follow up: LM3886 RC filter + DC offset questions.

[blackboy]

I try removing my input cap today, the BC 4.7uf mkp. Now the DC offset back to 10mv and -30mv (was -50mv and -90mv)... why?? even higher DC offset with cap? I know it's dangerous without this cap. Wish my discman not giving out any DC... Thx again, guys!

 

[Juergen Knoop]

a schematic would help
regards

 

[blackboy]

I totally follow Mick's design.
http://dogbreath.de/Chipamps/ThreeResAmp/ThreeResAmp.html
then, I added a Zobel network which is "100nf + 2.7r"
finally a 10k volume pot after signal input and then RC filter "4.7uf + 15k" after the pot, thats it, nothing else... thx!

 

[Juergen Knoop]

the circuit has a high dc-gain
removing the cap lowers resistance seen by the input bias current, which means lowering current induced offset voltage
maybe this could explain the observed behaviour
regards

 

[AndrewT]

Hi,
a fully DC coupled power amp. You have a lot of faith.

Have you now re-fitted the DC blocking cap at the input?
And is it after the passive pot?

I hear that pots become noisey if they have a long term DC applied to them (from the discman). A DC block before the pot may also be required for long pot life.

I do not see any mention of RF filter at the input to your power amp.

 

[blackboy]

Hi,

I do not see any mention of RF filter at the input to your power amp.

ya, I put the RF filter AFTER the pot, is there any difference from BEFORE the pot? I have tried before the pot, but the DC offset is just the same.

 

[markiemrboo]

Are you testing this with a source connected? Or with the inputs shorted to ground?

If it's with the source connected - perhaps a long shot - but might it be possible that the source is putting a tiny amount of 'positive' DC out and giving you those values?

E.g assuming a gain of 20, if the source has a DC offset of 3mV:

+3mV (source offset) * 20 (gain of amp) = +60mV
-50mV (the offset that the amp itself is producing) + +60mV (3mV from source multiplied by the gain) = +10mV

When you put the input cap back in it just appears worse because the source is not 'cancelling' out the offset, and you're only seeing the offset that the amp itself is producing.

Just a thought

 

[blackboy]

When you put the input cap back in it just appears worse because the source is not 'cancelling' out the offset, and you're only seeing the offset that the amp itself is producing.

Just a thought [/B]

WOW, I think you got the point! thx! however, how can I reduce these high -voltage offset?

 

[Juergen Knoop]

WOW, I think you got the point!

easy to check! redo the offset measurement without a source connected. But leave the pot in place.
regards

 

[blackboy]

I test again today by remove the input caps... no matter there is input or not, the offset is arround 20 and -30mv... I am confused again... after I have put the caps back, it come back to -90.. - 50.. omg!! whats the problem.....???