GK 250ML combo setting bias

[Randall]

A GK 250ML is in for loud buzzing. The amp has two power amps, one for each speaker. I found the outputs on one side shorted, Q27 (TIP 33) fully, and Q31 (TIP34) partially. I replaced them along with drivers Q26 and Q30. Now the buzz is gone and both channels A and B produce sound, but with no signal the amp produces what looks like an HF oscillation of about .4 mVpp. When a sine is applied it looks like this distortion is riding on the wave and moving along it.

I wanted to check bias, the manual states:Turn speaker load off and bias. Adjust R114 to get a 5mV bias across R106 and R123 (these are the .33R emitter resistors), Adjust R147 to get a 5 mV bias across R139 and R157.

So I disconnected the speakers and measured "across R139 and R157", assuming this means the points connected to the outputs emitters, and get watched it quickly climb to over 200mV and then slowing drop down to some other value. I did not wait to see what that value is, I already know it's too high. This was the amp I did the repair to, when I checked the other side of reference, I found it similar in that it climbed to about 100mV before starting do drop. I have not touched that amp, so what could be the reason?

Also, both rails are low, ~40/-40v. Not sure where to begin. If the bias issue was just one amp I would question it, but since it is on both I wonder do I have a power supply problem. Or am I doing it wrong?

https://www.manualslib.com/manual/10...page=11#manual

PS about to get wet down here in SWFL

[nosaj]

A GK 250ML is in for loud buzzing. The amp has two power amps, one for each speaker. I found the outputs on one side shorted, Q27 (TIP 33) fully, and Q31 (TIP34) partially. I replaced them along with drivers Q26 and Q30. Now the buzz is gone and both channels A and B produce sound, but with no signal the amp produces what looks like an HF oscillation of about .4 mVpp. When a sine is applied it looks like this distortion is riding on the wave and moving along it.

I wanted to check bias, the manual states:Turn speaker load off and bias. Adjust R114 to get a 5mV bias across R106 and R123 (these are the .33R emitter resistors), Adjust R147 to get a 5 mV bias across R139 and R157.

So I disconnected the speakers and measured "across R139 and R157", assuming this means the points connected to the outputs emitters, and get watched it quickly climb to over 200mV and then slowing drop down to some other value. I did not wait to see what that value is, I already know it's too high. This was the amp I did the repair to, when I checked the other side of reference, I found it similar in that it climbed to about 100mV before starting do drop. I have not touched that amp, so what could be the reason?

Also, both rails are low, ~40/-40v. Not sure where to begin. If the bias issue was just one amp I would question it, but since it is on both I wonder do I have a power supply problem. Or am I doing it wrong?

https://www.manualslib.com/manual/10...page=11#manual

PS about to get wet down here in SWFL

When you put it back together make sure you put the short screws in the back. You put a long one in the wrong place and you'll short an output.

Best of luck for you. Looks like we're gonna get it pretty good here in Pensacola too..

nosaj

[Jazz P Bass]

"across R139 and R157", assuming this means the points connected to the outputs emitters"

What they are saying is measure across R139 or R157.

Are the power supplies ripple free (both the high voltage & the opamp supplies)?

Are both power maps doing that HF thing?

250ml.pdf

[52 Bill]

If the amp is oscillating, that alone could be the cause of the high current draw and the reduced supply voltages.

Does interrupting the preamp signal going into the power amp change the oscillation?

[Enzo]

You had a blown output stage, you replaced the transistors. Was the resistor in the zobel network burnt open?

[J M Fahey]

Transistors can be fast to turn ON and that is prominently displayed in datasheets; what most conveniently ignore or bury behind other data is that in general most are sloooowwww to turn OFF.

Ok, such time may be measured in fractions of a millisecond but at ultrasonic frequencies (the S.O.B.´s never oscillate at audible ones) turn OFF delay is an appreciable part of the cycle.

What does it mean?: that , say, top transistor (the one connected from +V rail to speaker out rail) is still ON, and bottom transistor , the one from speaker out rail to -V turns ON faster than the top one turns OFF, so for a fraction of time : both are ON, shorting straight from +V to -V , a catastrophic situation.

Yes, that "simultaneity" only lasts a few microseconds ... but is repeated every cycle at any frequency from 20 to 200kHz ... resut is catastrophic overheating and failure.

That´s why power stage oscillation is a bad thing.

[Randall]

"Are the power supplies ripple free (both the high voltage & the opamp supplies)?"

The + 45v rail has ~600mV AC p-p, the -45 Rail has ~ 400mV AC p-p.

"Are both power maps doing that HF thing?"

Yes. The side I did not work on has ~800mV p-p, the side I worked on has ~ 400mV p-p.

"Does interrupting the preamp signal going into the power amp change the oscillation?"

No. Plugging into the Send or Return jacks has no effect.

"You had a blown output stage, you replaced the transistors. Was the resistor in the zobel network burnt open?"

If I understand either the 5W .33R or the 5W 5R to be the Zobel, then no, they both look fine. However there is no capacitor between the 5R and ground in this one.

"result is catastrophic overheating and failure."

Yes, this is getting warm fast. I believe the chassis becomes part of the heatsink when assembled, but I do believe it is getting hotter than it should. I don't dare leave it powered up for long.

[Jazz P Bass]

"However there is no capacitor between the 5R and ground in this one."

Is there a spot on the board for C50 & C54?
That 101 (100pf) cap & the 5R resistor are the zobel network.

[Randall]

Now I see there is actually a C50 and C54, but they are located not on the ground side of the 5R, but on the other side. The amp I did not work on has a little blue device with 104 printed on it, I take this to be a 100 nF cap, it measures 94 nF. The other amp has what I thought was a resistor in the same position, with black/yellow/black stripes followed by a wider brown stripe. It measures 102 nF.

Odd they wouldn't use the same components on either side? I see both types scattered all over the board.

[Jazz P Bass]

Can you see where the odd behavior is starting?
Like maybe at the opamp output. (U21)

Are the .33 r's to ground good (on the speaker -)?

[Randall]

"Can you see where the odd behavior is starting?
Like maybe at the opamp output. (U21)"

I have not done signal tracing yet as I am reluctant to leave it powered up for fear of burning something else up.

I get .5 ohms from both speaker "-" tabs to board ground, so I would say that is OK.

[Jazz P Bass]

"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.

[J M Fahey]

Now I see there is actually a C50 and C54, but they are located not on the ground side of the 5R, but on the other side.

When two componebts are in series, the order is irrelevant.
C+R or R+C is exactly the same.
You should know that after *years* repairing amps.

The amp I did not work on has a little blue device with 104 printed on it, I take this to be a 100 nF cap, it measures 94 nF.

Yes.

The other amp has what I thought was a resistor in the same position, with black/yellow/black stripes followed by a wider brown stripe. It measures 102 nF.

Well, here order does matter.
Rather than "black/yellow/black stripes followed by a wider brown stripe" try "wider brown stripe followed by black - yellow - black" and consider the unit to be picoFarads ... what do you read now?
Is it compatible with what you measured?

Odd they wouldn't use the same components on either side? I see both types scattered all over the board.

I can easily visualize a fine lady assembling these boards on a workbench, surrounded by small plastic bowls filled with parts, noticing the .1uF one almost empty and asking for more.
Same *value* , not necessarily from same supplier, brand or colour.
New bag is opened and fresh caps are dropped into the bowl, mixing at random with the ones already there, she picks them "first found first used" style, and keeps assembling.
VERY MUCH doubt she will waste extra seconds per part just to match colours ; what for?

Of course in 60 years Forumites will *agonize* and fight bloody (virtual) battles about *why* the blue ones were used *here* and the banded ones were used *there*.
All kind of fairy tales and wholly made up BS will be written about cap colour , writing, shape, influence on Mojo.

Or as Chuck´s customer said: "love".

EDIT: now to serious matters: 101 as written on schematic is either a typo (most probable) or false flag to confound cloners, that value is useless as a Zobel cap.

.1uF is standard, and will equal 5 ohms at 300kHz or so, which is useful .... just imagine at how many *giga* Hertz will 101 start acting.

[Randall]

"When two componebts are in series, the order is irrelevant.
C+R or R+C is exactly the same.
You should know that after *years* repairing amps."

Of course I know that. I just didn't see it at first. If you have something you could offer to help solve the problem, I would be grateful.

[nosaj]

"When two componebts are in series, the order is irrelevant.
C+R or R+C is exactly the same.
You should know that after *years* repairing amps."

Of course I know that. I just didn't see it at first. If you have something you could offer to help solve the problem, I would be grateful.

He did say this in his post
Nosaj

EDIT: now to serious matters: 101 as written on schematic is either a typo (most probable) or false flag to confound cloners, that value is useless as a Zobel cap.

.1uF is standard, and will equal 5 ohms at 300kHz or so, which is useful .... just imagine at how many *giga* Hertz will 101 start acting.