pseudo-random & slow random ?

[Martin Czech]

>> Actually it's (2**N) - 1, but what's that amongst friends.
	
True, because one "forbidden" value has to ommitted (it is 0 i think), 
because the whole machine would get stuck there.

> As Horowitz and Hill points out, a 33-bit register clocked at 1MHz would 
> last over an hour. A 100-bit register clocked at 100MHz would have a cycle 
> time a million times longer than the age of the universe. I don't think 
> that this is going to a problem repeating somehow, not even for the longest 
> sequence I could listen to.
 
For the above reason, one needs a kind of reset circuitry, that will
set up all bits correctly at power up to avoid the forbidden state.
With a determined reset all sequences would start at the same point,
regardless of their lenght.  I think in this case a 100 bit register
makes no sense, after power on it's always the same,  unless someone
comes up with a good idea for a random number reset circuit.
Perhaps some relays switching noise or other stuff a power up.
Or sampling analog noise with at least one bit forced to 1.
Or no reset at all, with a "false mode detection".
This detects the forbidden state and gives a kick to the machine to start it up.

Now we have nice, almost white noise. But how to get low frequency
random from that? DACs are quite cheap today. What about connecting a 14 or
16 bit DAC to the shift register and modifying the clock rate?

Any ideas ?

m.c.