Launching the v1

One of his first decisions proved crucial to success: to use binary arithmetic instead of decimal). One of the friends whose help was enlisted, Walther Buttmann, was asked to research the pub­lished work of Gottfried Liebniz in the Berlin University library. It was Liebniz who had first studied binary arithmetic in the I7th century.

So in l936 Zuse started making the component parts of his first all-mechanical machine: using metal pins and slotted metal plates, the ends of the slots representing ones and zeroes. The memory was to hold 64 binary numbers of 16 bits each and he successfully completed it with help from friends who laboured to make the thousands of parts by hand. However, the more complex arithmetic unit required greater manufacturing precision than they could achieve. Programs were coded by punching series of up to eight holes into discarded 35mm movie film, which was far cheaper than the commercially–available paper tape.

This machine was named the Versuchsmodell-1 (experimental model 1) or V1 for short. It was followed by a V2, both of which were later renamed the Z1 and Z2 to avoid confusion with the V1 - flying bomb and the V2 - rocket.

The Z2 re-used the successful memory of the Z1 but with an arithmetic unit made from second-hand telephone relays. Here another friend, Helmut Schreyer, came into his own. Like others, Schreyer had done his share of cutting out metal plates for the Z1. Now he suggested using electromechanical relays instead of the mechanical pins and slots.

New relays were expensive, and since funding was coming out of their own pockets and those of friends, every penny counted. A fully mechanical computer had proved im­practicable and a full-sized relay machine would need thousands of re­lays; so a test model was built using just 200 second-hand relays.

By this time, Zuse had developed the design of his future computer to the stage where he had achieved the yes–no (binary) logical structure for the machine and recognized that it was independent of the physical methods used to build it.

An Electronic Computer

The possibilities for a relay computer looked optimistic when Schreyer sud­denly suggested using electronic valves instead. Though they were not then commonly employed for switching be­tween two states, valves could be used in that way and would be far faster than relays. "At first I thought it was one of his student jokes – he was always full of fun and given to fooling around", Zuse has recalled.

About 2000 valves would be needed. Asking for them, and getting them, were two different things in a Germany then at war. Private enterprise stood no chance⁴ so they talked to the German Army Command. Whilst the initial reaction was favourable, the idea foundered when they said it would take about two years to build. "And just how long do you think it'll take us to win the war?" they were asked.

So little help came, but by the end of the war Schreyer had built an ex­perimental computer with just 100 to 150 valves, and gained his doctorate on the way for his work on valve switching circuits. Like the other computers, this too was a casualty of the war. After the war the development of electronic equipment was banned in Germany and so Schreyer emigrated to Brazil. It was there that he died in 1985.

Whilst Schreyer worked part–time on the electronic machine Zuse completed the electromagnetic relay computer, the Z3, encouraged by the Experimental Aircraft Institute. The Z2 had con­vinced the Institute of the usefulness of Zuse's ideas and so it financed the Z3, though Zuse still had to work alone and at home. And he had to escape a recall to active duty for service⁵ on the Eastern Front.

The Z3 was the first general–purpose digital computer in the world. It was completed in 1943. It employed binary numbers, floating-point arithmetic and a 22-bit wordlength, and it has been estimated that it used around 2000 re­lays (and eight uniselector switches) and cost the equivalent of between $6000 and $7000. "The most important thing", says Zuse, "seemed to be to keep the frequency absolutely even, so that one cycle equaled one addition". This he achieved using a rotating disc or roller, each revolution defining one operation. As the disc's speed could be varied, so too could the operating speed of the computer. Sparking at the relay con­tacts was eliminated by making or breaking them before any current flowed, so increasing reliability. Post­war Zuse machines are said to have been "legendary" for their reliability."

Although the Z3 was completed (with the help of friends) it served mainly as an experimental machine and it never went into routine use probably because of the limited capacity of its memory. There are no doubts, however, that it was fully functional, because there are several witnesses to its operation. Though the original Z3 was blitzed out of existence⁶ a reconstruction was made years later, based on the surviving patents, and is now in the Deutsches Museum in Munich.