starting my relay computer project

[Brent Hilpert]

Randy Dawson wrote:
>
> > Date: Sat, 23 Jan 2010 10:45:57 +0000
> > From: philip at axeside.co.uk
> > To:
> > Subject: Re: starting my relay computer project
> >
> > Have you read "The Computer - My Life" by Konrad Zuse (New York:
> > Springer Verlag, 1993.  ISBN 0-387-56453-5)?  Mostly historical /
> > autobiographical, but some good technical bits on the evolution of relay
> > logic as Zuse's designs got more mature.  Interesting how the number of
> > relays per bit he needed in his adder circuit got less and less...
> 
> Phillip,
> 
> Thanks for the lead on the book, and no I have not seen it.
> The Harry Porter relay computer and others seem to all use Zuse 4PDT relay design for the alu.
> 
> If you have not been to his web site, take a look at the youtube, It appears that there are a lot of us relay computer guys out there (when the you tube finishes there are related links presented).
> 
> Somewhat related, and also my interest, are the marble machines that do logic.  Of course there is the Edmund Scientific NIM plastic 3 bit machine, analog computers and dont forget Martin Gardner and the deck of cards paper computers.
> 
> There is more to classic computing than collecting microprocessor based machines.

A relay computer is something I too would like to build if ever I were to come
across enough suitable relays.

A couple of years ago I fleshed out the Simon design and wrote it up, as well
as breadboarding a small portion of it.
  (http://people.cs.ubc.ca/~hilpert/e/simon/index.html)
Ultimately, I felt it would be unsatisfying to implement it fully, due to it's
limitations. 

Funny how the adder circuit keeps coming up, one of those perennial topics in
computing. I too worked on optimising it.
(http://people.cs.ubc.ca/~hilpert/e/simon/imp.html#ALU). Later I saw Zuse's
design and noted that I had -almost- arrived at the same design. The number of
contacts and basic circuit were identical, but I missed one final optimisation
of flipping some contacts in the sum circuit so the relays required would be
4-pole & 4-pole rather than 5 & 3 (the former tending to be a little more
practical in the real world, of course).