From: "John Jardine"
Subject: Re: Trig circuits
Date: Tue, 24 Dec 2002 00:25:49 -0000
References: <email@example.com> <firstname.lastname@example.org>
NNTP-Posting-Date: 24 Dec 2002 00:12:42 GMT
X-Newsreader: Microsoft Outlook Express 5.00.2615.200
Bill Sloman wrote in message
> "John Jardine" wrote in message
> > Spehro Pefhany wrote in message
> > news:email@example.com...
> > > On Mon, 23 Dec 2002 14:11:47 -0000, the renowned "John Jardine"
> > > wrote:
> > [clip]
> If you really were a hair-shirt masochist, you'd program one of Babbage's
> analytical engines into an FPGA, which is to say, you'd set up a pair of
> accumulators and start one at +1.0 (cos 0) and the other at 0.0 (sin 0)
> subtract sin from cos, while adding cos to sin - step size and accumulator
> length depends on accuracy required.
> I did a concept design once - and turned out to need a few extra
> insiginificant bits of preload to compensate for the bias in the rounding
> error. Fun.
> Bill Sloman, Nijmegen
FPGA, bah!. A cam controller and relays ;-)
Bill that sounds a really weird but neat 'machine' method. No 'times' no
'divs', cant' believe it's so simple. Babbage was though, interested in
'differences' Sounds like a basis for some Cordic algorithms. I learn
something new every day!. Just read a New Scientist article on the odd
Golden Rectangle ratio and the Fibionacci series, Trancendentals can do
strange things!. I've used in the past for sines, a limited Taylor series
approximation and also a look up table with linear guesses
("interpolation"). Most times I'll very gladly exchange speed for
simplicity. I'm going to try yours (lazily, in Basic!) and see what happens.