Can power supply filter capacitors affect DC offset voltage?

Schematic attached: Can power supply filter capacitors affect DC offset voltage?

For no good reason other than they were cheap to buy, I wound up with 3 of the same power amps. I only wanted one (which was found) and after following advice received on this forum, is in good working order. It was while searching for a companion preamp when the other 2 power amps were offered to me at next to no cost.

I had surplus capacitors and transistors from the first amp rehab so went ahead with replacing the same parts in the other two. The only parts I did not have surplus of were the large 50 volt 10,000mF filter caps so the originals were left in place.

I was instructed to adjust the bias/idle voltage to 10mv and achieve DC offset as near zero as possible. With the first amp, adjusting for 10mv bias produced offset voltages minus 3mv & minus 4mv.

The other 2 amps with exactly the same parts replaced except the original filter capacitors and bias adjusted to 10mv, produced about +12mv on left channels and +25mv on right channels.

Even though DC offset is low enough to be considered acceptable, could the disparity be because of the filter caps?

Those caps are fairly expensive at about $20 per pair locally. That is more than what both of those amplifiers cost and why I did not buy new ones.
Last edited:
" could the disparity be because of the filter caps?"

No, not for that circuit.

OK then, thank you.

When I received one of the amplifiers, it's bias was was well in excess of 1 volt and the heat sink was too hot to touch for more that a second or so.

I understand that heat is detrimental to electrolytics.

Since I have no idea how long it may have operated in that condition and I do not have the means to test capacitors, I believe I shall proceed with replacing those caps in that one.

Ian Finch

Paid Member
2010-04-11 4:22 am
Coffs Harbour, NSW
.....When I received one of the amplifiers, it's bias was was well in excess of 1 volt and the heat sink was too hot to touch for more that a second or so.
Was the current similar in either channel and were you able to reduce it? It's not unusual to find previous owners have played with bias (or anything that is adjustable) in the hope of getting things to work or sound better.
the output stage is a double EF ClassAB.
Optimal ClassAB bias is usually substantially higher than 10mV across the two emitter resistors.
D.Self suggests Vq=54.8mV between the two test points when 0r47 is used. (this must be a "hot" value since we normally see 26mVre for 25°C operation)
At only 10mV (=5mVre across each emitter resistor) there will be significant crossover distortion when driving a load.

The output stage will run much hotter if you adjust up from the 10mV value.
Will the heatsinks keep the transistor sufficiently cool to permit reliable operation?
Thanks Andrew. This is the reason why I set the bias to 10mv:

Some weeks ago I started a thread pertaining to the meters in the first of these amps I acquired. There was a response from forum member Irribeo who kindly provided a schematic on which were hand written notations.

A subsequent thread about the power supply filter capacitors (again started by myself) elicited further responses from Irribeo who it seemed is well acquainted with this particular amplifier. That thread went on to include suggestion for changing out six transistors which were cited as being failure prone. Four of them were both input pairs and also TR5 & TR6. It was mentioned again in that particular post that "idle current should be adjusted to between 10 and 12 millivolts".

I went ahead with those replacements and followed that instruction (which was one of the handwritten notations on the schematic) to adjust bias to 10mv. Subsequent dialoging focused to some degree on this figure in order to achieve the objective of 0 volts DC offset.

That is why I set the bias current to 10mv. It seemed wise to take direction from one who knows these amps well.

Perhaps Irribeo will spot this and elaborate.

With regard to the heatsink, it seems substantial enough considering the compact dimensions of the chassis. To me, it appears to be adequate for 4 output transistors provided that air flow is not impeded.