[sdiy] analog loses another niche

[Veronica Merryfield]

On 2010-06-30, at 4:38 PM, Tim Parkhurst wrote:
> 
> In theory, only the analog source is TRULY random. However, digital

Hmmm - I am not sure that any source is truly random, just that we don't understand it enough.

For instance, if one were to be able to construct a coin flipping machine such that all the parameters were repeatable (flick position, flick force, starting conditions, wind/air etc), that the results would be predictable.

The Intel source in the original post is really an analogue source. They are sampling the analogue signal during the state switch, which for given starting conditions and well known parameters, is likely to give the same result if sampled at the exact same time - effects such as trace length, capacitance and so on aren't changing providing the environment is steady, but it isn't - temperature, other gate activity affecting power rail levels and so on.

Junction breakdown, avalanching, zenering and so on look to be random but are they, or is it just that we don't understand it enough. Even at the atomic level, I very much doubt it is truly random. The very fact that some devices are better than others would suggest that that design and manufacturing can control it, just unpredictably.

However, this all said, the effect is useful and looks to be random.

As to pseudo digital sources, very long sequence systems can make very credible noise sources for synth uses, as can FM. But then so does a modem negotiation :)

Vrnc