Ampeg BA115HP is driving me nuts!

[glebert]

Posted in a different thread about this amp burning up output transistors while I was working on it, now something else is stumping me.

With output FETs removed and no input I am trying to get the gate voltages to be in the right ballpark. If I have the op amp U5 pulled the voltages look really good, the gate voltages move symmetrically when the the bias pot is turned, seems like a reasonable range of Vgs can be had. The node at the pin 7 of the op amp sits at about 0 volts when it is pulled.

When U5 is installed though the voltage at pin 7 drops to -10V, and the gate voltages no longer move with the bias adjust, the gates for Q2 and Q7 are 10V below the positive rail, the gates for Q5 and Q8 sit at the negative rail. There is no voltage across R35, Q1 is not turned on at all. There are no DC voltages on the op amp pins other than pin 7 and the two supply pins which are at +/-16.5V. I have scoped for oscillations and see nothing. Measuring some of the resistances in this part of the circuit has not turned up anything off.

Q1, Q3, and Q4 have been replaced. Multiple U5s yield the same results.

Any thoughts?

TIA,
Greg

ampegba115hpschem.pdf

[Enzo]

You have the MOSFETs out? DO you have a speaker load? Without a load, the output bus floats, and I may be all wet, but I don't see a way to get a good refernce on pin 5 of U5, so I don't doubt the output drifts around.

I could be wrong, better heads can correct me. But it is a hunch. But then I do have a glass of moonshine next to me.

[Jon Snell]

How do you expect to fault fing with no feedback source?
Go back and have a recap.

TP11 is the output that feeds back to U5 B, that sets the DC level on TP11. No output fets = no control.

The most important measurement is the collector/emitter voltage across Q3.
You have established that Q3 is biased on, (or short circuit) because you state zero volts across R35.

I would say it was safe to add a single pair of fets after checking the voltages across R31 and R34, they must be less than 1volt when AP1 is fully clockwise.
With a pair of fets wired in, there will be feedback to U5 and as long as U5 has +- on the supply, will compensate for errors.

A fault at this point will be ellementary to find.

Don't over think it, It is only an amplifier, meaning, vary the DC voltage on pin 5 of U5 and that will vary the DC voltage on pin 7 which will vary the current amplifiers Q1 + Q4 collectors of which vary the voltage drive on the fets to give the desired output. TP11 then tells U5 all is or is not well in triving for 0v at TP11.

Always fault find with no load.

[glebert]

Thanks for the lesson guys, I appreciate it. I know that I tend to think too much in terms of stages of a circuit, and in this case you can't separate this section because it is a big loop. I think I was also gun shy on the output FETs because the amp had freaked out on me before.

[glebert]

OK, hopefully last question on this amp. I have the power transistors in and the bias voltages are acting as expected. I can get clean signal through the headphone jack.

I have a mismatch in the currents on the IRFP140 transistors though. One was reading about 2.5x the other. I replaced the higher current one with another new part and it got better but still significant mismatch. Now the hotter one is about 50% more current. The 9140 parts seem to be very well matched. Do I keep trying different parts? How close do they need to be matched (is it just the 5mV diff from the schematic)? If I do replace try other parts should I replace the hotter one (again) or the colder?

TIA

[Enzo]

2.5x the other? OK, is that 1mv and 2.5mv? Or 1v and 2.5v? COntext is everything.

[glebert]

2.5x the other? OK, is that 1mv and 2.5mv? Or 1v and 2.5v? COntext is everything.

Measuring across the 0.33 ohm resistors, one was at 21mV and one was at 8mV. Now one is at 12mV and the other at 18mV.