Soundcard interface.

[Mr Evil]

A device to make a soundcard more useful as a measuring instrument: http://www.cheese83.freeserve.co.uk/amp/scope.htm


Perhaps someone has some ideas for a better bandpass filter? Something that uses more reasonable component values, ideally small enough to wire p2p between the sockets, like the current one.

 

[SY]

Thought about a gyrator to replace the inductor?

 

[jan.didden]

SY,

You been to the beauty farm? Your avatar looks even younger than me!

Jan Didden

 

[SY]

Difference is, mine is dead.

 

[Mr Evil]

Thought about a gyrator to replace the inductor?

That's a possibility. An op-amp with twin-T notch filter feedback would also work, but before I try anything like that I want to try to find a passive solution since that will introduce the least noise, and if I can avoid inductors (or at least use only air-core ones) then distortion will likely be lower too. It's unfortunate that there is no bandpass equivalent to the twin-T (as far as I know).

 

[Mr Evil]

I've been doing some experiments with bandpass filters trying to improve the resolution of distortion measurements and have come to the conclusion that the -96dB 2nd harmonic I see is actually the distortion of the soundcard's input, not the output. Thus it will never be able to measure anything lower than that directly. I should have paid more attention to your notch filter suggestion, SY...

After playing around with a notch filter in place, I reckon I could get maybe 20dB improvement, which would mean the noise floor would become the limiting factor.

My findings on bandpass filters are:

• Passive LC: Very high Q. Very high distortion.
• LC with L replaced by gyrator: Low Q. Low distortion.
• Op-amp with twin-t feedback: High Q. Low distortion. Stability problems.
• Passive RC: Very low Q. Very low distortion.