[sdiy] Uniformly distributed noise generator?

[Tom Wiltshire]

Hi Damian,

On 10 Jun 2013, at 12:37, "cheater00 ." <cheater00 at gmail.com> wrote:

> Reason why I asked about analogue implementations:
> - they're very likely simpler

Unlikely, in my view. A digital implementation could be one chip and costs pence. It doesn't get much simpler than that.

> - they might function via some inherent property of a part or circuit,
> which is then ideal

If you get lucky. Or you don't.

> - you don't have to code anything, and in ten years it'll actually be
> possible to repair or replicate it

Meh. In ten years I'll use my desktop 3D printer to print a NoisePart from a file that I downloaded. Or I'll use a bit of spare processing on one of the other ten processors that a circuit will consist of by then.
If I can replicate a NMOS noise generator from the 1980's on a PIC now, I don't see why I won't be able to replicate the same function in ten years.

> - parts count very likely much smaller..

What, less than 1?

> - analog noise is truly random, not pseudo-random

In theory. In practice, analog noise is unlikely to be totally pure, and a good digital implementation would give better quality.


So, there you go! Have I convinced you yet?! I'm just playing devil's advocate, really. I don't care either way. Do it with analogue if you can find a good way.

Regards,
Tom