Uneven DC offset

[62vauxhall]

I duplicated, or nearly duplicated, what I did to a Toshiba SC335 power amp on a second one. With the first I replaced all the electrolytics and followed a recommendation to replace 2 pairs of transistors with 2 matched pairs of new ones. Since I have nothing but a DMM, I read of a down & dirty measuring technique to check transistors which was what I followed to match them.

The end result with the first amp was a DC offset of about 12mv on one channel 13mv on the other.

With the second amp I did not replace the large power supply filter caps. After adlusting bias to ~ 10mv as was recommended, the DC offset is about 25mv & 12mv.

I understand that this is considered to be within acceptable limits but I am curious as to why one side is double that of the other?

I can increase / decrease the bias to equalize offset but that means the idle currents are unequal.

Could those filter caps, old as they are affect DC offset?

This is probably no big deal as both offsets seem to be low enough but it bugs me that they are not closer.

 

[Lazarus Short]

Avoid being too OCD, the best advise I can give here...

 

[laatsch55]

I duplicated, or nearly duplicated, what I did to a Toshiba SC335 power amp on a second one. With the first I replaced all the electrolytics and followed a recommendation to replace 2 pairs of transistors with 2 matched pairs of new ones. Since I have nothing but a DMM, I read of a down & dirty measuring technique to check transistors which was what I followed to match them.

The end result with the first amp was a DC offset of about 12mv on one channel 13mv on the other.

With the second amp I did not replace the large power supply filter caps. After adlusting bias to ~ 10mv as was recommended, the DC offset is about 25mv & 12mv.

I understand that this is considered to be within acceptable limits but I am curious as to why one side is double that of the other?

I can increase / decrease the bias to equalize offset but that means the idle currents are unequal.

Could those filter caps, old as they are affect DC offset?

This is probably no big deal as both offsets seem to be low enough but it bugs me that they are not closer.

Offset is usually a product of the input differential pair losing tolerance. Yes, adjusting bias to compensate for offset is a bad idea, it does mess with idle current. What was your down and dirty technique of matching transistors??

 

[62vauxhall]

Offset is usually a product of the input differential pair losing tolerance. Yes, adjusting bias to compensate for offset is a bad idea, it does mess with idle current. What was your down and dirty technique of matching transistors??

Red to C, black to E + B, count to two & note the reading.

Having to explain this just now has made me realize something. I had the meter at DC volts not the diode setting. So I am going to re-check the batch of transistors I have and most likely be changing out those I just installed.

I think if you looked up OCD in a dictionary, you'd find my picture.

 

[laatsch55]

It works better on resistance....

 

[laatsch55]

You need to measure the B-C, B-E junctions....Red on base for NPN, Black on Base for PNP....

 

[Gepetto]

You will have a difficult time with just a DMM. These are microvolt level readings and you need to bias the transistor as it will be used in the circuit to get very low offset.

You will also have to wait longer than count to 2.

I use a master transistor as a constant and a to be measured transistor. That way you can achieve balance within a couple hundred microvolts of the master.

The offset should be negative in a PL400 or 700 with the NPN input stages. Typically 9-13 millivolts negative.

 

[62vauxhall]

Check transistors for resistance? I thought resistance was futile?

I'll use your method and mark them accordingly in case I run into another situation that needs these.

I'm going to leave that second amp as it is since things seem to be within acceptable limits.

The third amp was a little different.

Prior to measuring bias and offset, I turned it on to warm up. After a half hour or so, it was throwing off some serious heat. I checked bias which was about 1.25 volts then turned it off to cool down. Once done, I started adjusting and the absolute lowest idle current I could achieve on one channel was 24mv and that is with the pot hard over. The barest nudge farther causes the bias to shoot up hard. That particular pot is at it's limit although there is more travel available with the other channel. I adjusted that channel's bias to 10.5mv.

Offset in this case is in the negative with minus 13mv & minus 24mv.

Seems like it could be considered OK but that pot out of adjustment room is a bit concerning.

 

[Gepetto]

Check transistors for resistance? I thought resistance was futile?

I'll use your method and mark them accordingly in case I run into another situation that needs these.

I'm going to leave that second amp as it is since things seem to be within acceptable limits.

The third amp was a little different.

Prior to measuring bias and offset, I turned it on to warm up. After a half hour or so, it was throwing off some serious heat. I checked bias which was about 1.25 volts then turned it off to cool down. Once done, I started adjusting and the absolute lowest idle current I could achieve on one channel was 24mv and that is with the pot hard over. The barest nudge farther causes the bias to shoot up hard. That particular pot is at it's limit although there is more travel available with the other channel. I adjusted that channel's bias to 10.5mv.

Offset in this case is in the negative with minus 13mv & minus 24mv.

Seems like it could be considered OK but that pot out of adjustment room is a bit concerning.

The original bias pots used are absolute garbage so it is not surprising. It is likely that this pot is oxidized an no longer linear. You should be able to determine that with your DMM.

 

[grapplesaw]

Check transistors for resistance? I thought resistance was futile?

I'll use your method and mark them accordingly in case I run into another situation that needs these.

I'm going to leave that second amp as it is since things seem to be within acceptable limits.

The third amp was a little different.

Prior to measuring bias and offset, I turned it on to warm up. After a half hour or so, it was throwing off some serious heat. I checked bias which was about 1.25 volts then turned it off to cool down. Once done, I started adjusting and the absolute lowest idle current I could achieve on one channel was 24mv and that is with the pot hard over. The barest nudge farther causes the bias to shoot up hard. That particular pot is at it's limit although there is more travel available with the other channel. I adjusted that channel's bias to 10.5mv.

Offset in this case is in the negative with minus 13mv & minus 24mv.

Seems like it could be considered OK but that pot out of adjustment room is a bit concerning.

I am with Joe on this one. I had two with incorrect pots from the factory.

 

[62vauxhall]

If that's so, I shall get some new pots.

The local part shops's website shows they stock these in 330 ohms and what's on the board are 300 ohms.

Is that close enough?

I may be in their neighbourhood later this morning and if so, will pick up a couple. They're only about $1 so no great loss if not suitable.

EDIT: The parts list describes these as "partially fixed". What does that mean?

 

[laatsch55]

Who makes the pots you want to buy?? Get some decent Bourns pots. Post a pic of the pot you are replacing...

 

[62vauxhall]

Thanks Lee.

I'll post those pics but don't get home for another 10/11 hours.

 

[laatsch55]

Amen Gary....

 

[62vauxhall]

Sixteen hours behind the wheel of a tow truck later, here is that 300 ohm trimmer pot:



I was nearby the parts store today so for 50 cents each, got some of these - 330 ohms not 300:

 

[62vauxhall]

There was an hour or so to kill around lunchtime so I went home and replaced those pots. What brand they are is anybody's guess but they would have come from somewhere in China.

They were drop in replacements and work properly, which it seems the originals (or one of them anyway) did not. I was easily able to obtain bias of 10mv and neither pot was cranked hard over - lots more travel available.

All that happened to the DC offset was that it went from minus readings to plus readings but the numbers are about the same - 12mv for left side and 24mv for right.

Later tonight, I'll let it idle for an hour or so and measure again - dialling in as necessary.

I read a thread elsewhere which said that for 10mv bias after warmup one should see 2.5mv immediately after amp turn on. Does that seem right?