Marshall rack power amp Valvestate 8008

[Marco Pancaldi]

Hello,
my 8008 has always had a slight but annoying crossover distortion problem, probably due to the cold biasing originally designed to not overheat the final transistors.
(schematic: https://schematicheaven.net/marshall...2x80w_8008.pdf)

I believe that the problem can be solved by varying the ratio between R25 / R24. I would like to fine tune the bias to the minimum necessary to make the low level distortion almost inaudible.
I think that using a trimmer with the wiper connected to the R25-R24 junction is not very reliable and that a better practice is to replace only one of the two resistors with a trimmer and a series resistance.

Maybe someone can explain better and suggest some trimmer/resistor values? I know it's a simple modification but I don't want to risk overheating the circuit, so I would like to use a low value trimmer to experiment (ear and oscilloscope) without fear of overheating the amplifier.
Many thanks

[g1]

So you are seeing the crossover distortion on the scope?
Measure the voltage across the output transistor emitter resistors (0.33 ohm 5W) with no signal to give you a reference for what the setting is now. Then, when you have an adjusting trimmer, you can bring the idle current up by a millivolt at a time. That will prevent overheating.
Will look closer at the circuit as you will want a fail-safe for the trimmer/series resistor in case the wiper goes open circuit.

[g1]

I think replacing R24 (680R) with a 500R resistor in series with a 500R pot (wired as rheostat) should work.

[loudthud]

Using a 1K pot in place of the 500 fixed and 500 pot is not recommended because if the pot is rotated to a short, output stage current would go to a very high value.

[g1]

Using a 1K pot in place of the 500 fixed and 500 pot is not recommended because if the pot is rotated to a short, output stage current would go to a very high value.

Thanks! Somehow I got the logic flipped for TR7 with my last statement, so I have deleted it.

[Marco Pancaldi]

Hello all, thanks for your answers!

I haven't the scope on the bench now, but I will use it to check the bias fine tuning - I am reasonably sure that the problem is crossover distortion because the amp plays fine at mid volume, while the distortion is mostly evident on low levels, ex. clean string decays, the same way for both channels.

g1: Checked voltage across resistors: 1.2mV - 0.9mV (R53 - R54) for channel A, 1.3mV - 0.9mV (R66 - R65) for channel B. Qs:
loudthud:I am very afraid about overcurrent: my idea was to replace R24 (680R) with a series 650R + 50R trimmer (about 10% value of original 680R), so I have some +/- margin.

What about the trimmer value and choice (multiturn, wattage...)? The original resistor is 1/4W.

Thanks again!

[g1]

To increase idle current of output devices, you need to reduce the conduction of TR7. That requires a slight decrease in the R24 value.
So you can't use a 680R in series with the pot as that will only allow you to increase the amount of x-over distortion. The series resistor has to allow something less than the original 680R.
An alternative would be adding a fixed resistor in parallel with R24, just tack it on across the leads. A 10K in parallel will drop R24 down to about 630R. That would be a good place to start. The remeasure across the emitter resistors.
Right now you have approx. 1mV across 0.33R, so about 3mA current. Somewhere in the 5 to 10mA range will probably get you out of the x-over distortion.