Page 2 of 3
The next step was to build the Z2. In this Zuse replaced the mechanical switches with electro-mechanical switches - relays.
This is the earliest known use of relays in a computer. However the memory remained mechanical and while the Z2 was faster it still wasn't reliable.
It uses a 16 bit fixed point arithmetic unit and the clock rate was increased to 5 cycles per second and it could perform slightly less than one addition per second.
Fortunately one time that it did work perfectly was during a demonstration to Alfred Teichmann, a leading aircraft designer. He was impressed enough to arrange for backing to build a third machine.
The next task was to build a new machine - the Z3 At the time of the Z3 an important new influence entered the scene - King Kong. Yes, the hairy giant ape that climbed the Empire State building and was shot down by biplanes!
King Kong - an affect on computing history?
Zuse was a fan of the film and so were many of his friends, enough to form a fan club. They decided to produce a stage version of the film and Zuse hoped for the part of Kong - but in the end the honour fell to an even larger man - Helmut Schreyer. The play ran for weeks in a Berlin theatre. Night after night Schreyer would lurch about the stage smashing papier mache skyscrapers and being shot by toy planes.
What has this got to do with computing?
The answer is that Schreyer was an engineer and he and Zuse talked endlessly about computers during the rehearsals. Schreyer suggested that Zuse used valves as the switching element in the logic gates. The fact the Zuse's machine was based on logic gates implied that you could build it using any technology suitable for implementing logic gates.
Valves were fast switches but they were expensive and with the limited backing that Zuse had they were well beyond his reach. However the pair did realise that a valve computer was a possibility and so saw the future while playing King Kong. Schreyer did use the ideas in his doctoral thesis in 1938 even though Zuse had to return to relays for the Z3.
The Z3 (rebuilt in 1961)
When the war started in 1939 Zuse was drafted into the army. As a leading specialist developing machines vital to the war effort?
No as a simple soldier.
The German government took no interest in computers and saw them as irrelevant.
The reason may have been that the German military thought that the war was almost won and so any military research needed to be of short duration. What was the point of building a military computer when its completion date would have been beyond the end of the war.
Zuse and Schreyer did indeed propose a propose a high speed computer specifically for code cracking in 1940 and they were asked if it would take more than a year to build. When they admitted that, yes, it would take more than a year the authorities dismissed it as irrelevant to the war effort!
The Z3 was completed in 1941 and it worked despite the lack of government funding. It was built for the German Aircraft Research Institute and intended for stress analysis.
It used 2600 relays, cost $6,500 and was the first fully programmable calculator and a Turing complete machine (meaning that given enough time it could compute anything a modern computer could).
The arithmetic unit could add, subtract, multiply and divide but it was slow for the tasks that it was designed for. It took about a third of a second for an addition and three to five seconds to multiply two 22 bit numbers. Its memory could store only sixty-four 22 bit numbers but for a relay machine this was a lot of storage.
The machine was built into three cabinets, had an operator's console and was controlled by a 35mm film tape reader. Its main task was to solve simultaneous equations by evaluating the determinant of a complex matrix. The purpose of the calculation was to predict wing fluttering - hardly something that would win or lose the war. Still Zuse continued to design special purpose computers to help with the problem of aircraft design.
Zuse with the Z3