[onerng talk] Possible magnetic vulnerability

[Paul Campbell]

On Fri, 23 Jan 2015 06:58:03 materribile wrote:
> After looking over your project info, I see what might be a vulnerability to
> outside influence, and have joined to tell you.  It's possible you've
> addressed this already; if so, please excuse the interruption.
> 
> Your noise source is driven by a higher voltage generated by a switching
> supply.  If the inductor in that supply is anything but air-core, it is
> potentially vulnerable to saturation by a constant or very-low-frequency
> -magnetic- field.  Your copper shielding will have a very limited effect on
> such a field.
> 
> I don't know whether it would be possible to influence the RNG in a covert
> way, or whether the effect would always be detected, but I wanted to bring
> this possibility to your attention, and for that purpose I have registered
> here.
> 
> If an air-core inductor will not serve, and cannot be isolated from VLF
> interference, you might consider replacing the switching supply with a
> voltage multiplier. --

I think you're right - the  main shield is steel but the base is copper - 
theoretically someone could place a coil close under the OneRNG and do that - 
it would however be obvious and need to be placed extremely closely and 
precisely - it would be obvious to someone looking at the device

But if they're doing that they could also physically connect to the USB and 
mess with the data that's flowing.

The solution is simple - mount it inside your (steel) server case - we're 
already addressing that by working on a variant that can be mounted internally

	Paul