Spec4 Bias question

[sawdust]

I have several Pioneer Spec4 amps all of them the bias swings several mV from zero one of them swings over 020.0 mV is this a characteristic of the design . All of them have been recapped and new transistors . The last 2 amps I checked the bias before I restored them and both of them did swing several mV from zero . my question is this normal for this type of design any input would be helpful . I read a thread a few years ago about a modification to the amp boards to stabilize the bias .

 

[EchoWars]

Are you talking about DC offset, or idle current?

 

[sawdust]

DC offset, idle current is fine

 

[EchoWars]

http://www.audiokarma.org/forums/index.php?threads/spec-4-failure.272617/

Bottom line is the feedback path of the Spec 4 allows for ALL DC variations to be amplified. Never been crazy about that.

On a few units, with the owners consent, I've cut a trace and added the 330uf cap. Offset goes from flaky to solid.

 

[rcs16]

All of them have been recapped and new transistors

the stock input pair,Q1 are 2sa798, which are a matched pair that are thermally coupled.
What did you use as replacements for Q1? It is important for DC offset drift? If you did not use 2sa798 and used something like 2 of ksa992, did you make them thermally coupled to each other?

one of them swings over 020.0 mV is this a characteristic of the design

The last 2 amps I checked the bias before I restored them and both of them did swing several mV from zero .

I think you answered your own question,
20mV of DCV offset is not much to be concerned about. It is when it gets into the 100's of mV that I believe it is more of a concern.

 

[sawdust]

Q1 was the only transistor I did not change ,I was going to match 2 ksa992 and glue them or shrink wrap them to couple them together. I always check the original transistors I change that are gain matched pairs and I never had a pair that was matched not even close

 

[markthefixer]

Gain changes with collector current, and no two devices have exactly the same Ic/Hfe curve.

Technically gain matching should be done with the same current as the application.The meter has it's own test current fixed, and the gain is the other current. Like a set base current and a measured collector current, or a set collector current and measured basfe current - depends upon the meter. The thing is that the Vbe differences between the transistors of the pair will also affect things.

That's why I mentioned in-situ matching long long ago. Just like we "roll op amps" in some equipment, swap tubes in others.

 

[sawdust]

http://www.audiokarma.org/forums/index.php?threads/spec-4-failure.272617/

Bottom line is the feedback path of the Spec 4 allows for ALL DC variations to be amplified. Never been crazy about that.

On a few units, with the owners consent, I've cut a trace and added the 330uf cap. Offset goes from flaky to solid.

Glen thanks for the link and the info, I guess it been more then a few years since I read this thread.

I will try out the mod, I have a extra amp board and a UPW 330uf 25v cap and a 1uf film and it looks like R52 resistor was changed to 100k I have them also and I will cut the trace between R52 and R18.

Not sure why this amp board has the 2 diodes none of the other amp boards I got have them is this something I half to worry about doing this mod they look original .

 

[EchoWars]

I forget the purpose of the diodes, but I'd not worry about them.

 

[ymarko]

... from zero one of them swings over 020.0 mV ....

20mV of DCV offset is not much to be concerned about.

The original poster see 020 mV, which is not 20 mV. 020 reading is probably shows confused voltmeter.

I measured my Spec-4 DC Offset:
One channel has 0.000 mV.
Second channel has 020 mV (exactly as original poster).

I'm not a technical at all. And my both channels are equal and work perfect.
I read that Spec-4 is "coupled" amp that DC Offset somehow cannot be measured in conventional way. I have no worry and I don't touch Amp.