-l

'l
'|

r The prototype computer. 2 The latestvesion redesigned

according to ergonomic principles. uexen E M I ElectronicsLtd.

Ergonomics
for a,
Gomputer

AN ANALocuE coMPUTER is an assembly of many,

'building bricks' which can be consimilar, electronic
nected together in a multitude of different ways. The
input and output leads of the amplifiers, integrators,
and function potentiometers are all brought out sepa'patch panel', which is thus very
rately to sockets on the
similar to a telephone exchange switchboard. By making the appropriate interconnections the operator can
'patch up' the computer to solve most problems which
can be defined in suitable mathematical terms.
Analogue computers are particularly useful to simulate and optimize the design and working of complex
devices with several interdependent variables, such as
suspension systems of motor cars or chemical processing factories with continuous flow.
Flexibility is an important feature of an analogue
computer, so that it can quickly and accurately be

reset from one problem to solve another; easeofoperation and avoidance of operator mistakes are therefore
essential.Such mistakes are far more troublesome than
loss of speed or machine inaccuracy' A correct setting,
but on a wrong potentiometer - or a switch left in the
wrong position - can ruin the solution of a problem
and cause much delay while the fault is found. Even
worse, the error in the solution may not be discovered.
Therefore the risk of these mistakes must be reduced
to the minimum by proper attention to ergonomlc
factors.
II, at EMI ElecDuring the redesign of EMIAC
tronics Ltd, advice from the firm's engineering
psychology laboratory was sought on various aspectsof
the operational layout; the prototype, r, was already
fairly satisfactory, but the project team wished to make
every possible improvement.
DSIGN

I2O

I
I

ElfIAC

Potentiometer panel
The major problem in this study was the panel of z4 potentiometers, all of which may have to be selected one at a time by an
appropriate switch and adjusted very carefully to an exact value
on the meter. Unless the layout is clear and simple, there is a risk
of setting up the wrong potentiometer or of starting to adiust one
which has already been set, with consequent annoyance and risk
of error.
On the prototype, to economise in space, the potentiometers
were staggered at 45" to each other and the jack sockets, through
which electrical connection is made to the potentiometers' were
set in between them. The result is somewhat confusing and there
is no room for numbers to identify each iack socket, 3. Despite
added electronic complication and slight increase in panel size it
was agreed to use push-pull switches with a number on the knob
to replace the iack sockets; the design problem therefore was to
- frsr, find
frnd a layout with which the sequence of operations
third,
and operate the switch, second, frnd the potentiometer,
adiust it, and, fourth, turn off the switch - would be done most
easily and with the least error.
The method of tackling this, as with all ergonomic problems, was

DESIGN

I2O

first to refer to results of all relevant research I ; it was found, however, that the evidence in the literature was not adequate to allow of
reliable extrapolation to solve the problem. Therefore an expcriment2, had to be made which resulted in the optimum layout, 4.
It was suggested initially that a rotary selector switch should be
used instead of the iack sockets. This proved impractical electronically, but the discussion and explanation brought to light a way in
which part ofthe basic engineering circuit could be improved. This
is quite a common experience during ergonomic redesign studies;
explaining the machine to someone with a different specialist
approach

stimulates

the

engineering

team

to

reconsider

its

system from a fresh viewPoint.

r See for instance: Applied experinental psycholggy,,A. Chapanis'
\. R. Garner and C. T. Mbrgan. Chapman and Hall Ltd, f,z tzs
Hwnan engineeringgtide fot equipnent designets'W. E. Woodson' CUP
{r 6s.
Data on hunpn perJornnnce fot engineeting designers,K. F. H. Murrell'
reprint from Ezgineering, 3s

8L'Y;i;.51'X'
Y'f:{r:if:;o,'"r;9*W,i;*h.3X'J""f

H. K. Lewis & Co Ltd, dr Is.

zAnote on Danel layout for nuntbets of identical irerts, B, Shackel, to be

published i Ergononics. Taylor and Francis Ltd

continued

Iirgonomics lbr a corntrruler.

I'ltutogrtil.lt

ol tlk'ftilEI\

r 11. O/irrr

u.'t

ff'l

ES

;p

il,Jb
d l i b d b d f r u " rd N t ' o r ' . " ' !d " * ' d ' r i * r
t r'l r'l,i l,dl
d"'fi
'r
'r
r i i i ..1:'r r r i r;r.* , i.l , i, *,o'.i
* i :u* r
i i t.r.:
.' i i !1.. r r r
" ''r
-r'
'
i
'
r
- r c i
r i " r . r r | | . f ' r . n r " . o
t r a o i r a i'," ir
r . a''a a
.
"';r"*
n r i*i"*.13''i r **;ii.'.3-'.'a'|;.*
r'r'iwir
tll,l"" r *-*r r'i
.",.
r -,r i { l r i
n'C'n,r'. l r''l'3'i..
r * ' n i ' i , i ' - 3 t ' r ' ' . . . . . ' li ' r r *
r i n n u'tii'rI r,
. r . o t t * . j ' c | . o . ' ' i r i ' , r , . . ri " * ' r i ' . r r i " ' , 3 ' . r r , r ;
r'r .ll . . |
'ii'"., 'i r
ir. i i
i
;l
r ,i r i .'lll'!
-t . . . l * r .
i * * *lil'r. r I * n 3"'r
r
"'?'i
"
*'.'i. o **;,.i3'.'i'*;i
a';';';'.*;
r'ij'* r'i*-r illl . r r
t d i;r r.*"*.i*"
'r'.
.,r' . r'i't'J
. r-.i,,.'. .
. * r * *:d r I r .* o J . ! . . ..J-i i,a, s .,.'i..
i ' *r | ; ' i
i .
. r ' , . ; . ' " : ' .r i
. r i i . l d d o r r . , 3 , i r r i r . . .

d
I
.
r
i
.
I
i
.
i

t.

ciq

rffifl

'"ryn

'%,;,"''

'-u-''u'

!$_ '

. ? C

'

,3rro; Sft CfpmSC *rrn qQri *90 qCp SC *Ki#S

1

lrclt

i I rr
! j .
, ia

l l !
{ r ln
s Ir
IIA

Ll.

p I u
v
. i a
.
rl
snp

. i . t l t
/ If''
r -o a

ua

3 t

i o f a
n*"ptrr
'
l. fx-r-]
i
i { . . i*rc"-5n
. . +1+e
-,1-i
- i - i *2 ./ . a oJa' {.!
I

{i 'J'
lis
-a
a

l'\{{

r _! 1

*.?

t
' I

a
3
a
5

a
a

ao'a

jl .. j

k:

a
t

a a
- t l s
Ti
&l .

) s

i :

l a
? i <

li
*r
1 a

rl .
ii

[;

a'ra
a
2l 'slr
Fa

.L-.

74

"-i*I*ffiw.*.:.

.1.
3 !

.&,
f,--t
f-i*----l
?1a{ri
a
a"

.lt
-l
s
a

d:r
a----l
l
a

"*

r i l r : L . ?,

\''Jt "J\''J
Ti----:

].1/w@
I n t ' l r
'--:ol
'

,
'\

' r @
tqt

:h;
"i-r:
a

a 1a

/@,|E+
l r : ?

ffi

+rotr Frt

I a

J 4

a
3

; ;7; "
ffii+l
t
I
Ia
a ''r , a a
O

t|*
a a

I I

2 r +t

'fF

l a

.'t'

i[3

lg
r;
ffil: N.:.
ilY
i:tEffi

3 a la
la:Y$
a

a',,Frt
!

ii;hiffi,i il
fry

*->a

J-ilIl3 37i :

it-I

f *
Fr
,l
a' I

ffir:
!):: ;:ih:ffi

J
3

,:'

-^,

Patch panel
r\lthoush

apprrcntll'coniusing

irnd clcn

lirrhtcning

to

rhc

u n i n i t i a t c t l , r h i s p a n c l , 5 , s i m p l v c o n s i s t so i i n p u r a n d o u r p u r

38

casl'sclcction ol thc riqht sockct cvcn uhcn manv plugs obscurc

n u c h o i t h c p a n c l s u r l a c c . ( l o b u r - c o c i i n g * a s c o n s i d c r c d ,b u t

sockets lirr thc t8 ampliiicrs and z-1 potcntiometers. riong rritlr

'l'hc
brsic eroupine into t8

r b o u t s i x c l i f l c r c n t h u c s u o u l d b c n c c d c d ; t h e r c s u l t i n g d a z z l ci n

som!' intcrconnccting iunction lincs.

l crorr,dcrllrca uould probabll,bc

b l o c k s o i r l m o s t i t l c n t i c a l s o c k c t s , a l o n g l i t h t h c : ; p c c i a ls s i t c h i n g

h c l p . l L a r i o n l l i s l t i o n o 1 a n u m b c r o l r l c t r i l s t o m i n i r l i s e r h c r i s ko l

g r ( ) u p r t t h c l c { t h i r r r t l c n d . r c c l u i r c e ie x t r a c m p l r r s i s . , t r i c n s u r e

crror ctn bc scen bt eomplpjllg thc lrvo lltouTs.

rnorc o1 r iiistmction thanr

l)list(i\ I 20

Control panel
Thc ull opcrating sequcncc must bc knorvn belbrc clccidirrg rm_
portant panel layouts; only bv analysing thc various sub-rourrncs
canrelated controls bc properly grouped together. Critical adjust_
ments and master controlling actions arc usuallv made with thc
prefened hand (thc right hand for most of the population; and
therefore such conrrols should be placed casily for the right hancl,
leavingassociatcd less critical switchcs to thc lei.
It bllowcd that the main changc, T ancl g, u,as to intcrchange
the Meter F S D switch and the accuratc potcntiometcr with rhe

graduated dial; previously the opcrator rvould tcnd ro opcratc

thc
metcr switch u'ith thc lelt hand, thus obscuring the meter he was
about to read, so as to use his right hand br critical adjustmcnt.
Thc master control is Rorv A switch and is thercforc in thc corrccr
thc otlter groups of amplifiers are usually undcr its
control, and thercfore Row B and Row C switchcs arc in thc less
rmportant corncrs of the pancl. Other changes in dctails improvc
position;

thc idcntification

of su'itch positions and group

rclated ircms

togethcr.

tE-

t
r.
t'

ft

"s_

,t*:
ri! $e
n*Brulf

'-fr

r&
l

1::'J

{il"

Pt}'
tt

RlV

px:{

w \ '

gE

Conclusions
In this brief example of how an electronic machine is studied rn
orderto improve the ergonomic aspects of its design, only the
moreobvious points have becn mentioned. To the design engi_
neersand technologists, who are at prcsent the chief users of this
kind of machine, some of the more detailed aspects u,ill be
obviousfrom the illustrations,

Some time after the study had finishcd an opportunity arose to

ask the recipient of the prototvpe model, the Hawker Siddeley
Nuclear Porvcr Co Ltd, or criticisms of its operation. It rvas
pleasing, if not unexpected, to find that comments from
at least
three months' operating of the machine had all been predicted
by
rhe ergonomic redesign study.

_$t