[sdiy] analog loses another niche

[Colin f]

> So...I went and looked to see if Cavium had a patent on their 
> TRNG, and indeed they do.  They use a bunch of ring 
> oscillators (64 in one example), built out of very weak 
> transistors, that are in a digitally noisy environment, 
> thereby causing the oscillators to be extremely jittery, as a 
> source of entropy.  They sample the oscillators and feed the 
> result into a cryptographic hash such as SHA-1 (although the 
> patent indicates you could use an LFSR).  The output of the 
> cryptographic hash by it self is 'random' enough to pass most 
> any test without the ring oscillators being fed into its 
> state variable.
> 
> Rings a bell - doesn't someone in the sdiy community make an 
> LFO whose output is the sum of six inverter based oscillators 
> set to different frequencies?

Wow, they combined the cymbal noise from a 606/808 with a rough equivalent
of the snare/clap/tom noise from a 909 !
How on earth did they get a patent on that anyway ?
Surely there is a ton of prior art along the lines of unstable oscillators
in a digitally noisy environment...

Cheers,
Colin f