Help requested: broken amp

[nowater]

Hello, I have a basic AB MOSFET amp that I built from an article some years ago. It has one 2SK134 / 2SJ49 output pair per amp, no parallel output FETs. It has worked well but recently the right channel blew. The output of this channel is now jammed to the negative voltage rail, about -37V. There is zero current through either output device (formerly approx 500mA when the amp was working).

I am assuming the problem is one of the output devices. Is that reasonable?

If so, to push the output voltage to the negative rail, is it more likely the 2SK134 (connected to the +ve rail) has failed, or the 2SJ49 (-ve rail)?

Are there any simple diagnostic measurements I can take, to locate the fault?

Many thanks[IMG]Schematic[/IMG]

[Tube_Dude]

I am assuming the problem is one of the output devices. Is that reasonable?

Yes..it is reasonable...i suspect one or both outputs are bourned.
Inspect also the source resistor , it can be also bourned.

Regards

[Centauri]

Mosfets are fairly hardy devices, and not easy to damage under normal operating conditions (a short on a speaker line whilst running at high power could). If one or both of the output devices failed, more than likely it would have blown the rail fuses.

Check the voltages either side of the bias pot - if these are also at or near the negative rail, then problem is further forward of this.

Take a number of voltage readings and post the results referencing the schematic, and we may be able to help track it down.

Cheers

[janneman]

Most probably the neg side device has failed, shorted. But then you should have SOME current through it. Measure the gate voltages of the mosfets, that should give you an initial pointer.

Jan Didden

[cunningham]

Also check the voltages on the VAS transistor and verify that it is saturated and that it is suppose to be, due to circuit function, [or nonfuntion](check the base & emitter voltages.)

If above is true, check the state of the differential transistors to verify that they are set to cause saturated VAS transistor, and then check the negative feedback loop to see if the "neg. input" transistor in diff. circuit reacts to the 'now -60V' output.

Someware in this loop you should be able to track down the mishap.

Unlikely, but one of the current source devices could have shorted.

overall, this is not a really complex circuit...you will figure it out

[anatech]

I'd expect a high current flow if a FET had failed. I am looking at the current source(s) as a possible problem.

[nowater]

Thanks all, I will check out your suggestions tonight and report back.

[nowater]

OK I have some measurements.

Originally posted by Centauri
...Check the voltages either side of the bias pot - if these are also at or near the negative rail, then problem is further forward of this.

On the good channel, +1.77V and -1.55V
On the bad channel, -35V and -35V

Originally posted by janneman
Most probably the neg side device has failed, shorted. But then you should have SOME current through it. Measure the gate voltages of the mosfets, that should give you an initial pointer.

On the good channel, +1.9V and -1.7V
On the bad channel, -34.7V and -34.7V

Originally posted by cunningham
...on the VAS transistor ...(check the base & emitter voltages.)
...check the state of the differential transistors to verify that they are set to cause saturated VAS transistor, and then check the negative feedback loop to see if the "neg. input" transistor in diff. circuit reacts to the 'now -60V' output.

Sorry, but I'm not sure which one is the VAS
Base voltages on the differential inputs:
Good channel, 0V pos input, 0V neg input
Bad channel, -40.2V pos input, -39.5V neg input

Originally posted by anatech
I'd expect a high current flow if a FET had failed. I am looking at the current source(s) as a possible problem.

The CCS is set for 200uA. I checked the voltage drop across R12 (should be 0.66V) and R11 (should be 9.4V):-
Good channel, 0.74V and 10.5V
Bad channel, 0.0V and 0.0V

So, it seems the problem is not the output MOSFETs but some kind of short near the input? I checked the input RCA voltages: for the bad channel I measured -40.2V at both inputs (signal and shield).

I get the feeling I'm missing something blindingly obvious!?

[anatech]

Looks like you are missing your input ground (top side R7). That would shut down your current sources and pull the output negative. This kills your bias as well.
-Chris

[Centauri]

On the left portion of the schematic (on the previous page of the magazine article - and not shown in the attached scan) is a note saying "shield must be connected to 0V BEFORE applying power".

It would seem that this connection has come adrift.

If your main power supply earth is connected to chassis earth and mains power earth, then I would recommend connecting the input earths to main earth via 10 ohm 5W resistors anyway to avoid earth loops.

Cheers
Graeme