Re: Vector Volume 23 =?windows-1252?Q?N=B03?=

Stephen Taylor <editor@[EMAIL PROTECTED]
> wrote:
> 
> Keith Smillie fancied a Differential Analyzer build from Meccano
> parts, but found it easier just to model it.

golly -- I never built it it, mind, but I've still got the original 
plans for that

reading through said plans, it transpired that the thing was accurate to 
about 1 or 2 sig figs, the problem being the differential unit itself 
which consisted of a rubber shod wheel being rotated by a flat 
(cardboard?) disc operating in a plane tangential to the rim of the wheel

basically, the wheel slipped if the flat disc was rotating too fast

I used to use a Brunsviga during the early days of an abortive actuarial 
career, but in the end I gave up on mechanical means of computation

Heathkit used to do a kit for an analogue computer -- something else I 
never built

nowadays, of course, it's all FFT and DSP   . . .   /phil