An edited transcription of a talk by P. J. Walker, delivered on 7th February 1995 to the North West Group of the Computer Conservation Society, at The Museum of Science and Technology in Manchester, England.

(There is a abridged version of Jeremy's talk in PDF format at RESURRECTION The Bulletin of the Computer Conservation Society Issue Number 14 Winter 1995/96 ---

[NOTE: Figures in Square Brackets refer to the Slides listed in Appendix 1 of this paper]


Good evening, Ladies and Gentlemen. To an extent Peter Hall has stolen some of my thunder because I wanted to say first of all who I am and something about why I'm qualified to be here. [1]

I'm 60 at the end of this month, but when I was 20 I was the "young lad" who joined the late-development team of DEUCE at English Electric Company's (EEC) Kidsgrove Laboratory.

Initial development of DEUCE had taken place at NPL, the National Physical Laboratory, and at the English Electric Company's Nelson Research Laboratories, and though I subsequently spent a good deal of time at NPL, associating with the early luminaries, the Pilot ACE and eventually ACE itself, I joined the DEUCE team in late-1955 as it was recruited at Kidsgrove. Our task was to take a design proven on several prototypes and make it fit for routine production and reliable operation. We also undertook the on-going development of the System, adding further peripherals, more high-speed store and additional computing functions - delivered as new "Marks" of machine.

I've been somewhat exercised about the title of this talk. When I talked to Ben Gunn I decided on 'The Digital Computer DEUCE' and then, after a few months of research had brought memories flooding back, I decided something like 'DEUCE - its life and times' might be a better description of what I'm going to talk about tonight. For your interest, I gave my first talk entitled 'The Digital Computer DEUCE' to the IEE in 1959 - very nearly 40 years ago!..

I'm delighted to be here, in part because Manchester was a hotbed of the computer-activities of Ferranti and Manchester University - our competitors at the time - and I hope to set straight part of the historical record. In most of the literature, the people who get the credit for developments in British Computing invariably tend to be associated with Manchester and perhaps this is an opportunity to put English Electric's place in history a little more positively.

I must make the usual apologies because shortly I'm going to put up a slide of notables in the English Electric DEUCE business at the time and I shall inevitably have missed someone. So I apologise to them although they're not here tonight; however, acknowledgements are due to two particular people: Colin Haley - ACD Haley just to distinguish him from Gordon Haley late of ICL Manchester, and George Davis; Colin on the hardware side and George on the software side of the DEUCE development. English Electric didn't seem to get the publicity that we think we should have had - in retrospect probably for two reasons. Firstly, English Electric had decided that, as a big engineering company itself, it needed computers; furthermore, people at the Royal Aircraft Establishment, Farnborough, were quite certain that they also needed this new tool. So English Electric's first objective was to use it and to sell it to a small number of specific customers thus, initially, there was no requirement to 'Market' the machine nor seek special publicity. Secondly, Ferranti took "computers" a lot more commercially and did a very much better job of marketing and publicising their activities.

Each time I see something reported about early machines there is almost always no mention of DEUCE. In fact, if any of you have read the Official History of ICL, you will find it an interesting book, but here too, I could wish that the author had looked a little more into what the English Electric side had achieved.

My next Slide [2] is just a little background of what English Electric had achieved because I want to set DEUCE in the continuum of our computing-developments. From NPL came the Pilot ACE which was intended to be a pilot of the ACE (the Automatic Computing Engine) and which we engineered as DEUCE and was developed through three particular stages: the Mark 1, 2 and 2a.

At the same time, English Electric developed independently of DEUCE, at Kidsgrove, a process-control machine called KDN2, realised that with a little tarting up it could become a commercial machine, and developed it into KDF6 - later KDF7. KDF9, that well known 'stack machine', was also developed at Kidsgrove, technologically independent of the other machines, and found a certain favour with computing-intensive customers. We then got together with RCA and engineered, or re-engineered, the RCA501 to become the KDP10, and later the KDF8, aimed entirely at the emerging marketplace of 'commerce'.

So that's the spectrum of machines that we were involved with during those 10 years.

I've got a couple of slides, one of which shows a typical DEUCE at the time - probably the largest installation that we might ever have sold [3] - and one of a KDP10 at Commercial Union's establishment at Exeter in 1961.[4] Perhaps you can distinguish that young lad sitting at the console as me; the young lady is Jean Dace - nee Thomas, who married David Dace, lately Technical Director of ICL but at the time Chief Programme Analyst for Commercial Union in Exeter. Jean previously worked for English Electric in Marconi House in London as an early Programmer but she was drafted-in for this 'Photo-call', handed an Operators' yellow coat and asked to change a tape reel!.

I said earlier I've been somewhat exercised about the content of my Talk tonight - "The life and times...." seems to me to be a better description than the one I selected because it's certainly not my intention to send you out of here able to fix a broken DEUCE (even if one still existed) nor to be able to programme one, which I suppose is still, in theory, possible. [5] I've talked about my qualifications to be giving the talk and I've said something about the computer developments we undertook during that ten years. I'm going on to talk about the background-activity at NPL just to give some indication of where it sits in time, what we did at NRL at Stafford and in London and at Kidsgrove, and then something about the machine itself.

I've cribbed the next Slides [6] [7] from a recent Guardian in which Turing's memory and achievements are honoured, because, after the War, Alan Turing returned from Bletchley Park to NPL with a "stack of documents which effectively was the specification for ACE". So the first project at NPL was to be a pilot of an eventual ACE, decided upon by Harry Huskey (of the National Bureau of Standards) and Sir Charles Darwin (of NPL). They put together the necessary development team, starting with three people charged with setting up an Electronics Section to build it: Messrs Wilkinson, Donald Davies and Newman. Sir George Nelson, (later Lord Nelson of Stafford) Chairman of EEC and at the time on NPL's Advisory Council (a quango) offered English Electric's help in this particular endeavour, and Colin Haley was recruited in 1949 to 'make it happen' and indeed it was he who coined the name of DEUCE - for Digital Electronic Universal Calculating Engine and the 'logical' successor to ACE!.

Colin describes the ACE pilot as a "dog's breakfast" and from 1949 through to 1951/1952 they set about engineering what was very much a 'laboratory model' into a more robust entity, to be known as DEUCE. Wilf Scott, who, later, was Managing Director of English Electric Computers, together with a major interested customer - RAE at Farnborough - decided that English Electric had a real requirement to deploy this computing power in their businesses. So, seven DEUCEs were built at NRL, the Nelson Research Laboratories at Blackheath, with a first delivery in 1952. Some time after that, it was realised that DEUCE was no longer a laboratory curiosity and that it should be sold commercially by EEC.

To this end, continued development and readying for production was transferred in 1954 to the fairly recently formed Industrial Electronics Department (IED) at Kidsgrove. At the same time, I joined the IED (though not initially in the computer lab), from English Electric at Bradford and my then boss, Derek Royle, was tasked with the job of recruiting a team to take the machine forward. Despite the re-engineering of the product that had taken place at NRL, it was hardly reliable and so a good deal of work was to be done on making it reliable and enhancing it as necessary, with all further development and output taking place from Kidsgrove.

It's worth looking at some of the typical customers - this Slide [8] is as complete a list as I and several other people have been able to put together. The asterisk signifies "Built at NRL", the others at Kidsgrove. They're not necessarily in order of delivery and you can see from the list that with very few exceptions, they're all aiming at the scientific community. Farnborough was interested in airframe stability: English Electric's Atomic Energy Division at Whetstone was interested in Monte Carlo methods of neutron capture and so on. As you can see, there are a number of multiple sites - there were three at NRL but that was still very much in a laboratory environment. More relevant multiple-sites were Farnborough where there were two, Bristol Aeroplane Company (now BAe) at Bristol and the other ones further down. MAFF (Ministry of Agriculture, Fisheries and Food, at Guildford) was unique in having three machines and, as I understand it, entirely related to commercial data processing. As I said earlier, EEC were tardy in realising the commercial applications which Ferranti were by then addressing on a worldwide basis.

I've already mentioned a couple of the scientific applications - airframe stability and flutter, Monte Carlo methods of neutron capture prediction and statistics: we also sold one to BP in Aldgate, London where they were analysing seismic studies for oil exploration.

On this Slide [9], at Farnborough, among the earlier deliveries, are the two DEUCEs: their initial configuration included only 32 -column card Input/Output (I/O). We used to refer to them fondly, (though fondly was perhaps the wrong word in view of the trouble they gave us!), as "Hollerith" machines - though they were from BTM. They were the Balancing Tabulator, reading cards at 200 per minute and the Gang Punch outputting at 100 cards-a-minute - they were our sole input/output devices. We had 400 odd 32 bit words of High Speed Store in ultrasonic mercury delay lines and 8192 words on magnetic drum. On the left of the console [10] is the card reader, the punch over the other side, the monitors, on the right, in which one could look at all the contents of mercury-store.

As an aside, the absence of such a facility on subsequent machines I encountered came as a horrible shock to me, because this is the only machine-type on which one could sit there, looking at the Monitor, and 'see it all happening' in the Store as computing proceeded. I believe you could do that on the Ferranti Mark 1 or Mark 1* and, for all I know on the Pegasus, but if you think about it we haven't been able to do that since!.

This Slide [11] is rather out of focus and there's too much incident light: it's another I took myself at Farnborough and the only reason I show it is because it includes the Power Supply at the back of the room - the PSU is very nearly the size of the machine itself!. I shall say a little more about that in a few minutes, under a heading of the 'foibles and faults' which we experienced.

This Slide [12] shows the Magnetic Drum situated at the back of the machine and me adjusting it: I became very, very expert on the Drum system, as, for a year or so, the servos which controlled the Drum's phase-locked rotation and it's head-positioning were, to put it mildly, somewhat marginal in their operation.

On this Cartoon-slide [13], please note the wide 'paper tape' - this is really by way of leading into the staff involved - and on this Slide [14] I have listed all the people that I could think of who were associated with the machine. In fact, as the number of people involved grew and grew as time went on, I decided in the end that I would draw the line under the last person to join the Development Group rather than those more involved with the Delivery Program. For example, Ann and Noel Wesson who are here tonight contributed a great deal to that Program but are outside that timeframe. I have put six names in large capitals: Turing of course, Colin Haley, Cliff Robinson, George Davis, Derek Royle and, of course, myself. A handful of those people are no longer with us but the vast majority are alive and kicking. The titles at the top of the Slide?. NPL is fairly obvious, NRL is Nelson Research Laboratories, not only Blackheath staff but also those at Marconi House on the Strand where most of the software people were sited.

On the right-hand side (of the Slide) is the team that Royle recruited at Kidsgrove to continue with the machine after initial-development was completed. These Slides [15][16][17] are just general views of the machine and people - myself, Jack Richardson and John Boothroyd. Here is another early shot [18]taken in the labs: that's Jeremy Walker nearest the camera, then John Newman, Derek Savoury, located until relatively recently at ICL Bracknell, and Vic Matthews.

So, how did it work? I introduce this section with a Cartoon Slide [19]and a close-up of the Operating Console [20]. There were two noteworthy attributes possessed by DEUCE: one was that because acoustic delay lines are serial devices and inherently slow, the method of structuring the Instruction word was chosen so as to enable one to optimally programme the order of Instructions, so that if you did it correctly, you could actually pick-up Instructions in sequential Minor Cycles - which were 32 micro-second timeframes. This, it was believed, would overcome the problems of the slowness of access to the Lines.

The other of the attributes was Simultaneity. Simultaneity really came big in perhaps the 1960s or 1970s, but in fact DEUCE was able to operate all its I/O independently of Instruction-processing, whether it was card or drum and later paper tape and magnetic tape. So all those operations could proceed together and this was made good use of in two particular areas.

Firstly, although, initially, only binary-data could be read from the 80 column cards, there was enough time between card-rows to carry out binary-to-decimal conversion, so you could do your decimal conversion before the next row came along. Similarly. in the Multiplier-Divider Unit, it was possible to calculate the sign and do any rounding that you wanted to do, (neither of those operations being automatic) whilst the autonomous Mult/Div. Unit was actually operating. This highlights an interesting anecdote of the times: a Programmer called David Ozanne, wrote a very, very good Test Program for Multi/Div but during it's use it was eventually discovered that it ALWAYS failed and continued to fail, even if there was no hardware malfunction..... he had realised you could access a partial result of the multiplication or the division before the operation had completed, and he used this facility to make his programme run faster. That was fine so as long as the machine was running at it's normal high speed but, if you were 'single-shotting' instruction-by-instruction through the programme looking for some real fault, by the time you had got to the Instructions which would have corrected the rounding, the machine had finished the calculation with a result not anticipated by the Programme!. He had built in an inherent automatic-fail mode...he was not popular for a period!

Now let's come on to the machine-specification itself [21]. This, I would emphasise, was the first version of the machine, the Mk. 1 - we subsequently added more delay lines, Magnetic Tape and so on. I was speaking a few minutes ago about multiplication and division times - this was the so-called LONG multiplication and division - and took two milliseconds: this is the operation where, if you weren't very careful, it had finished what it was doing before you were ready to see what it had done.

Basically this was a serial machine - little different to any other machine of its type, which I introduce with Slide [22]. This illustrates, in Block-structure, how Instruction Sequencing, (referred to as 'Control') interfaces with Input/Output, the Magnetic Drum and arithmetic circuits and the mercury delay lines. For completeness I include a Slide [23] showing the Logic Diagram of Control: these are rather detailed Slides and probably not visible at the back of the room - however, I don't propose to go through their operation, you'll be pleased to know!. This particular slide came into my hands in an interesting manner: some years ago, ICL Kidsgrove decided that it would transfer most of its reprographics work to ICL West Gorton. A consequence was a great 'throwing-out' of old documentation. The guy running this exercise knew I was interested in saving-for-posterity this documentation and I was delighted to acquire two pallets, piled high with all the aperture cards - the microfilm aperture cards - containing (amongst other products) all the Manufacturing and Test drawings for DEUCE: including the Logic Diagrams and Circuit Diagrams.

I've given these filmcards to Manchester University and received assurances that all would be carefully looked after: they will no doubt be of interest to Research Students of the future.

This is the list of the Instruction Set of the machine [24]. Note the Sources and Destinations. If, for example, you wanted to add into the Short Accumulator you would send data from the Source containing it to Destination 25 where it would be added into that particular accumulator. One of the more interesting Instruction-groups is this one where there was a Destination 24, known as the 'Triggers'. If you wanted to cause the Long-multiplier to operate, you used the instruction 0 to 24. I'll show you an example of a small programme in a moment. This was the instruction word [25]: it got changed over the years with, for example, P1 (Pulse 1) becoming used to signify the address of a different bank of High Speed Mercury Store rather than the normal High Speed Store. DEUCE was really a 'two-and-a-half address' machine because there was not only a Source and a Destination but also a Next Instruction Source. The Characteristic Number (2 digits) indicated the SIZE of transfer one wanted to take place. The Wait Number (5 digits) enabled one to specify WHEN the transfer was to take place, so that if, say, one wanted to transfer the data in one particular Delay Line into another or to the Drum then the Wait Number controlled when that would happen - remember, this was a Serial Machine with data passing sequentially in time past any given logical point. In later MARKS of machine, the 4 digits between Wait and Timing Numbers were known as 'Auto Joe' and were used to perform Automatic Instruction Modification (AIM). The Timing Number controlled access to the Next Instruction Source and thus which Next Instruction you selected. I mentioned earlier that if you picked your timing number with care, you could optimally program the instruction sequence for fastest operation. The GO digit, P31, signified by it's presence that that Instruction was to be obeyed without further ado: for example, every time the Card Reader was asked to read or the Punch to punch, each card-row generated a GO digit (P31) which enabled the machine to press on to the next Instruction in the Programme. It was possible to inhibit the GO digit so that one could single-step - "single-Shot" - through the Instruction-sequence by hand from the Console Keys.

So in this Slide of a simple program [26], all the Instructions are in Delay Line One as the Next Instruction Source, the Source 0 to D13 implies data will be taken from Source 0, (which was in fact keys on the Control Panel) and sent to the Short Accumulator at Destination 13. So you could put a number up in binary on the keys and send it to D13, then, as Source 27 was a source of Ones, (P1s), and Destination 26 was the Accumulator itself, the obeying of Instruction '27 to 26' subtracted a 1 from the accumulator.

'13 to 28' took the contents of the Short Accumulator 13 and presented it to the Discriminator, Destination 28, which decided whether the Sum was zero or non-zero and branched accordingly: this repeated until eventually the content of 13 became zero, the Branch lead to '7 to 24' which Stimulated an Alarm - a buzzer - and the programme stopped.

So, a very simple little program where I haven't bothered to specify the Characteristic, Wait or Timing Numbers in the interests of simplicity. This Cartoon-Slide [27] says "In case of emergency, break glass" and it has obviously amused you. I should say that, many years ago, I started to collect cartoons published in the magazines and the like, and it occurred to me that they might make suitable points in this talk tonight.

I mentioned that we had Mercury Delay Lines: well at this point I unwrap this - an almost-complete Long Delay Line - which is not here just to repel borders, (massive though it is). This is one of two lines that I possess, not full of mercury, but more or less complete. One of the other DEUCE-components, which I'm very sorry not to have accumulated, was one of the Drums because they were marvellously engineered devices - I shall say more about the Drum, later.

This Long Delay Line, with its fellows was housed in a thermostatically controlled enclosure known, from its appearance, as the 'Mushroom'. This Slide[28] shows the Enclosure - without the mushroom-shaped top which reduced the effect of draughts: sited round the side of it, are the transmitters and receivers by which the data got into and out of the Delay Lines. It was temperature controlled at 50 degrees Celsius and a source of some worry from time to time, as the Line-length and thus the apparent data-capacity of the Storage Lines was very susceptible to temperature variation. We also had a number of Short Delay Lines - four Single Word, three Double Word and two Quadruple Word - which gave much faster access to the data because the Single Word Line gives you the Single Word more or less instantaneously whereas if you wanted one word out of the Long Lines, containing 32 Words, you had to wait for the other ones to 'go past'.

So this Slide [29], shows the signal-path to and from the mercury and in this Delay Line in my hands, at the top, is an assembly here which could be adjusted up and down, varying the position of a crystal-transducer immersed in the mercury, and thus the length of the acoustic-path. Thus pulses would travel down this tube, reflect off these two angle plates at the bottom and back up the other side to the receiving crystal. This Slide [30], shows one of the straight, Short Lines - a Quad Line, for example, was about 600mm long - with the circuitry (common to all Lines) necessary to get data in and out.

The whole idea behind a Serial Store, is that it is necessary to pick the right moment at which to access or interrupt the pulse sequence in order either to insert new data or to extract the data that you want. So access was rigorously sequenced and clocked by the Control Logic and was critically dependent on the maintenance of an exact physical-length of the path-length in the mercury. I've put up this Slide [31] of the Store Circulation Circuit not to explain it in any sort of detail but to illustrate to those of you not brought up on valves, the type of circuit-diagram we used: there was one of this set of circuitry for each Delay Line. Turning to a Slide [32] of an actual circuit, the most common circuit in the machine was the traditional long-tail pair. It's a bi-stable device which was fitted with a large cathode-feedback resistor R7 which effectively stabilised the current through the valve. Because current was switched either to flow in the left-hand or the right-hand half of the valve and not halted, the effect was to provide a more constant load on the Power Supply so that the PSU did not need to be regulated nor was Regulation ever a problem.

Turning to the Drum which I mentioned earlier [33], of course the modern term for it is "Backing Store". As the Slide [34] shows, the actual rotating drum was about 150mm in height: it had 256 tracks and rotated at about 6500 rpm. As I remarked earlier, there were a superb couple of servos associated with it, which, until they were tamed, gave a great deal of trouble. One servo controlled the rotational speed and phase of the Drum, locking the phase of information recorded on it to the basic 1Mhz clock, with position-sensing from a very precisely-cut toothed wheel with 1024 indentations, on the top of the Drum.

At Farnborough, just the other side of the canal, there was a Company known as Pyestock which used to test big turbines: on Start-up, the associated motor starting-current used to depress the mains voltage over the entire area and the Drum dropped out of synchronization and...... you'd had it because phase-position had been lost and your data couldn't be recovered!.

The other servo was the Head Shift Servo: there are two sticks each of 16 heads, one there, on this Slide and another you can't see on the opposite side of the Drum. These were driven by moving coils in these magnets and with feedback of vertical-position by contacts on two arms sliding up and down a linear-potentiometer, called the 'Reset Pot'.

The way that position-feedback was obtained was by applying a voltage across the Potentiometer such that the sliding-contact connected to the head-stick, would pick-off a voltage representing the position of the Stick, and thus the heads. Initially the tip of this arm was just a copper blade which had two bad effects. First of all, the knife-edge of the copper blade fairly quickly wore and therefore the position was indeterminate and, secondly, the wear-debris shorted out the turns of resistance wire on the pot, so making the positional-feedback voltage non-linear. We eventually found a palladium alloy called JM77 from Johnson Mathey which overcame that problem. I have a Reset Pot here and if you talk to any ex-DEUCE Engineer he will probably spit and swear as he remembers the problems it gave us!. On this Slide [35] is, again, the drum itself: all these racks round here contained it's circuitry and that is a Perspex cover, necessary for the following reason. Although the machine, in theory, needed no forced-cooling, it actually emitted some 79 kilowatts of heat and made it rather uncomfortable for people in the room. So the machine itself was cooled by pumping in lots of air... but you know what comes in with air. However well-filtered, over a period of time dirt accumulates and it wasn't long before we were discovering fine lines on the drum because dust was getting in between the heads and the oxide-surface: so the Drum was put inside it's own little enclosure separately blown through a filter. Here's a further Slide [36] of another 'Attendant of the Drum' - Frank Thompson, one of those unfortunately no longer with us.

The Power supply Unit is worthy of further mention: it actually incorporated, inter alia, six big power supply units which produced plus/minus 100, 200 and 300 volts and lots of Amps. If you wanted to examine circuitry up here, (in this Slide [36]) the best way of doing it was to get hold of the handle of the chassis there and there and stand on the baseplate. Unfortunately, on this chassis it's possible for your fingers to touch the terminals, there, and you could quite easily find yourself across plus/minus 300 volts..... this made one jump, to put it mildly!. I mentioned earlier that the PSU was not stabilised, however in areas where there was long-term drift in the mains voltage, we had a mechanical voltage-sensing device and a big auto transformer about one-and-a-half metres tall that kept the mains within its normal bounds but did nothing to provide current-stabilisation. I've already shown a number of Slides in which the circuitry is visible: this Slide [37], of an individual chassis, shows the reverse-side on which valves and other bulky components were mounted. The chassis would be some 600mm long and 300mm wide. All the connections were hand-soldered and here is a Slide[38] of a production line at Kidsgrove in the 1950's, with girls busy soldering things into place.

The initial machine, rather than it's later developments, used some 1400-odd valves of nine types of which the most common was the ECC91 common-cathode double-triode. I've brought along some examples of four of the various types of valves used in the machine, in case any of you are so transistor-bound that you've forgotten what a valve looks like!. This one in fact is an EL81 - it was used to drive the head movement on the Drums. The failure rate of valves was typically about 40,000 hours, so with 1400-odd valves in DEUCE it doesn't take a genius to work out that they're going to fail fairly frequently.

I mentioned card input/output; this was initially by the Hollerith devices but we later replaced them with the IBM Accumulating Reproducer.

Next I am going to talk about Test and Maintenance philosophy. I noticed in your Program of Forthcoming Events that somebody is going to come along to one of your Meetings and talk about Maintenance in the 1950s - so I'm perhaps pre-viewing here something of what he may have to say. In those early days, maintenance and faultfinding, though disciplined, was very much a case of "Hello, I think I'm on to something here" - as this Cartoon Slide [39] shows. Marginal Checking on DEUCE was the order of the day, recognizing that valve-characteristics were very prone to drift. Checking was done with a thing called a Bias Box. If the definition on this Slide [40] had been a little better, it would illustrate how one could go round the chassis' and insert a plug, driven by a cathode-follower, into any of the valve stages. A potentiometer in the Box could change that valve's grid-level up or down thus showing how safe was the operation of that stage. A resistor-value was then changed if it wasn't: there were a lot of Stages to check but they were fairly stable and were all checked on a Monthly basis. There was also a key on the Control Panel which permitted one simultaneously to alter the bias-points in the entire machine up or down, positive or negative, and thus see whether or not an operating-programme was going to fail in the extreme positions - so providing a margin of safety in the normal position. We discovered, quite late on, that an ordinary battery was actually quite a good way of carrying out this biasing and we tended to use these in order to avoid an unfortunate glitch, which could take place when one plugged in the Bias Box. There could be enough of a surge actually to cause a Test Programme, say, to fail and, if one was actually trying to establish whether it failed because of marginal drift, this wasn't very useful. Eventually the result of this freedom was to have half-a- dozen battery-units and frequently, visiting a Site, you would find all the doors open and see these things plugged-in all over the place. This was usually because "so-and-so" must get his 'Rolling pull-out Programme' through the machine, and offsetting biases in this way was the only 'quick-fix' - otherwise needing a lot of time-out to achieve a 'permanent' repair.

The machine had no parity, no other automatic checking and no automatic error-recovery, though in the later additions of Paper and Magnetic Tape, parity was included in those I/O circuits. Parity was not included in the Drum System because the track length was 1024 bits and the argument was that to have a single parity bit in 1024 isn't terribly helpful. As a consequence, all sorts of very clever programmes were put together which used 'Check Summing' to ensure any errors were highlighted. Towards the end of the Production Programme a form of automatic 'double-read-and-compare' was introduced in the Drum circuitry - a tremendous benefit: if the Compare failed then a (programme-controlled) re-read was forced.

The Maintenance Period was 2 hours a day. I was reading some documents last night in preparation for this talk and I came across something that I'd written back in the mid-1950s. This was an adjuration to maintenance engineers round the country - "do not be talked out of your 2 hours just because the machine was faulty yesterday and your customer lost time, surely it needs more maintenance not less".

We used to age the valves, (I mentioned they'd got a 40,000 hour failure rate) and we had a little test rig in which we used to operate valves for 100 hours rejecting those that failed - drifted - prematurely. We had a very interesting maintenance technique and it's related to these 2 valves I am holding. Those of you who can distinguish it will observe that I've got a length of coloured-wire wrapped round each one - this one's red and this one's green. I'm holding in my hand, Ladies and Gentlemen, things that were called "Johnnies": a Left-hand Johnnie (the red one) and a Right-hand Johnnie. Rightly, you all look puzzled; "what the hell is this?", I sense you thinking!. What we did was to cut off the left anode of one half of a common-cathode double triode and in another one the right anode, so that if one wished to force current up a particular path one took out the existing valve and plugged in the appropriate Johnnie. This led to a whole new philosophy of how you prevented the damn things from being left in-place - particularly if there wasn't a piece of coloured wire wrapped around them. I came across instructions I wrote saying, "... never use an unmarked Johnnie, never leave it in the machine, never put back the valve-can that held it in place because then you can't see it anyway". However the technique became a favourite method of training people. I'm sure Noel there (in the audience) who went through one of the Training Courses would have been asked to find faults which were caused by that technique being used by his Instructor, although it was in fact a fault-finding technique.

And that brings us on, talking of valves, to vibration testing [41]. Now vibration testing has, of course, become well known in any electronic gear. At the time we used to refer to some intermittent faults as "tappy" faults....... and we would beat the machine 'to death' in trying to reproduce a failing condition.. Anne Wesson (here in the audience) was there when we delivered a machine to Glasgow University - you were a Maths Post-graduate Student at the time, weren't you, Anne?. She was appointed by Dr Gillis (who managed the new Computer Centre), to adjudicate on the Customer Acceptance Test. The Acceptance Test at that time required three days of totally fault-free running - remember, there was no automatic error-recovery and precious little diagnostic ability. We learned in the end that the only way of getting these machines through such Tests was to subject every valve to a test for Microphony - very heavy vibration, i.e. hit it - and many's the time that one has banged too hard and shattered the glass. The blow also shorted out the internal electrodes and that naturally took the entire machine off with a bang and a crash. However, it was the only way you could get it to run for about four days. The valves started to become microphonic again in due course but that is another story.

This Slide [42] shows three machines being tested in the Factory environment at Kidsgrove: the large cabinet in the foreground is one of the (very large) oscilloscopes we used at the time.

This Slide [43] illustrates the formation of our "Expert Help" mobile service unit, the Digital Computer Mobile Service Unit, later becoming DCMSU/North and DCMSU/South. Jack Richardson, who sadly was killed in an air crash about 15 years ago, is on the left. He and I formed this Unit in about 1958. This was our van, in which we used to dash up and down the country, packed with oscilloscopes, signal generators and spare parts; the Slide was taken near Farnborough. We added Magnetic Tape to the machine using Tape-drives from Decca and here's an example of a Decca deck [44]and, before you ask, the girl's name was Jean Birchall. She was my girlfriend for several years though she subsequently married Norman Annis and both are still living in Alsager.

As I have said, further developments went on based on the DEUCE Mark 1 which started off with only 32 column binary input and output by cards but was fairly quickly changed to 64 column though still binary. We then attached the IBM 528 Accumulating Reproducer, modifying it that so we could have 80-column decimal input/output. Subsequent to that, we added further High Speed Storage in the form of another Mushroom full of Delay Lines. We added the Automatic Instruction Modifier already referred to and a Paper-tape reader and Paper-tape punch.

I thought I'd say something about the most common faults that we had. Funnily enough it wasn't 'dry joints'; dry joints plagued us enormously on KDP10 - our early use of printed circuit boards. No, the problem wasn't dry joints, it was unsoldered joints. The predecessors to Peter Murphy's production-people at Kidsgrove inevitably missed a wire-joint - not your fault Peter!. Another fault with which we had a lot of trouble initially was that all the power lines to the individual circuits went through feed-through de-coupling capacitors and these used to fail spectacularly, going off 'bang' from time to time. I've mentioned the Drum servo synchronism - whenever we had really severe mains-spikes the thing would go out of sync and this could be extremely annoying if it happened in the last hour of a three-day Acceptance Test, because it was by definition a failure. Glasgow didn't have many of THOSE problems, I'm glad to say. The troubles we had with the Head Servo and Reset Pot I've already talked about and also with temperature-drift affecting the length of the Delay Line. I mentioned that the Line-enclosure was stabilised at a temperature of 50 C and precautions were taken in its design to mitigate against draughts. Well that was fine unless it was in a STRONG draught and we had one or two occasions where people would come in and leave doors open - showing it wasn't proof against that!.

When the Lines were first put together by our colleagues at NRL, the crystal-holders immersed in the mercury were made of Perspex but Perspex has a high Thermal Coefficient of Expansion so, as the thing got hot, and even within the limits of the temperature controller, the air-gap between the electrode driving the crystal and the crystal itself would vary so that the Line-length became wrong. We subsequently changed to an Araldite crystal- holder.

Card reader brushes were a problem and card jams. Now we all know about card jams: hours and hours were spent trying to prove that one particular colour of card - you may remember that they had colour stripes - was the cause of the trouble, somehow affecting the stiffness of the card. I remember at Farnborough we spent days convincing ourselves that grey stripe cards were the reason for having card jams - and it was TRUE!. It turned out that the 3 boxes of grey stripe cards in the Card Store were close to a leaking radiator and they were damp. So we were right, the grey stripe cards WERE causing the fault...... . As I've already mentioned, the valves became extremely microphonic as time went on - perhaps our worst problem in normal operation.

Day to day operations.

I've mentioned that we used the Bias Box to map the ability to work of any given valve-stage and so one of the things one did was to see whether the bias-range was still what it had been - was that particular valve still biased to sit nicely in the middle of the range?. We had a facility called Request Stop on the Control Panel which enabled one to set up a particular Instruction and it's Source and then, running a programme, it would stop at that point: so at least it was a method of stopping appropriately during a programme. There was also a telephone dial on the Control Panel by which, in single-step mode, one could 'dial' a specific number of Single Shots and thus enable one to step-on up to 10 steps in a controlled way. I used to covet one of those Dials - it was 'high tech' at the time and I thought I could use this for something: I did eventually acquire one....!.

We had a Program Display facility which was very powerful. It enabled one to punch-out all the instructions in the programme - this was the forerunner to MOP-testing and the like. The Programmer could go away with his entire programme punched-out onto cards and sort-out his problem off-line. I've mentioned Multi-Div, Simultaneity, the sign and partial results which you had time to calculate between successive operations.

I shall finish this talk - with some 'funnies' - hoping that the humorous will stick in your minds.

This Slide [45] is a close-up of the cathode ray tubes on the Control Panel. This, left hand, display was a permanent display of all the Short Lines: four Single Word, three Double Word, and four Quad Word - for these latter, for example, you see four lines of digits. This, right-hand, display displayed all 32 words in any one of the long Delay lines selected by a switch. We had some early 'graphics' - in about 1956: this 'lady' was known as "York's Flossy" because the Chief Mathematician at Farnborough, whose name was York, created this outline and then animated it by a programme which called-up successive different images stored in each of the Long Delay Lines - you could vary the speed of presentation so she would dance from side-to-side - York's Flossy, Ladies and Gentlemen [46]. DEUCE had a door at the back of the cabinet-frame and one went inside to change the valves or to 'beat them to death' or whatever. It was also quite nice and warm and fairly private. I have to confess that I did some of my courting inside these machines and, on one occasion, after I had married, I went late one night to fix a machine at Warton, near Blackpool: I took my wife with me. The Machine Room itself was cold for the reasons I've given but it was warm inside the computer so she took a book and a chair inside!. [47]

One of the favourite tricks, played by those amongst us who smoked, was, when somebody was peering into the Monitors trying to work out what had happened to one of his programmes, to go inside and puff cigarette smoke into the back of the Control Panel so smoke would issue from the front round the sides. This resulted, on more than one occasion, in a man going horizontally across the room to the Power Supply Unit to turn off the machine before it burst into flame...!.

I've mentioned the 600 volt chassis-shock: that was a very real problem but no one was ever actually killed by the shock. However we had some informal training for the experience: people used to charge-up capacitors to this voltage, and throw them, without warning, to someone and if, catching it, you happened to hit the terminals you had yourself a nice little shock!.

Another problem with mercury which isn't always obvious is that it was fairly easy to spill when you were trying to mend a Delay Line, and many's the time I've chased blobs of mercury around the inside of the Mushroom, the Delay Line Enclosure, with a straw with flux on it's end - gooey flux - because (and not many people know this, folks) you can pick up mercury with sticky flux. It was an aluminium enclosure and the stories at the time were that if you left loose mercury in there it would amalgamate with the aluminium. Now I don't know whether that was true but when you look back and think of mercury vapour and the Health and Safety at Work Act........ . I used to chew pieces of PVC- covered wire too and there are carcinogens in PVC...!..

One of the other Funnies - those of us to whom it happened weren't too pleased - was related to the Drum mechanism. It took something over 20 milliseconds to change head-positions on the Drum, so between initiating a requirement for data off the Drum from a particular Head-position and the time you got it took 20 milliseconds. If you tried to access the buffer store during this period you would get garbage. Therefore, a valve-stage known as the Control Magnetics Interlock - CMI - valve V5 in Unit N, provided a 35ms interlock. In order to carry out a check of the Head-shift time, one removed the CMI bottle - sorry - valve, in order to disable the Interlock. It was very easy to forget to put it back, so when you handed the machine over to the Customer saying, "It's alright now, it's shifting heads within, say, 25 milliseconds, plenty of safety", it wouldn't work and one could spend hours looking for a fault....and the fault was that the CMI valve was out!.

I well remember what I can only describe as the thrill that I experienced of positioning the one-third microsecond Clock Pulses within the one microsecond Digit Pulses - such fine-tuned precision: we didn't call them nanoseconds then!.

One of the other interesting things that happened was that ITV, (I think it was), decided they would use DEUCE for Election prediction and, of course, since the machines were far from portable the Election Team had to come to us. I remember one event taking place at NRL in Stafford and one in Kidsgrove. About this time, small transistor radios were just becoming available and I'd made one from a kit, a Sinclair kit, and I was using it in the computer room. It was obvious that it was picking up electro-magnetic interference from the machine and the guy in charge of feeding the data back to ITV was so intrigued with this that we positioned the radio inside the machine - for better 'reception' - and placed a microphone so that as the programme was running, this "Radiophonic" noise was broadcast live. This has to be the earliest example of Electro-Magnetic Incompatibility.

And that brings me to the final group of Slides.

Back in 1956 or 1957, I drew these cartoons to illustrate some well-known phrases of the time - so this one [48] is "His first fault" - you'll notice the 'Tapping Instrument' which he has in his hand - a hammer. One of our problems was, "Digits creeping in" [49]- that word, halfway up the ladder, says "Sssh, Acceptance Tests Start Tomorrow" [50]. The corollary was, of course, we had "Digits Falling Out" [51].

So there it is, DEUCE - some Reminiscences, it's Life and Times. I hope I've given you some idea of what can only be described as "thrilling times". I very nearly brought along my Log which I dug out to help me in preparing this Talk. I kept it for several years whilst running the Field Maintenance Unit; I don't think I was ever in Kidsgrove (my base) for more than a day without dashing off to Ireland or Oslo or Scotland or wherever - and remember we didn't have Motorways then. An interesting thing here: because I was responsible for all the machines north of a line from the Wash to Bristol, the traditional line, it followed that the two machines in Belfast and one in Oslo were mine - so I spent quite a lot of time in Oslo during the skiing season - my computer and skiing skills improving in the best SERIAL tradition!.

Acknowledgements: my thanks to Colin Haley and George Davies for assistance in historical research and to Jenny Weston, Curator of Science, Museum of Science & Industry in Manchester and Jackie Bull, Secretary, for the initial transcription from audio-tape.

Copyright: P J Walker 13/11/1995

Appendix 1 to "The Digital Computer DEUCE"

List of Slides referenced in the Paper

1 Opening27 Cartoon-"break glass"
2 Deuce machine families28 Mushroom (DL enclosure)
3 Large DEUCE installation 29 Circulation system - diag.
4 KDP10 at CU, Exeter - '62 30 DL Circe's. system - diag.
5 Structure of talk 31 Circe's. Unit dia.-Unit S
6 Turing - Portrait 32 Trigger circuit - diagram
7 Turing - Wartime genius 33 Cartoon - "Backing Store"
8 List of Installations 34 Magnetic Drum - close-up
9 Double-machine Site 35 Drum
10 Console + CR & CP 36 Drum & FT
11 System front-inc.PSU(colour)37 Typical chassis- close-up
12 Drum & PJW (colour) 38 Assembling chassis'
13 Cartoon-"Computer Staff" 39 Cartoon-"Hello.."
14 Dramatis Personae 40 3 Mc/s in lab. PJW, LM, FT
15 PJW at Console (colour) 41 Cartoon-"Hit it...stick"
16 JWAR at Console 42 Testing a line of DEUCES
17 JB at Console 43 DCMSU van with PJW & JWAR
18 VM,JN,DAS,PJW in lab. 44 Magnetic Tape Unit & JB
19 Cartoon-"Counting techn'y"45 Console CRTs - close-up
20 Console - freestanding 46 York's Flossy
21 DEUCE Mk 1 Specification 47 Cartoon-"Ms Brown.. "
22 Logical structure - diag. 48 Cartoon-CompTerms .No.1
23 Logic diagram of "CONTROL"49 Cartoon No.2
24 Instruction Set 50 Cartoon No.3
25 Instruction Word 51 Cartoon No.4
26 Simple Programme

© - P J Walker 13 November 1995