Re: [linux-audio-user] Pre-Amp hum (related to jack, somehow)

From: Vince Werber <ka1iic@email-addr-hidden>
Date: Sun Oct 29 2006 - 23:15:52 EET

The individuals 'hums' can cancel each other out when mixed iff.. both
hums are 180 degrees out of phase with each other and are of the same type
and intensity...

The terms 'singing capacitor' has been keeping me out of this
discussion... perhaps the correct phase should be 'microphonic'
capacitor? The only time I have known a capactitor to sing is via 1,000
plus volts and the cap is dried out and is well into a need for
replacement... the term explosion works here... :-)

A 'microphonic' capactor is noticable from vibrations from an outside
source that causes the cap to 'ring'

In any sense though both conditions come from inferrior components,,,
ie... dried up floor sweepings...

vince

ka1iic@email-addr-hidden

> The irony of this is killing me. I had a singing capacitor in my
> notebook computer, which I was able to solve using a load generator at a
> low scheduling priority. However, that same capacitor is what had
> triggered me to get a higher-end sound card in the first place.
>
> Now this higher end sound card has quite a loud hum when the phantom
> power for its mic pre-amps is on. Interestingly, when I mix the signals
> from both pre-amps by connecting them both to the same outputs in JACK,
> the noises from the Mics cancel each other out.
>
> The load generator has no influence on the hum; however, the sampling
> rate does. The higher the sampling rate the higher pitched the hum is.
> The inverse is true for buffer size: The larger the buffer, the lower
> pitched the hum.
>
> My current theory is this: Jack causes a surge of power in the CPU
> through an interrupt. This surge gets transformed into current in the
> grounding cable through electromagnetic induction, which in turn induces
> a current in the microphone pre-amps, which causes the hum. The hums can
> cancel each other out because... uh, not the slightest on that one.
>
> Help appreciated.
>
> Carlo
>
Received on Wed Nov 29 00:15:11 2006

This archive was generated by hypermail 2.1.8 : Wed Nov 29 2006 - 00:15:11 EET