Why does my IGC distort on very low volume settings?

[Dr.H]

Hi All,

Thanks to all the inspiration here, I have built a stereo IGC using the following :

LM3875T
220k feedback, 10k input
300VA 22VAC (about 31.5 VDC) per channel, no regulation
1000uF per + and - rail


When I used a cheap 470k linear pot, I could turn volume to lowest setting (i.e. dead quiet) and there was no distortion.
I have now fitted a 100k Blue Alps pot and am finding that on its lowest setting, I hear much distortion from the IGC, muxh like a loud hissing.

Any ideas?

Thanks
Ryan

 

[r0cket-]

Could be a bad pot. A lot of the so-called Alps Blues around these days are Chinese immitations with a stepped attenuator inside. If that's the case here, there could be a bad solder joint or resistor in the attenuator.

 

[Dr.H]

Thanks for the reply. This is a good pot however, used in a valve based pre before.

 

[PMA]

Has anybody ever measured THD and IMD spectra of this kind of amp? On different output levels?

 

[Raka]

Has anybody ever measured THD and IMD spectra of this kind of amp? On different output levels?

Mr Rasmussen did it in the thread with the tube, have you read it?
Actually, I can't remember if it was done with the buffer or without it, but it inlcudes different power values.

 

[PMA]

Mr Rasmussen did it in the thread with the tube, have you read it?

No, I have not. I will try to find it.

 

[Raka]

http://www.diyaudio.com/forums/showthread.php?s=&threadid=9557

 

[PMA]

Thanks 😉

 

[peranders]

I think the problem can be too high value of the pot. I wouldn't not use more than 47 k, if it's, possbile 10 k! Good for noise and stability.

One other trick is to pull down the input after the pot with some resistor, like 5-10 time the pot value. This resistor will take care of bad wiper contact. Normal opamps are sensitive if the don't have any current path for their bias current. I don't how it is for LM3875?

 

[ronc]

I find that for the best performance and lowest noise i use a 50k alps pot shunted to ground with a 10k carbon in the signal path.
ron