Ampeg BA115 HPT help!!!!!!!!

[nanodocl]

I am trying to repair an Ampeg Ba115 HPT that blown output MOSFETS. When I change the MOSFETS approximately 2.5V DC appears in pin 7 of the operational U3 NE5532 and in a few seconds the fuse and the MOSFETS Q11 and Q8 blows.
I have been fighting with it for a week and already in my despair, I have changed all the components of the power stage, matching Q7 and Q6, and the MOSFETS.
I have tried to repair it with signal, without signal, with load, without load ... The voltages 55V, -55V, 16V and -16V are fine and stable.
I do not know what to do.
Any ideas?

Thanks in advance

ampeg_ba-115hpt.pdf

[Jon Snell]

My first question is why just Q11 & Q8, why not the other pair. Is it that these have been changed and persist on blowing perhaps?
What is the quiescent current? Why don't Q4 & Q5 blow.

[DrGonz78]

Quick Question.... Where did you buy these mosfets?

[nickb]

If output transistors keep blowing then we can probably say the quiescent current is too high and so that strongly suggests something is wrong with the drivers. Pull the board and remove all four FETS taking care to avoid blowing them up with static as they are very sensitive. Disconnect the speaker and add a wire link from the output to pin 7 of U3. This should allow us to get the thing into a state where we can make some measurements. Next we need to power it up and take some DCV measurements.

Measure
U3 pin 7. If not close to zero stop and find out why. If OK then
Q7 E, B C
Q6 E,B, C
Q3 B ( we should already have E and C from above)
Voltages across R67 and R58

[nanodocl]

I bought the MOSFETS from Mouser.
I have checked Q3, Q7 and Q6 off the board and they are fine just like the diodes.
I have jumped from pin 7 of U3 to pin J3 of the output.
By injecting a 1Khz and 170mV signal the voltages change.
The voltages have been measured with respect to the chassis, without load and with the BIAS adjusted to the minimum.

No signal
U3 pin7 = 8V

Q7 emitter = 7V
Q7 base = 2.4V
Q7 collector = 56V

Q6 emitter = -1.8V
Q6 base = -2.3V
Q6 collector = -45V


With signal
U3 pin7 = 1.2V

Q7 emitter =4.5V
Q7 base = 2.4V
Q7 collector = 56V

Q6 emitter = -3.6V
Q6 base = -2.3V
Q6 collector = -50V

Voltage across R67 = 0V
Voltage across R58 = 10V

Q6 gets quite warm

[Jon Snell]

Looking at the results, I would say that Q7 was faulty. Base Emitter voltage exceeds 0.7volts.

[nickb]

I realise I said something dumb. I was forgetting that the output stage provides an extra inversion so the attempt to get 0v on U3 pin & didn't work out. Remove the jumper and simply connect pins 7 and 6 with a 1k (or so) resistor. That should make pin 7 close to 0V. Please redo the measurements and don't forget about R58 and R67 as mentioned above. These are all DC so no signal.

Q7 BE junction is reverse biased so there nothing wrong with that. At first blush, nothng jumps out as being wrong. Therefore please also check the source to gate voltages for each FET in case one or more of the gate resistors are open

[Jon Snell]

Whoops, didn't notice the polarity, assumed again. Doh!
R56, check it is conducting to hold the base of Q7 positive with respect to its emitter.

[nanodocl]

Q7 and Q6 are OK. Tested with a PEAK.

I removed the jumper and put a 1K resistor between pin 6 and 7 of U3.
Voltages now seem correct.

U3 pin7 = 0.2V

Q7 emitter = 1.7V
Q7 base = 2.4V
Q7 collector = 56V

Q6 emitter = -1.7V
Q6 base = -2.3V
Q6 collector = -55V

What FET do you mean? Q10 and Q9?
Q10 gate has voltage.
Q9 gate is grounded. No jack insert on headphones

[nickb]

Q7 and Q6 are OK. Tested with a PEAK.

I removed the jumper and put a 1K resistor between pin 6 and 7 of U3.
Voltages now seem correct.

U3 pin7 = 0.2V

Q7 emitter = 1.7V
Q7 base = 2.4V
Q7 collector = 56V

Q6 emitter = -1.7V
Q6 base = -2.3V
Q6 collector = -55V

What FET do you mean? Q10 and Q9?
Q10 gate has voltage.
Q9 gate is grounded. No jack insert on headphones

Looks OK.

Your forgot about R58 and R67 AGAIN. They are critical measurements.

Which FET ? All of them i.e Q4, 6,8 & 11 (but not Q9, it's not a concern). Really just a sanity check in case of open circuits. Make sure you measure right on the pads of where the FETS were.

[nanodocl]

Yes now. Sorry.

R58 = 2.5V
R67 = -1.8V

Q4 S=56V D=0V G=54V

Q5 S=-56V D=0V G=-54V

Q8 S=-56V D=0V G=-54V

Q11 S=56V D=0V G=54V

I measured it with the 1K resistance at U3

[nickb]

Thanks. The average of the voltage across R58 an R67 is 2.15V which exceeds the minimum gate threshold for the FETS of 2V. My guess would be that Ampeg specified or selected FETs that had a higher threshold than that. We can modify the bias network to fix it. But first double check you have the bias trimmer set so the voltage across R58 and /or R67 is the minimum.

[nanodocl]

Yes, it's at minimum. All the way to the left. When I turn it to the right, the voltage across R58 and R67 raise.

[nickb]

OK great. Change the 4.7k to 3.3k . That should be enough to do it. Still keeping the current configuration test that the voltage across R58 added to the that accross R67 is less than 4 volts. If OK replace the FETS, take care to avoid any chance of static. Remove the 1k and put it at all back to the way it should be ( keeping the 3.3k of course). If you haven't got a variac with ammeter or a bulb limiter ( 60W) now is the time to get / make one. Power up using the preffered device, gradually raise the volatge in the case of a variac owhile checking the current, or the bulb should be dim. Noiw adjust the bias accroding to the notes.

[nanodocl]

OK, thanks. I will test it with a Variac. I have ordered the MOSFETS. As soon as they arrive, I test it and post the results.
Thank you very much for your valuable help nickb.