Marshall valvestate 100r bias problem

[Demian2020]

Optimal value requires trial and error. Start with 120R and increase to 150R or 180R if necessary. Or wire 100R trimmers in series with existing resistors. All emitter resistor voltage drops should be equal.

it's very strange. Currently I am with R18 and R20 at 200ohm and when I turn on the amplifier the bias start at 0.48v and little by little it goes up to 1.08v the transistors T10 and T11 in their ballast resistors rise from 0.01v to 0.10v the transistors T9 and T12 in their resistors ballast measures 0.01v this problem is very rare. the output transistors have the hfe quite similar

[Demian2020]

Don't R18, R20 set the output current limit?
R12, R13, TR6 set the bias

which value suggests that I modify so that the transistors do not heat up at idle?

[Helmholtz]

Are T9/T10 and T11/12 same types respectively? Are all transistors well attached to their heat sinks using thermal grease where required?

Unfortunately I'm not familiar with the construction of this amp so I can only give general advice.
Maybe someone else knows this amp better.

[Helmholtz]

Don't R18, R20 set the output current limit?
R12, R13, TR6 set the bias

Missed your post.
I think you're right.
A closer look reveals that T7/T8 are inactive at lower output currents.

[Demian2020]

Are T9/T10 and T11/12 same types respectively? Are all transistors well attached to their heat sinks using thermal grease where required?

Unfortunately I'm not familiar with the construction of this amp so I can only give general advice.
Maybe someone else knows this amp better.

Thanks!

[Dave H]

which value suggests that I modify so that the transistors do not heat up at idle?

I would do the following first.
Disconnect any load (speaker, dummy load) from the output.
Measure the voltage on the speaker output. Is it close to zero and stable?
Measure the voltages across the ballast resistors. Are all voltages around 0.01V and stable? (If any rises to 0.1V something is broken)
Measure the voltage between Tr6 E and Tr6 C. What is it?

To reduce the idle current you could either Increase the value of R12 or reduce the value of R13.
Try tacking 12k in parallel with R13

[Helmholtz]

...and restore original 100R values for R18/R20.
My circuit interpretation was wrong, sorry.

[nickb]

Please see my edit above and post all 4 emitter resistor voltage drops.

You should use identical darlingtons on each side.

Just to clarify as I think the importance of this point has been missed.

T9 and T10 need to have the same part number and be matched i.e. have the same characteristics. Same goes for T11 and T12.

[Demian2020]

as BORAT says! Great success! I change again the 4 output transistors for the worst and cheapest tip142 & tip147 in the to220 packaging ... below I detail the modifications R13 1k3 and R12 1k6, tip31c by tip29. that's it! thanks a lot to everybody who help me!

[J M Fahey]

Please see my edit above and post all 4 emitter resistor voltage drops.

You should use identical darlingtons on each side.
Bias can be adjusted by the values of R18/R20. Higher resistance should lower bias.

Sorry but R18/R20 do NOT adjust bias.
They are part of short protection circuit, completely unrelated.

Bias is adjusted by R13/R12 ratio.

Interesting bias circuit, because instead of the typical TO92 transistor they used a way more expensive end needlessly powerful TO220 Darlington as heat sensor.
But then I see it is an *isolated* case one, which simplifies mounting, and also the wide metallic back of a TO220 allows for excellent contact with heatsink.

It is a classic "Vbe multiplier" circuit, with the added detail that it being a Darlington transistor, we have TWO "diode drops".

As shown CE voltage, which sets bias voltage for output transistors, should be around (R13/R12)+1 * 1.3V (2 diode drops) = 2.17V
That from T9-10 base to T12-11 ones.

Evenly split so 1.09V for eact power Darlington base.
*Slighly* underbiased or *just* biased, drop across emitter resistors as suggested by Helmholz should be quite low, a few mV at most.

IF bias voltage is fine (within these values), higher emitter currents (hence high drop across emitter resistors) may mean bad/fake/lossy power transistors.

Or Nobel Prize fakes, where supplier sends NON Darlington transistors labelled as such.

In that case they will be ALWAYS overbiased, overheat and burn.

I have seen similar cases where amp Tech didn´t realize power transistors were Darlington and used regular ones instead.

EDIT:

Ok R19 0.02V R21 0.03V R24 0.13 and climbing R25 also climbing 0.6v then I ear a little pop an goes to 0v

Ouch!!!!
R19-21-24 voltages normal and what I calculated above,
R25 is GROSS, either a very weak/fake T10 or it is, as I have already seen elsewhere, aNON Darlinton transistor.
FWIW emitter resistor drop is almost exactly "one diode drop" , a BE diode junction we are missing.


EDIT 2:

Originally posted by Demian2020 View Post

What resistors do you suggest to replace with R18 and R19?



Optimal value requires trial and error. Start with 120R and increase to 150R or 180R if necessary. Or wire 100R trimmers in series with existing resistors. All emitter resistor voltage drops should be equal.

PLEASE do not do that, those resistors do NOT adjust bias, see above.

Besides, "bias is right" since all other emitter resistors show proper drop around 10mV, so instead of a Bias problem which woukld affect all, I suspect bad power transistors , not all but those which show abnormal voltages.meaning they pass WAY too much current EVEN with proper Bias)
600mV across 0.33 ohm means almost 2A so an idle dissipation of 2A*42V (rail voltage) of 84W !!!!!!!

That for a transistor which in that heat sink can barely dissipate 25W or so.

Forget the 100W or whatever number is quoted in the datasheet, that applies to a transistor bolted to an infinite heatsink (think a 1 cubic meter block of aluminum), absolutely unrealistic.

No wonder it goes "pop"

EDIT 3:

as BORAT says! Great success! I change again the 4 output transistors for the worst and cheapest tip142 & tip147 in the to220 packaging ... below I detail the modifications R13 1k3 and R12 1k6, tip31c by tip29. that's it! thanks a lot to everybody who help me!

Cool!!! CLEARLY the earlier TIP142-147 were as fake as a rubber Dollar and the new ones, even if the smaller case version, are the real thing.

Congratulations

[Helmholtz]

Sorry but R18/R20 do NOT adjust bias.
They are part of short protection circuit, completely unrelated.

Dave H already pointed that out and I apologized for my error/wrong advice.

[Demian2020]

Sorry but R18/R20 do NOT adjust bias.
They are part of short protection circuit, completely unrelated.

Bias is adjusted by R13/R12 ratio.

Interesting bias circuit, because instead of the typical TO92 transistor they used a way more expensive end needlessly powerful TO220 Darlington as heat sensor.
But then I see it is an *isolated* case one, which simplifies mounting, and also the wide metallic back of a TO220 allows for excellent contact with heatsink.

It is a classic "Vbe multiplier" circuit, with the added detail that it being a Darlington transistor, we have TWO "diode drops".

As shown CE voltage, which sets bias voltage for output transistors, should be around (R13/R12)+1 * 1.3V (2 diode drops) = 2.17V
That from T9-10 base to T12-11 ones.

Evenly split so 1.09V for eact power Darlington base.
*Slighly* underbiased or *just* biased, drop across emitter resistors as suggested by Helmholz should be quite low, a few mV at most.

IF bias voltage is fine (within these values), higher emitter currents (hence high drop across emitter resistors) may mean bad/fake/lossy power transistors.

Or Nobel Prize fakes, where supplier sends NON Darlington transistors labelled as such.

In that case they will be ALWAYS overbiased, overheat and burn.

I have seen similar cases where amp Tech didn´t realize power transistors were Darlington and used regular ones instead.

EDIT:

Ouch!!!!
R19-21-24 voltages normal and what I calculated above,
R25 is GROSS, either a very weak/fake T10 or it is, as I have already seen elsewhere, aNON Darlinton transistor.
FWIW emitter resistor drop is almost exactly "one diode drop" , a BE diode junction we are missing.


EDIT 2:


PLEASE do not do that, those resistors do NOT adjust bias, see above.

Besides, "bias is right" since all other emitter resistors show proper drop around 10mV, so instead of a Bias problem which woukld affect all, I suspect bad power transistors , not all but those which show abnormal voltages.meaning they pass WAY too much current EVEN with proper Bias)
600mV across 0.33 ohm means almost 2A so an idle dissipation of 2A*42V (rail voltage) of 84W !!!!!!!

That for a transistor which in that heat sink can barely dissipate 25W or so.

Forget the 100W or whatever number is quoted in the datasheet, that applies to a transistor bolted to an infinite heatsink (think a 1 cubic meter block of aluminum), absolutely unrealistic.

No wonder it goes "pop"

EDIT 3:


Cool!!! CLEARLY the earlier TIP142-147 were as fake as a rubber Dollar and the new ones, even if the smaller case version, are the real thing.

Congratulations

Thanks juan! In a day or two I will buy you 2 10" speakers. They sound amazing

[Demian2020]

Dave H already pointed that out and I apologized for my error/wrong advice.

do not worry! Thank you very much for worrying and responding so quickly