[OneRNG-Discuss] LED frequently going dark for various lengths of time

[bsr]

On 7/26/22 15:10, Paul Campbell wrote:
> On Wednesday, 27 July 2022 7:16:08 AM NZST bsr wrote:
>> I'm having an issue with one of my OneRNG units.  For the past few
>> months I have observed the LED occasionally going dark.  When I first
>> started to notice it would be dark about 3-4 seconds but lately the
>> duration is sometimes longer (10 seconds) and the frequency of
>> occurrence has increased as well.  My issue seems very similar as
>> described in a previous discussion, except Hans was using v1 hardware
>> and mine is v3 -
> 
> there is/was a firmware bug I found a while back that did cause behavior like
> this - there's a fix for it (not a firmware update) in the last couple of
> software releases
> 
> 	Paul
> 
> 

I meant to post this a long time ago... but here is what I found.

I had my onerng config set to cmd3 operation (raw avalanche plus RF - no 
whitening) and was experiencing the behavior I described.  Switching to 
cmd0 (default) my ONeRNG functions normal.

Checking with a scope when set to cmd3 operation, I didn't notice any 
change in the waveform at T3 as the light would go dark.  It was hard to 
catch but I did get a couple of direct device dumps via dd that showed 
some really strange data.

I posted one of those dumps here:
https://dpaste.com/6AGA3KZE9.txt

The spamming of zero in the output stream seems to hit around offset 
0x340 and seems to resolve around 0x1838.  Since I didn't notice any 
change on T3 (using cmd3) and I am not seeing this behavior using cmd0 
(no RF) it seems probable that RF is somehow responsible(?).  I haven't 
dug deeper but could RF interference really cause such issues?