Marshall mosfet 100 - power amp issues

[Intripped]

Hello, I'm a new member of this forum - I'm not a native English, so please be forgiving with my writing.
I've found this place while I was searching for infos about this Marshall mosfet 100 amp I'm trying to fix
- schematic -

This amp had issues with volume levels, that were not constant, and sometimes sound cut off.

After a visual inspection, I found that all the three ground wires supposedly bolted to the chassis had loose bolts, and the bolts were oxidated as well.
I proceeded to restore good ground connections, and fired up the amp with no speaker plugged, in order to check voltages.
I also started looking around for info about solid state amps, since I'm not expert (at all).
I've learned that the internal trimmer (BAL-RV1 1k) is for D.C. Offset adjustment, and i noticed that I couldn't reach 0V (I measured around 2,5V(!) with the trimmer in the original position); so I ordered a pair of matched MPSA06, in order to sub TR1 and TR2.

After substitution I can now adjust the D.C. Offset properly, but I've found a new issue: TR3 and TR6 are getting quite hot, so hot that I can barely keep the finger on them for just some seconds (...yes, I know this is not a safe procedure, but I pay a lot of attention while doing it)

Voltage drop across R13 and R15 is around 5,5/6mV, so it should be ok, meaning ~20mA bias current on output mosfets
Voltage drop across R7 and across R14 is around 690mV

Questions:
Is it normal that TR3 and TR6 get so hot?
I think not, and I suspect TR3, am I on the right path?
What other measures can I make in order to address the problem?
How can I be sure that the mosfets are ok?

Thanks to everyone, and please excuse for this long first post

[Intripped]

I've measured some voltages, with D.C. offset trimmer set to mid position (930mV on output)
TR3 and TR6 still getting quite hot
Do you see anything strange?
As I said I know nothing about solid state amplifiers ...please help.

[g1]

Your numbers show that TR3 & TR6 have about 7mA current through them. (680mV/100R at R7 and R19)
They are rated for 500mA so I don't think you have any problem.
Set the balance pot for 0VDC on output and work on the volume issues.

[Intripped]

Thank you for the answer g1,
so TR3 & 6 are supposed to get so hot? I mean, is this normal?

[Dave H]

so TR3 & 6 are supposed to get so hot? I mean, is this normal?

They will get hot. They are dissipating 300mW in a TO-92 case.

I'd also check C3. The way it is on the schematic is reverse polarity. If the output is set to zero then the base current drop across the 22k will make the base of TR2 -ve so the -ve of C3 should be facing the base of TR2.

[Intripped]

Thanks Dave,
I've bought a new mpsa93 for TR3, and a bipolar capacitor for C3.
I'll try and substitute them and report back.

About C3, anyway: On the schematic there's a 22uF, but on the board there is a 2.2uF.
This seems to be the original cap, it's identical to the others on the preamp board...
What do you think? Assembling error or schematic error?
What is this cap for?

EDIT:
If I'm not wrong this cap forms an high-pass filter in conjunction with R6 (560 Ohm), in the NFB path.
With 2,2uF the corner freq. is 129Hz, and with 22uF it's 12,9Hz, so with 2,2uF we have less bass freqs at the output

[Intripped]

So,
I've installed a new C3 capacitor - no change: TR3 & TR6 still getting hot, also no change in D.C. Offset
Installed a new mpsa93 for TR3 - again same as before: no change.

Anyway I've plugged the speaker, after D.C. Offset adjustment, and played for half an hour: the amp works very well, no thump at switching on and off, stable volumes, the mosfets getting warm but not too much.

The conclusion seems to be that those transistors are actually supposed to get so hot...
I'll try and sub also TR6, but I have to wait a little because the local shop hasn't got mpsa06 at the moment.

[Dave H]

TR3 & TR6 still getting hot
The conclusion seems to be that those transistors are actually supposed to get so hot...

I'm sure you are right. I calculated from the data sheet that they will be running at about 70 deg C case temperature.

[J M Fahey]

Unjustly despised, yet very good amplifier.
Distortion is cheesy, but clean channel is killer and you can always add your own pedals, so ....
IF possible, try to run it on 8 ohms only and it will last forever; some experienced overheating when running it full blast on 4 ohms.
If at all possible, please post a couple pictures showing the heatsinks, both from inside and outside.
I think the cooling fins are buried inside the cabinet so won´t work well; I bet a 24V PC type fan, with a series resistor so it can run from the 49V rails might be the answer.
Also post a full amp picture from the back, showing ventilation slots (if any at all).

[Intripped]

I'm sure you are right. I calculated from the data sheet that they will be running at about 70 deg C case temperature.

Thanks for confirming this (could you please post here how to do this calculation? I'd like to learn it, thanks!)

Unjustly despised, yet very good amplifier.
Distortion is cheesy, but clean channel is killer and you can always add your own pedals, so ....
IF possible, try to run it on 8 ohms only and it will last forever; some experienced overheating when running it full blast on 4 ohms.
If at all possible, please post a couple pictures showing the heatsinks, both from inside and outside.
I think the cooling fins are buried inside the cabinet so won´t work well; I bet a 24V PC type fan, with a series resistor so it can run from the 49V rails might be the answer.
Also post a full amp picture from the back, showing ventilation slots (if any at all).

I think the fins are ok
the mosfets aren't overheating, I was talking of TR3 and TR6, two transistors in TO-92 case mounted without heat sinks on the power amp board.

Here are some pics anyway:

[Dave H]

Thanks for confirming this (could you please post here how to do this calculation? I'd like to learn it, thanks!)

OK, here you go

First calculate the dissipation of the device.
TR6 is set up as a constant current sink with a voltage reference of 2 x diode Vf. One diode Vf is offset by the Vbe of TR6 so the voltage across R19 is 1 x diode Vf. The current is therefore (1 x diode Vf)/100 = 0.6/100 = 6mA. The amp's output is at 0V so the voltage across TR6 is a little lower than the negative rail (say 50V). Therefore the dissipation of TR6 is 6mA x 50V = 300mW.

From the MPSA06 data sheet the thermal resistance junction-to-ambient is Rja=200 deg C per W
The junction temp rise above ambient is therefore 200 x 0.3 = 60 deg C
Assuming a 40C ambient inside the amp gives a junction temp of 60 + 40 = 100C

From the MPSA06 data sheet the thermal resistance junction-to-case is Rjc=83.3 deg C per W
The junction to case temp difference is therefore 83.3 x 0.3 = 25C
Subtracting 25C from the junction temp gives a case temp of 100 – 25 = 75 deg C

Hope that helps.

[Intripped]

WOW, it's all very clear, thanks a lot!

[loudthud]

There are little clip-on heat sinks for TO-92 transistors. The problem is making them stay on the transistors and not end up bouncing around inside the amp.

One example: https://www.aavid.com/products/standard/575200b00000g

[g1]

If the board is not brown, you should be fine.
(and sometimes even if the board is brown) I seem to recall some similar marshalls with browning around some small signal transistors under normal conditions; no defects. Maybe reverb drive transistors?

[Intripped]

Well, those heat sinks look interesting, I've also seen that RS-components and Mouser have them.
Maybe I could give them a try...
anyway the board is not brown, and the amp has been working fine for many years.
After all that has been said here on this topic, now I think that the main problem was loose bolts on ground-wires, on the chassis. Not transistors' overheating.
So for the moment I go without heatsinks, but thanks, because I didn't know they existed