dc offset help!!!

[GiulioSL]

Hello All,

I'm new on this forum and is for me a pleasure to partecipate at this community.
I need some help for understand if my power amp is good or have few problem:
My amp when is cold give around +60mV of dc offset on output, then after around 20-30minutes, when heatsink reach 27-30°C the dc offset drop down until 5-10mV,when is very hot less than 5mV
What do you think about this?
Can someone let me know if my amp is ok or if there's some problem?


Many thanks for help

 

[lanchile]

After warm up, 5 to 10mv is good. I would not worry about it.

 

[Bone]

Nothing wrong with that.

 

[east electronics]

behavior like that will mean one amplifier that is depending very much on the hfe of the transistors that form the circuit .

most of elektor amplifiers are like that

also will mean that it could be one amplifier with plenty of output transistors but with marginal or low bias so its expected that before amplifier equalize to an ""operating""temperature both bias and offset values will not be correct .

If the amplifier hasn't plenty of outputs it could be a minor error in the thermal design

Would be nice to know what are we talking about ...

Kind regards
sakis

 

[GiulioSL]

behavior like that will mean one amplifier that is depending very much on the hfe of the transistors that form the circuit .

most of elektor amplifiers are like that

also will mean that it could be one amplifier with plenty of output transistors but with marginal or low bias so its expected that before amplifier equalize to an ""operating""temperature both bias and offset values will not be correct .

If the amplifier hasn't plenty of outputs it could be a minor error in the thermal design

Would be nice to know what are we talking about ...

Kind regards
sakis

Hello All, and many thanks for help!!

Sakis, if i try to increase a little bit the bias curruent (F.E.from 12 to 13mV) the temperature of heatsink increase exponentially and i don't know if there's some risk to burn my PA

 

[Mooly]

Altering the bias current will not affect the DC offset. They are two different things.

I doubt bias would be stable to 1 mv anyway. Measuring with the amp hot and then cold will give different results. Always use the manufacturers setup procedure and values as a starting point.

 

[GiulioSL]

Yeah that is true..another strangers is if i measure the bias on mosfet N (4in parallel) there's always some difference of bias current..12,18,22,16mV...and the same for P rail

 

[AndrewT]

or does that difference reflect the 5% tolerance of the source resistors used in the build.

You cannot rely on 5% tolerance resistors to give better than a 10% spread of source currents.

 

[GiulioSL]

or does that difference reflect the 5% tolerance of the source resistors used in the build.

You cannot rely on 5% tolerance resistors to give better than a 10% spread of source currents.

Yes the power resistor have 5% tolerance

 

[GiulioSL]

Yes the power resistor have 5% tolerance

And if i add for example an utomatic dc offset null circuit, i can ontain best performance::?,there's someone with experience with this kind of "problem"? Or is bettere don't touch nothing?

 

[magnoman]

Giulio,

What kind of amp is it?
If it is typical with global negative feedback, then the offset is dominated by the front end stage and not the output stage.
I would expect the output bias current to be dependent on matching and thermal equilibrium, but with feedback this shouldn't show up as an offset.
It sounds like this may be a fairly high power amp and the offset change may be due to the entire amp warming up and causing the input stage to drift (and possibly it was trimmed warm thus explaining the decreasing offset with rising temperature).

Thanks
-Antonio

 

[GiulioSL]

It sounds like this may be a fairly high power amp and the offset change may be due to the entire amp warming up and causing the input stage to drift (and possibly it was trimmed warm thus explaining the decreasing offset with rising temperature).

Yes is an high power amplifier and i have performed the DC Offset setting with hot Heatsink, but everithink appaer too sensitive at the temperature variation
Thanks

 

[magnoman]

Giulio

Would really help to see the design to determine if the change you are measuring is typical.
It's not so easy to have low offset drift, requires very precision matching within the input stage, and or compensation techniques. With what you've described I would guess that nothing has gone wrong with the amp, and it's just the way it is.

Thanks
-Antonio

 

[nigelwright7557]

Yeah that is true..another strangers is if i measure the bias on mosfet N (4in parallel) there's always some difference of bias current..12,18,22,16mV...and the same for P rail

Unless your going to the expense and time of matching the mosfets exactly then will be a difference.

I designed a mosfet amp with 12 transistors on the output and never bothered matching the mosfets yet sold loads and never had any problems.
The source resistors will help balance things up and spread the load.

 

[djk]

Some old Electrocompaniaet amplifiers ran 700mV cold, you adjusted them to be 0V after 20 minutes or so.

I don't get excited until it goes over 50mV~100mV (warmed up).