On Sat, 2006-11-11 at 13:51 +0100, Daniel Wagner wrote:
> > But the most intriguing part of the whole story: Not only did the noise
> > get louder at a lower latency, it also seemed to precisely double its
> > frequency when running at halve the number of frames. I even tried to
> > measure the frequency of the noise. As far as I remember, when running
> > jack for instance at 44100Hz, periodsize 32, the noise was at exactly
> > 44100/32 = 1378Hz!
>
> Is an interrupt fired at each period? If so the cpu will use more power
> (or woken if it was asleep); a small power 'surge' which can of course
> have an influence on any analog hardware near by.
Really sounds like the power supply, the high frequency transformers can
"sing" with changes in load if the windings are a bit loose. I think
capacitors can also do that.
Somebody I know had a fanless computer in his room (a very old machine
that really did not need fans if the case was left wide open) and the
only remaining sound was some component singing with the network
traffic...
-- Fernando
Received on Mon Nov 13 04:15:16 2006
This archive was generated by hypermail 2.1.8 : Mon Nov 13 2006 - 04:15:16 EET