Yamaha G50 output biasing problem

[glebert]

Have a Yamaha G50 112 (first gen) that plays well at first, and then starts to get distorted and quiet after 5-10 minutes of playing. Voltage in the middle of the totem pole output section is 20V when powered on, drops down to 6V when distorted (this is measured at output cap to speaker). Voltage rail is 89V and solid the whole time. 20V seems pretty low, I would think it would be close to 1/2 of the voltage rail, especially since 20V wouldn't give enough swing to get 50 watts with an 8 ohm speaker. Haven't measured individual transitor bias voltages yet. Thoughts?

yamaha_g50-112_guitar_amplifier_sch.pdf

Thanks,
Greg

[The Dude]

The schematic link doesn't work for me. Anyway, I'd try some freeze spray after the amp begins to distort. I'm thinking a component is failing when it warms up. You might be able to narrow down which part is causing the issue.

[glebert]

Don't know why pdfs aren't loading. Here is a picture of the output section.

[DrGonz78]

You have to pack them up in a zip file or something of the sort. Or if you can add something to the drawing to increase the file size that will work too.

[lowell]

If that output cap is an electrolytic it is suspect. Does this happen with no speaker connected? That could allude to the output cap leaking DC. (If the problem is fixed with no load connected.) Ofcourse you'd be getting hum from the speaker as well.

[glebert]

Been spending more time with this amp, and it is really confusing me. The voltage on the amp side of the output capacitor comes up VERY slowly, like 20 minutes to get up to about 1/2 of the +B voltage (which is about 86V), and is fluctuating quite a bit the whole time. At that point the amp will play, but the voltage will start drifting down until it starts to fuzz out. The slow startup happens with the output capacitor disconnected as well. Thought the 10uF feedback cap might be leaking but the voltage across the 4.7k resistor to ground drops when the voltage is crashing, if the cap were leaking I would expect that voltage to go up since the current would be going through it. Basically all voltages at the collector of TR14 and beyond all rise and fall together (other than the +B and TR15 voltages which are solid).

Have tried compressed air can to cool things down, nothing seems to respond. I can make the voltage crash very quickly however if I hold a hot iron near the 0.47 ohm resistor on the emitter of TR19. This resistor has some signs of being very hot at some point but ohms out correctly at cold. I am planning on replacing this but don't think this explains all the behavior.

When I pull TR19 and TR20 (easy since they are TO-3) the totem pole midpoint voltage doesn't get above about 17V (although it does come up quicker and more linearly than with TR19 and 20 in place).

Additionally, this board is a PITA since it is on a module that hangs under the chassis such that probing the foil side of the board is impossible, and many of the component leads are covered with some sort of braided cloth tubes.

Don't want to start shotgunning and pulling components but not sure where to even get started. There are the three trimpots that are probably important as far as getting the bias right, but without some (legible) test point voltages I don't want to just start tweaking those.

[Enzo]

TR17-20 are a current amplifier, whatever voltage is on the bases will be at the output, more or less. TR16 controls them. Whatever is on the base of that is about what is on the output. Or the emitter. SO start there, is that also fading low?

Is the main supply staying up and clean?

The trimmer by TR14 I think is the DC offset, which in this amp sets the output at 1/2B+. Is the voltage at TR13 base stable or drifting?


Oh, and voice of experience here: resolder all the pins on the 6-pin connector the module connects to the chassis with.

The two e-caps over by TR11 are suspects to me.

[glebert]

TR17-20 are a current amplifier, whatever voltage is on the bases will be at the output, more or less. TR16 controls them. Whatever is on the base of that is about what is on the output. Or the emitter. SO start there, is that also fading low?

Yes, TR16 base and emitter voltages fade

Is the main supply staying up and clean?

yes

The trimmer by TR14 I think is the DC offset, which in this amp sets the output at 1/2B+. Is the voltage at TR13 base stable or drifting?

Thought it stable was but will have to recheck

Oh, and voice of experience here: resolder all the pins on the 6-pin connector the module connects to the chassis with.

The two e-caps over by TR11 are suspects to me.

Good ideas, will resolder and check voltages by TR11

[glebert]

OK, reflowed solder on the board connector (both sides), no change.

TR13 voltages come up right away and are stable (emitter 86.3V, collector 85.4V, base 85.5V) Test point voltage on the schematic says 89.7 at the emitter but all my voltages seem a couple volts lower than the TPs.

TR10 and 11 voltages come up right away and are stable, but at higher than 1/2 of +B (source 65.9V, gate 63.9V, drain 84.6V). If I squint hard the test point voltage at the source of TR10 might say 68.3V, which would be in the ballpark.

TR14 and TR15 seem to be where it gets interesting.
TR15 (emitter 85.4V, base 85.0V, collector 65.7V) all come up right away
TR14 (emitter 85.4V, base 84.8V, collector voltage rising slowly, was 18V when I did measurement, but tracks with voltages through rest of output)

These were taken with TR19 and 20 in place BTW

The resistors in the emitter legs of TR14 and TR15 aren't that much different, and when you consider TR14 is also driving the base of TR16 I would expect those voltages to be pretty much the same. Then again at that point and beyond all voltages are rising and falling together so a leak anywhere seems like it could drag this down too.

Diode D1 (which is mounted to heatsink and presumably thermal compensation to prevent runaway) I am getting about 1.5V across, but I think this is actually 3 diodes in series, correct? Part number STV 3h Varistor. Don't know what kind of voltage to expect across this. Found one chart on audiokarma.com that says that 1.5V would be for a current of 500uA, which seems very low.

[glebert]

TR16 voltages are about 90V at collector, base follows TR14 collector, and there is a 525mV drop from base to emitter. It seems like TR16 is just barely turning on. This makes me think there isn't a "leak" after TR16 since that would turn it on harder.

Just thinking out loud: Assuming the collector of TR20 (i.e. at the output capacitor) should sit close to 45V (the middle of the 90V +B rail ), then the collector of TR14 should be three diode drops higher than that. The collector of TR15 is quite a bit higher than that (65V). Assuming TR13 is a constant current source then it seems like too much current is going through TR15 and not enough through TR14? I'm assuming with no signal TR14 and TR15 current should be equal and voltages similar.

[glebert]

Update on this: Since the voltage problems seem to start at TR14 I pulled TR16 to isolate it from the rest of the output circuit. Basically it didn't change things much, with the TR15 collector voltage coming up right away and the TR14 voltage coming up very slowly to about 40V. While measuring the TR14 collector voltage I blew some canned air on it and saw that the voltage actually dropped. I cooled down TR15 and the TR14 voltage climbed rapidly up close to 50V. Spraying both transistors got a voltage of about 45V which is about where I expect it to be. I have orderd replacement transistors for TR14 and 15, they sure seem suspect, but I really do prefer things that are failing to just get it over with and die completely.

[glebert]

Replaced TR14 and TR15, no change. Noticed a small change when cooling the 4.7K pot connected to the FET current mirror. By adjusting that (quite a lot) while monitoring the output stage voltage I can get it to come up more quickly to the range that I expect, but it is very unstable, fluctuating by several volts over a few seconds and probably drifting higher than it should given enough time. I did check with my scope to very there were no oscillations happening other than the meandering voltage.

Even though I had to adjust the 4.7k pot quite a bit I don't think that is changing the voltages a lot (since the 150k resistor south of that should dominate) but a small voltage change to the TR14 base will increase the current substantially. I will spend some time checking out the voltages and components around the FETS. Also finally got a variac so I can try to set the voltages to match up with the test point voltages on the schematic.

[g1]

Wonky trim pots can be a real pain. You might want to try some fixed resistors in there just to rule it out as a possible cause for your instability.

[glebert]

Well, for probably the 10,000th time on MEF, I have to say it seems that Enzo guy is right. Went back to his hunch about the caps around TR10 and 11. Pulled them all and now the voltage comes up almost instantly, settles down a couple volts after a few minutes, but then the variation is tenths of a volt rather than multiple volts of wander. Unfortunately won't get a chance to get the new parts and put it back together before my vacation, oddly enough, to within spitting distance of Lansing.

Thanks for the assistance, will update in a few weeks hopefully with a success story.

Greg

[Enzo]

Well, don't spit, but you are welcome to come to town. Been very hot and fairly humid lately.

But then I live next door to a brewery.

https://www.badbrewing.com/

(In the video on the home page, right after the MASON sign is a view from across the street, and in the background to the left is the old High School building. I live in that building.)