[Cryptech Tech] Seeking comments on a proposal for changes to the Cryptech RNG design.

[Manuel Domke]

Hi,

> There's even a PSoC 5-based project on Github:
>
> https://github.com/alwynallan/redoubler

I know - and this may be a great start to integrate it here.
But for my standalone RNG I prefer the dumb "off-the-shelf" component
version from Bill Cox.
-> https://github.com/waywardgeek/infnoise

The Redoubler PsoC 5 performs very well, built my first one some days ago.
It puts out 7,968319 bits of entropy per byte as shown in my first tests
of a 25MB sample.

> For ease of understanding: “modular entropy multiplication” is essentially a 
> successive approximation ADC, which runs forever amplifying its internal state 
> instead of actually sampling some input, an as a result, only converts its own 
> noise into binary data.

Thanks for the short explanation - I should have included one too.
It actually gathers most of its entropy from thermal noise from the
resistors connected in the opamp loop.

Greetings,

Manuel


On 24.03.2018 14:10, Bernd Paysan wrote:
> Am Donnerstag, 22. März 2018, 18:07:07 CET schrieb Manuel Domke:
>> Currently I'm focusing on spreading the word about modular entropy
>> multiplication - and will start a crowdfunding campaign to sell some
>> Infinite Noise very soon. So I won't have the time to work much on this
>> project soon - but its on my todo list.
> There's even a PSoC 5-based project on Github:
>
> https://github.com/alwynallan/redoubler
>
> For ease of understanding: “modular entropy multiplication” is essentially a 
> successive approximation ADC, which runs forever amplifying its internal state 
> instead of actually sampling some input, an as a result, only converts its own 
> noise into binary data.
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cryptech.is/archives/tech/attachments/20180327/8b9c2862/attachment.html>