GK 250ML combo setting bias

[Randall]

Since I measured both caps to be reasonably close to .1mF, can we consider the Zobel network to be not the problem?

[Jazz P Bass]

I thought the zobel network was there to alleviate issues with the speaker impedance changing with frequency.

So if you are driving a dummy load, there are other issues with the amp.

[Randall]

I'm driving the speakers.

[nosaj]

"The + 45v rail has ~600mV AC p-p, the -45 Rail has ~ 400mV AC p-p."
Not to be jumping around but that is an awful lot of ripple.

What symptoms will a lot of ripple cause?

nosaj

[g1]

Excess ripple could indicate poor filtering which could result in instability. But the ripple could also be a symptom of excess current due to the oscillation. So no simple answer, could be cause or effect.

[nosaj]

Excess ripple could indicate poor filtering which could result in instability. But the ripple could also be a symptom of excess current due to the oscillation. So no simple answer, could be cause or effect.

Well if they're the big caps that were like the ones in the one I had. I'd replace them in a heartbeat.
nosaj

[J M Fahey]

I´d get rid of oscillation first, and then recheck whether we still have ripple.

Or from other point of view:
* ripple is caused by caps discharging
* will be higher with higher current or lower capacitance
*so I would measure idle current, not only across emitter resistors "which is what we use when everything else is fine" but at a more primordial way: straight cutting (with an X Acto knife or similar) the track feeding one power amp, previously scratching solder mask to expose copper and plain measure it with an ammeter.
Repeat on the other channel.
After measuring you can bridge across the cut with a drop of solder or even add a little piece of tinned wire across the gap to be even safer.
In the good old days (60´s, early 70´s) PCBs already included 2 side by side pads across a gap specifically for that purpose, then we started saving a few seconds of assembly time by measuring across emitter resistors, which is also more precise, but the all way catches anomalies that the new one does not, because we measure the grand total: power transistors, drivers, and the rest of the circuit.

As of ripple, it can be calculated knowing current and capacitance , and compared with what´s actually there.
In principle not sure 400 or 600mV ripple is too much or too little, capacitance is fine or reduced, amp is pulling normal or too much, all 3 values are closely related to each other, let´s measure actual current, actual capacitance, calculate ripple and compare.

We have 3 variables there, knowing two we can calculate the third.
I expect each power amp channel to eat, say, from 40 to 80 or 100mA tops, anything above might be suspect.
Only then we can say "no big deal" or "WTF???"

[Randall]

"*so I would measure idle current, not only across emitter resistors "which is what we use when everything else is fine" but at a more primordial way: straight cutting (with an X Acto knife or similar) the track feeding one power amp, previously scratching solder mask to expose copper and plain measure it with an ammeter."

Juan, I am unsure where you would like me to cut and measure. Could you be more specific? Thank you.

[J M Fahey]

Sure
On page 11 of the schematic, bottom right, you have the power supply.
You see a "+45V" labelled rectangle on top of 2200uF capacitor C51? (it´s somewhat fuzzy on my screen).
On top left of same page you see the matching "+45V" labelled rectangle feeding the top power amplifier.
"Something" joins those two rectangles, maybe an end to end track, or a short track at the supply board (if separate), some wire and connectors, again another short track at the other end.
I´d make a narrow cut at such track , previously scratching both sides of it, to measurevthe current going to the top amplifier.
Choose a point where current goes only to the top one.
After measuring join both sides of the cut with a drop of solder.

Then look center left, you´ll see another "+45V" labelled rectangle feeding the bottom amplifier, do the same.

IMPORTANT: both measurements made without speaker connected.

Post results, then we´ll see whether current consumption is too high or not, measuring emitter current is only part of the picture.

Also scope output *without* speaker load, to check whether it oscillates with a speaker connected and not without OR maybe some other stage is oscillating and power amp just reproduces that.

We do not know (yet) whare that oscillation comes from, so we test various possible causes.

[Randall]

Thank you so much. Will do tomorrow and report back.

[Randall]

Well then. I made the first current measurement as described and got an initial reading of 153mA that drifted up to 253mA over perhaps 10 seconds. I repaired the trace, and lifted a jumper to make the same measurement on the other power amp. This time I got an initial reading of 83mA and as I watched it drift up my 2 A line circuit breaker tripped, and now I have blown output transistors on the amp that I did not replace them in.

Called customer and explained this could keep happening, and he said to abort, he will sell it for parts. Can't say I am sorry to hear that actually.

Thanks for the advice and help as always.

[nosaj]

Well then. I made the first current measurement as described and got an initial reading of 153mA that drifted up to 253mA over perhaps 10 seconds. I repaired the trace, and lifted a jumper to make the same measurement on the other power amp. This time I got an initial reading of 83mA and as I watched it drift up my 2 A line circuit breaker tripped, and now I have blown output transistors on the amp that I did not replace them in.

Called customer and explained this could keep happening, and he said to abort, he will sell it for parts. Can't say I am sorry to hear that actually.

Thanks for the advice and help as always.

Wonder why the 1amp fuse in the amp did't blow instead?

nosaj

[Enzo]

1A fuses do not blow right at 1A magically. All fuses have blow curves in the data sheets. 2A in a 1A fuse takes X-much time to blow a fuse. The transistors might fail, and he sees 2A mains draw, but didn;t leave it on long enough to take the fuse.