close

Вход

Забыли?

вход по аккаунту

?

Патент USA US3079532

код для вставки
Feb. 26, 1963
P. H. MCGARRELL
3,079,522
AUTOMATIC MACHINE TOOL CONTROL
Filed March 51, 1958
TARE
12 Sheets-Sheet 1
Feb. 26, 1963
P. H. MCGARRELL
3,079,522
AUTOMATIC MACHINE Tool. CONTROL.
Filed March 31, 1958
@FAD/lv
f
l2 Sheets-Sheet 2
Feb. 26, 1963
F. H.4 MCGARRELL
AUTOMATIC MACHINE TOOL CONTROL
3,079,522
Feb. 26, 1963
3,079,522
P. H. MCGARRELL
AUTOMATIC MACHINE TOOL CONTROL
Filed March 31, 1958
12 sheet‘s-sheet 4
www„Nn„En«wm
smh
3%
CNN
Feb. 26, 1963 _
P. H. MCGARRELL
3,0 79,522
AUTOMATIC MACHINE TOOL CONTROL
Filed March 31, 1958
l2 Sheets-Sheet 5
Feb. 26, 1963
P. H. MGGARRELL
3,079,522
AUTOMATIC MACHINE TOOL CONTROL
Filed March 31, 1958
12 Sheets-Sheet 6
Mmum.ub
FI
wm
m)
»v
QQW
-n___________________ _____________________._
Ww
25g
/
Feb. 26, 1963
P. H. MCGARRELL
3,079,522
AUTOMATIC MACHINE TOOL CONTROL
Filed March 51, 1958
12 Sheets-Sheet '7
Feb. 26, 1963
P. H. MCGARRELL
3,079,522
AUTOMATIC MACHINE TOOL CONTROL
Filed March 31, 1958
l2 Sheets-Sheet 8
idf@ E#
Feb. 26, 1963
P. H. MGGARRELI.
3,079,522
AUTOMATIC MACHINE Toor. CONTROL
Filed March 5l. 1958
f7@
12 Sheets-Sheet 9
Feb. 26, 1963
P. H. MCGARRELL
3,079,522
AUTOMATIC MACHINE Toor. coNTRoL
Filed March 3l, 1958
l2 SheeilS-Shee‘l‘.V` 10
E_îl‘- /4
„38
I.. _ _ _
_
|
’86
_
_
_
_
_ _
’f2
„een |~
""VF ~
SQ-
:5.1mm
AMR
cwi.'
_
_
_ _
_m-
|
_
_ _
____
85
»vzsfr
FF
657-
` ‘
_______________|
»meer
/84
'
I
_) g4'
l
I
/88
I
su"
l
l
mmv.
|
P. 4-
|
I
| /22
@s
l M_, @Q
-FL
/
I
i
I
#40M DIRECTOR
[_ _
_
_
_
_
_
_
»4M/a
_
_
I
|
/87
fu
/
s”
‘E
„555,
_ _ _ _ _ _ _ _ _ _
/aa
, I,
,5o
»f1/MMM@ ` „nur
6‘57'
_
_
_
_ _
CVRc‘a/l'
_ _ _ _ _ _ _ _ _
f-*rî
I
'
l
|
rv Ac: ¿ne I
_ _
_
_
__
I
Feb. 26, 1963
P. H. MCGARRELL
3,079,522
AUTOMATIC MACHINE Toor. CONTROL
Filed March 31, 1958
12 Sheets-Sheet 11
Erg- /Ó
m85
_I
u
l- @
l\
¿cf/87
@0M '_' l n n [_l---acura
E IQJJ
(
c)
Riso; vn? ¿ua/„ç
v
ï
Í
I
i
|
|
"`
l l
'
s |'
5S
l
l
àëo
uw
'
ë _ l :i
|
: 5
,g l
l
l
‘
|
SQIL i
Í
|
§52
vb
z
Y
à
km
A
Il'
but
ma
°
tu
l
'
I
~
ì'
l |
I l
‘s»5g
gf
` :'
g
I
’ i
,I
Í
kl*
l
l
|
|
@u
o
»
|
|
Í`L
n (ej
-
acura
.s
(F)
°
Y
|
t, u2 tJ t4
@ema/¿w ¿,fç‘çwvç
Y
Aff/a5
_1
(C)
l
I’ î
'u ‘L' 0
Í
__
|
|
=
l
(d)
v
r
|
n
sun
l |
â?
~
>
iff/a7
r
Í
s
a
4
l
|
i
l
‘
l
_
‘
l
l
l
l
_I_ì
,ff/a.;
(b)
I
l
Q
(a)
l
|
Üì
5
,
| Il
l
°c
\
«
|
FLA-1_
,
#nel m (A)
(d)
77M!
au”
,
am [ya
/
C
/zv #Af/16v
TIME
(à )
Feb. 26, 1963
le». H. MGGARRELL
3,079,522
AUTOMATIC MACHINE TOOL CONTROL
Filed March 3l, 1958
l2 Sheets-Sheet l2
.E_zîg/ó’
,24
,J
2,5)
l`
.
o
o
`
/5/0. @s ng :En
0a
°
c»
Erg- /9
/46/7
___________ __ __ _iL ___ _ __I
i|
:|
I
ß//45zé'
I
@fw/'f
I
I
l
7.0 ¿V45?
afwwM/äf/x
|
l
I
_Elzezzfar
United States Patent C ” cICC
3,@79,52È
Patented Feb. E6, 1953
l
3,979 522
AUTQMATIC MACHHÍ‘QE 'EMBL CGNTRÜL
Paul Harry McGarrell, South Euclid, Ghia, assigner to
Thompson Ramo Wooidridge Enc., a corporation of
(litio
2
complex and so delicate, as not to be practical for use
immediately adjacent to the controlled machine tool in
the dirt, vibration and varying ambient temperatures and
humidities normally encountered on a factory floor.
It is, therefore, an object of this invention to provide
a director for an automatic machine tool control sys
Fiied Mar. 31, E58, Ser. No. 725,414
15 Claims. (Ci. did-M2)
tem which itself provides an analogue output signal from
chine, so as to automatically produce a desired machined
door adjacent to the controlled machine tool and oper
digital input data and which utilizes electrical logic
This invention relates to data processing equipment.
circuitry which is simple, ei’iicient, and adaptable to com
More particularly, this invention relates to digital data 10 plete implementation by transistors or other similar small,
processing equipment adapted, for example, to control
rugged active devices and logical elements so that the
the operation of a machine tool, such as a milling ma
director may, in practice, be placed directly on the factory
product in accordance with a predetermined digitally
coded numerical input program.
Many systems have in recent years been devised for
the automatic control of machine tools of the type which
ated as a unit therewith.
It is a further object of this invention to provide a di
rector for an automatic machine tool control system
which is adapted to convert digitally encoded input data
were originally controlled by a human operator to pro
into a phase modulated output signal suitable for direct
duce a desired product. In general, such systems are
ly operating an analogue servo-mechanism controlling a
based upon the use of a coded program which has been 20 machine.
previously derived by using a generaal purpose digital
lt is yet another object of this invention to provide a
computer to determine the steps necessary to be executed
director for a machine control system which will pro
‘oy a given machine in order to produce a particular de
vide a continuing output signal of fixed value so as to
sired product. This program is then stored on a paper
hold the machine in position when the input program is
tape or other storage medium in a code adapted to the
stopped.
requirements determined by the structure of the particular
it is a further object of this invention to provide a di
control system and of the particular machine being con
rector Which is adapted to automatically execute a pro
trolled. The stored program is then fed to a director
gram which specifies one of a plurality of operating speeds
wh'ch converts the program information into electrical
for each command and to provide in the dfrector manu
signals suitable for controlling a servo-mechanism or other 30 ally adjustable means for modifying the program speci
control apparatus to operate the machine tool in accord
lied operating speeds.
ance with the instructions or commands contained in the
lt is still a further object of this invention to provide
coded program.
such a director having improved and simplified logic
’ Most such systems have in the past used a digital
circuitry which may be implemented through the use
servo-mechanism rather than the simpler and less expen 35 of standard commercially available transistorized “plug
sive analogue servo-mechanism. This choice has been
in” modules any one of which can be readily replaced
largely governed by the fact that known directors pro
so as to facilitate maintenance and insure maximum reli
vide a digital output which must be converted to analogue
ability of system operation.
form if an analogue servo is to be used. In practice, both
It is yet another object of this invention to provide such
the director and the converter have been extremely ex 40 a director wherein the registers storing the data in the
pensive, complex, large, and delicate equipments. There
director each utilize the same easily read binary-decimal
fore, when an analogue servo was used it has been com
mon practice, as a matter of economic necessity, not to
code as is used to encode information on the input tape
so that the particular channel in which any error or de
apply the electrical output signal of the director direct
fect in operation may occur will be readily and easily
ly to the analogue servo-mechanism operating a machine 45 apparent to an operator monitoring these registers from
on a factory floor, but rather to record this output sig
a control panel.
'
nal on a storage medium such as magnetic tape. A
it is a still further obiect of this invention to provide
single computer, director and converter combination at
a director which can be economically manufactured at
a remote location may thus -be used to produce mag
a cost competitive with the cost of the magnetic tape
50
netic tapes which, in turn, are used in conjunction with
equipment alone required in presently used systems.
a suitable tape play-back mechanism to provide an out
It is yet another object of this invention to provide
put to control the servo-mechanism in direct proximity to
for such a director, circuitry utilizing a particular binary
the machine tool on the facto-ry door.
decimal code which leads to simpliiication of the cir
The procedure of interposing a magnetic tape between
cuitry required to convert a digitally encoded number
the director output and the servo-mechanism input, how 55 into a train of pulses equal in number to the magnitude
ever, has a number of disadvantages. First of all, the
of the coded number and occurring substantially uni
tape recording and play-back equipment adds additional
formly over a specified time interval so as to perform
expense to the system. Secondly, if the magnetic tape
a linear interpolation between commands.
lt is a further object of this invention to provide im
is stopped by a human machine operator, there is no con
tinuing output to hold the machine tool in position and 60 proved circuitry for digitally phase modulating a stand
the continuity of the program is lost. Similarly, the
ard reference rectangular wave by means of a train of
speed at which the machine tool is programmed to oper
ate is normally ñxed and can not be changed to allow
for diiîerences in hardness of the material being cut or
for differences in cutting tools used in diiferent opera
tions. Furthermore, the use of magnetic tape leads to
an inherent inflexibility in that other functions performed
by the director are not under the control of the human
o erator while the machine tool is in operation. In spite
of these diiiiculties, magnetic tape has, in the past, been
necessary since directors for processing the program out
put data of the computer have been so expensive, so
pulses of predetermined number occurring in a prede
termined time interval.
it is another object of this invention to provide ap
paratus for generating a predetermined number of pulses
in a predetermined interval of time in accordance with
sto-red information.
It is a further object of this invention to provide im
proved electrical circuitry for synchronizing each of a
70 train of randomly occurring pulses with one of a train
of pulses of fixed frequency.
Briefly, in accordance with one aspect of the present
3,079,522
3
invention, a paper tape reader, a director, analogue servo
mechanisms and a control console, are placed immedi
ately adjacent to a machine tool to be control-led on a
factory floor, there being one servo for each axis of mo
tion of thecontrolled machine. In normale-poration,
»a program which has previously been derived from a
computer or by other means, is encoded on «a paper tape
which is read by the tape reader, Athe output of which is
supplied to the director. The director, which is prefer
ably, a fully transistorized unit, converts the `digitallyen
coded information derived from the tape reader to phase
modulated output signals which «are -applied to phase de
modulators or detectors forming the input stage of each
servo system.` The phase detectors derivek unidirectional
voltages which drive servo-amplifiers which, in turn, op
FIGURE 19`is a schematic diagram of a feedbïaclrre
solver used in the system.
Turning now to the drawings, there is shown in FIG. l
a block diagram of an automatic machine tool control
system in which a punched paper tape Ztl is fed to a tape
reader 2l to control a director 22 which controls servo
`systems 23 connected to a machine «tool 24. Various
controls for the machine may be mounted in a control
console 25 which is preferably physically positioned ad
jacent the machine tool Z4, as shown in the perspective
view of FIGURE 18. Ille illustrated machine tool 24
may be a milling machine having four types of move
ments x, y, z and b, such as is shown in simpliñed or dia
grammatic manner in FIG. 2. The various circuits and
components of the director 22 and `also the tape reader
2l may preferably be mounted within the housing of the
era-te hydraulic valves controlling rams or other mecha
console 25. The punched paper tape Zllcontains in
nisms controlling `the yoperation of the machine. Al
coded form, a program defining a sequence of operations
though the operation of these components is normally
to be performed by :the controlled machine in order to
fully automatic, each of these components may,` in ac
cordance with the present invention, also be manually 20 produce a particular desired product. Tape reader 2l
senses the information encoded in the punched tape and
`controlled in some or all of their functions by an operator
converts this information to electrical impulses corre
seated at a control console to which each of «the fore
Is_ponding to the punched holes. The tape reader, which
going components `are electrically connected. The human
may for- example be a commercially available photoelec
intelligence of the operator may thus be interposed to
modify the automatic operation of the machine when this 25 'tric tape reader, reads the tape »one digit at -a time. Since
the reader is not provided with storage faiclities to ac
cumulate the information derived from the tape, it de
livers the information, o-ne digit vat `a time, to the director
272. The director, which will b_e described in greater de
While the novel and distinctive features of the inven 30 tail below, is provided with both intermediate and final
storage facilities, and first starts »the Itape reader to ac
tio-n are particularly pointed out in the -appended’claims,
cumulate anl entire command in intermediate storage.
a more eXpository treatment of ¿the invention, in principal
is desirable sincethe arrangement of the equipment is
such -that the operator has all of the foregoing kcompo
nents within his immediate filed of view and under his
control.
l
y
»
and in detail, together with additional objects and ad
When one command has been executed, thepreviously
vantages thereof, is afforded by the following descrip
tion and accompanying drawings in which like reference
read command is transferred from intermediate to final
characters are used to refer to’ like parts :throughout and
A command may conveniently consist of twenty-live
wherein:
'
storage.
'
' digits 4there being six digits in each of the signed num
bers x, y, `z and b which specify the magnitude of the
motion along the plus or minus direction of one of the
FIGURE 2 is a simplified perspective view of a milling 40 axes x, y, z or b shown in FIG. 2 and one digit specify
ing the lengthof time in which such motions is to take
machine showing diagrammatically the axes of the con
place. When one command has been read and accumu
trolled motions of the machine.
lated in intermediate storage,` the tape is instantaneously
FIGURE Sis a diagrammatic View showing the format
stopped. While one command «is being read into interme
of the information encoded on fthe paper tape controlling
diate storage, the director simultaneously starts to convert
the system of FIG.`1.
>
the previously read command which has been transferred
FIGURE 4 is a ch'art showing a preferred tape coding
f VFIGURE'l is =a`block diagram of an automatic ma
chine tool control system.
.
for the system.
'
.
:inparallel from intermediate `storage to iinal storage, into
.
four separateipulse trains, one to control each axis of mo
FIGURE 5 is a more detailed block diagram of one
tion.- The number of pulses in each train is determined
embodiment of the 'director shown in FIG. l.
FIGURE 6 is a block diagram of a decade counter of 50 lby the corresponding number x, y,'z or b, in ñnal storage;
the type used in the system of FIGURE 5 and includes
The time duration or interv-al for each pulse train is, how'
waveform diagrams »illustrating the operation thereof.
FIGURE 7 is a detailed block diagram of the clock
ever, the same and is determined separately for each com
mand by the clock cycle time encoded as the twenty
cycle control and of the pulse distributor of the `system
fifthV character of each command on the tape. Each of
of FIG.' 5.
55 ythe pulse trains is then directed into one of eight chan
nels, dependent on the sign of the respective number con
FIGURE 8 is a detailed block diagram of the output
trolling the" pulse train for x, y, z and b, respectively, and
is used lto phase modulate one of four fixed frequency
rectangular wave outputs, «to a degree proportional to
FIGURE 9 isa detailed block diagramV of the Iphase
60 the number of pulses in the train, «thernovel phase modu
modulator of the system of FIG. 5.
lator used thus itself serves ias ,a digital to analogue con
FIGURE lO is a volt-time waveform diagram illus
gating matrix and of the sign gates of ’the system of
FIGS.
l
*
tratingthe operation of the phase modulator- of FIG. 9.
verter.
Since the director is such as to be suitable for on-line
FIGURE l1 is a detailed block diagram of a servo
operation,
its output signals Vare applied directly to the
system suitable for use in the system of FIG. 1.
analogue servo-systems or other actuating means or 'con
FIGURE l2 is a detailedblcck diagram of -a modilica 65 trol apparatus controlling the operation of the machine
tion ofthe director shown in FIG. 5.
tool. The phase modulated outputs for the x, y, z and br
vFIGURE 13 is a detailed blo-ck diagram of the syn
channels, respectively, are eachl applied to a phase de
chronizer shown in FIG. 12.
modulator which is the input stage of each of four ana
FIGURE 14 is a detailed’block diagram of the phase 70 logue servo-systems 23 controlling the machine tool 24..
demodulator used in the analogue servo' systems.
Each 'servo-system is, of course, also supplied an un
FIGURES l5, 16 and 17 are volt-time wave-form dia
modulated output which may be used as a reference so
grams illustrating the operation `of the circuit of FIG. 14.
that the servo can determine the degree of phase modu
FIGURE 18 is a perspective view showing a control
lation for each channel and drive the machines- distance
console and a director adjacent a milling machine.
75 along each of its four axes which is proportional to this,
la'
n)
3,079,522
_phase modulation. The degree of phase modulation is,
of course, in turn determined by the program numbers
encoded on the tape.
At the end of the clock cycle time associatted with
the command in final storage, the command will have
d
for both vertical and transverse motion with respect to
table 26 and work piece 3l as indicated by the arrows
y and z, respectively. When power is applied to operate
cutting tool 32, the cutting point may be brought into
contact with work piece 3i by means of manual controls
been converted and the machine operation called for will
to be described below. rIherear’ter, the relative motion
have been executed. Final storage is then automatically
between work piece 3l and cutting tool 3d, made up of
reset to zero, the contents of intermediate storage (the
components of motion along some or all of the four axes,
`cornniand which was read in while the lirst command
x, y, z, and b, determines the shape to which work piece
was being converted from final storage) is transferred 10 3l will be cut. Or" course, it will be understood that
from intermediate storage to tinal storage and the inter
the milling machine shown in FlG. 2 is merely part of
mediate storage register is reset. The tape reader is then
a preferred illustrative embodiment of the system and
started again to read the next command into intermedi
that other machines and/or other aires of motion could
atc storage while the command just transferred to linal
be controlled by the data processing equipment of the
storage is being converted and executed. Thus the di
present invention.
rector assirnilates and converts the information read sen
quentially and incrementally from the tape and provides
in order to facilitate a detailed explanation of ythe
structure and operation of the preferred embodiment of
an output suitable for controlling a plurality of servo
‘the invention vshown in the drawings, a set of particular
systems to operate the machine tool in a continuous man
quantitative values will be given and used in tracing
ner it desired. However, the program tape may, if de 20 through the execution ot a command by the system.
sired, also call for stopping the operation of the machine
While the circuit structure of the system is related to
as will be explained in detail below. Furthermore, each
the logic of the coding to be described, it should be un
ot the units shown in
l, tape reader 2l, director
derstood that both the particular coding and the quanti
22, servo-systems
and machine tool Z4, are con
nected to be manually controlled (as for initiating opera
tion, resetting, etc), from a control console 2:3' in a
manner also to be described in greater detail below.
lt should be noted here, however, that in the present
system tl e director is preferably an all transistorized unit
which is suliiciently small and rugged to be used on a
factory floor immediately adjacent one or more machine
tools being controlled. 0f course, the output of the
director could be recorded on magnetic tape and this tape
then used to control the machine tool as has been the
practice in the past. However, the director ci the pres
ent invention can be manufactured at a cost competititve
tative values to which the system is calibrated are merely
illustrative ot a presently preferred embodiment and
could readily be varied in accordance with the needs of
a particular application as Will be apparent to those
skilled in the art.
in
3 there is diagrammatically shown the gen
eral format of the twentydive characters in a single com»
mand digitally encoded in a single block of information
on the paper tape 2li which is fed to tape reader 2l.
The chart of FIG. 4 shows the articular coding used to
encode any desired character. For the purposes ot this
speciñcation the term “character” will be used synony
mously with the term “digit” and shall be understood to
lean any single letter, symbol, or decimal digit. In the
present apparatus at 1--1-2-5 binary-decimal code is
with the cost of the tape play-back equipment required
for this procedure and, as noted above, is rugged enough
to operate in the presence oi factory dirt, vibration and
varying ambient temperature and humidity. lt is, there
fore, preferred to position all ot the units shown in EEG.
used wherein each decimal digit is represented as the sum
of tour ‘weighted binary “bits” The term “bit” or “bi
nary bit” is, in conformity with general usage, herein
1 in close physical proximity on a factory lloc-1', so that
used to mean a single binary number, that is, a one or a
an operator seated at the control console can see all
zero. ln pure binary notation, of course, each binary
other units and monitor the operation ot' the entire sys
bit implies a weight multiplier or coenicient which is
tem, Of course, it will also be understood, however,
some power of two, the power dependi g on the position
that a plurality of similar machine tools executing the 45 of tl‘e bit in the binary number. Similarly in pure deci
same program can be operated in parallel from a single
mal notatiton each digit implies a coefiicient which is
director under the control of a single operator.
in FIG. 2 the appropriately labeled arrows diagram
matically illustrate the axes of motion, x, y, z, and b,
ot a machine ZA. controlled by the numbers a', y, z and b
lread from the tape. in practice the controlled machine
some power or” ten, the power depending on the position
of the digit. Thus, the three digit decimal number 237
may be explicitly written as (2X162-l-3Xl01~l~7>< 10°).
ln the l-l-~2-5 binary-decimal code used, this same
decimal system is used, but each digit is expressed as a
may, for example, be a commercially available milling
binary number wherein the weights or implied coefficients
machine such as ti e Cincinnati Milling Machine Corr
«of each bit are not ascending powers of two as in pure
pany’s l-lydrotel model 30 X 16 milling machine which
has been provided with four hydraulic valves operated
inary, but rather the decimal numbers l, l, 2 and 5.
ln FEiG. 3 it will be noted that tape 2@ may conveniently
by the servoeysterns 23 to drive rams or other meche
nisms which cause the relative motion between the worlt
piece El and the cutting tool EZ along the tour axes x,
have eight parallel channels arranued longitudinally of
the tape and respectively labeled i, 1", 2 feed holes, 5,
y, z and b, respectively. rl`he hydraulic valves and the
representing any given digit or character are arrange
transversely of the tape, one bit in each of the channels
i, i', 2 and 5, w1 creas the sequence of digits in a decimal
number is arranged longitudinally of the tape as shown.
On the tape, the presence of a hole in a given position of
a particular channel is used to indicate a binary one for
that bit, whereas the absence of a hole indicates a binary
zero. It will be noted that only four of the available eight
channels on the tape are used in the binary-decimal code
referred to above. One of the remaining7 four channels
contains the tape feed holes which feed the tape over a
sprocket in the reader or” the other three channels either
rams or other actuating mechanism may be of any con
venient conventional construction and not shown in de~
tail since many such ’are known in the art. Thus, the table
26 oi the machine 24 moves with respect to the iixed
frame Z? and cutting tool 352 or“ the machine and car
ries the work holder
and work piece ffl with it along
the longitudinal axis as indicatetd by the arrow x. Work
holder 3G is in turn mounted for rotary motion in a plane
perpendicular to table Z5 and about a horizontal axis
lixed with respect `to table 2d as indicated by the arrow
b. Cutting tool 32 ifs mounted in a vertically movable
holder 23 which is, in turn, mounted in a transversely
movable member 29. Member
is mounted for trans~
verse motion (along the axis) in a ñxed upright portion
427a of 'frame 2'7. Cutting tool
may thus be actuated
CK, (D and X.
The above mentioned four binary bits
the X or the O channel may be used to encode a program
stop command in a manner to be indicated below while
the remaining channels may, it desired, be used in a more
elaborate code to control a tape encoding machine, such
l envases
d
7
as the Friden “Flexowriter” which may conveniently be
time corresponding to each clock cycle code letter is
used to prepare and check the punched tape ‘20 from
computed data. It will be understood, however, that
only the format of FIG. 3 and coding »of FIG. 4 are
given in the chart below.
Clock Cycle Tape
Code
pertinent to the present invention.
It will be note-d from FIG. 3 that the ñrst digit of the
twenty-tive digit command in a block to be read by the
tape reader is labeled “sign x.” That is, this digit indi
cates whether the component of motion to be executed
along the x axis shown in FIG. 2 is to take place in the
Clock Cycle
Clock Cycle Tape
Time
Code
(seconds)
plus or min-us x direction. By reference to the detailed
coding given in the chart of FIG. 4 it will be seen that
so far as the ’5, 2, l’ and 1 channels are concerned, a plus
Clock Cycle
Time
(seconds)
%
20
1
40
2
5
10
80
160
320
It is thus apparent that the machine may be pro
grammed to complete the motions -called for in any one
command during a time interval having, for example,
any one of the above ten lengths in seconds. The logic
of the electrical circuitry of the director is such as to
sign is coded as 0110 (holes punched in channels 2 and
1’ only) and a minus sign is coded as 0010 (hole in chan
nel 1’ only). These sign characters thus diifer by only
a single bit in one channel, channel 2, and only this single
effectively perform a linear interpolation in taking the
channel need 'be read. Hence, so far as storage require
machine from the point it was left at by the last com
ments are concerned, only one bit rather than the usual
mand to the point called for by the increment of motion
four bit provision per character need be made for the 20 specified in the command being executed. Thus, if x
four sign characters in each command. The twenty-five
only is specified and y, z and b are zero for a given com
digit command will thus consist of eighty-eight bits. The
mand, for example, the machine will not only move in a
second digit to be read is the ten thousands digit of the
straight line along the x axis but also will move at a
number specifying the magnitude of the motion to be
substantially uniform rate so as to complete the required
executed along the x axis. The other four digits of this 25 x distance at the end of the time specified by the clock
five digit decimal number follow immediately in -se
cycle code letter. The detailed manner in which this
quence. Each decimal digit is, of course, encoded ony
interpolation is accomplished will be described below.
the tape in four-bit binary code in a manner shown in
It is, however, mentioned here in order to point out that
detail in the chart of FIG. 4. Thus a decimal 1 is writ
in the operation of the machine any given curve or
30
ten 0001; 2, 0100; 3, 0101; 4, 0111; 5, 1000; 6, 1001; 7,
curved surface is approximated by a series of straight
1100; 8, 1101; and 9, 1111. For the sake of clarity it
line or rotary component motions. Of course, the smaller
should be pointed out thatthis representation of 9, for
the distance traveled during' each command, that is, the
example, may be explicitly written as
smaller each ‘straight line segment or circular arc b of the
35 approximation is, the better the approximation can be.
It is, of course, assumed that the detailed program of
It will thus be seen that the coded representation of
moti-ons necessary to cut any desired surface to any
each decimal digit is a simple one which can be quickly
specified tolerance will have been computed beforehand
read visually by an operator either directly from the
as 'by the use of a general purpose digital computer. The
results of these computations are encoded on the paper
tape or on banks of neon lights on the control panel
representing the contents of the various registers of the
director which are all arranged in the same coding pat
tern for ease and simplicity of operation, monitoringand
maintenance.
Returning to FIG. 3 it will be noted that the first six
digits read by the tape 'reader are the sign and the ñve
digits of the number indicating the motion required in
the x direction. The units of magnitude of moti-on repre
sented by this number may conveniently -be ten thou
tape 20, in any convenient manner as, for example, by
use of a “Flexowriterf’
Thus, in use, a library of re
producible paper tape programs may «be built up and
kept on hand ready to feed to the tape reader and direc
tor in order to produce any desired item. The computa
tional techniques used to compute a program to be stored
on the paper tape are well known in the art and do» not
form a part of the present invention.
This stored program, encoded on tape 20 in the man
sandths of an inch. Thus, the largest motion which a sin
ner
discussed above, isr sensed by the tape reader 21 and
gle command can call for in either the plus or minus x di 50 converted to an electrical -output in which a binary one
rection is 9.9999 inches. The next six digits to «be read,
(corresponding to-a hole in the tape) is represented by
digits 7 through 12 on FIG. 3, are similarly the' binary
the presence of a pulse on a given channel at a prede
coded representation of the signed magnitude of the
termined point in time and a binary zero (corresponding
motion required in the y direction. Similarly, digits 13
tothe absence of a hole at a given point on the tape) is
through 18 specify the signed magnitude of the motion
represented by the absence of a pulse at the corresponding
required in the z direction, and digits 19 through 24
predetermined point in time. This output is applied to
»specify the signed magnitude of the rotary motion re
director 22 over cable 35 as shown in FIG. 5 which is a
quired in the b direction. The units of the b number may,
more detailed block diagram of one embodiment of
for example, conveniently be hundredths of a degree of
60 the director Z2. In the embodiment of FIG. 5 no pro
rotation with, clockwise rotation being taken as positive
vision is made for manually varying the'clock cycle time
as indicated in FIG. 2. Thus, the largest rotation which
speciñed in the commands of the program. A modifica
a vsingle command could call for would be 999.99 degrees.
tion of the director of FIG. 5 in which such provision for
The last or twenty-fifth digit of the command or block
varying programmed time is made is shown in FIG. 12
to be read is one of the letters A through J which is used
which will be discussed in detail below. It should be
as a c-lock> cycle code to indicate the length of time during
noted here, however, that the two embodiments have
which the machine'tool is to «simultaneously execute the
four moti-ons having magnitudes and directions specified
by the x, y, z and b numbers.
That is to say, each of
these four motions is to take place simultaneously and 70
at such a rate that each will be completed at the end of a
many features in common and that like reference char
acters have -been used to indicate like parts in. the two
embodiments.
In both embodiments, all of the binary bits in a single
digit or character' on tape
are sensed simultaneously
by tape reader Z1. Hence, cable 35, though shown as a
single line, will, of course, be understood to include one
channel for each of the-bits so sensed. This output 'from
A. through J is shown at the beginning of the chart of
FIG. 4 and will not be repeated here. The length of 75 tape reader 21 is applied over line 35 to a data distributor
time interval having a length specified by the clock cycle
code letter. The binary-decimal coding for these letters
3,079,522
control unit 3d, the output from which is in turn applied
over a plural channel cable 37 to the intermediate storage
to its zero indicating condition. The term “set” will
similarly e used to indicate that a bistable circuit is
unit
Again, assuming the exemplary quantitative set
placed in its one representing condition. Which of the
of values mentioned above, the data distributor and in
termediate storage must have the following characteristics
two stable states of any particular bistable device is
and act as a pair to perform the following functions.
nection required are matters of convenience and con
chosen to represent a one or a zero and the exact con
First, they must accept and store information from the
tape reader. The data distributor must route each of
vention which will be obvious to those skilled in the art.
The reset pulse on line 5l is also applied through a
t le twentydive characters to a separate storage slot such
second delay element 52 and lines 53 and 4i to inter
as an individual register in intermediate storage. The 10 mediate storage element 3%. The pulse is so applied to
tape reader must be stopped after a block of information
intermediate storage 33 as to cause the parallel transfer
has been read. The intermediate storage circuitry must
of the contents of intermediate storage 3S to final storage
be such that a complete block can be transferred in paral
43 through line or cable d2 thus leaving intermediate
lel from intermediate storage to linal storage on a pulse
storage 3d in the reset or all zero condition ready to have
command, preferably in less than 50 microseconds. 15 a new command read into it from tape reader 2l. and data
Overall, the data distributor and intermediate storage
distributor 3o. The reset pulse is further applied through
must be such that a complete block cau be read and be
a third delay element 5d and a line 55 to a program
ready for transfer to final storage so that the apparatus
stop hip-flop 56 which will have been placed in the reset
is ready to start to read another~ blocl; in less than one
condition by a pulse emitted from tape reader 2i. over
halt second, that is the smallest clock cycle time used.
line 59 when the tape reader was stopped after reading
in practice, the pulse distributor may, Íor example,
the previous command if that command did not contain
consist ot two cascaded ring counters each of which counts
a program stop code. The progra-m stop coding and
up to 5 and which together, therefore, count up to 25,
operati-on will be described in detail below. The pulse
that is, up to the number of characters in a complete
applied to flip-liep 56 over line 55 is therefore a setting
block. This counter is driven by pulses generated by
each separate incremental motion of the sprocket Wheel
of tape reader 2i. These pulses are applied to the pulse
25 pulse or one applied so as to normally change the state
of the liip-'llop from its reset or zero representing to its
one representing condition. Output is taken from ilip
liep 56 so that only a change from reset to set condition
results in a pulse output which is applied over lines 57
gate each so as to enable or open one set of gates for 30 and 53 to start tape reader 2l and cause it to read the
each particular count. The one ‘bit sign characters are
ext command into intermediate storage through the data
distributor counters over line 39. The counter, in turn,
.controls 2l sets of four gates each and four sets of one
of course fed through the associated single gates whereas
the numeric digits and time letters are each Íed through
its associated four gate set. The output of each set of
distributor. The output pulse from flip-dop 56 which is
v ln operation, the iirst ot the twenty-live sets of gates
intermediate storage into iinal storage. Delay 54, ini
tiated at the end of delay 52, allows intermediate storage
applied over line 5S to start the tape reader 2li is also
simultaneously applied over line dt) to start pulse dis
gates is ted to one of twenty-tive sets of hip-flops or 35 tributor 4S which simultaneously initiates the process of
converting and executing the previous command now
bistable circuits, each set forming an individual slot or
stored in final storage 43.
register in the intermediate storage unit 3S. Twenty-one
it is apparent that delay elements 50, 52„ 54 and iiip
of these registers have four liip-tlops each to accommo
iiop 55 provide between cycle timing. Delay 52 should
date four bit digits whereas four of the registers are
each single flip-flops to store the sign bits. The memory 40 provide ample time for resetting of iinal storage before
the end of this delay signals for a reset and transfer of
thus stores a total of 88 bits.
is enabled or opened while the tape reader is reading the
íirst digit. Since the output of this set of gates is con
nected to the first register in intermediate storage, the
iirst character is thus routed to the proper slot. This
_process is then repeated for each of the succeeding
twenty-four characters. At the end of one block, the
to be reset and transferred before iiip-ilop 56 is set to
again start the tape reader. The sum of delays 52 and
55.- should preferably not exceed 100 micro-seconds. In
practice, these delays may be one-shot multivibrators of
the type which are also commercially available as tran
sistorized plug-in packages, from the above noted source.
pulse over line di), which »is applied to the tape reader 50 Of course, dip-flop 56 is also such a commercially avail
able module.
Sil to stop the reading process.
The structure of i’inal storage unit 43 is quite similar
rl`he ring counters and gates forming the circuitry for
to that of the intermediate storage 3d in that it also con
data distributor 35 and the bi-stable circuits or hip-flop
sists of twenty-live individual registers composed of bi
registers used in intermediate storage 3S may, of course,
be ot' any suitable well known design so tar as the indi 55 stable circuits or flip-flops providing storage facilities for
a total or” 88 bits. The four bits comprising the cloclo
vidual stages are concerned. lt is, however, preferred
cycle code letter are stored in one such register, the out
to torni these stages from plug-in modules consisting of
put from which is applied over a cable d5 to control a
solid state components such as transistors, diodes, etc.
clock cycle control unit do which, in turn, has an output
mounted on a printed circuit board which is supplied
with appropriate terminals. Such basic modules are 60 applied over cable 47 to the pulse distributor 4S as will
be described in greater detail below. As noted above,
commercially available, for example, from the Computer
the sign of each of the numbers x, y, z, and b can be de
Control Company, Inc., Wellesley, Mass., and are fully
termined from a single bit. These four individual bits
described in their various catalogues including., for eX~
are stored in separate registers. Zero indicating and one
ample, Catalog M. Eince these and other equivalent basic
counter having reached a count of twentyh've, emits a
indicating outputs from each such sign register are applied
.stage circuits for performing the functions hereinabove 65 over an eight channel cable 61 to control eight sign gates
described are wel known in the art, they will not be
described in further detail.
When one complete command has been executed, a
indicated by element 62 in a manner and for a purpose
which will be described in detail below.
The remain
ing S0 bits, comprising four bits for each of the tive digits
pulse distributor ¿53, to be described in detail below, 70 of each of the four numbers, x, y, z, and b, are stored in
emits a pulse over line It@ which is applied to a delay
twenty separate ¿l-bit registers, the outputs from which
element 5d and thence through lines 5l to 44E» to final
storage »t3 to reset the bistable circuits of final storage 43.
The term “reset” is, of course, here used to mean that
Aeach of the bistable circuits 'in linal storage 43 is returned
are applied over line d3 to output gating matrix 64 so as
to control the output applied from pulse distributor 48
over line 65 to the matrix in a manner to be described
below. it should be noted that the intermediate and final
3,079,522
Il
storage units each provide for parallel read-in and parallel
read-out from each of their respective registers. The
data -distributor control 36 provides a sequential or serial
read-in as between different registers within intermediate
storage, but the basis storage circuitry within both inter
mediate 'and final storage is parallel in both input and
output.
A further input to the clock cycle control 46 and pulse
l2
representing the command to be executed and applied
over line 63 -to output gating unit 64. At the .end of
these trains of pulses, pulse distributor 48Yemits the pulse
over line 49 which is ltransmitted through delay elements
50, '52 and 54 and, «in tu-rn, resets linal storage, transfers
the previously read command ~from. intermediate to linal
storage thus resetting intermediate storage, again starts
the ‘tape reader in reading the vnextcommand »and simul
taneouslyfrestarts'the-pulse distributor 48- in converting
distributor liiS-is derived from a master oscillator 66 which
’
may, for example, be a free running multivibrator con 10 the next command.
nected to provid-e a square wave »output having a fre
The detailed manner in which the clock cycle control
quency of 1GO kilocycles. This output is applied over
leads 67 and 68 lto pulse forming circuits 69 which'ditler»
entiate the rectangular wave output in -'order to ‘derive
46, the pulse distributor y43, the output gating’unit 64,
v‘andthe sign A,g-ate's62 co~act to produce the above noted
train-s of pulses in accordance with the >information stored
spiked pulses at vthe leading and trailing edges of each 15 in ñnal vstorage '43 Ymay be seen more clearly'by referrectangular pulse. rln practice, this may most conven
ence to FlGURES’óI 7 and 8.
iently be accomplished by deriving a rectangular wave
The pulse distributor .48 which i-s shown in greater
detail in the dashed line block 4'8iofFIG. 7'may, for
output over line 68 which is 180` degrees out of yphase with
example, consist .of ñve >cascaded vdecade counters each
theY output derived over line'67 as illustrated in 4BIGJS
being of thetype shown in greater detail inFIG. 6. Each
in the Waveform inserts 71B and 7-1, respectively. -Pulse
forming circuits 69 then diiierentiate the leading ->edge-of
decadeis constructed of four interchangeable plug-in and
preferably transistorized’ module flip-flop .or vbi-stable cir
each of these Wave forms'to Vderive pulse outputs‘as shown
in FIG. l'5 in the wave form inserts- 72 and 73, respectively.
cuits, plus the external feed-'hack connections shown in
-FIG. 6. 'I‘lfieseV ñour flipañops are indicatedfor each dec
Thus, the point “A” in Wave form 71 coincides with' pulse
A in waveform B and similarly point “B” in 'wave form 25 »adecounters in FlG.ï6 .by lthe blocks FFI, FP2, FFS
76 coincides with'pulse “B” in wave form 72. By difter
«.andFFßl, respectively. Each of these yflip-flop circuits, as
entiating lthe leading edge of these out-of-phase wave
is well kno-Wn in the art,.has two stable electrical states
and may Ibe-triggered from one state to the other by an
forms, pulse forming circuit 69 provides one output over
input ypulse whichmay =be applied to PF1, for example,
line 74 which consists ofthe “B” pulses illustrated in wave
form v’72 and asecond output over line ’75 which consists so over line '80, or,xas -shown in FIG. 7,«over lines 30a and
.18%. The binary zero representing statey of `the flip-dop
'of the “A” pulsesillustrated in wave form v73|. It will
may he termed its reset condition whereasthe binary one
be apparent that'there will be one “A” pulse and -one “B”
pulse for each cycle of the 100 kílocycle mast-er oscillator
representing state of the. flip-flop may betermed its set
condition. The line 8u is »connected to the “binary input
output. Each “A” pulse will thus occur ten microseconds
after the previous “A” .pulse,~ar1d each “B” pulse will 35 terminal” of‘FFl, whereby this phrase isfmeant the ter
minal so connected internally that an applied pulse will
occur ten microseconds after vthe previous “B” pulse.
change the state of the flip-flop regardless of which of
Furthermore, each “B” pulse will occur halfway between
vits two states it is in. When PF1 is >in the reset or zero
two successive “A” pulses, that is, `live microseconds
representing condition, »an input‘pulse applied over line
after the previous “A” pulse.
The “B” pulse output from line 74 is applied'over line 40 Sûwill flip the circuit to >its set condition and in sodoing,
76 to a frequency divider 77 which may conveniently be
a count `down circuit dividing the frequency ofthe “B”
pulsesby 5 so as to derive from the l0() kilocycle input
ya 20 kilocycle output which is applied over line 78 to
an input gate 93 and thence over line 94 to the clock' cycle
controllltî. Of course, it will be understood ‘that master
-the circuit will emit what may be termed a non-carry
.output pulse over line 81. yThat is to say, line 81 is con
nected to the “set” »output terminal which is a terminal
so connected that Ia >pulse appears at it only when the
Hip-flop is changed `from its reset to its set condition.
This iirst output pulseis yindicated as pulse 81a in the
wave form diagram at the right of FIG. 6 which shows
Yoscilla-tor 766 provides the basic timing- synchronization
‘all lof the non-carry output pulsesemitted from each of
`for the entire system in a manner which will be described
'the four `cascaded iiip-tlops asthe .decade counts the in
below. >This master oscillator is preferably >a ’crystal
’controlled vrectangular Wave oscillator or multivibrator 50 put pulses froml tolti. When asecond inpu-t .pulse is
applied to PF1, over line 86,. it returns the FFI from its
which, like the other components of the system, is com
mercially available as a vtransistorized module and will
.rset .condition to its reset or zero representing condition,
.andthe PF1 emits a carry pulse over line 82». That is to
not be further described herein. lt is valso understood,
say, line 82 is connected to the “reset” output terminal of
of course, that the stated frequency of 100» k.c. is exem
plary only'and that> any convenient frequency could >be
EF1 whichis a terminal at which a pulse appears only
when the flip-nop is changed from'its setto its >reset con
ïused.
dition. ì ’ T_his carry pulse is applied to the binary input
It is the function of lthe clock cycle‘control-46and
terminal of FP2 thereby’changing its' state from reset to
pulse distributor 4,3 to derive from the 20 kilocycle input
set. FP2, since vit was changed from a Zero representing
applied over line 78 a train of output pulses foreach
'of the four axes of motion, x, y, z, and b, which trains 60 _to a one representing state, emits'a non-carry pulse over
line 8'3 which -Ísshcwn in vthe :Wave-form diagram as
`of pulses Vare applied over output cable 65 through out
pulse 33a. `The carryY pulse emitted by FP1 whenever
put gating matnixï6á‘, and sign gates 62 to a phase modula
_i'ts state ischanged from a one to a zero ’representing
'tor 125. Each of these trains `of pulses is initiated by
condition is also applied over a line 34 to the reset input
the pulse emitted from program stop >flip-llop 56 simul
rterminal *of PF4 for apurpose to be describedbelow.
taneously over lines 58 and 6d to start tape «reader 21
By “reset input terminal,” of course, :is meant the termi
-and to set a flip-flop in pulse distributor 48 which opens
nal Vso connected that an applied pulse will change'the
gate 93 to start the pulse distributor 48 to convert and
state of the flip-flop only if it is already in a set condition
‘execute the previously read command stored inñna‘l stor
so that it may be reset. Similarly, the term “set input
age 43. Each of these trains of pulses further has a
time -duration the length of which is determined by the 70 terminal” will be used herein to mean the terminal so con~
nected that an applied pulse will change the state of the
clock -cycle'code letter s-tored in linal storage '4S-.and
flip-Hop only if it is already reset so that it may bel set.
applied over line 45 to the clock cycle control 46.
The carry pulses lemitted yfrom PF1 when the second
Each of these trains 'of pulses further consist of a'series
y andsucceeding even numbered pulses areappliedare in
lofpulses equal in number to the corresponding axes
dicated in the Wave-form ‘diagram .at `the left of FIG. 6
number `x, y,'z, 'or b, respectively stored in linal storage
3,079,522
13
labeled “carry output pulses”.
lili
lt will, of course, be
to it of a trigger pulse such as 82e. The detailed cir»
understood that the reset or carry output pulse terminal
of FFî is connected by line S2 to the binary input ter
minal of FFZ, the carry output terminal of FFZ being
connected by line do to the binary input terminal of
FFS, the carry output terminal of FFS being connected
by line itil to the binary input terminal of FF@ and the
is well known in the art and need not be further described
here.
`lt is thus seen that the four cascaded Ílip-dops are
Connected so that when they are all initially reset to
zero by any conventional means not shown and ten input
reset or carry output terminal of FFâ providing an out
pulses are applied to line 3G, the lirst nine output pulses
put along line 9th lt will be noted that the term “carry
output terminal” is used synonomously with the term
“reset output termina ” and that the term “non-carry out
put terminal” is used synonomously with the term “set
output termina ”.
The output pulses appearing at non-carry output lines
dit, d3, 37 and 89' from the respective flip-flops as a se~
quence of ten input pulses is applied to yline Sti are illus
trated in the appropriately labeled wave-form diagram
at the right of FlG. d. rÍhus, pulses Sla, ërlb, Slo, 81d
cuitry by which this latter action per se is accomplished
will change the internal states and produce nine non
carry Output pulses, whereas the tenth input pulse will
not produce a non-carry pulse but will produce an end
carry output pulse such as pulse 9de at output line 90.
In FIG. 7 five decade counters of the type shown in
detail in FlG. 6 are arranged in cascaded relationship.
That is to say, the units decade is provided with a carry
output pulse line Qtlb which is connected to the binary
input terminal of the tens decade. Thus, after ten pulses
have been applied over line âtlb from the clock cycle
and @le appear on line 8l, pulses 83a and @3b appear on
control ¿o to the units decade, the units decade at the
line 33, etc. rEhe output pulses appearing on carry out 20 tenth pulse will emit a pulse over line @Gb which is up
put lines S2, 86, S3 and 9@ of the respective flip-ilops are
plied as an input to the tens decade. The tens decade
similarlyillustrated in the appropriate labeled wave form
is, in turn, provided with a carry output pulse line 90e
diagram at'the left of FlG. 6.
which connects to the binary input of the hundreds
v'vt/hen the entire decade is in its reset condition, all
decade. Similarly, lines 90d and 9de connect the hun
of the flip-flops arc in the zero representing state and 25 dreds, the thousands and the ten thousands decades in
the decade is ready to begin counting. The first pulse
cascade, Whereas the ten thousands decade is provided
applied, as noted above, changes FFl to its set state
with a carry output pulse line rdf which connects to
and provides output pulse dla. The second pulse ap
line e9 for a purpose discussed in connection with FIG. 5.
plied changes FFl back to its zero or reset state and
From the foregoing, it is apparent that when input
the carry pulse 82a along line S2 changes FFZ to its
pulses are applied to line Sill), the units decade will
et or one state, thereby providing non-carry pulse dSa.
emit a carry pulse in the tenth pulse, the tens decade
The third pulse applied changes FFl to its one state,
will emit a carry pulse at the one hundredth pulse, the
thereby providing non-carry output pulse Sib. The
hundreds decade will emit a carry pulse at the thou
fourth pulse applied changes FFl back to zero, its carry
sandth pulse, the thousands decade will emit a carry
pulse changes FFE back to zero and its carry pulse,
pulse at the ten thousandth pulse, and the ten thousands
in turn, changes FFS from zero to one, thereby provid
decade will emit a carry pulse at the hundred thousandth
ing non-carry` output pulse 237e. The fifth pulse applied
simply changes FFîl to its one state and provides non
carry pulse die. The sixth pulse applied changes FFl
pulse. Thus, if for example, the frequency of the input
pulses applied to line Seb is 312.5 cycles per second,
the end carry pulse emitted from line Stlf will be emitted
from one to zero, thereby providing a carry pulse which 40 320 seconds after the ñrst input pulse is applied to the
changes FF?. from zero to one, hence providing non~
units decade over line Sub. lt should also be noted
carry pulse S315.
The seventh pulse applied simply
change FFl from zero to one, thereby providing non
carry output pulse Sid. At the count of eight, FFl
goes from one to zer-o, providing carry pulse 32d which
is applied over line d to the binary input of FFZ and
is also applied over line dft to the reset input terminal
of PF4.
By reset input terminal, at noted above, is
that input line Stia from the clock cycle control is ap
plied as sho-Wn in FIG. 7 directly to the trigger input
of the tens decade rather than to the units decade.
Hence, for pulses applied over line Sila, the pulse dis
tributor counting chain will count to ten thousand rather
than to one hundred thousand, as is the case when the
input is applied directly to the units of decade over
line Sub. For example, if the frequency of the input
pulse will change the state of the flip-liep from one to 50 pulses applied over line Stia is twenty kilocycles per
meant the input terminal so connected that an applied
zero only if the ilip-llop is originally in a one state.
¿FF/t1, however, has already been reset and is in its Zero
state, therefore neither pulse 82d nor previous carry
pulses have any effect on FFßt. However, at the count
of eight, the pulse 82d on carry line d2, changes FFZ
second, the counter will count to ten thousand and the
output pulse on line 93j will appear one-half second
`to zero which, in turn, changes FF3 to zero and emits
8% of the pulse distributor counter so that the counter
will provide an output pulse over line 4% at the end of
pulse
which is applied to the binary input of PF4
`to change it to the set or one representing condition
after the ñrst input pulse is applied. lt is the function
of the clock cycle control to apply input pulses of the
appropriate frequency to the appropritae line Stia or
the clock cycle time called for by the clock cycle code
and provide nou-carry output pulse @9a. At the count
letter stored in iinal storage. The precise manner in
of ninc,rFFl is simply changed from its zero to its one 60 which` the clock cycle control accomplishes this will be
representing state and provi es non-carry output pulse
described below.
Sie. At the count of ten, FFî is changed from its one
lt should first be noted here, however, that while the
to its zero condition, thereby providing carry output
decade counters of the pulse distributor are counting
pulse 827e. rthis output pulse, which is applied over
line ’de to FFd, now iinds
in its one or set condi-
tion and resets FF@ to provide output carry pulse Sida
on line
The carry pulse
which is applied over
line 84 to FFdl is, of course, also applied over line S2
to FFZ and would normally change its state from zero
to one.
'from one to ten thousand, or from one to one hundred
thousand, as the case may be, at a rate determined by
the frequency of the input pulses from the clock cycle
control, each stage of each decade counter will also
provide a non-carry output pulse at the respective ter
However, this action is prevented 'ny applying 70 minals 8l, S3, S7 `and S9 thereof, Whenever the par
a signal over line
from FFd to FP2.. Line 3S is
connected to sense the one representing state of *Fd and
to apply a voltage Vto FFZ such that whenever' FFli- is
>in its one representing, or set state,
will be held
in its zero representing state in spite of 'the application
ticular stage is set or changed from a zero to a one con
dition as explained in detail above in connection with
FIG. 6.
Furthermore, referring to the graphical illus-tration of
the time distribution of these non-carry output pulses
3,079,522.;
15
‘
shown in FIG.l 6„it will be noted that the .total number> of
non-carry output pulses provided by each respective ilip
`hop of the Idecade as the decade counts from one through
ten is indicated by the numbers in the blocks at the right
of FIG. 6 as 1, 1, 2, andl 5, respectively. Thus it will be
noted that’FFl produces tive such non~carry output pulses,
FFZ produces two such non~carry output pulses, FFS
produces one such non-carry output pulse and that FF4
`also produces one such non-carry output pulse. Further
more, no non-carry pulse from any given flip-flop is ever
time coincident with any other nor1~carry pulse from any
other Hip-flop or" the same decade and no non-carry pulse
from any ñip-ñop of a given decade is ever time coinci
dent with that decade’s- iinal output carry pulse. Since
the carry ouput pulse fromone decade is the trigger input
pulse to the next decade, it follows that as the complete
16
z, and b axesv for each decade. Thus, the matrix will
consist of L rows and M XN columns, where L is here 4,
the- number of separate axes of motion M is 4, the num
ber of bits per digit, and N is ñve, the number of digits
per character,
Turning now to FIGURES 5 and 7, it will be noted
that output carry line 90f of the last decade of the pulse
distributor is connected to the reset input 4of a flip-flop 91
and that the line 60 to which the pulse which starts tape
reader 21 is also applied, is connected to the set input
terminal of ñip-ñop 91. Outputs are, in turn„taken from
flip-flop 91 over line. 92 to control a gate 93y to which
the 20 kilocycle output derived from the master oscilla
tor through divider 77 is applied as an input over line 78.
Thus,- the pulse emitted fromv flip-flop 56 which initiates
the operation of the tape reader to read a single command
also sets íiip-ñop 91 and thereby enables o-r opens gate 93
so that the 20tkilocycle signal on line> 78 is applied to the
clock cyclecontrol over lines` 7S and 94. When the cas
some decade for each input trigger pulse» counted except
the last which produces the final output carry pulse. Fur 20 caded counters of the pulse distributor have completed
thermore, no two non-carry output pulses will be time ' their count, the final end carry output pulse which is ap
plied from line 9W to line 49 is also applied to the reset'
coincident. Even more important, these non-carry out~
input terminal of iiip~ilop 91 thus changing its state and‘
put pulses will be numerically distributed among the
thereby closing gate 93 so that the 20 kilocycle signal on
stages of each vdecade as-indicated by the num-bers assov
line 78 is not applied to the clock cycle control while the
ciated with lines 81, 83, 85 and 87 of FIG. 7. Tha-t is,
apparatus is between cycles, that is, while the output pulse
during .a count of one hundred thousand pulses, FFI of
on line 49 applied through the various delay elements is
the units decade produces fifty thousand non-«carry output
resetting final storage and transferring the previously read
pulses, FFZ twenty thousand, etc. Of course, in the ten
counter> goes through a complete count there will> be one
and` only one non-carry output pulse from some stage of
thousands decade, FFI produces only 5 non-carry pulses,
FFZ produces 2, FFS and PF4 one each. The fact that 30
these non-carry output pulses have this particular numer
command from intermediate storage.
s
When the next cycle starts, after the above noted d ~
lays, gate 93 is again openedand the 20 kilocycle signal
ical »distribution and the foregoing logical properties is
is applied over line 94 to a count-down chain of cascaded
used to instrument the above discussed binary-decim-al
l-~1----2--5 code.
Thus, the fact that the total> of the non-carry output
pulses from the flip-hops of a given decade is arranged in
the 1_1-2-5 pattern illustrated in the graph of FIG. 6
is the basic rea-son why the input information is encoded
on tape 20 in the l--l--2-5 binary-decimal code dis
ofthe frequency of .its input since each flip-flop provides
flip-flops 95, 96, 97, 98, 99 and 100 forming a part of
the clock cycle control 46. Each of these flip-flops, of
course, has an output the frequency of which is one-half
one carry output pulse for every two input pulses. The
harmonically related `signals thus -available at the input
and outputs of flip-flops 95 through 100, respectively, are
cussed above. It will be seen that the non-carry output 40 applied to a group of ten clock cycle control gates indi
cated in FIG. 7 as the “A” gate, “B” gate, “C” gate,
line, from each flip-flop or stage of each decade as con
“D” gate, “E” gate, “F” gate, “G” gate, “H” gate, “I”
nected to the input terminals of each of four gates, as
gate, “J” gate, resp. Thus, the 20 kilocycle input to-ñip
illustrated in FIG. 8, and that the control termin-als of
flop
95 is applied over line 101 to the “A” gate and the
each of these gates .are connected by control lines C tothe
registersof final storage element 43 in such a manner that 45 “D” gate while .thevlO‘ kilocycle output of flip-Hop 95A is
applied over line 102 to the “B” gate and the “E” gate.
the gate connected to any storage element containing a
The tive kilocycle output of flip-dop 96 is applied over
line 103 to the “C” gate and the “F” gate. The 2.5 kilo
cycle output from flip-flop 97 is applied over line 104 to
closed. The output gates are thus effectively connected
the “G” gate, the 1.25 kilocycle output from flip-flop 9S
in. a matrix array wherein the non-carry output terminal 50
is applied over line 105 to the “H” gate. The 625 cycle
of one stage of thecounter is connected to all of the
output from flip-flop 99 `is applied over line 106 to the “I”
gates in one column of the matrix and the bits encoding
gate, and the 312.5 cycle output from ñip-ilop 100 is
one of the four numbersY x. y, z, or b in final storage are
applied over line 107 to the “J” gate.
respectively connected to successive-gates in each respec
The clock cycle code letter which is associated with
tive row of the matrix. Furthermore, it should be noted 55
each
command and which is stored in ñnal storage ele
that'the comiections for any given number such as X are
ment 43 for the command being executed, is applied over
made in inverse magnitude relation. That is to» say, the
linev 45 to a decoder 108. The output fromdecoder 108
bits of the ten thousands digit in final storage control the
is applied over a ten channel cable 109, one channel of
output gates connected to the units decade of the pulse
distributor. counter, the bits of the thousands digit control 6,0 which is connected to the control terminal of one of the
respective clock cycle gates as by control lines C-1, C-2,
the hundreds decade of the counter, etc. Finally, of
C-3, C-4, C-5, C-d, C-7, C-tl, C-9 and C-0. Thus, if
course, the bits of the -ten thousands digit control the
the command in final storage includes the clock cycle
units decade of the counter. The open gates will then
one is enabled or open, whereas a gateconnected to any
storage element containing a zer-o is not enabled, o-r is
pass a total number of non-carry output pulses corre
tape code letter “A,” the inputs to the decoder 108 are
so actuated that the output of decoder 108 will supply a
sponding Ito the number represented in linal storage.
65
signal
which holds gate “A” open but which holds all
Furthermore, since none of the non-carry output pulses
of the other clock cycle gates closed. That is to say, only
`are ever time coincident, the outputs from these gates may
the one particular clock cycle gate called for by the clock
vbe connected to a single line such as one of the lines 110',
cycle
code letter in ñnal storage will be held open during
112, 114, 116 so that the sum total ofpulses appearing
on -a given line as the decade counter runs through its 70 the entire period of execution of the command, while
other clock cycle gates are closed. Decoder 108 is es
complete count will be equal to the associated number
sentially nothing more than a binary-decimal to decimal
x, y, z, or b controlling the gates feeding tha-t line by the
converter and includes circuitry all of which is well known
electrical states of the registers representing the number
in the art. For example, the decoder may consist of a
inthe final storage element. Of course, it is necessary
to provide a set cf four such gates for each of the x, y, 75 diode logic network including “and” circuits and “or” cir
Документ
Категория
Без категории
Просмотров
0
Размер файла
4 026 Кб
Теги
1/--страниц
Пожаловаться на содержимое документа