Trace Elliot AH250 smx gp12 - oscillation on the gate of 2sk135 mosfets

[martyman4050]

Re-post to the correct forum - sorry...
Greeting...need some expert opinions/help....
I recently was playing my favorite Trace bass amp when I noticed a drop in the output.
The amp still worked but it just did not sound right and the heatsink became quite hot to the touch. I opened it up and did not see and obvious burnt components. I powered it up on the bench with a suitable dummy load and started probing around the output with a 1khz sine wave at the input. The DC on the output was about 40mv which did not appear excessive. The amp is working and the signal is getting amplified but when I looked at the gates of the 2sk135 output MOSFET I noticed a sizeable high frequency signal riding on the sine wave when the output is turned up just a little.(which I assume is causing the hot heatsink) The HF does not appear on the output or on the gates of the 2sj50's. What could cause this? I figured I would just replace the output transistors....any ideas? - thanks in advance - Martin

[R.G.]

Does this thing have gate stopper resistors right at the gates of the MOSFETs? If not, it needs them. If it does but they're not RIGHT ON THE GATE PIN, then put gate stoppers on the gates right at the gate pin.

[martyman4050]

Ah250 smx - gate resistors and schematic

thanks for the response..
Yes the gate resistors are 220 ohm and I only see the HF on the side that is connected to the gate. I have attached a schematic pin 1 of tr6 and tr7 is where I see the HF noise.

[J M Fahey]

I'd cut the tracks if necessary and add 1k 1/4W in series with each gate, soldered to the actual transistor gate leg.
Don't trust the scope very much because its probe might actually be triggering the oscillation, trust the actual transistor overheating or the lack of it.
If you connect your scope there, do it through a 2k2 or 4k7 isolation resistor, with on end soldered to the gate and clipping the scope probe on the free end.

[Jazz P Bass]

These Trace amps do run hot.
Even at idle.

[martyman4050]

I bought this amp new in about 1994 and it has been working fine until two weeks ago.It never ran this hot.. (I would check the heatsink during and after loud gigs).
Something happened during a recent jam (the output dropped down and it sounded kind of thin, and when I touched the heatsink it was very hot.) The heat sink barely gets warm when it's idling and the gate does not show the oscillations until the output control gets to about "1".
Do you think it's possible that the transistor had some internal break down which changed it's capacitance causing oscillations or just caused it to run hotter but still able to operate? or do these things just totally fail when they go?
R.G brought up a good point ...maybe the scope probe induced the oscillations I saw... it's kind of weird that I did not see the oscillations at the output... or would the zobel network kill these?.. I will try using a resistor to probe the gates...but something still is making this run way too hot...
Thanks again...

[J M Fahey]

Maybe your Zobel opened !!!
Tack solder a new one in parallel with it.
It won't hurt your amp if redundant.

[Jazz P Bass]

Check the power supply voltages for adverse Vac ripple.
There are two supplies.
The output power rails & + - 15 Vdc.

[R.G.]

Maybe your Zobel opened !!!
Tack solder a new one in parallel with it.
It won't hurt your amp if redundant.

I like this one. It explains the issue with it starting this suddenly.

[martyman4050]

After reading through the replies (thanks) I went back and probed the gates utilizing a 1k resistor and lo and behold I did not see the HF oscillation on the gate...I guess they are quite sensitive to probing...
Could anyone tell me the correct procedure to set the pot that appears to balance the drive to the transistors?

[MarkusBass]

Marty,

I don't know what you you mean by:

the pot that appears to balance the drive to the transistors

but the pot sets the quiescent current of the output transistors and it has nothing to do with balance. First, check whether it is not open or set to maximum resistance (because in this case the quiescent current will be very high and output transistors will get got - maybe it is the case you have). Then, measure the current of output transistors (just put a multimeter between the power supply and the amp). If it too high, decrease the resistance of the pot.

Mark

[km6xz]

Does the amp still get hot without loads and will it swing full rail to rail output open circuit? An oscillation that is due to phase shift in the output will be seen at the output also.
If it is not generating excess heat with no loads, and the problem occurred during a performance, could it all be traced back to problems with the load? What is the dc resistance of the cabinet compared to the past if you ever measured it? The sound dropping in quality but no offset and full unloaded swing all suggest looking at the speaker carefully. Do you have a dummy load or access to another speaker you know works properly?