cycle. R96B
points set up the circuit to R97P, which picks as
follows when P40 makes: post 3, through fuse 8,
R97P coil, R96B
N/O,
R95B
N/C,
R99B N/C, P40,
to post 1. R97 holds through its R97H coil, R97A7
and P41 until 11.5 of the same cycle. R97B sets
up a circuit to
R98P, so that at X time of this
cycle, R98 is energized as follows: post 3, through
fuse 8,
R98P coil, R25AL
N/O,
R97B, R100BU7
through CB's, to post 1. R98 holds through its
R98H coil, R98AL, and P39; R98AU points set
up a circuit to
Rl OlP so that RlOl picks when
P40 makes.
RlOl then holds until 11.5 through its
RlOlH coil, RlOlA, and P41. RlOlB points set up
a circuit to energize the error
relay at X time as
follows: post 3, through fuse 7, R57P coil, R31B7
to R3 8A
N/O,
through RlO lB (Section 14B), R-
100BU, CB's to post 1.
When
R57 picks, its R57B points open, causing
the machine to stop as previously explained. Also,
the error
light will glow to indicate an error.
Observe
tha,t if the X to identify a negative
product had been punched, R99 would have
picked
and held through its R99H coil, R99A, and P39;
then when P40 made, R99B would be transferred
along with
R96AL, and R97 could not pick up. If
R97 does not pick up, nothing further happens in
the checking
opera,tion. The R99B points in combination with the R95 and R96 points provide a
means of completing a circuit to R97 for proper
X
punch and X-check control.
Since it is recommended that X's be punched
over the units column of a field, it is necessary ro
suppress the X check when using the double punch
check. This may be done by wiring the units
column through the column
split, so that only the
0-9 positions of the card column are checked for
double punch.
Class Selectors (Optional)
Class selectors are optional features on this ma-
chine.
If one class selector is, installed, relays
SO
through 86 are used. The second class selector
50
TYPE
603
ELECTRONIC MULTIPLIER
RI~O~"P
37
R80PU-CBIX hole
in
cord)
".'-
- - - - - - -
-------
--
=-t
.PI
R 8lPU -'4'4~22
'i
I..O
RE1
H
I
'r4p23
I
1s.e
CS
Rsloys PU R81 AUIPU
Hub
1)
R84,85,86
R 82
H
""
Claaa Salactor
I
-
wlrad to
Hub
I
ia Tranrtam
I
Tronatwrad
Figure
52.
Sequence of Operation of Class Selectors
uses relays 87 through 93.
The circuits will be
traced for number
1 class selector.
The sequence of operation of the selector cir-
cuits is shown in Figure
52. This sequence should
be carefully studied in connection with rhe circuits to enable a clear understanding of the circuits.
The selectors are arranged for either X or
D
pickup and for normal or delayed operation.
If
it is desired to transfer the selector during the cycle
following the reading of the X, the
c
PLG
hub is
wired to the
1 hub. Wiring the
c
PLG
hub to the
2 hub causes the selector to rransfer during the second cycle following the reading of the X hole.
When an X is sensed at the first set of brushes,
a circuit is completed to
R80P through the X pick-
up hub and
R15AL.
R15AL allows only X's to be
recognized by the X hub.
If digit control is de-
sired, the
D
hub is used to pick R8O. R8O holds
through its
R80H coil, R8OA, and P21 until 14
of the cycle following its
pickup. At 13.6 of this
same cycle,
R81P is energized through R80B and
P22. R8 1 holds through its R8 lH coil, R81AL,
and P23 until near the end of the cycle.
Then if
the selector is wired to operate on this cycle,
R84,
R85, and R86 are energized through the R8lAU
points during this entire cycle. The points of R84
through R86 are the actual selector points, as indicated by the chart on the wiring diagram.
If the selector is wired to operate during the
second cycle following the X, R83 must
be
ener-
gized before
R84, R85 and R86 can pick up. At
12.2 of rhe first cycle following the X, R82 is
energized through
R8 1B and P28, and a holding
circuit is completed through
R82H coil, R82A,
and P25 until 14.5 of the second cycle. At 13.6
of the second cycle, R83 is energized rhrough
R82B and P22. R83 holds through the entire second cycle through its R83H coil, R83A points,
and P23. R83B points complete a circuit during
this cycle to energize the class selector relays R84
through R86.
Distributor (Optional)
A convenrional 12-segment distributor can be
installed as an optional feature on this machine.
The distributor can be used as a digit emitter by
wiring the CB hub to the C hub of the distributor.
ELECTRICAL PRINCIPLES
5
1
Any timed impulse from 9 through 12 is available
at the distributor hubs labelled accordingly. The
distributor can also be used as a digit selector by
wiring from the brushes to the C hub of the distributor. Only the desired digit in any card column can be recognized by proper wiring of the
9-12 hubs of the distributor.
POWER SUPPLY CIRCUITS
Main Transformer and Selenium Rectifiers
Power is su2~lied to this machine by two transformers, the main transformer and the tube power
supply xransformer, which can be seen in the
power
supply circuit diagram (Section 3 for 11
5
-
2 3 0 volt A.C. operaxion and Section 4 for 1 1 5 volt
A.C. operation). The main transformer is mounted
directly behind the control panel. The only difference between the
11 5 volt A.C. xransformer
and the
11 5-230 volt A.C. transformer is that the
latter has two primary coils. All following discussion will assume a 11
5
volt A.C. transformer.
Three taps are provided in xhe primary to permit
constant secondary output voltages at 105,
11
5,
or 12 5 volt primary voltages. The secondary is
provided with six taps to
supply 115 volts A.C.
for the tube
filamenxs and to supply the selenium
rectifiers which provide D.C. outputs of 40 volts
and 140 volts. The 140 volt supply is the anode
voltage supply for the
25L6 read-out power tubes
in the electronic
compu~ing unit. The 40 volt
supply furnishes the power for all relays and magnets in the punch unit which are usually supplied
by a generator.
From the punch running circuits it will be re-
membered that when the main line switch is
ON,
the A.C. from xhe main line appears across posts
9 and 10 to pick up HD3 and HD4.
The HD4
points in turn complete a circuit to the main transformer primary across the two
20 amp fusetrons.
The circuit through the primary of xhe main line
transformer is from post 8, through 20 amp fusetron 3, HD4 points, primary terminal 17, primary
coil, terminal 19 (for
11 5 volt supply), fusetron
4, to post 7. The RMS or effective
volxage pro-
duced at
fz~ll load
by transformer action between
each of the secondary taps is shown in Figure
53.
(RMS is the abbreviation for the root mean square
value of alternating current or voltage, also called
"effective value." It is the square root of the mean
values of
xhe squares of the instantaneous values
taken over a complete cycle.) All tube filaments
in the electronic computing unit are heated by the
11 5 volt A.C. between terminals 13 and 14. The
filamenx supply uses cable connectors CNp18
through CNp20 and CNp38 through CNp40.
Three connectors are used on each side to distribute
the load. It is advisable to adjust the primary tap
so that xhe filament supply is approximately
110-
11 5 volts. Although the tubes will require slightly
longer to heat, their life will
be considerably
longer. A
1
5
%
increase in voltage will reduce
the life of the tubes 30%.
40 volt D.C. is supplied
by the full-wave selen-
1s rec-
ium rectifier across terminals I
I and 12. Th'
tifier is mounted on the
right end gate. The cenxer
connection of the selenium rectifier is the positive
connection while
xhe transformer center tap be-
tween terminals
11 and 12
(terminal 15) is the
negative connection. Terminal
15 serves as the
reference point for
all
voltages, i.e., it is the point
of zero reference and is the ground connection.
Four
2000 mfd, 40 vol~ electrolytic capacitors connected in parallel act as the filter for the 40 volt
supply. These capacitors are mounted on the
lower section of the left end gate. Post
1 and post
3, which are the 40 volt terminals, are connected
across the
2000 mfd capacitors.
The actual D.C. voltage produced at the output
terminals is greater than the A.C. voltage at the
xransformer secondary because the capacitors
charge to peak voltages. The actual D.C. voltage
is equal to the
peak voltage of the A.C. (42 times
the
RMS voltage) minus the voltage drop across
the selenium rectifier stack, which is approximately
6-10 volts at full load.
52
TYPE
603
ELECTRONIC MULTIPLIER
White
I
I
I
I
I
I
I
I
I
I
1.0 Amp.
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
Block
I
AII valtogee ore
RMS
voluer at full load
-
-
-
Figure 53. Main
140 volt D.C. is supplied by the half-wave selen-
ium rectifier connected to transformer terminal 16
through R4B contacts. The purpose of R4 was
discussed in connection with the start and bias in-
terlock circuits. A half-wave rectifier is sufficient
for this supply because of the very light load. The
two 200 mfd
200 volt electrolytic filter capacitors
in parallel are able to maintain a
fairly constant
supply with the light load on this line. A
5000 ohm
bleeder resistor is connected across the capacitors
to maintain a minimum current flow and thus prevent the high peak voltages which would exist at
very light or no load. With this bleeder resistor,
better voltage regulation results. All the components for the plus 140 volt
supply are mounted on
the left end gate.
Tube
Power
Supply Chassis
The tube power supply chassis contains all components enclosed in the dash-dotted section in Section
4 of the main wiring diagram. The tube
power
supply transformer for 11 5 volt A.C. oper-
ation is provided with taps on the primary in
5-
Transformer
volt steps to permit constant secondary output
voltages at primary
voltages of from 100 volts to
12F volts. The taps are connected through a tap
switch to post
CHI0 and to post 10. The extreme
right-hand setting is the
100 volt tap. When the
tap switch is set to the extreme left-hand setting,
the transformer primary is disconnected. The tap
switch is mounted on the power supply chassis and
is accessible from the front of the machine
by low-
ering the left front cover.
Note that the circuit to the transformer pri-
mary passes through the
HD2 contacts. The operation and purpose of this relay was discussed in
connection with the time delay circuits. The circuit through the primary is from post
9,
through
HD2 contacts, post CH7, ~rimary coil, tap switch,
post
CH10, to post 10. The secondary of the
transformer is in two equal sections, one delivering
3
88
volts RMS approximately, and the other ap-
proximately 364 volts RMS. These values vary
with the setting of the tap switch. Each section
is centered-tapped to allow full-wave rectification.
The upper section shown on the circuit supplies
ELECTRICAL PRINCIPLES
53
+
15 0 volts for the anode voltage and
-2
5
0 volts
for bias voltage, while rhe lower section supplies
-100 volts for bias voltage.
A
voltage replating
circuit is provided between the + 1 5 0 volt and the
-100 volt lines to maintain a constanr ratio between anode and grid voltages. This circuit will
be discussed shortly.
The tube power supply transformers
along with
the rectifiers and chokes are shown in Figure
54.
Note that all transformers are supplied with
grounded electrostatic shields, indicated as
dotred
lines between the primary and secondary windings
in Figure
54. The purpose of these electrostatic
shields is to prevent any high frequency transients
in the
A.C.
line from being passed to the secondary
through the capacity coupling between windings.
With these shields any transients are by-passed to
ground.
Full-wave rectification to obtain
+
150 volts and
-100 volts is accomplished by directly heated, pas
filled, twin diodes. Heater current is supplied to
both rectifier tubes
by a separate filament trans-
former as shown in Figure
54. The chokes are connected to the center raps of the corresponding filament transformer secondaries to provide a more
nearly constant output.
Figure
54.
Tube
Power Supply Transformers, Rectifiers, and Chokes
54
TYPE
603
ELECTRONIC MULTIPLIER
The anode supply rectifier is a type EL-3C di-
rectly heated twin diode manufactured
by Elec-
trons, Inc., of Newark, N.
J.
The
-100
volts bias
supply rectifier is a type EL-IC.
Both are xenon
filled and have an internal drop of approximately
12
volts when conducting. Characteristics of these
tubes are shown below:
Type EL-IC
DC Current (Maximum Rated Output)
Continuous 1.0 ampere
Surges
(3
seconds or less)
1.5 amperes
Oscillograph peaks 4.0 amperes
(continually recurring)
Peak
Inverse Voltage 725 volts
(maximum instantaneous)
A.C. Volts per Anode
12-250 volts
Tube Drop 12 volts
Filament Voltage 2.5 volts
Filament Current 6.5
30.5
amperes
Type EL-3C
2.5 amperes
3.7 amperes
10.0 amperes
725 volts
12-250 volts
12 volts
2.5 volts
11.5
31.0
amperes
To provide rhe best voltage regulation, choke
input filters are used in the
+I50
volt and
-100
volt lines. The ratings of the chokes and capacitors are shown on the circuit diagram and on Figure
54.
The chokes
L1
and
L2
are mounted beside
the power
supply transformer. Choke L3 and all
capacitors are mounted under
rhe chassis and are
accessible only when the chassis is removed.
The purpose of the
0.1
mfd capacitor across the
electrolytic capacitors is to protect the
electrolytics
by by-passing high frequency rransients. Electrolytic capacitors
by their nature do not readily pass
high frequencies.
Note that the output across the filter capacitors
of both
rhe
+I50
volt line and the
-100
volt line
is
150
volts
D.C.
This is arrived at by subtracting
the voltage drop across the rectifier tube and across
the choke from the RMS voltage across one-half
the secondary and converting rhe resulting
RMS
voltage to
average
voltage, since the filter capaci-
tors charge to the
average
voltage when a choke
input filter is used. (RMS voltage
=
0.707
peak
voltage, and average voltage
=
0.636
~eak volt-
age.) To obtain
100
volts for the
-100
volt line,
the
voltage ratio regulator circuit provides a regu-
lated
50
volt drop.
Btforc discussing the voltage replator, the rec-
tifier circuit for the
-250
volt bias supply will be
discussed. The load on this line is very light and
consequently a vacuum tube is used. The tube
used is a type
2526
twin diode. Note in Figure
5 5
rhat the cathodes of the
2526
are connected
across the upper section of the transformer secon-
dary. It is necessary to reverse the connections in
this case because the output from this rectifier
must be
negarive with respect to ground, and the
same common ground as the
4-1
50
line is used.
Since the load is very light on this line (approximately
10
ma), satisfactory voltage regulation can
be obtained with a capacitor input filter. In rhis
manner a higher output
voltage can be obtained
since the capacitors charge to peak voltage. This
is necessary in this case because only about
190
voks RMS are available at the transformer yet
250
volts
D.C.
are needed at the output. The
50,000
ohm resistor across the
16
mfd electrolytic capaci-
tors serves the usual function of a bleeder resistor,
i.e., .to maintain a minimum current and thus prevent excessively high peak voltages across the capacitors. All the components for this rectifier are
mounted under the chassis and are accessible only
by removing the chassis.
In order to provide an adjustment of the
-250
volt supply, a rheostat is mounted between the anodes of the
2526
and the choke L3. By adjusting
the amount of voltage drop across rhe rheostat, the
output can be maintained at
-250
volts. The rheo-
stat adjusting knob is accessible
through a hole at
the rear of the power
supply chassis.
Constant Ratio Voltage Regulator
The purpose of rhe voltage regulating circuit in
the
-100
volt line is to maintain a constant ratio
between the anode voltage and rhe grid voltage. It
is very important in the operation of the electronic
computing section that the same relation be main-
tained between the anode voltage and the grid
voltage.
If
the anode volrage drops sharply owing
to a suddenly applied load, the bias
voltage must
ELECTRICAL PRINCIPLES
55
+&
161nfd 16mfd
each
-
2526
0
6
5,OOOn.
30
Millihenries
Figure
55.
Schematic of
-250
Volt Rectifier Circuit
be made less negative to compensate for this. Any
increase in anode voltage requires a more negative
bias voltage. If the
A.C.
supply voltage changes,
both anode and grid bias voltages change, and the
regulation circuit merely adjusts for the proper
ratio between the two.
Note that a resistor network consisting of a
36,000 ohm resistor, a
10,000 ohm potentiometer,
and a
20,000 ohm resistor is connected between the
+I 50 volt and the -100 volt line. Also note that
a 10,000 ohm resistor by-passed by a
0.1 mfd cap-
acitor is connected across the
-100 volt terminals.
The
10,000 ohm resistor serves as a bleeder for the
-100 volt line, and the 0.1 mfd capacitor serves as
a by-pass across the
-100 volt line for high fre-
quency transients. The regulator tubes consist of
three type
25L6 beam power tubes connected in
parallel between the filter output and the ground
line.
The bias voltage is originally adjusted at
-100
volts by the potentiometer.
The adjusting knob
of the potentiometer is accessible from the rear of
the machine through a hole in the rear of the
chassis. The 3,300 ohm screen grid resistors and the
10,000 ohm control grid resistors are both current
2 5L6's. Note thax the heater filaments of all 2 5L6's
along with the heater for the 2526 rectifier are
connected in series across post
12 and post 13. 100
volts
A.C.
is available across these terminals of the
main transformer. All
componenxs of this regulator circuit except the tubes are mounted under
the chassis and are accessible only if the chassis is
removed.
In order to simplify the explanation of the regulator circuit, a schematic drawing showing only
the essential elements is illustrated in Figure 56.
The output from the filters is represented by bat-
teries of the corresponding polarity, and the three
25L6's have been replaced by a single triode. In
this circuit the tube acts as a variable resistance
controlled
by the grid voltage. The value of tube
resistance is varied so as to maintain the proper
voltage across the
-100 volt line at all times. The
voltages existing, together with
polarities, are
shown in Figure 56 when all conditions are normal.
The voltage drops across
R1, P, and R- can be com-
puted by determining the current flow through
the resistors, then calculating the IR
drop across
each. Or the voltage drops may be computed by
the ratio of resistors as follows:
limiting resistors and do not enter into the opera-
E
=
2$0
X
R1
,
etc.
tion of the regular circuit other than to stabilize
R
1
RI
+
P + R?
the tube operation.
Observe that the screen grid
For sake of illustration, it is assumed that the
resistors are tied to the
+I50 volt line. This ar-
tap on the potentiometer is set so that there is a
rangement provides increased current through the
95 volt drop from the -100 volt line to the grid.
56
TYPE
603
ELECTRONIC MULTIPLIER
-
-v
-400
Figure
56.
Schematic of Voltage Ratio Regulator Circuit
Because of the polarity of the IR drop across the
resistors, this
95
volts can be subtracted from -100
volts to obtain the grid potential (cathode is at
ground potential, or at zero potential). Th'
is means
that
sthe grid is at
-5
volts, and sufficient current
flows through the tube to provide a
50 volt drop
in the line. This 50 volts subtracts from rhe bat-
tery voltage of
150 volts and leaves 100 volts across
the
-100 volt line. As long as conditions remain
normal, nothing further happens.
Assume now that a heavy load is suddenly ap-
plied across the
+I 50 volt line. The tendency will
be to drop the voltage across this line. The job
of the regulator is then to make the
-100 volt line
less negative to compensate for the drop in anode
voltage. For sake of analysis assume
%he anode
voltage drops to
+I30 volts as indicated in paren-
thesis. Then the new voltage drops across the resistors on the basis of 230 volts across the network
are shown in
parenthesis. This change will produce
a voltage of -12.5 volts at the grid (100 - 87.5),
and the current through the tube will be consider-
volt line, thus producing the desired result.
Another way of looking at this is that the decreased
tube current results in a decreased IR drop across
R3 which is across the -100 volt line.
Of course the
opera'tion will be reversed if the
anode voltage increases.
If both anode and grid
voltages change owing to variations in the A.C.
supply, the regulator again adjusts for the proper
ratio between the two. An analysis in this case
is more difficult, but the end result is the same.
The circuits shown outside the chassis on the
main wiring diagram have already been discussed
in connection with the punch unit circuits,
Screen Grid
Supply
Although it is not included in the power supply
chassis and not shown in the punch circuit diagram, there is a separate
+65 volt supply for the
.
screen grids of all tetrodes and penrodes in the
electronic computing unit. This power supply is
located on the B chassis (Section
30B), and the
circuit is shown in Figure
57. It was necessary to
ably decreased.
The decrease in current means an
place this power supply on the B chassis because
increased voltage
drop across the tube.
An
in-
the original design did no% call for a +65 volt sup-
creased tube
drop leaves less voltage across the -100
ply.
When it was decided that this supply was
ELECTRICAL PRINCIPLES
57
necessary, there was no room on the power supply
chassis for the additional tubes and components.
The
+65 volt supply is in reality a voltage regulator circuit. Since +I 50 volts were already available in the machine when this supply was added,
the easiest method of obtaining
+65 volts was by
means of a voltage regulator circuit which provided an 8
5
volt drop across the 1 50 volt line. This
provided a 65 volt supply be,tween the gound line
and the cathodes of the regulator tubes. In order
to provide a constant voltage, a voltage regulating
circuit under the control of a
6SJ7 penrode is pro-
vided. The operation of this circuit follows:
A
study of the circuit in Figure 57 will reveal
that four type
25L6 beam power tubes are con-
nected in parallel between the
+I 50 volt line and
the
+65 volt line. Four tubes are necessary to
handle the current requirements of this supply.
Obviously, there must be an 85 volt drop across
these tubes. Since the
D.C.
voltage drop across a
vacuum tube varies with the amount of current
passing through it, it is necessary to regulate the
amount of current so that the proper anode current flows at
all times to maintain this 8 5 volt drop.
This is done by adjusting the grid potential of the
25L6's.
Note that all the grids are connected to-
gether to the anode of the
6SJ7 pentode.
Any
change in anode current through the 6SJ7 will produce a change in potential at its anode owing to
the change in IR drop through the anode resistor.
Any change in potential at the anode of the
6SJ7
will be felt at the grids of the 2 5L6's. Consequently, by adjusting the current flow through the 6SJ7,
the current flow through the 25L6's can be ad-
justed. This provides a means of obtaining very
close control, since the
6SJ7 has an extremely high
amplification factor.
A
very slight change in grid
potential on the
6SJ7 will produce an appreciable
change in grid potential at the 25L6's.
The grid of the 6SJ7 is normally biased at ap-
proximately
-1 volt by means of the potentio-
meter.
This setting
provides the proper potential
on the 25L6 grids to provide +65 volts between
ground and the
25L6 carhodes. This 65 volt po-
tential is maintained at a constant value as follows:
The voltage divider R4 and R:, provides a volt-
age drop of 19 volts across
R5, thus placing the
cathode of the
6SJ7 at $19 volrs. Between the
gound line and the +65 line there is a resistor net-
work consisting of resistors
RI and R2 and poten-
58
TYPE
603
ELECTRONIC MULTIPLIER
tiometer
P.
The voltage at rhe top of each com-
ponent is shown in Figure
57. As previously shown,
these values are easily
computed by the ratio of
resistances. The voltage at the top of
Rz is +12.5
volts and at the top of P is + 23 volts; hence by
adjusting the potentiometer tap, the grid can be
set
anywhere from $12.5 volrs (-6.5 volts with
respect to cathode which is at
+
19 volts) to +23
volts
($4
volts with respect to cathode). This ad-
justment
is
made only once when the 65 volt sup-
ply
is
being adjusted to the line voltage. Thereafter
adjustment
is
automatic. Assume, for example,
that the +I 50 volt line
.tends to drop in voltage.
This will in turn reduce the voltage across the 65
volt supply, meaning a lower voltage
drop between
ground and the tap on potentiometer
P.
Less volt-
age at this tap means a more negative grid on the
6SJ7, since the cathode remains at essentially +19
volts at all times. The more negative grid means
a decreased anode current with a consequent reduction
in IR drop across R3, the anode resistor of
the
6SJ7. A lower IR drop across Ra means a less
negative voltage at the anode of the
6SJ7. This
is
the voltage of the 25L6 grids; hence the 25L6
grids will be less negative than before and more
anode current will flow with
athe consequent re-
duction in
drop across the 25L6's and increase the
65 volt line.
In this manner the 65 volt
supply
is restored to its proper value.
Any change in the opposite direction will make
the
6SJ7 grid more positive which, in turn, will
make the
25L6 grid more negative. This will in-
crease the
drop across the 25L6's and again 65 volts
will be maintained across this supply.
The
100 ohm screen grid resistors in rhe 25L6
circuits are current limiting resistors. Low values
were selected to allow maximum current without
damage to the screen grids. The 10,000 ohm control grid resistors are also current limiting resistors.
These resistors also provide more stable tube operation. The
8
mfd capacitor across the 65 volt line
serves as a filter with the
10,000 ohm resistor
Re
as its bleeder. Rc, serves the purpose of a dummy
load in case the operating load is removed. An
extremely high resistance
is
used at R3 in the
anode circuit of the
6SJ7 to provide very close
control. A very slight change in anode current
produces a large change in voltage
drop across R3.
For example, a
I
micro-nrrzpere increase in anode
curren,t increases the IR drop across R3 by 5 volts.
NOTE:
The 6SJ7 may be replaced by a 6SK7 in an
emergency.
A
slight readjustment of the
potentiometer may be necessary.
On some early models the 12,500 ohm
resistor, which is part of the voltage divider
which establishes the cathode
potential of
the
6SJ7 tube B-26 (Section 30B on wiring
diagram) is a
0.1 watt carbon resistor. The
load across this resistor is considerably more
than
0.1 watt, consequently this resistor
overheats and changes in value, thereby
causing fluctuations in the
$6 5 volt supply.
This carbon resistor should be replaced by
a 12,500 ohm,
5
watt, wire-wound resistor
as indicated on the wiring diagram.
Voltage Adjustments
When it is desired to adjust the voltage supplies,
a definite procedure should be used to insure thar
one change does not
affkct the other. First, the
fiIament voltage should be checked for 110-11
5
volts. If a change
is
necessary, the voltage tap on
,the main transformer primary can be moved to
either post
18, 19, or 20. The series filament
strings are designed for approximately
11 3 volts,
but a slightly lower voltage will considerably in-
crease the life of the tubes.
However, a slightly
longer time must be allowed for
the tubes to heat.
Changing the primary tap on the main transformer
will
change the voltage output from the selenium
rectifiers slightly, but these voltages are not critical and are usually higher than necessary.
The first adjustment in the tube power supplies
is the
$110 volt supply. This output is adjusted
by varying the position of the tap switch in the
primary of the power
supply transformer. By
ELECTRICAL PRINCIPLES
59
varying the position of the tap switch, the voltage
across the secondary is changed. Since the voltage
drop across the tube and the choke does not vary
to any extent, any change in the secondary volt-
age will result in a change in the
+I50 volt line.
147 to
152 volts are acceptable.
After the
+I 50 volt line is adjusted, the bias
voltages may be adjusted in any order. The
-100
volt line and the -250 volt line are both adjusted
by means of potentiometers. Both potentiometers
are accessible from
athe rear of the power supply
chassis. The
+65 volt supply is also adjusted by
means of a potentiometer in the B chassis.
PRINCIPLE
OF
MULTIPLYING
Multiplication is performed in this machine by
over-and-over addition of the multiplicand factor.
The value of the multiplier digit determines the
number of times that the multiplicand is to be
added into the product counter. The order of the
digit
(i.e., units, tens, etc.) determines which posi-
tions of the product counter will receive the mul-
tiplicand factor and also during which column
shift cycle the adding will take place. An example
of multiplication
by over-and-over addition is
shown in Figure
58. Observe that no adding oc-
curs in the
12th position of the product counter;
this position is reserved for carry-overs. This
method of multiplication is the simplest possible
since only three counters are required. This
mehod
is not feasible for mechanical machines, because
the time required to complete a problem is too
great. However, electronic addition is so rapid
that it permits use of
rhis simple method. The time
required
by the electronic unit to complete a problem by this method corresponds to slightly over
one-half a cycle point of mechanical movement of
the card through the punch unit.
Observe that the column shifting is performed
in the reverse order to that customarily done when
multiplying by hand. The order of column shifting is immaterial, as the final result will be the
same regardless of rhe order in which the multiplier
factors are taken.
In the 603 multiplier, column
.
shifting is from the 6th multiplier position to the
first simply because the earliest engineering model
required it. When changes were made which no
longer required reverse column
shihing, it was felt
that it was not feasible to make the necessary circuit changes to establish normal column shifting.
Note from Figure
58 that no addition occurs
during column shift cycles in which rhe multiplier
is
0.
Mechanical machines usually have some ar-
rangement to test for
0's so that all multiplier posi-
tions containing
O
may be slripped. This compli-
cates the machine but saves time. Where time is
not a factor, this unnecessary complication may be
eliminated
by completing all the column shift
cycles regardless of the value of the multiplier
digit. This is done in the 603 electronic comput-
ing unit; six column shift cycles are taken for
every multiplication regardless of the value of the
multiplier.
Also observe from Figure
58 that the multipli-
cand is added into the product counter as many
times as the value of the multiplier digit by which
multiplication is taking place. This means that
during any given column shift cycle, the multi-
plicand may have to be added as many as
9
times.
Again, to avoid unnecessary complications a fixed
number of adding cycles are taken each column
shift cycle regardless of the value of the multiplier digit.
In the 603
compu~ing circuits ten
adding cycles are taken during each column shift
cycle rather than nine for reasons explained later.
This means that each multiplication consists of six
groups of ten adding cycles, or a total of 60 add-
ing cycles.
The computing section of the 603 multiplier is
arranged so that a definite number of voltage
pulses represent an
adding cycle. For an analogy
to a mechanical machine, one pulse may be con-
sidered as one cycle point of the basic adding cycle.
The basic
adding cycle consists of 16 pulses; 10 are
reserved for adding,
4 for carrying, and the other
60
TYPE
603
ELECTRONIC MULTIPLIER
0
COU8.S
cTumn ahifl
Only
Product
Figure
58.
Principle of Multiplication
2
for setting up certain circuits.
It is obvious then
that one multiplication will require a total of 960
pulses, since there are 60 adding cycles of
16
pulses
each. In order to establish these cycles electronically it is necessary to provide
pulse counters.
Three of these pulse counters are
provided--one
to count pulses and establish the basic adding
cycle; one to
c0un.t adding cycles and determine
when to column shift; and one to count column
shift cycles. These three
pulse counters are called
the primary timer, secondary rimer, and tertiary
timer. Figure
59 is a block diagram showing the
operation of the three timers in schematic form.
A careful study of Figure
59 is essential for proper
understanding of the electronic cycle. All pulses
for rhe electronic computing operation are generated
by a multivibrator which is an oscillator gen-
erating essentially square waves of
3
5,000 cycles
per second frequency. The multivibrator generates pulses continually as long as the power is
ON,
ELECTRICAL PRINCIPLES
6
1
Pulse
Generator
Multlvibrator
P-w
dl
Cam Contact
Compute Start Pulse
start- Stop
Compute Stop Pulse
after 6th column shift
(
960th pulse)
Counts pulses
from MV to estoblish
basic adding cycle
of
16 pulses
I
Primary
Timer
Counts adding
cycles to establish
column shift cycle
of
10
adding cycles
Counts column shift cycles
Advances column shift and
causes computation to stop
ofter completing 6th Column
shif t
Carry pulse
after 16 pulses
Figure
59.
Block
Diagram
of
~ikrs
but these pulses cannot pass to the timers until per-
As an example of the multiplication operation,
mitted by the electronic start switch.
The start
4
x 8 may be used as a problem. The multiplier
switch is controlled
by an electronic start and stop
4
indicates that the multiplicand 8 is to add into
control which opens the switch and closes it.
For
the product counter
4
times.
The 8 in the multi-
an
analogy, the electronic switch may be visual-
plicand counter indicates that the product counter
Secondary
Timer
ized as a valve. An open switch permits pulses to
pass while a closed switch prevents passage of
pulses. After the factors are read into the counters
from a card passing the entry brushes, the compute
start contact
P24
makes (at 11.5 on the punch index) to start computing. This is done by opening
the start switch and
permiming pulses to pass to the
electronic timers (or
pulse counters). On every
16th pulse entering the primary timer, a carry
pulse is
passed to the secondary timer which ad-
vances
1 to indicate that one adding cycle has been
completed. When
10 adding cycles are completed,
the secondary timer carries, and a pulse is passed
to the ,tertiary timer to
signal a column shift.
When 6 column shift cycles have been completed,
a pulse from the tertiary timer passes to the compute stop control, and the start switch is again
closed, thus stopping the pulses from passing to the
timers after 960 pulses have been counted and indicating the completion of the
problem. This entire
operation requires only
27
milliseconds during
which time the punch index moves 6 teeth.
is to receive
8
impulses during each of four adding
cycles for a total of
32
pulses.
Figure 60 indicates the machine operation in
schematic form. Since the only digit in the multiplier is in the units position, not until the sixth
column shift cycle will there be signals to cause
addition in the product counter. Then during the
sixth column shift,
8
pulses are added into the
product counter during each of the last four adding cycles. Observe that the machine goes through
all adding cycles and all column shift cycles even
though addition occurs only during the last four
adding cycles of the last column shift cycle.
For a more complete illustration, the same prob-
lem shown in Figure
58 is worked out in detail in
Figure 61, showing all the
adding cycles through
which the machine goes to complete one multiplication. Notice that
during the entire first column
shift cycle consisting of ten adding cycles no adding takes place because there is a
0 in 6th position
of the multiplier. During the second column shift
cycle the multiplicand is added into the
product
-
Carny Pulse
otter
10
adding
*
cycles (160puIses)
Tertiary
Timer
62
TYPE
603
ELECTRONIC MULTIPLIER
Multlpl~cond Counter
654321
00000e
Mull~pller Counter
654321
000004
oddlng cycles
lo
I
1st Column Sh~ft Pos~t~on
of t6 pulses No oddit~on
10
~n 6th pos~t~on MIPller)
each
2nd Column Sh~tt
No oddltlon 10 In 5th posltlon M'Pl~er)
3rd Column Sh~tt
No oddlt~on
(
0
In 4th posltlon MIPller)
4th Column Sh~ft
No oddltlon (0 In 3rd poslt~on M'Pl~er)
5
th Column Sh~fl
No oddlllon
(
0
In 2nd pos~t~on M'Pl~er
1
6th Column Shltt
8
pulses added In Products Counter 4 ttmes
(
4 In Is1 porlllon MBPller)
Figure
60.
Block Diagram
of
Multiplication
by
Over-and-Over Addition
counter only during the last four of the ten add-
advanced
1
during each adding cycle. Since there
ing cycles, because the multiplier digit in the 5th
are ten adding cycles in each column shift, each
position is a
4,
etc.
position will advance through
O
and back to where
If half-entry is being used, a 5 is added into the
it started. When the counter position goes from
Proper Position of the Product counter during
the
9
to 0, a signal is provided to start adding the rnul-
first adding
the
last
tiplicand
on
the next adding
cycle
and to continue
Observe that except for half-entry, no adding takes
through the tenth adding cycle. Figure
61
shows
place in the products counter during the first
add-
this operation for all six column shift cycles. There
ing cycle of each column shift.
One reason for
are no carry circuits in the multiplier counter;
having ten
adding cycles in each column shift cycle
hence adding ten pulses to each position will return
is to permit half-entry during a cycle when no
other addition is taking place.
the counter to its original reading. The chief rea-
Figure
61
indicates how the machine determines
son for having ten adding cycles in each column
the number
of
times to add the multiplicand
shift instead of nine is to permit restoration of the
amount in the product counter during any given
counter
to its original reading at the completion
column shift.
One position of the multiplier is
of a multiplication.
This is necessary when group
Multiplicand Multiplier
413079
Product Counter
12111098765432I
0 0 0 0 0 0 0
2
000000000000
3
000000000000
4 000000000000
5 000000000000
0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0
8 000000000000
9 000000000000
I0 000000000000
I 000000000000
2 000000000000
3
000000000000
4
000000000000
5 000000000000
6
000000000000
7 004130790000
e
00826 1580000
9 01 2392370000
10 01 6523160000
I
01
65231 60000
2 01 65231
60000
3
01 65231 60000
4
01 65231
60000
5 01 65231 60000
C
S
6
016523160000
7 0165231 60000
8 01 65231 60000
9
016523160000
I0 016523160000
01 65231 60000
0652360000
01 65231 60000
01 65231 60000
4 th
01 6564467900
C.
S.
01 6605775 800
01 6647083700
8
016688391600
9 016729699500
I0 01 6771 007400
016771 007400
5th
C
S
6
016771007400
016771007400
e
016771007400
nolf entry
I
Cycle
-
<
:.ti.
1
a
9
I0
016771007400
J
9
I
-Start
0 1 6 7 7 5 I 3 8 1 9 0
o
'7(Stop
/~n
3rd pos
,-Of
6775 138690>
0
Figure
61.
Principle of Multiplication Showing Detailed Operation
64
TYPE
603
ELECTRONIC MLlLTlPLlER
1-
One adding cycle
-4
Pulse Number
(
I6 pulses counted by Primary Timer)
J-
Pulse to Multiplier Counter
0-
Pulse to Multiplicand Counter
X-
Pulse to Product Counter
Figure
62.
Schematic Showing Details of Operation
of One Position
multiplying, because the same multiplier is used
for a large number of multiplications.
For a more complete understanding of the multiplying operation, the last column shift cycle in
the
illustra~ion in Figure 60 is analyzed in detail
in Figure 62.
This schematic shows all that takes
place in one position of the multiplicand and product counters during one column shift cycle. All
column shift cycles are alike except for the digit
in the multiplier; consequently, a careful study of
Figure 62 should illustrate the principles employed
by the 603.
Observe that Figure 62 illustrates
10 adding
cycles (laid
OUT
horizontally) divided into 16
pulses each (vertical divisions).
Also observe that
the pulses are not numbered from zero; the first
pulse is number
IF. This method of counting the
pulses is used for the same reason that the first
division on
(the punch index after unlatching is
14 instead of
9.
A short time is necessary for setup
before addition starts, and it is desired to have the
pulse number correspond to the actual number of
pulses added into the mul'tiplicand counter.
As indicated in Figure 62, near the end of each
adding cycle one pulse is added to the multiplier
position corresponding to the column shift cycle.
This pulse is the number
12 pulse of each cycle
(14th pulse from the start of the cycle).
Since
there are ten adding cycles, the multiplier position
will return to its original reading in the tenth adding cycle. During the
6th adding cycle the mul-
tiplier position advances from
9
to 0. This provides
a signal to the multiplicand to add into the product
once during each of the
4 remaining adding cycles.
The method of transferring the
mul'tiplicand to
the product counter is also shown in Figure 62.
When the signal is received from the multiplier
counter, each position of the multiplicand counter
receives
10 pulses during
each
of
the
remaiming
adding
cycles
of ,the column shift cycle. This
means that each position of the multiplicand is
"rolled" once each adding cycle and returned to its
original reading.
Since there are no carry circuits
for the multiplicand counter, the counter may be
rolled as many
*times as desired without changing
the reading of the counter. When a position of
the multiplicand counter is rolled, it passes from
9
to O sometime during the adding cycle, as deter-
mined
by the value of the digit in rhat position.
When this occurs, it is a
signal for the product
counter
'to receive all the remaining pulses of the
group of
10. This efiectively transfers the multi-
plicand
digit to the product counter; this transfer
will occur once each time the multiplicand counter
is rolled. An examination of Figure 62
will reveal that in the illustrated problem the multiplicand is rolled
4 times, and rhe product counter
receives 8 of the 10 rolling pulses during each adding cycle that the multiplicand rolls. Since it requires
2 pulses to bring the multiplicand to 0, it
means that the
8
remaining pulses pass 'to the pro-
duct counter.
Figure 62 illustrates only one position of the
multiplicand counter. Remember that all six positions will be
rolling simultaneously during an ac-
tual multiplication. From Figure 61 it will be
remembered that the multiplicand is transferred to
the product counter many times
during the multi-
plication. However, each transfer is identical
ro
ELECTRICAL PRINCIPLES
65
Multiplicand
Counter Position
654321
413079
=
PYIses to M'Cond
524180 stort here
PuIasa lo Product Countrr
atorl otler M'Cand passes
from 9 lo
0,or indicoled
by double
lines.
(
No corry in M'Cand counlrr)
PuIres slop here
Tronataring 413079 from M'Cond lo Producl Counler
Figure
63.
Operation of Multiplicand Card Counter
During Transfer to Product Counter
the other except for the positions of the product
counter which receive the multiplicand.
Figure
63
illustrates in detail one adding cycle during
which the multiplicand is transferred to the product counter once. Observe that each position of
the multiplicand is rolled through
O
and back to
its original reading. In each position a double line
is drawn at the pulse number where that position
passes from
9
to
O.
The corresponding position of
the product counter will receive all
,the pulses retnaining of the lo rolling pulses. In this way six
positions of the product counter, as determined by
the column shift position, will receive the proper
number of pulses to add the multiplicand amount
in the product counter.
A block diagram showing all the operations taking place when multiplying in the computing section is shown in Figure 64.
All slow-speed pulses,
which travel through the cable either to or from
the punch unit in synchronism with the card
movement through the punch, are shown in heavy
lines to distinguish from the high-speed computing
pulses in the electronic unit. Dotted lines indicate
locking circuits, and
light solid lines indicate puls-
ing circuits.
The master timer for the computing section is
the multivibrator, which is a square wave oscillator
generating essentially square waves of
3
5,000 cycles
per second frequency. The clipper clips the multivibrator output to square the top of the wave,
and its output is fed to the "A" and
"B"
pulse
tubes which produce almost perfect square waves
180" out of phase with each other. In other words,
"A" pulses are going positive at the same time
"B"
pulses are going negative, and vice versa. These
pulses are used throughout the computing section
for high-speed computing.
"A" and
"B"
pulses
are being produced continuously as long as power
is on the tubes, but nothing happens until the start
switch is opened to allow the pulses to start a computation.
The reason for two pulse sources will
become apparent when the actual circuits are discussed.
The multiplier and multiplicand factors are read
into their corresponding counters from the card as
it passes the entry brushes. The principle used in
reading into electronic counters will be discussed
later in the circuit description.
After the factors are read into the counter, P24
cam contact makes
.to start the computation. P24
trips the start trigger and opens the start switch
to start the primary timer by feeding a stream of
"A" pulses to it. As previously explained, the primary timer is an electrical timer which counts 16
pulses to establish the basic adding cycle. The op-
eration of all computing circuits is based on this
timer. The primary timer counts 16 pulses and
then returns to zero to produce a 16-point cycle
effectively. Thus the
reading of the primary 'timer
is equivalent to the index used in the punch unit.
The primary timer in turn operates the secondary timer which in turn operates the tertiary
timer. For every primary cycle (or adding cycle)
the secondary timer receives one pulse and advances
one. After
10 pulses the secondary timer returns
to zero and starts over, at the
Sam.: time advancing
-I-
Llttrr8 in circlas rrirr lo
challis
Read-in and rrad-out pul8r8 in lynchronilm
with cord movrmrnt
-
High spcrd pulse8 originating at multivibrotor
------
Unlocking ond locking control8
Figure
64.
Block
Diagram
of
Type
603
Circuits
66
ELECTRICAL PRINCIPLES
67
the tertiary timer one; hence, it is obvious that the
tertiary timer advances once for each
10 primary
cycles. The secondary timer provides I0 adding
cycles in each column shift position while the ter-
tiary timer controls the column shift.
During each primary
cycle, a pulse
is
passed to
the multiplier counter by
,the multiplier advancing
control through the multiplying control switch,
thus advancing the proper position of the multi-
plier
1. The multiplier coun,ter position receiving
the pulse depends upon the column shift position
as indicated in Figure 64 by the dotted line from
the column shift control to the multiplying con-
trol switch. Thus, at the beginning of a computa-
tion the pulse is passed to the 6th position of the
=multiplier counter, since column shifting is from
left to right.
During the adding cycle that the
6th position passes from
9
to 0, a carryover pulse
through the multiplier output inverters opens the
10-pulse switch through the multiplicand output
control to start rolling the multiplicand counter.
For example, if the 6th position of the multiplier
contains a 3, this counter position would pass from
9
to O during the cycle in which it received its
seventh pulse from the multiplier advancing control. This
m'eans that there are 3 adding cycles
left in the first position of column shift, and the
multiplicand counter will be rolled three times in
this column shift position.
Rolling the multiplicand counter is accomplished
by adding
10 pulses to all positions of the multipli-
cand counter. Since there are no carry circuits for
the multiplicand (or multiplier) counter, after ten
pulses, each position will return to the same reading
from which it started; a signal carry pulse will be
provided from each position through
.the multipli-
cand output control when that position advances
from 9 to
O.
This carry pulse opens the column
shift switches which have been conditioned for the
first column shift position, and
"B"
pulses start
adding into the product counter.
The
10th rolling
pulse then closes the column shift switches through
the stop pulse control, and the product counter
will have received as many pulses as the reading of
the corresponding position of the multiplicand
counter. In this same manner the figures in the
multiplicand are rolled into .the products counter
three times during the first column shift cycle if
there is a 3 in the 6th position of the multiplier.
All carry pulses in the products
counter are stored
in the carry control section, and carry takes place
at the end of each adding
cycle (primary cycle).
When 10 primary (or adding) cycles have been
completed, the tertiary timer is signalled to advance. The column shift control then feeds pulses
to a new set of column shift switches to shift the
,
output of the multiplicand over one column to the
right in the product counter, and the 5th position
of the multiplier counter starts receiving advancing pulses. Again when a carryover occurs, the
multiplicand starts receiving
rolling pulses. After
this operation has been performed six times, rhe
computation is complete and the primary timer is
stopped.
Five pulses are added into the products counter
during the first adding cycle of the 6th column
shift cycle, if half-correction is wired on the control panel. An explanation of this operation is reserved for the section on circuit description.
After the end of the sixth column shift cycle,
the product is retained until the card in the punch
unit moves to
punching position. To read out the
answer from the product counter,
CB's in the
punch unit start pulsing the product counter at
9
on the index, and one is added to all positions of
the product counter for each index poinr at the
line of index. When a position in the products
counter carries over, the carry
pulx trips a power
tube which energizes the corresponding punch
magnet to cause punching of the proper
digit in
the card. A more complete description of this
operation will be found in the circuit explanation.
Computing Circuits - General
The circuits for electronic computing are shown
in Sections 21 through 80 of wiring diagram
68
TYPE
603
ELECTRONIC MULTIPLIER
213639A.
The diagram is laid out by chassis to
permit a close tie-in
with the machine.
None of
the cable wiring is shown, but each cable wire or
jumper is indicated at each chassis terminal on the
wiring diagram.
The terminals on the chassis are numbered from
1 through 20 on the right side (facing the rear or
wired side of the chassis). On
,the left side terminal
strip the terminals are numbered
21 through 40.
Across the top or bottom the terminals are num-
bered 41 through 60. Since no chassis has both
a
top and a bottom terminal srrip, the same numbering is used for either top or bottom. Unused termi-
nals retain their number and must be counted to
locate the proper terminal. To prevenr
possible
trouble from capacity coupling in the cnbles in the
electronic unit, all cables carrying high-speed
pulses traveling from one chassis to another
'Ire
shielded.
A color code is used to identify all power lines
zuithilz
a
chassis.
This code does
nor hold for the
cable wires. The wiring diagram indicates the color
coding used.
For reference
it
is repeated below.
Ground-Black
f
65 volts-Slate
+
1 5 0 volts-Red
-1
00 volts-Green
-100 volts-White
(Cancel)
-2
5
0 volts-Red-White
11
5
volts A.C.-Blue
(Filaments)
NOTE: On some chassis on early models the
-2 50
volt lines were purple instead of red-white.
The $65 volt supply was added after the orig-
inal circuit layout was made, hence the
+
65 volt
line is
placed above the +I
YO
volt line on the cir-
cuit diagram.
In addition to the color
coding used on all the
power lines, all controlled grid and suppressor circuits are indicated
by yellow wires.
In some cases
a
slight departure from this rule is necessary, but
usually a yellow wire indicates a
grid or suppressor
circuit. The cathodes of all tubes in the computing
circuits (except rhe multivibrator) are connected
to the ground line (zero which is shown
as a heavy line throughout all the electronic circuits.
There are only three types of tubes used in the
compu~ing circuits, namely, types 12SN7, 6SK7,
and 25L6. The 12SN7 is a twin triode, used as a
trigger, as a blocking tube, and as an inverter
switch; the
6SK7 is a pentode used as an electronic
switch; and
,the 2 5L6 is a beam power tetrode used
as a power tube in circuits requiring power. Figure
65A shows
the symbols for the tubes used, along
with the pin connections on the tube, while Figure
65B shows the socket connections. The tube symbols are also shown in Section
22B of the wiring
diagram. Observe that the
25L6 symbol is shown
on the wiring diagram as a
pentode with the suppressor internally tied to the cathode, whereas Figure 65A shows a symbol for beam-forming plates.
The beam-forming
plates act as a suppressor and
for that reason the symbol often shows a suppressor
tied to
rhe cathode internally. In practical cir.
cuits, however, neither is shown in order to sim-
plify the symbols. All circuits show the
25L6
with only two grids.
@
Pin Connections ( From bottom of tube)
TOP Bottom
@
Socket Connections
Figure
65.
Tube and Socket Connections
ELECTRICAL PRINCIPLES
69
Since the three tube types mentioned are the
only types used in the computing circuits, they are
not labelled
by tube type. Merely remember that
all triodes are
12SN7's; all tetrodes are 25L6's;
and all pentodes are 6SK7's. As an aid to servicing,
it is recommended that the pin connections of these
three tube types be drawn on a card and the card
kept in the bottom of the electronic unit for ready
reference.
Spare tubes for replacement purposes are
mounted in unused positions on the
L,
M, and
W
chassis. These spare tubes are readily identified
by white tube sockets. The spares in the L and M
chassis are
12SN7's, while the
W
chassis carries
spare
25L6's, 6SK7's, one 6SJ7, and one 2526.
All the tubes used in the computing circuits
draw the same filament current of 0.3
amperes,
consequently the tube filaments can be wired in
series in any arrangement.
In order to allow connection to a common source, the filaments are connected in series
strings providing a total potential
drop of approximately 113 volts. In some cases
a string of filaments may include tubes from two
different chassis. The exact arrangement will be
found on the circuit diagrams for the individual
chassis. The filaments do
no,t directly affect the
circuit operation and for this reason no further
reference will be made to the filament circuit. In
several instances dummy tubes are used to fill out
a filament string. These can be readily identified
because the tube socket is painted red.
In describing the circuits, the control grids of
all tubes will
simply be referred to as
grids,
while
the suppressor grids of the
~entodes will be called
st~ppr~ssors.
The screen grids of all the 25L6's and
6SK7's are connected to a +65 volt supply line
which is shown above the
+
150 volt line on the
circuit diagrams. The screens of
25L6's require
a resistor in the circuit to limit the screen current.
Resistors are used in the screen circuits of
6SK7's
only to prevent parasitic oscillations when several
screens are wired in
parallel. The screen grids may
be ignored in the description of the circuits unless
they serve a functional purpose.
The anodes of all the tubes are connected to the
+I
50 volt line through load resistors ranging from
5000 ohms for the
25L6's to 20,000 ohms for the
12SN7's and 6SK7's. The +I50 volt line is shown
parallel to and directly above the ground line. The
5000 ohm anode load resistors for the
25L6's are
5
watt wire-wound resistors to dissipate the energy
in the anode circuit of the
2 5L6 power tubes, while
the 20,000 ohm resistors are
0.5 watt carbon re-
sistors of the BTS type.
The circuits are designed and the tubes are con-
trolled in a manner to assure positive operation and
to overcome chance or unavoidable differences between tubes of the same type. For this purpose, a
tube is held non-conductive, when so required, by
a
po,tential on one of its grids which is considerably
below cutoff. To drive a tube to a conducting
state, its grid potential is given a tendency to rise
considerably above cutoff. The grid resistor has
a high value, so that grid current flow will bring
the
potential of the grid to the cathode potential.
The
potential necessary to cut o'ff
he
various
tubes with
+If0 volts on the anode is as follows:
12SN7
..
-
9
volts
2 5L6
_-----.-----..----...---..
_------
-2 5 volts
6SK7
_--...-- ~ ...---..----..---
-17 volts (grid)
-40 volts (suppressor)
To insure safe operation, the following poten-
tials are maintained at the grids of the various
tubes when they are cut off:
12SN7
.
.
5 volts
2 5L6
..
..
-3 5 volts
6SK7
.-----------.----------
-3 5 volts (grid)
-5
0 volts (suppressor)
These are maintained through resistor
networks.
The exact values of resistance in the
voltage divider networks can be found in the circuit diagrams for the individual chassis. It must
be remembered that exact voltage values are not
always attainable because standard available re-
70
TYPE
603
ELECTRONIC MULTIPLIER
sistors must be used.
The values given above are
cause improper operation of triggers. To eliminate
approximate.
If a tube is to be maintained in a conductive
state, its grid is connected to a high positive potential through a
high value resistor. The grid cur-
rent flowing through the grid resistor will reduce
the grid
~otential to the ca,thode potential. Any
tendency of the tube current to
drop will be accompanied by a reduction in grid current, hence
the grid potential will rise and overcome
the ten-
dency of the tube current to drop. This method
of connecting the grid in normally conducting
tubes overcomes any difficulty which is due to var-
iations in individual rubes. All normally conducting tubes are indicated on the wiring diagram by
placing an X under the cathode symbol.
In many
parallel arrangements, grid resistors are
used to prevent
parasitic oscillatiotzs.
A parasitic
oscillation is any undesirable oscillation in a
circuir
which interferes with its normal operation, or
lowers its efficiency. Parasitics may result from a
momentary fluctuation of current through a tube
which causes grid
porentials on other tubes to
change. This would cause current through these
other tubes to change, which in turn changes the
potential on the grids of the first tube, and creates
a
con,tinuous interaction or oscillation. By placing
resistors in the
grid circuits, the effect of potential
variations in one tube on the grid of another is decreased to a point where interaction is
eliminared.
An example of parasitic suppressing resistors is
shown on the circuits for the
C
and D chassis. On
the
C
chassis (Section
31)
the suppressors of six
6SK7's are in parallel, while the D chassis (Section
34)
shows six gids in parallel.
By-pass capacitors are also used in many cases
throughout the computing section to prevent un-
desirable effects by
tm?zsie~zt z~oltages.
Transient
voltages are very rapid voltage changes resulting
from load changes, fluctuation in supply voltage
arcs resulting from breaking inductive
circuirs, etc.
Transients can cause parasitic oscillations if pro-
per
precautions are not taken. Also transients can
possible transients caused by arcs, extreme care
must be exercised to eliminate all arcs in the
punch
unit.
The trouble resulting from arcing at
HD1 relay
points may be used as an example. Although these
points in the motor circuit are on the shielded side
of the transformer, a heavy arc across its points
will cause the product counter read-out trigger to
operate and add
1's in the product counter. Of
course this will happen only when the motor is
being started. This trouble was eliminated by
suppressing the arc with a
2
mfd capacitor across
the
HD1 points.
Transient voltage pulses may be transferred
from one circuit to another by capacity coupling
between wires in a cable. To minimize this possibility, all high-speed
pulses
raveli in^
from one
chassis to another not adjacent to it are carried by
shielded cable.
An example of by-pass capacitors used to eliminate trouble from transients can be found in the
circuit for the
C
chassis (Section
31).
The
.OT
mfd capacitors at the input to the read-in triggers
are by-passes for transients.
All resistor values are shown on the circuit in
~rzegoh~rrs.
All capacitor values, shown as
zvhole
~zz~t~zbers,
are in micro-microfarads, while capacitor
values shown as
deci~nals
are in microfarads.
All tubular capacitors and wire-wound resistors
are stamped with their value. However, the value
of molded mica capacitors and carbon resistors is
indicated
by color code only. The code used is
shown in Figure 66.
As an example, a resistor with a yellow stripe on
the end, followed
by a violet stripe, another yellow
stripe, and a gold stripe will have a resistance of
470,000
ohms, and the value will be within
-t
5
%.
For another example, assume a molded mica capac-
itor with a yellow dot at the upper left, followed
by two black dots. The lower left has a green dot
followed by a red dot and a gold dot. This code
ELECTRICAL PRINCIPLES
7
1
ABCO
First significant digit
Number ot zeros
@
Resistor Color Coding (Values In ohms)
Third sipniticont digit
Voltoge ~est'
Tolerance
DecimOI Multiplier
@
Copocitor Color Coding (Volues in micro-microtorods)
Black
Brown
Red
Orange
Yellur
Crwn
Multiplier Tolerance
0
1
2
3
4
5
0
1
2
3
4
S
Black
Brown
Red
Orannc
Yellow
Creen
Rlack
Bmwn
Red
Orange
Yellow
Green
-
10
100
1.000
10.000
100.000
not
Color
Black
Brown
Red
Orange
Yellow
Green
Blue
Violet
<;ray
White
Cold
Silver
No
Color
Blur
Gray
Gmcitance
in
micromicrofarads
(uuL)
None
Silver
Cold
f
20%
t
10%
t
5%
6
7
8
9
Figure
66.
Color Coding Used on Carbon Resistors and Mica Capacitors
Signiricant
Figures
ARC
Blsck
Brown
Red
Orsnac
Yellow
Cmn
Blue
Violet
k?Ze
Cold
Silver
No Color
Blue
Violet
(;my
White
will signify a 40 mmfd capacity accurate within
the anode circuit since
,the grid is at -25 volts.
A2
7C
and tested for operation up to 5 00 volts. With no current flowing through the load resistor
R there will be no potential drop across it, and
I.OOO.000
10.000.000
-
-
Dccimol
Multnplier
D
BASIC CIRCUITS
point A will be at the same poten.tia1 as the battery
+
terminal, i.e., +I50 volts. In Figure 67B the
Before proceeding with the actual electronic
cir-
switch S has been transferred over to the cathode,
cuits, several elementary principles will be reviewed
and several basic circuits which will facilitate later
cxplana~tions will be discussed.
In all the electronic circuits, voltages are referred
to the cathode.
The cathode will thus
be
at zero
voltage and all voltages positive with respect ro the
cathode will be considered
above
the cathode and
will be indicated with a
plus sign
(+).
All voltages negative with respect to the cathode will be
considered
below
the zero voltage of the cathode
Csparitivr
Tolerance
E
thus placing the grid at carhode potential.
Con-
duction will take
place through the tube and
through the load resistor
R.
Assume that 10 milliamperes of currenr pass through the tube. With
the given value of
10,000 ohms for the load resis-
tor, the potential drop across
R
will be 100 volts.
The direction indicated is the direction of electron
flow in conformity
wirh the direction of flow
through the tube. The potential drop across the
tube will be E
=
Ea - IR = 150 - 100 = 5OV,
and will be indicated with a minus sign
(-).
For
since the sum of the potential drops across the
example, in the simple triode circuit of Figure
67A,
series components of a circuit must equal the po-
the grid is at -25 volts and the anode is
at
+I50
tential of the source. The potential at the midpoint
volts.
tap
M
in Figure 67A is +I50 volts before con-
In Figure 67A it is assumed that the triode will
duction starts.
However, when sthe tube is con-
cut off at
-1 5 volts; therefore, no current flows in ducting, the potential at point M is +lo0 volts.
-
DC
Test
Voltnge
F
Dot
Color
72
TYPE
603
ELECTRONIC MULTIPLIER
0
0
Figure
67.
Basic Notation
This is evident from the fact that ,the two halves
The potential of this battery must be the
potenstial
of resistor R can be considered as two equal resis- at point A when TI is conducting plus the desired
tors with a 50 volt drop across each. negative bias for
T2.
Assuming the same tube as
This change in potential at the anode of a tube,
shown in Figure 67, athe potential at point A when
when conduction takes place, can be utilized to
TI is conducting is
$50 volts.
It
is desired to
a4
control another tube by connecting the grid of the
bias
T2 at -25 volts; hence EgZ must be a 75 volt
second tube to the anode of the first as shown in
battery. As long as TI is conducting, the grid of
Figure 68. The grid battery Eg2 is necessary to
TZ
is at -25 volts and T2 cannot conduct.
How-
reduce the
po,tential on the grid of T2 to the de-
ever, when the switch
S
is transferred to the -25
sired negative potential when the tube is cut off.
volt tap, TI is cut off, and no
current flows
A
r
Eg
2
81
+
~IIIIIII/III
-
75
v.
I\
L
I Ill
Figure
68.
Control of One Triode
by
Another
ELECTRICAL PRINCIPLES
73
Figure
69.
Direct-Coupled Triodes
through RI.
With no
IR
drop across R1, the po-
tential at point A is the same as the
150 volt bat-
tery potential at the plus terminal,
i.e.,
+I
TO
volts.
This means that the grid of
T2 will now be at +75
volts (1 50 volts - 75 volts) and conduction will
take place through T2. Therefore,
TZ
cnjz con-
duct only
if
TI
is
?lot co?zducti?zg.
In this manner
T2 can be controlled by controlling TI.
The circuit of Figure 68 is not a practical one.
A
more practical circuit using a resistor network
to obtain the grid bias of
T2
is shown in Figure 69.
In this circuit a voltage divider consisting of
R1
and R2 is connected between the anode of TI and
the
-100 volt line. When TI is conducting, point
A
1s at
+lo
volts as shown in previous illustrations.
Under these conditions, there are
150 volts across
R1 and R2 (from
+5O
volts to -100 volts is 150
volts). Since R1 and R2 are equal, there is a 75
volt drop across each.
The IR drop across
RI and
R2 is of a polarity opposite to the -100 volt supply; therefore, the potential at point
G
(grid of
T2) is -100
+
75 or -25 volts. Assuming T2 cuts
off at
-IF volts, it is evident that T2 is cut off,
proziding TI
is
conducting.
When TI is not con-
ducting, point A is at
+IF0 volts, and the total
potential across
R1 and R2 is 2 50 volts
(
+
1
FO
to
-100). This means a 125 volt drop across each
resistor and a potential of
+25 volts at the grid of
T2 so that T2 conducts.
Actually, the grid of
T2
will not go very much positive; it will only
tend
to rise to 1-25 volts.
As soon as
thc grid reaches
cathode potential, some grid
currefit starts to flow
through
R1 from cathode to grid, through R1, to
point
A,
through the 10,000 resistor to + 1 5 0 volts.
The resulting IR
drop across RI due to the grid
current is of a polarity opposite to the positive potential at the
grid, and the grid potential is thus
reduced. The grid potential will stabilize at ap-
proximately cathode potential (zero).
The values of
RI and R2 must be very large
compared with the load resistor R otherwise the
potential at points A and
G
will not be of proper
valuc. In the preceding calculations involving
Figure 69 it was assumed that
R1 and R2 were so
large compared with R, that R could be ignored
in calculating the grid potential of
T2. This
is
not strictly true, because there is always a current
flowing from the
-100 volt line through R2, R1,
R
and back to the + 1 5 0 volt line, even though
tube
TI is out of its socket. A rigid analysis would
show that there is a 2.48 volt drop across
R
at all
times. However, compared with the total poten-
tial of
2 50 volts across the resistor network, the
potential
drop across R due to this current is so
small
that it can be ignored for practical calcula-
tions.
74
TYPE
603
ELECTRONIC MULTIPLIER
+
I
50
,--
,++100V.
pulse)
Figure
70.
Capacity-Coupled Triodes Showing Phase Inversion
Often it is not desirable to couple one circuit
directly to another. In these cases capacity coupling
is used to pass a pulse from the anode of one tube
to
,the grid of the next. In studying capacitycoupled circuits it is important to remember that
the
reartaizce of a capacitor decreases as the fre-
quency
of
the
applied pote~ttial i~zcreases.
Direct
current cannot pass through a capacitor, since
direct current has a frequency of zero and results
in an infinite reactance. Conversely, an infinite
frequency will pass through a capacitor with zero
reactance. Infinite frequency
implies an i?zsta~z-
ta7zeous
change of voltage. Obviously, this is im-
possible in practice; but very rapid changes can
be
obtained. Very rapid changes in voltage are
equivalent to very high frequencies; consequently,
very rapid changes in voltage can easily be transmitted through a capacitor, even an extremely
small one.
Figure 70 shows a capacity-coupled circuit
wherein changes in the anode potential of
TI con-
trol
T2. The anode of TI is coupled to the
grid of
T2 through capacitor
C.
Only chajzges
in
potential will be transmitted through
C.
Once
point
AI reaches a steady value of potential, no
further effect is felt at point
G2.
With TI conducting, the steady state potential
at
A1
is
+SO
volts (assume the same circuit con-
stants as previous illustrations). Point
G2 (gid
of T2) is connected to the cathode through a
300,000 ohm
resistor and to the -100 volt line
through a 700,000 ohm resistor. This voltage
divider places point
G2 at -30 volts normally, thus
cutting off
T2 (assume -1 5 volt cutoff).
Assume now that contact
S
is suddenly trans-
ferred to the
-100 volt line as shown in Figure 70.
If
it is assumed that the transfer takes place in-
stantaneously,
,then there will be an instantaneous
shift of
potential on the grid of TI from O to -100
volts. This shift in potential can be represented
by the square wave shown at the grid of TI. This
shifi in potential is toward a more negative point
and is thus a negative pulse. This shift of potential causes
TI
to be cut off with the resultant in-
crease in
potential at point A1 from
+5O
vol~s to
+I50 volts as shown by the square wave at point
A1. This voltage shift passes through C as a positive
pulse to point G2, thus rending to drive point
G2 to
+70
volts. Only the change
in
voltage,
i.e., 100 volts will be felt through the capacitor.
Since the grid of
T2 is at -30 volts, it will tend to
approach
$70 volts. Actually, grid current will
ELECTRICAL PRINCIPLES
75
I
1
I
0-100
Figure
71.
Use
of
Blocking
Tubes
start to flow through the grid resistors as soon as
the grid tries to go positive, the resultant IR
drop
opposes the tendency for the grid to go positive,
and the grid
poten'tial will not go much above
cathode
It is important to note here
that although the potential at point
A1 rises to
+
1
SO
volts and remains there, the pulse through
capacitor
C
is of short duration. The actual duration of the pulse is determined by the capacitance
of
C.
This means that although point A1
vetnuills
at high potential, point
G2
will be at high potenrial for only a short instant, i.e., for the time required to charge the coupling capacitor
C.
Tube
T2
will then conduct as long as point
G2
is at cathode potential or above, with a resultant
IR drop across load resistor R2. Thus the
poren-
tial at point
A2
drops from +I
SO
to
+fO
volts.
This illustrates the inversion of pulses by a tube.
A negative shift in potential applied
.to the grid of
TI causes a positive shift in potential at the anode
of
TI. In turn, the positive pulse at point
G2
produces a negative pulse aT point
A2.
This is a
most important fact to remember in the study of
electronic circuits.
On the circuits illustrated so far one tube has
been controlled by
anorher. Sometimes it is desirable to block the action of the controlling tube
under certain conditions. This can be done by connecting two controlling tubes in
parallel with a
common load resistor and providing separate grid
controls, as shown in Figure
71.
The value of the
load resistor R is chosen so that the tubes are operated on the portion of their characteristic curve
where most of the
drop is across the load
resistor and where changes in potential at the anode
are very
slight with a change in anode current.
This means that the potential at point
A
in Figure
71, is
essen~ially the same whether one tube is con-
ducting or both are conducting. Hence with
eifher
TI or
T2
conducting, point A is essentially
at
+
50
volts and point
G3
at
-2Y
volts, since the
grid resistors between
+YO
volts and
-100
volts
are equal. This means that as
long as TI, T2, or
both are
conducting,
T3
is cut off. Only when
~zeifher
T1
?tor
T2
is co?zducting
does point
G3
go
positive and allow
T3 to conduct. When neither
TI nor T2 is conducting, point A rises to
+IS0
volts, and the grid of
T3
tends
KO
rise to
+25
volts. Thus, TI can nullify the action of T2, or
vice versa. As indicated in Figure 71,
TI
and
T2
have separate grid controls in the form of other
tubes which are indicated as tube
Y
and rube
2,
which are not shown in this figure.
Another method of obtaining dual control is to
use a
pentode and provide grid control on both the
control grid and the suppressor grid as shown in
Figure
72.
Since the suppressor is spaced much
farther from
he
cathode than the control grid,
a
76
TYPE
603
ELECTRONIC MULTIPLIER
To
Anode Tube
I
I
A'
Suppressor
Cutof t
....
-
40
V.
Grid Cutoff
...
-
17
V.
Figure
72.
Two-Circuit Control of a Pentode
greater negative potential is required to cut off the
300
X
R2 300
X
680,000
-
tube by means of the suppressor than is required RI
+
R2 1,010,OOO
=
200V (ap-
for cutoff by means of the control grid.
In this
proximately)
case, assume rhe tube can be cut off by -17 volts
on the
control.grid or by -40 volts on the suppressor. Either the control grid or the suppressor grid
can stop conduction. In order for conduction to
take place both
grids
nus st
be nboue cutof joten-
tin!.
In Figure 72 it is assumed that the grids of the
pentode are controlled by the anode potential of
preceding tubes. Assuming
,that both tubes con-
trolling the
pentode are conducting and that they
are of the same type illustrated in previous exam-
ples, the
potential ast their anodes will be
+YO
volts.
This will place the suppressor at
-50 volts, which
is determined as follows:
Total potential across Rl and R2 is 300 volts
(from
+
50 volts to -2 50 volts).
Suppressor is at
poinr S and the potential at S
is determined by the ratio of the resistors
R1
and R2 as follows:
Hence point
S
is 200 volts above the -2
5
O
line,
or at
-50 volts.
The control grid is normally at approximately
-3
5
volts as determined by the voltage divider
R3 and
R4 between the cathode and the -100 volt
line.
With both grids negative no conduction can
take
place.
Now assume that the tube controlling rhe sup-
pressor stops conducting. Point A2
will rise to
+I 50 volts potential and point S will tend to rise
to approximately
+17 volts. Thus rhe suppressor
has been conditioned to allow conduction through
the
pentode; but the control grid is still blocking
conduction, since it is below cutoff.
When point
A1 rises in potenrial, a positive pulse will be passed
to point
G
and the tube will conduct for an in-
stant, providing the suppressor is still conditioned
to conduct.
ELECTRICAL PRINCIPLES
77
Figure
73.
Basic Eccles- Jordan Trigger Circuit
When the tube conduas, the
at point
tion of a trigger circuit will be ~resented. Although
A drops because of the
IR drop across the load re-
a detailed knowledge of the theory of operation
sistor of the
pentode. In this case the output is
of a trigger is not necessary to repair rhe ma-
taken from a midpoint tap
M
on the load resistor. chine, a thorough knowledge of the theory will
This would be done if the voltage shift desired is
assist in analyzing trigger troubles. A trigger cir-
only half the voltage shift at point A. Assume
cuit is one which has two stares of equilibrium for
the potential
a,t point A changes from $1 50 volts
to
+YO
volts. When conduction starts, a 100 volt
negative pulse is at A. However,
at
point
M
the poten~ial only changes from
+I
50 volts to
+
100 volts, producing a 50 volt negative pulse at
M.
The screen grid in Figure
72
is shown at a fixed
potential of
+65 volts supplied by a 65 volt
screen voltage supply. The capacitor
C
across RI
is necessary to balance the imerelectrode tube capacity so that the grid of the tube can follow the ap-
plied
potential without any time delay.
TRIGGER CIRCUIT
fixed values of supply potential and circuit components. The trigger circuit derives its name from
the fact that it can be made
.to "trigger" abruptly
from one state of equilibrium to the other
by
means of small controlling potentials. The trigger
circuit used in this machine is basically the Eccles-
Jordan trigger circuit shown in Figure
73.
The
use of this circuit is based on the fact that current
can flow through only one tube
a.t a time. A
change in grid potentials or anode
potentials can
be made to transfer conduction abruptly from one
tube to the other.
As mentioned above, rhe trigger circuit is a de-
vice using two triode tubes so interconnected that
Theory of Operation
one tube is conducting while the other is non-con-
The
most important and basic circuit used in the
ducting.
As shown in Figure
73 the grid of T2
electronic computing section is the trigger circuit.
is coupled to
The anode of TI through the coupling
For this reason a detailed analysis of the opera-
resistor Rc and capacitor
C.
The !grid of
TI
is
78
TYPE
603
ELECTRONIC MULTIPLIER
Figure
74.
Practical Eccles- Jordan Trigger Circuit
similarly coupled to the anode of T2. The resistors
RL are the anode load resistors while the
grid bias
is normally furnished through the grid resistors
Rg. The values of RL, Rc, and Rg are not critical
as to
,their exact value but
it is essential that the
circxit be sym?~zetrical.
Rc and Rg must be
matched within
2':L
and they should be approxi-
mately ten times as large as
RL.
The value of
C
determines ,to a great extent the speed of response
of the trigger circuit to an applied pulse.
A suitable voltage pulse applied at the proper
points causes the conducting tube to stop conducting and the non-conductive tube to start conduct-
ing. A second pulse restores the original condition.
This cycle may be repeated at will at any speed
from zero up to speeds in the low radio-frequency
range, depending upon the circuit constants used.
Thc tubes used in the circuits illustra,ted in Figure
73 are 6JY's or the equivalent. The two triodes
can just as well be the two halves of a twin triode.
Figure
74 shows the Eccles-Jordan trigger circuit in a form more suitable for analysis. A close
study of Figure
74 will reveal that it is the same
circuir as Figure
73
with only the input circuit
added. The resistance and capacitance values shown
in this figure are the values used in the
actual
triggers in the electronic computing circuits:
In
this illustration the two triodes are the two halves
of a twin triode, Type
6SN7.
Figure
75
shows how T2 can control TI. As-
suming that both tubes are non-conducting,
,the
potential at the grid of TI can be determined by
the ratio of the resistors between the +IYO volt
line and the
-100 volt line, or by determining the
current flow and computing the IR drop across
each resistor. In this manner the potential at GI
is found to be
+
19 volts; or 19 volts positive with
.""
-
.-..
Figure
75.
Anode of T2 Controls Grid of
T
ELECTRICAL PRINCIPLES
79
.02 Meg.
I
+40
i
-------------------__---------------
A2
1
I
A
l
J
1
+
136
I
1
I
2MW.
G2
0
I
.2 Meg.
I
I
I
-
-
- - - - -
- -
-.
-
- - -
-
-
J
-90
Figure
76.
Retroactive Coupling
respect to the cathode since the cathode is a.t zero
potential. Actually, point
GI will only tend to
reach $19 volts, since this positive
potenrial will
cause TI to conduct, and sufficient grid current
will flow through Rc to hold the grid down to
approximately cathode potential.
Now suppose that T2 is made .to conduct by
somc means as shown in Figure 76.
Assuming the
potential drop across
T2 when conducting is 40
volts, point
A1 will have to be at $40 volts. Another method of looking at this is as follows: when
T2 conducts, approximately 5 milliamperes of current flow as determined by the anode potential and
the load resistor. This current flow through RL
causes an IR drop of
,005 X 20,000 or 100 volts.
Naturally, .this is in the same direction as and
in
~tdditio~z
to the drop resulting from the bleeder
current always flowing from the
-100 volt line to
the
+
150 volt line. Therefore, the potential drop
across RL is increased
by 100 volts, making point
A1 100 volts more negative than it was previously.
In practice, this potential at point A1 is $40 volts
instead of $38 volts with
T2 conducting, since
the current is not
exactly
5
milliamperes.
A
new analysis of the potential at point GI will
show that the grid of TI is at
-30 volts when T2
conducts, thus cutting off TI. Since Rg and Rc
are
equal, GI is halfway between $40 volts and
-100 volts.
From $40 to
-100 volts is 140 volts, and the
drop across Rg is 70 volts. Therefore, GI is 70
volts above the
-100 volt line or at -30 volts.
From this analysis it is obvious
thax if T2 is con-
ducting, it prevents TI from conducting.
If the
grid of T2 is connected to the anode of
TI by means of a similar network (Figure 76),
TI can control T2 in the same manner described
above,
and the desired condkion in which only one
tube can be conducting at a time is obtained. In
Figure 76 point A2 is shown at $136 volts instead
of
$138
volts. This is because the grid current
rcduces the poteritial at point G2 to zero. On
analyzing the potential between point
G2 and the
-+I50 volt line, it is found that point A2 is at
$1 3
6
volts.
In order to provide a means of applying a trig-
gering pulse, both grids must be coupled to a
source providing the pulse as shown in Figure 77.
80
TYPE
603
ELECTRONIC MULTIPLIER
1
0-100
-100
Instantaneous
Potentials
in Porenthmsis
Figure
77.
Use of Capacitors in Trigger Circuit
The potential values shown outside parenthesis are
For
.this reason a square wave is used as a source of
those existing when
T2 is conducting and before
pulses.
a pulse is applied.
This circuit is designed
,to trig-
When a negative pulse of 20 volts with a steep
ger on the application of
a
20 volt negative pulse
wave front is impressed across the input, there is
to both grids. In order to operate
properly, the
a sudden drop of 20 volts at point
X.
Since a
triggering pulse must have a
very steep wave front.
steep wave front is equivalent to an extremely high
Figure 78. Voltages Existing after Ci Discharges but before
Cc Discharges with Neither Tube Conducting
ELECTRICAL PRINCIPLES
8
1
g
t
10.-
0
+
4
Time
--c
Conduction starts
I-8V)
Grid TI storts from - 50 V
(Fig
77)
Grid T2 starts from
-20V
(Fig
77)
Figure
79.
Calculated Rise of Grid Voltage as Circuit Voltages Change
from Those in Figure
77
to Those
in
Figure
78.
frequency, the input coupling capacitors Ci will
offer practically no reactance to the
pulse and the
sudden drop of
20
volts will be felt at points
GI
and
G2.
The po,tential values shown in parenthe-
sis are the instantaneous values obtained when the
-20
volt pulse is applied. This pulse will render
both
,tubes non-conducting for an instant.
In order for the trigger to operate,
the Cc capa-
citors must be considerably
larger than the Ci
capacitors so that there will be very little change
in the
potential across the Cc capacitors in the
time required for the smaller input coupling capacitors to reach a steady state condition. To
sim-
be held non-conducting by some external means,
the potentials shown in Figure
77 in parenthesis
would soon change
TO
those shown in Figure 78.
The shown in Figure 78 exist with the
input coupling capacitors Ci in equilibrium, with
capacitors Cc not yet discharged, and with neither
tube conducting. If it is assumed
th2t the Cc cap-
acitors are
no~t discharged, they will maintain the
potential across them, and they can be considered
as batteries with a potential equal to the charge on
them at the time the pulse was applied. The potentials at the various points can then be analyzed
plify the explanation, assume that there is no
On this basis.
change in po~ential across the Cc capacitors in the
Since
,the tubes are not held non-conducting, the
time required for the Ci capacitors to reach steady shown in Figure 78 will never be
state values.
As shown in Figure
77, both tubes
reached. The grids will
only tend to reach the
are rendered non-conducting by the instantaneous
values shown. However, on the basis of the ten-
potentials
resul~ing from the
-20
volt pulse, since
dency of the grids to approach the limiting values
these tubes cut off at -8 volts. If both tubes could
shown in Figure 78, the actual rise of potential on
82
TYPE
603
ELECTRONIC MULTIPLIER
,<
/
'
/
/
'
/
/
Stor1 of
/
negative
pulse
/
/"
voltape increose if
/
:either lube conducts
Figure
80.
Oscillograrn of Actual Grid Voltages During Triggering
the grids can be determined as shown graphically
in Figure
79.
While these curves could be accurately calculated, they were actually obtained by plotting an
ex curve between the known limits of grid potential.
This
gives a theoretical capacitor discharging or charging curve. Because of the much greater swing in potential on the grid of
TI,
caused by
the relatively low
potential existing on the capa-
citor Cc between the grid of
TI
and the anode of
T2,
it
is obvious that the rate of potential rise will
be much greater than the rate of rise in the potential on the grid of
TZ.
It is evident that the grid
of
TI,
which has been non-conducting, is the first
to rise above the conducting point of
-8
volts.
These tubes cut off at
-8
volts, therefore conduc-
tion
will
start as soon as a grid goes above .this
potential. Hence
TI
starts conducting and blocks
'TZ
from conducting as previously explained, and
the trigger is reversed. The action of the capacitors
Cc produce the desired trigger action, and
the trig-
ger will now reverse
itsclf every time the gids are
given a negative pulse of
20
volts or more.
In Figure
78
it
was assumed that the capacitors
Cc were much larger than the grid capacitors
Ci.
As evidence that this assumption does not alter the
general shape of the curves of grid potential rise,
Figure
80
shows a sketch of the grid potential rise
in an actual trigger circuit. This sketch is adapted
from an actual oscilloscope pattern and shows
exactly what happens to the grids of
TI
and
TZ.
Originally
TZ
was conducting and
TI
was held
non-conducting by the
-30
volt potential on its
grid as shown
previously. At time t (Figure
80)
a
20
volt negative pulse is applied to both gids
through the input capacitors Ci. Owing to the fact.
that the square wave input is not perfectly square,
the
neggtive pulse as it appears at the grids is not
ELECTRICAL PRINCIPLES
83
Square wave
Input
Anode
Potential
TubeT,
Grid
PDtentiol
Tube Ti
Tube T2
+
40
Anode
Potential
Te
-30
Grid
Tube
T2
T2
Figure
81.
Oscillogram
of
Overall Trigger Operation
quite square, and at the grids the peak negative
dip is only
-1j volts. As soon as the maximum
negative is reached, both
gids start to
rise in potential. As
previously shown in Figure
79,
the grid of TI rises much faster than that of
T2 and reaches the conducting point of -8 volts
first. As soon as
TI starts to conduct, the poten-
tial at its anode starts to drop, forcing the grid of
T2 down and holding T2 non-conductive. With
,the circuit constants shown in Figure
74,
after a
time interval of approximately
3
to 5 microseconds,
the charges on
all
capacitors will have been equal-
ized and the circuit will be as before except that
TI is now conducting instead of T2. The dotted
lines
indicate what the rise in the grid potentials
might look like
if
the tubes could be held nonconducting by some external means. It is important to note that,
alrhough thc triggering action is
very fast, a definite time interval is required, hence
a
peaked pulse of extremely short duration (say
1
microsecond) may nor trigger the circuit.
Figure 81 shows
a
sketch of both grid and anode
potentials adapted from patterns taken directly
from an oscilloscope. The potential graphs repre-
sent the
po~entials when the trigger is triggered
or reversed continually
by
a square wave input.
Note that the shape of the grid
potential is the
same as shown in Figure 80.
So far nothing has been said
abour the ability
of this trigger circuit to
distinguish between posi-
tive and negative
pulses. The constants of this
trigger are such thar it is considerably more sensitive to negative
pulses than it is to positive pulses.
Therefore, if the input
pulse is kept within reason-
able limits, the
rrigger will respond only to the
negative pulses of a square wave (Figure
8
1
)
.
For
example, a
-20 volt shift in potential will cut off
the conducting tube, enabling the trigger to transfer;
but a +20 volt shift will not bring the grid
of the nun-conducting tube up to the conducting
point and thus cannot make the tube start to con-
duct. The only action of a
+2O volt pulse on the
conducting tube is to drive the grid slightly positive. Therefore, the trigger will transfer only on
a negative
pulse (or shift in potential), and the
trigger can be made to distinguish between negative and positive pulses. The limits within which
the trigger will respond only to negative
pulses for
the circuit constants given is approximately
20 to
80 volts.
Thar
is,
at least
-20
volts are required to
trigger, but around 80 volts the trigger responds
to positive
pulses as well as negative. For this
reason the triggers in this unit are operated
by 50
volr pulses, or roughly at the middle of the range.
Figure 81 shows why the trigger is not reversed
on a positive pulse which is theoretically
large
enough to bring the grid of the non-conducting
tube up to the conducting point.
Norice that at
point
a2 the grid of the non-conducting tube TI
actually appears to go negative although the square
wave
inpur is shifting in a positive direction. This
is because the positive
pulse acting on the grid of
the
conducting tube T2 drives the anode potential
of
T2 down almost 20 volts as shown at poinr a3.
I
84
TYPE
603
ELECTRONIC MULTIPLIER
Figure
82.
Application of Pulse to Trigger Circuit Without Cc Capacitors
Through this anode 'to grid coupling capacitor Cc,
this dip in anode
potential of the conducting tube
over-rides the positive pulse on the grid of the
non-conducting tube,
producing a negative dip as
shown at a2 (Figure
81).
The dip in the anode potential of the conducting tube shown at a3 is in
this case caused
by the same positive pulse acting
on its own grid.
For a positive pulse to trigger
The circuit, the
pulse must be sufficiently positive to overcome the
initial bias
plus the negative swing produced at the
grid of the non-conducting rube by the dip in
anode potential of the conducting tube.
If properly designed, triggers such as those described above are very stable, dependable, and independent of reasonable variations in
supply po-
tentials.
A
20%
varia~ion in either bias or anode
potential supply, or more if both vary together,
can be tolerated.
In order to illustrate the importance of
sthe Cc
1
Both Grids attemptino
/to reach +20V.
+
20-
+
to--
-
I
/I
/
Time
--
Grid TI $torts from
-50V.
Grid T2 starts from
-20V.
-
60~
Figure
83.
Rise in Grid Voltage as Capacitors Discharge
ELECTRICAL
PRINCIPLES
a5
Figure
84.
Coupling Trigger Circuits
+I50
.02
Yea.
Trigger
2
I
\
-
>
0
-
-
-
-
-
capacitors, an analysis will be made without the
Cc capacitors in the circuit. Figure 82 shows a
trigger circuit without the Cc capacitors; it will
be shown that this circuit is fundamentally incapable of reversing on application of a pulse to both
grids.
The potential values shown outside parenthesis
are rhose existing when
T2 is conducting and be-
fore a pulse is applied. When a negative pulse of
20 vdts is impressed across the input, there is a
sudden drop of
20 volts at point
X.
The poten-
Y
tial values shown in parenthesis are the instantaneous values obtained when
rhe -20 volt pulse is
applied. This pulse will render both tubes non-
.
conducting for an instant.
The new potential values
at
the anodes are determined by an analysis of the IR drops across Rc and
Rg using the new instantaneous values of grid potential and assuming both tubes are cut off.
However, both grids will immediately start to rise to
the resistor network limited value of
+I9 volts.
The rate of rise is determined by the time constant
of the resistor-coupling capacitor network. Since
these are the same for each tube, 'the time constant
will be the same for both grids.
The exponential rise of potential on both grids
will be as shown
graphically in Figure 83.
These
I!
--
-.
-
I\
P\
I\
*\
I\
IS
tubes cut off ar -8 volts; hence, conduction will
start as soon as a grid goes above this
potential.
Obviously, the grid of T2 will be the first to reach
the
-8
volt line, which means thar T2 will start
conducting first and prevent
TI from conducting
as before. In other words, the trigger has not been
reversed. Likewise, a positive
pulse will not re-
verse the trigger, since the only effect it
might have
on the non-conducting tube will be offset
by a
stronger effect on the conducting tube.
\
.\
i
r'
I
f
0
0
I
1
L
-
0-100
Coupling of Trigger Circuits
To couple two triggers together, ir is only necessary to tap one anode resistor of the first trigger
at approximately mid-point and couple it to rhe
input capacitors of the second trigger (Figure 84).
This provides a means of tripping the second trigger under control of the first
rrigger. Tapping
the anode resistor at one-half of its value serves
to furnish a voltage pulse of one-half the voltage
shift in the anode resistor. Point
A
in Figure 84
changes from
$1 36 when TI is nor conducting
to
+40 when TI is conducting, thus providing a
negative shift of approximately
100 volts; where-
as point
M
changes from +I43 when TI is not
conducting to
+95 when TI is conducting, thus
providing a negative shift of approxima~ely 50
92
TYPE
603
ELECTRONIC MULTIPLIER
Trigger l Triqqer 2 Triqqer 3 Trigger
4
Input
Pulse No.
Trigger l Trigger 2 Trigger 3 Trigger
4
Code
0
OFF
OFF OFF
OFF 0
I ON
OFF
OFF OFF
I
and add numbers on the decimal system. This is
done in the electronic counter
by means of special
coupling between triggers and by the use of a
special blocking tube, as illustrated
by the circuits
for one position of an electronic counter shown in
2
OFF
ON
OFF
OFF
2
Figure
90.
The triggers are designated by their
3
ON ON OFF OFF
1.2
digital value and the blocking tube is designated
OFF
0 N
OFF
ON
OFF
ON
OFF
0
N
OFF
ON
OFF
ON
OFF
OFF
ON
ON
OFF
OFF
ON
ON
OFF
OFF
0
N
ON
ON
ON
ON
ON
OFF
OFF
OFF
OFF
ON
ON
ON
0
PI
OFF
OFF
OFF
OFF
ON
ON
ON
ON
ON
ON
ON
ON
as
B.
Note that triggers
1,
2, and 4 are coupled in tan-
dem in conventional manner. However, trigger
8
has special connections to the input capacitors,
and tube
B
is connected in parallel with tube 1 of
trigger
2 to serve as a blocking tube. Tube B is
normally non-conductive and conducts only when
trigger
S
is
ON.
This arrangement operates exactly
as shown in Figure
89
up through the ninth pulse.
Then, on the tenth
pulse trigger 8 must be turned
OFF
and trigger 2 must be blocked from turning
16
OFF
OFF
OFF
OFF
o
ON
as i.t normally would on the tenth pulse. The
special coupling from trigger
1
to the left side of
pube
NO.
om-
trigger 8 turns trigger
8
OFF
on the tenth pulse,
while tube
B
blocks trigger 2 on the tenth pulse.
Trlggerl
-
-
-
-
-
-
-
-
Triqger
2
-
- -
Thus all .triggers are
OFF
on the tenth pulse, and a
Triqqer
3
Tr~qqer
4
negative pulse is available at the output for carry.
Figure
91
shows the operation of one position of
Figure
89.
Binary Counter Operation
an electronic counter in chart form, indicating the
special coupling from trigger
1
to trigger 8 and
Note from Figure
89
that the counter follows
the blocking of trigger 2 on the tenth pulse.
the decimal system up through
9.
Now if some Addition in an electronic counter is accom-
means can be provided to restore the counter to
0
plished by successive impulses, the number de-
on the tenth pulse, the counter can be used to store
pending on the value of the digit to be entered. A
Figure
90.
One Position of an Electronic Counter
ELECTRICAL PRINCIPLES
93
Number in Counter
0
1
2
3
4
5
6
7
8 9 0,
\
lat.Trigger
@
0
X 0 X 0 X 0 X 0 X
0
'\
/
4th. Trigger
@
0 0 0 0 0 0
0 0
X X
O(
\
Code 01214121810Earr~
2 442
8
4
Ncte
-X
Indicates Trigger ON
0
Indicates Trigger
OFF
Figure
91.
Electronic Counter Chart
carry to the left is effected each time a position of
the counter changes from
9
to
O.
Two tubes are
required for carrying from each counter position
except the last.
A
trigger is turned
ON
when the
counter advances from
9
to
O.
The trigger conditions a switch tube to permit a carry impulse ro
be fed to the next
higher order counter position.
If the position receiving the carry impulse stands
the carry impulse is of sufficient duration to complete the carry to all counter positions if necessary.
The carry impulse is made available after the adding portion of the cycle, and the carry triggers
that have been turned
ON
are restored to the
OFF
position at the completion of carry. The carry
operation is shown in block diagram form in Figure
92.
The details of the carry circuit are des-
cribed in a
la,ter section after the computing cir-
cuits have been described.
Pulses may be delivered to a counter at the comparatively slow rate of card cycle point timing, or
dt the
3
5.000
cycle per second rate as in the case
of counter transfers. Counters have been success-
fully operated at impulse rates up to
1
IO,OOO
cycles
per second. The impulse admittance is controlled
by tube
switches,
to be described later.
A
detailed explanation of the operation of one
position of an electronic counter circuit is pre-
.
sented below (Figure
90).
Triode B has its grid connected between the
u
at 9 and advances to
0,
its carry trigger is turned
anode of tube 1 of trigger 8 and the
-100
volr line
ON
to condition the switch and continue the carry
by means of a voltage divider
RIR2.
The anode of
to the next higher counter position. The carry is
tube
1
of trigger 8 is at
+
40
volts as long as trig-
not instantaneous to two or more positions, but
ger
8
is
OFF.
(See discussion of
Trigger Circzrits.)
Carry
Control
Suppressor
---------
i
I
Control
?
Input Counter
Pulses Position I
Corry
Suppressor
I
---------
1
I
Control
Position
2
Carry
Figure
92.
Block Diagram of Carry Controls
94
TYPE
603
ELECTRONIC MULTIPLIER
This means that the grid of B is midway between
+40
volts and
-100
volts, that is,
-30
volts as long
as trigger
8
is
OFF.
This is considerably below the
cutoff
poten,tial of tube B, therefore B is non-con-
ductive as long as trigger 8 is
OFF.
When trigger 8 goes ON, the potential at the
anode of its tube
1
rises to
+I36
volts. The grid
of tube
B
is rhen midway between
+
136
volts and
-100
volts, or
+18
volts.
Of course, the grid does
not
actually go to
+18
volts, because grid current
starts flowing through Rg as soon as the grid tries
to go positive. This current flow through Rg and
RI
is the reverse of the normal current flow
through
RI
resulting from the
+I36
volt and
-100
volt potentials; consequently the potential at
the grid is kept from rising above zero and
remains substantially at zero, or cathode potential.
With tube
B
conducting, the potential at its anode
is
+40
volts. Since ,the anode of B is directly con-
nected to the anode of tube
1
of trigger 2, the
anode of tube
1
of trigger 2 will be maintained at
+40
volts regardless of what trigger 2 a,ttempts
to do. From the theory of operation of a trigger it
will be remembered that the rise in
potential at
the anode of one tube is necessary to cause the
other tube to conduct. If trigger
2 is
OFF
when
rube
B
is conducting then
trigger 2 ca?znot be'
~?LIWP~
ON
regardless of the pulses applied, because
the anode of tube
1
of trigger 2 cannot rise above
+40
volts, and as a result, the grid of tube 2 of
trigger
2 cannot go above cutoff potenrial.
The input pulses to the counter are
50
volts
square wave pulses. From the theory of operation
of a trigger it will be remembered rhat the trigger
will recognize
only negative pulses provided the
voltage amplitude of the pulses is kept within certain limits. Hence, only the negative shifts in
volrage are recognized by the triggers. The entry
pulses are fed to both input capacitors of trigger
1
to turn it
ON
and
OFF
on successive pulses. Each
time trigger 1 goes
ON,
a
50
volt positive pulse is
produced at the anode resistor tap of tube
1 of
trigger
1.
However, this positive pulse is not rec-
ognized and consequently can be ignored.
Each
time trigger
1
goes
OFF,
a
50
volt negative pulse is
produced at the anode resistor rap of tube
2 of
trigger
1,
and this negative pulse is fed to both in-
put capacitors of trigger 2.
Assuming the counter position is
ar zero, all four
triggers are
OFF
and tube B is non-conductive. One
entry pulse turns trigger
1
ON, and the counter
I
position stands at
I.
The resulting positive pulse
I
I
to triggers 2 and 8 has no effect. A second entry
pulse turns trigger
1
OFF
which in turn trips trig-
ger
2 ON, and the counter stands at 2.
Note that the negative pulse which turned
trig-
S
ger 2
ON
also is fed to the left side of trigger
8.
Trigger 8 is
OFF
at this time, and the grid of tube
82
is already negative; consequently this negative
pulse has no effect. Also, the positive
pulse from
the anode resistor tap of tube
21 has no effect on
trigger
4.
A
third entry pulse turns trigger
1
ON
again,
and a positive pulse is fed to trigger 2. This positive
pulse has no effect; consequently, trigger 2 re-
mains in the
ON
status, and the counter stands at
3
(triggers 1 and 2 ON). A fourth entry pulse
switches trigger
1
OFF,
which in turn switches trig-
ger
2
OFF.
When rrigger 2 goes
OFF,
a negative
pulse is produced at the anode resistor tap of tube
21,
and this pulse is fed to both input capacitors of
trigger
4,
turning it
ON.
This leaves only trigger
4
ON, and the counrer stands at
4.
A fifth entry pulse again turns trigger
1
ON,
and the counter stands at 5 (triggers 1 and 4 ON).
A sixth entry pulse turns trigger
1
OFF,
causing
trigger
2 to go
ON,
and the resulting positive pulse
has no effect on trigger
4;
therefore, the counter
stands at
6
(triggers 2 and 4 ON). A seventh
entry
pulse again turns trigger
1
ON,
but has no
effect on trigger 2,
thus leaving triggers 1, 2, and
4
ON,
and the counter stands at
7.
An eighth entry
pulse turns trigger
1
OFF,
causing trigger 2 to go
OFF;
as trigger 2 goes
OFF,
it turns trigger
4
OFF.
When trigger 4 turns
OFF,
it produces a negative
pulse which is impressed on the grid of tube 1 in
ELECTRICAL PRINCIPLES
95
trigger
8.
This negative pulse stops conduction in
tube
81
and causes trigger 8 to go
ON
and the
counter stands at
8.
When trigger 8 goes
ON,
the anode of tube
81
is at high potential
(+
136 volts).
Since this po-
tential is applied ro the
grid of tube B through the
voltage divider
R1R2, it means that the potential
at the grid of tube B will rise above cutoff and tube
B
becomes conductive. This rube is used on the
tenth
pulse. A ninth entry pulse will again turn
trigger
1
ON
and have no effect on other triggers,
thus
learing the counter at 9 (triggers 1 and 8
ON).
A
tenth entry pulse turns trigger
1
OFF
which in turn passes a negative pulse to trigger 2
and to the left side of trigger 8. This negative
pulse tends to trip trigger
2
ON,
but such action in
trigger
2 demands a rise in potential at the anode
of rube
21. Since tube B is now conductive, it holds
the potential at the anode of tube
21 to t40 volts
and overcomes the attempred rise in
potential at
the anode of tube
21 and at the grid of tube 22.
Thus, with tube
B
conductive, trigger 2 is blocked
from turning
ON.
The negative pulse produced
by stage 1 is also applied ro the grid of tube 82.
This causes tube 82 to stop conducting; consequently, trigger 8 goes
OFF,
and the counter has
restored to zero (no triggers ON). When trigger
8 goes
OFF,
the potemtial at the anode of tube 81
drops abruptly from +I36 to t40 volts, while
the potential
at
the anode of tube 82 rises abruptly
from
t40 volts to
+
13
6
volts. A tap on the anode
resistor of tube 81 furnishes a
-50 volt pulse to
operate the carry trigger.
If
a positive pulse is
desired to operate a switch when a carryover
occurs, a tap on
the anode of tube 82 can be used.
Any digit can then be added into this counter
by applying the proper number of negative pulses.
If
6
pulses are applied, the counter will stand at
6,
since triggers 2 and 4 will be
ON.
A counter
will retain a reading as long as power is applied.
When a counter is to be cleared, it is merely a
matter of opening the
-100 volt cancel line. This
applies
+I
jO volts to the right side of all the trig-
gers and they are all turned
OFF,
thus restoring the
counter to 0. Cancelling is always necessary before
reading into a counter, because when power is first
turned
ON,
the triggers may assume any status, depending entirely upon chance or upon variations
in individual tubes.
Observe that on the
tenrh pulse when trigger 8
goes
OFF
and the potential drops at the anode of
tube 81, triode
B
is rendered non-conductive, and
its anode potential
immedia,tely rises, thus releasing
trigger
2 from the blocking action.
If this oc-
curred too soon, the tripping
pulse produced by
trigger
I
on the tenth pulse might still be effective
to turn trigger
2
ON.
To insure against this, the
blocking action of tube B is maintained for a short
time after
,trigger 8 goes
OFF
by maintaining con-
duction through tube
B
for a short period after
trigger 8 goes
OFF.
This insures that trigger 2 is
not unblocked until the tripping
pulse from trig-
ger
1 is spent. It is for this reason that the grid of
tube
B
is coupled to the anode of tube 8. through
capacitor C. During the reversal of trigger 8 to
the
OFF
status, the p'oten.tia1 at the anode of tube
8, is rising rapidly while that at the anode of tube
81 is dropping rapidly. The rising potential is
transmitted by way of
,the capacitor C to grid re-
sistor Rg of tube
B.
This rising potential counter-
acts the effect of the dropping
potential on the grid
of tube B and maintains the grid of tube B above
cutoff until capacitor C is fully charged. Thus, the
grid of triode
B
does not follow the anode of tube
81 immediately, but is held above cutoff potential
for a definite delay period determined
by rhe
charging time of capacitor C. Thereafter, the
low
potential at the anode of tube 81 is effective to hold
tube
B
cut off as long as trigger 8 is
OFF.
In practice the blocking tube is one-half of a
12SN7 twin triode. The other half is used as the
blocking tube for another counter position. For
this reason and to facilitate the handling of the
tube chassis, counter chassis are built with two
counrer positions per chassis as shown in Figure
93.
Each chassis also contains an indicator light block
TYPE
603
ELECTRONIC MULTIPLIER
Tube
1
Tube
2
Tube
4
Indicator
\
I
5th Counter
Position
\
Lioht Rlnrk
6th Counter
Position
Blocking Tube
Tube
10
A.
Front View - Showing Indicator Block
Carry Output
5th Counter
post 21
Indicator
Position Light Socket
Tube
Socket
4
/
Tube Socket
1
Input 5th
\
Post
1
Counter Position
\
Carry Output
Position
'
\
\
6th Counter post 26 Socket 10
Blocking
libe
Socket ~dst 6
0.
Rear View - Showing Terminals
Figure
93.
Two-Position Electronic Counter
(K
Chassis)
Loading...