PTT - Central Lab (Leidschendam)

The PTT (the Dutch mail and telephone company in those days) needed to know how much traffic they can expect and what kinds and amounts of equipment they need for that. Modeling the behavior of incoming calls was done with mathematical models and when that became too complex with traffic analyzers. These analyzers were built between 1940 and 1950 by a group leaded by Kosten. A visit of the board in 1947 to Harvard, news about the ENIAC and new technologies used to build computers lead to the decision to build a general purpose computer. This wasn't done right away, Kosten and his group first tried to use the new technologies to build a better traffic analyzer[1].

What did happen was the official founding of the Mathematical Department. The goal of this department was to do calculating work and give advise on solving difficult mathematical problems. This group was later renamed to the Dr. Neher Laboratory. The real switch to making a general purpose computer was with two events. One being the arrival of Van der Poel in 1950. The other was when Kosten found out after designing and building a load of 'die-throwing' machines (they called it 'Randomness apparatus') that this could easily be done with an algorithm[2].

Development

The early drums were nickel plated in a small factory outside of the lab. We once sent out a drum to be plated and got it back a week later. On measurement absolutely no signal came out even with the strongest current pulse we could apply! We asked the plating ship what they had done with the thing. "Oh," they said, "we have chromium plated the thing. That looks much nicer than nickel!"

The start had its difficulties and very momentarily successes as had all labs starting on computers. One of these was their attempt to make a memory grid of 32 by 32 out of Williams tubes. This turned out to be very sensitive so it had to be put in a Faraday cage to prevent interference. Once they managed to make it work for an hour without errors but never got it to work again.

Then they started working with more success on making a memory with a magnetic drum. Some design decision had to be made about them, like whether the drum would be positioned horizontally or vertically. Because they feared that when vertically precession would cause problems they decided to place it horizontally. For the coating there where a multitude of options as well, like nickel or ferrite. Ferrite would give more output but would be harder to bring to the right thickness and damages faster so they decided to go for nickel. The Mathematical Center chose ferrite.

The first tests with the drum to see if they could read back what they had written gave strange results. More and more pulses were read then they had written. Kosten found out what the problem was: the belt drive of the motor was working a bit like a Van de Graaff generator and carried charge to the drum and this discharged to the read heads, writing a new pulse. They solved it by earthing the bearing.

Another problem in the early fifties were oscilloscopes. The only widely available ones were for audio signals, only the radar people had pulse oscilloscopes. The lab therefore got the first pulse oscilloscopes from Dumont, and tried to design a special purpose oscilloscope themselves. But this wasn't a success. Much later Taktronix scopes with delayed sweep and vernier came to the scene.

The PTERA was still located on the second floor of an old house behind the building on the Kortenaerkade in The Hague. All good computers in those days were on the second floor of some old buildings. There were not many, and the Mathematical Center machine and the Cambridge University machine were also on the second floor, from which the conclusion (almost) follows!

Computer building projects

  • ZERO was the first project made by Van der Poel and Kosten, a testing project to test concepts to be used in the PTERA.
  • PTERA was designed by Van der Poel and Kosten as the first real computer of the PTT. Because of the budgetary limitations and time pressure it wasn't a particular innovative or creative machine. It was used for the calculating jobs and for telephone traffic simulations.
  • ZEBRA was the successor of the PTERA and designed by Van der Poel, Van der Mey and Kosten. The design was simple and flexible but still powerful. It was designed as a commercial computer and Standard Telephones and Cables in Great Britain has build them.

References

  1. Poel, W. van der , "The Early History of Computing at PTT", Delft Progress Report, vol. 5, pp. 174 - 184, 1980  .
  2. Kranakis, E. , "Early Computing in the Netherlands", CWI Quarterly, vol. 1, issue 4, pp. 61-84, 12/1988.