Fw: simple schematics

[Magnus Danielson]

From: "Roel Das" <Roel.Das at student.groept.be>
Subject: Fw: simple schematics
Date: Sun, 9 Jan 2000 18:43:04 +0100

> I actually meant, when you use a pulse-waveform to clock the random
> generator.
> Why would you need a freq much higher than the audio range? (I realize that
> the random generator will produce series of 2's and O's) Can't you just use
> a VCO
> tuned a few octaves higher with a pulsewave (I have built a working VCO a
> few months ago)?

First, you definitly want to overclock. According to the Nyquist theorem you
need to clock at least twice as high frequency as you want in bandwidth.
In reality you would have another decade of margin to get good noise.
You also need to lowpass filter the output in order to get something similar to
analogue noise.

You should not need to bother using a tuneable VCO. You can use an of the shelf
crystal oscillator with TTL/CMOS output.

> And, are these random generators available in logic components?

Yes, but they are a bit hard to fetch. I beleive Ferranti did such a thing.

You are probably better of using of the shelf CMOS chips, some are very well
suited for the task. The 4006 is a 18 stage shift register by which you
basically only need to add an XOR gate and you are in buissness. For longer
sequences just take two 4006s and you can do long stuff. Another neat chip is
the 4517 which is a 64 stage shift register with a tap at every 16th register.
Using that one it is simple to make a much longer shift register without
suffering high chip count. An 4517 and an 4015 with a 4030 XOR chip will
get you a 35 bit long shifter (tap the second feedback at 33) for basically
no money.

Hmm... maybe time to make some curcuits and actually test it out...

> And second, would the up down counter be worth the trouble?

No.

Cheers,
Magnus