Elliott 803 news, more tidbits from the archive and several generous donations to boot

Post this page to popular social media

Another bumper edition of 'tails from the museum!' ......

Elliott 803 update

It's been a while since we posted an update on the Elliott 803, maintained by PeterO. On the whole the system has been running very well with very few problems (other than some occasional startup problems when cold which often went away after the system had warmed up). It coped very well on the hot days although it did get a bit too hot one day so was shut-down to protect it. We still get the occasional core store fault which has been fixed by either adjusting or replacing a sense amplifier. More recently we had a problem with the paper tape punch 'mis-firing' and producing the odd error.

The most recent problem was something new, discovered while tracing one of the store faults. This was a fault in the floating point (FP) unit. I will digress a little first before coming back to the problem... Not all 803s were sold with the FP boards installed as Elliott decided to charge extra for the feature. Later they began shipping the 803 with the boards installed but disabled, thus making it much easier to 'field enable' the feature should the customer buy the upgrade at a later date. The system we have in the museum was actually used by Vaughan Systems to develop software for the 803 so had all the various features installed and enabled. Now, back to the problem...

Often after repairing a store fault, the Algol compiler would be run as it uses all of the system to operate and tends to show up problems even some of the test programs don't see.. In this case it became clear there was something wrong with the FP calculations. Running the FP test program X5 resulted in all tests failing, so a fairly major fault had developed, and according to John S, FP errors were one of the more complicated problems to find... that was a challenge PeterO would enjoy! I'll let PeterO explain the techie bits

"I started by checking that the FP result was at least getting put into the accumulator even if it was wrong. That was OK, so I started to run a very simple test programme.

+5120$

70 0:20 5119

60 5119:405120

)

This reads the keyboard into the accumulator and stores in in location 5119. Then it adds (in floating point mode) location 5119 to the accumulator, which should double its value. Since the 803 diagnostics unit was still connected I moved the connections to show the contents of the accumulator. Something very strange was going on... After experimenting with the value set on the keyboard I deduced that there was a problem with the logic that controls the pre-add normalising shift which is performed when the two values have differing exponents. With board 4 on the extender I confirmed that in the case where the two exponents were equal the output from the circuit that subtracts the two exponents was not zero. Tracing back from there quickly showed that the output of gate 4B4 was always zero, suggesting a dead OC84 amplifier. I tested the transistor once it was removed and it was indeed dead. With a "new" one installed all the X5 tests passed :-) I then proceeded to bore any visitors who would listen with the detail of my triumph :-) At that point (about 4pm) I realised I hadn't had anything to eat, so I was grateful of a couple of pieces of yesterday's pizza (courtesy of the microwave!)."

An unexpected donation

One of the peripherals that was often seen on the Elliott 803 and 503 system was a Calcomp 565 plotter. This was a small drum based pen plotter that took A4 sized paper (as well as tractor fed paper) and could be connected to one of the interfaces on the PTS (Paper Tape Station). It had a very simple interface and could be directly addressed to move the drum on the y-axis (up and down motion) and the pen on the x-axis (left and right motion) with an additional control to move the pen up and down (off and on the paper). We discovered one of the plotters a few months ago while searching for other equipment in one of our storage locations. It was in a very sorry state with dents in the drum and the control wire off the spools, but it was missing one vital part.. the pen holder.

Work began recently in restoring it to working order as it would be good to have something other than music or paper tape to show off the 803's capabilities. John W carefully took it completely apart and began an inspection of the mechanism and electronics. PeterO also made a start in creating an interface on the 803 PTS to enable the Calcomp to be controlled. PeterO had also done some research and found a website containing detailed drawings of how to make a replacement pen holder. This work was likely to take many months but it looked like we could get it working again....

A few weeks ago PeterO received an email from a visitor offering a Calcomp 565 (he may well have seen ours on a recent visit). What was even more interesting was it was complete... it had a pen holder and some pens. Apparently the plotter had originally been used on an Elliott 503 and was 'rescued' when it was no longer needed. It was last used on an Apple ][ in 1996 and has been in dry storage ever since. This is really good news and we hope to get it in a few months time. The one we currently have will now be used for spares.

Tidbits from the archive - Early Work on Computers at Bletchley

In 1976 Donald Davies, Superintendent of the NPL Division of Computer Science, commissioned a series of lectures from some of the people that took part in the germination of computer technology. The first of those lectures was by Professor I.J. ('Jack') Good on the Early Work on Computers at Bletchley.

Good, having just obtained his doctorate at Cambridge, began work in Hut 8 in May 1941, assisting Alan Turing's team in breaking the German naval ciphers. He obtained a degree of notoriety by sleeping on the floor of Hut 8 when on night-shift, much to the displeasure of Turing. However, all was forgiven when Good 'saw' the solution to a problem he was experiencing in breaking the doubly enciphered Offizier messages in a dream.

His paper, linked below, talks about his work with Turing and about life at Bletchley Park during those dark days. Some of his comments would be considered politically in-correct in this modern day and age.

After the war he moved to Manchester University to continue his early work on computers with Turing and Newman, followed by spells at GCHQ, Princeton University, IBM, Trinity College, Oxford, and the Atlas Computer Laboratory before moving, in 1967, to Virginia Polytechnic Institute and State University. In 1969 he was appointed University Distinguished Professor and, in 2004, Emeritus University Distinguished Professor. He died, of natural causes, in Radford, Virginia, in 2009.

The 'other' donation

Over the years it has been a constant struggle to find the necessary funding to pay overheads and to keep expanding and developing the Museum. BCS, PGP Corporation, IBM, HP and InsightSoftware.com have given tremendous support and individuals and visitors to the Museum have been very generous.

So, last week’s largest ever consortium [donation of £100,000 from Bletchley Park Capital Partners and its associates}(/news/news-releases/tnmoc-receives-its-largest-ever-consortium-donation) has been fantastic news. It is a marvellous morale booster for the Museum volunteers who over the years have given huge amounts of their spare time to help make it one of the top three dedicated computer museums in the world.

This new funding is tremendous recognition of what has already been achieved and enables us to plan with greater confidence for the future.

And finally.... How you can help the museum

Finding the necessary funding to keep the museum running is an on-going challenge. The Museum receives no government or lottery funding and is dependent upon the generosity of organisations and individuals who value our work.

As part of the process of securing funds for the future, two major initiatives are underway: The Corporate Foundation Sponsorship Programme which enables companies to donate £12,500 per year, for 3 years and the membership package for individuals which allows individuals to help the museum by donating £45 for a year’s membership of the museum. Both schemes offer incentives and preferential rates to exhibitions and our events. We have our first two companies signed up for the foundation scheme and almost 100 individuals have signed up to the membership scheme.

So blatant plug time again...

If you are a company that is looking to contribute to a very worthy cause, run by a very dedicated team of volunteers please get it touch with Kevin Murrell. If you are an individual who would like to help the museum, please sign up for the membership scheme.

Support us

The Museum has not received government or Lottery funding, so your help is needed.

Become a member »
Make a donation »
Donate equipment »
Sponsor us »