Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site fortune.UUCP
Path: utzoo!linus!decvax!harpo!ihnp4!fortune!rpw3
From: rpw3@fortune.UUCP
Newsgroups: net.wanted
Subject: Re: Wanted: random number gen. - (nf)
Message-ID: <2260@fortune.UUCP>
Date: Thu, 12-Jan-84 07:05:28 EST
Article-I.D.: fortune.2260
Posted: Thu Jan 12 07:05:28 1984
Date-Received: Fri, 13-Jan-84 06:32:42 EST
Sender: notes@fortune.UUCP
Organization: Fortune Systems, Redwood City, CA
Lines: 29

#R:cbosgd:-80800:fortune:12300010:000:1114
fortune!rpw3    Jan 12 02:09:00 1984

If you just need "random" noise, a sufficiently fast TTL clock will
appear random, especially if you use a pseudo-random method of picking
bits and of picking polarity.

A more complex (but not necessarily better) method is to run a D/A
into the "+" side of a comparator, run the output to a TTL Schmitt
trigger, which goes to you PIO port and... back to the "-" side of the
comparator. The comparator/Schmitt will oscillate like a banshee
(due to the hysteresis in the Schmitt trigger), and the D/A can be used
to adaptively bias the ratio of 1's to 0's.

The "most" random source I have seen used a VHF noise diode (yes they
make them deliberately noisy) in an analog version of the preceeding
circuit. That is, the (properly biased) noise diode goes into the
comparator and then to a Schmitt trigger, which goes to the sampler and...
back to an integrator which averages the 1/0 ratio and feeds a D.C.
voltage to the comparator.

Good luck.

Rob Warnock

UUCP:	{sri-unix,amd70,hpda,harpo,ihnp4,allegra}!fortune!rpw3
DDD:	(415)595-8444
USPS:	Fortune Systems Corp, 101 Twin Dolphins Drive, Redwood City, CA 94065