Sunn Coliseum Lead railitis

[Euthymia]

"Railitis" is my term for when a solid state amp has its output jammed up against one or the other power rail.

Got a Sunn Coliseum Lead on the bench, had blown up output transistors so I replaced all with MJE15022. This is a transistor I have used before in old Sunns and they usually work well.

This time, I replaced the blown transistors and all was not well.

With no load on the speaker jack, I am getting 36.7VDC. With an 8 ohm dummy load, the voltage drops down to about 4VDC, but R223 (see attached service manual) gets really hot.

If I disconnect all of the output transistors, I still get the 36.7 on the output jack. If I pull Q1 and/or Q207, the DC drops way down.

I have pulled and tested every transistor out-of-circuit (let's hear it for sockets), and in-circuit tested every diode and resistor on the driver board. I've replaced C202 and C205.

I've pulled both trimpots, and tested them across their ranges with an analog meter. They worked from end to end.

As you can see in the service manual, the PCB diagram and the schematic for the output driver board do not agree. The actual unit is closer to the PCB diagram. The big difference seems to be that the actual unit does not have the circuitry around Q205, Q206, CR213, CR214.

The bias adjustment has little effect on the voltage readings at the bases on the high side, but plenty on the bases on the low side.

Ideas?

http://www.rawbw.com/~emiller/Transfer/ColiseumLead.PDF

[Enzo]

Did you check all the ballast resistors for opens? Those are the 0.33 ohm 10 watt for each power xstr.

Did you test Q1 and Q2, the drivers? Open or shorted, either way woul dbe killer.

WHile there is R223 still on value? And R224? For that matter are predrivers OK and their resistors R221, 222 OK?

SO are we using page 1 or page 4?

If the bias adjusts the neg side but not the pos, then explore the voltage amp - the xstr just above the bias pot. And by extension the predriver to its right.

[Euthymia]

Here comes the cavalry

Did you check all the ballast resistors for opens? Those are the 0.33 ohm 10 watt for each power xstr.

Oh yes. One of them was open. I replaced it.

Did you test Q1 and Q2, the drivers? Open or shorted, either way woul dbe killer.

Yes, the Hickock says they are good.

WHile there is R223 still on value? And R224? For that matter are predrivers OK and their resistors R221, 222 OK?

I have checked and/or replaced them.

SO are we using page 1 or page 4?

Now that you mention it, I had been looking at page 4, but page 1 matches the unit on the bench better, so we are using page 1.

If the bias adjusts the neg side but not the pos, then explore the voltage amp - the xstr just above the bias pot. And by extension the predriver to its right.

Q203 and Q207, then. If I pull Q207, the problem seems to get much better, but it and the components around it test good.

What am I looking for at Q203?

Sheesh, one of the things that throws me about troubleshooting solid state is that a problem in one section can result in wrong voltages in a seemingly unrelated section. My troubleshooting skills rely on being able to isolate sections.

Every section in a tube amp is usually cap-coupled, which makes it relatively easy. SS amps tend to be direct-coupled, which I think helps throw me.

With SS, I never know when pulling something is going to cause a whole string to float or sink to a rail and make everything even more impenetrable (actually, it usually DOES do this).

I need to develop a new way of thinking to deal with that, but I haven't figured it out yet.

[Don Moose]

I need to develop a new way of thinking to deal with that, but I haven't figured it out yet.

I think solid-state amps require a plodding front-to-back approach; once you're comfortable that the power supplies are happy.

What I mean by that is that, by their non-isolated nature, a solid-state circuit requires that you verify each section, starting from the input. Where it starts to go awry, your problem is in that section +one/-two. The gotcha is that the entire power amp (predrivers, drivers and outputs is ONE section).

No practical help, here - SS circuits have me equally unmoored.

[Enzo]

Imagine a rubber band stretched out before you, with a spot marked in the center. If you pull harder or softer, the length varies, but the spot stays in the center. This is the amplifier at rest. The output is sitting at zero while the system is stretched between the two opposing power rails.

The amp circuits then move the spot back and forth... and then the analogy breaks down...

But if you let go of one end, the whole thing slaps over to the other end. If the negative rail cannot tug on things hard enough, then things will gravitate up to the positive rail.

You have + rail on the output, but a load draws it down to 4v. That means there is a resistance in the circuit, because the powr rail itself would not sag to 4v. (Well I assume anyway, is that tha case?) So I think the oyutput xstrs are not conducting, all the output current is coming through the driver Q1. That is why R223 gets hot. So why are the outputs not getting involved?

Here is a thought. The emitter of Q1 drives the base of Q2 et al by way of a connection through pins 8 and 9 of the circuit board. Verify ther is continuity from the base of Q1 AT THE TRANSISTOR ITSELF to the base of Q2 AT THE TRANSISTOR. Do not assume the tracework is intact. I am thinnking it might not be. Traces might have ben damaged when something burnt out.

For that matter, your rails might be OK, but are they actually getting to teh output xstrs?

The junction voltage of these transistors is roughly a half a volt. Normally the output is at zero DC, so readings to ground would work, but since the output is way offset, we need to take voltage readings with respect to the output bus.

In a normal amp, ther will be about a half a volt at the base of the output xstrs. This is a quasi complimentary, so the negative side will refernce the neg rail instead of ground. What voltage is at the bases on the pos side? SHould be about .5 at Q2, 1 at Q1, and about 1.5 at Q207.

The neg side is refernced to the neg rail, so we then expect the .5v at Q6 to then be across R218, and teh base of Q5 about 1v across R217 . But Q206 is right side up so to speak, so its base ought to be more or less 1.5v negative referred to the output bus.

I you want, you can lift CR207, 210 to disable the limiters, and make sure they are not involved.

So the bases of Q206,207 opught to have about 3v between them. In practice, it will be lower, more like 2-2.5v. But we are talking ball parks here, as compared to 30 some volts off. That 2-3 volts is the bias. The whole job of CR203-206 and the bais control in parallel with them is the maintain that voltage "space." Q203 will yank the whole thing up and down with the music, so the bases of Q207,206 will bob up and down, but always maintain that bias difference between them.

The whole totem pole of R212, Q203, CR203-206 (with control parallel), R215, R216 is the bias string. Q203 sets the current through the thing, and the diodes more or less set the voltage difference between those two bases - Q207,206. Q203 is the voltage amp, and when driven by the signal it will yank the whole totem pole up and down.

What could be wrong with Q203? DUnno. If it conducts to hard, then the totem pole is skewed positive. That the adjuster moves the neg side but not the pos, makes me think it is stuck. If not shorted, then look for an open CR201,202 or R210. Maybe shorted Q201, or shorted C202.

There is some stuff to chew on, check that trace.

[Euthymia]

Imagine a rubber band stretched out before you, with a spot marked in the center. If you pull harder or softer, the length varies, but the spot stays in the center.

Great analogy. Hadn't heard it before.

there is a resistance in the circuit, because the powr rail itself would not sag to 4v. (Well I assume anyway, is that tha case?)

As far as I can tell with the lightbulb limiter on it. Is there a safe way to check this? If I plug anything in, the resistor gets hot real quick and I don't want to fry it (again).

Verify ther is continuity from the base of Q1 AT THE TRANSISTOR ITSELF to the base of Q2 AT THE TRANSISTOR.

I assume you mean "emitter" of Q1? Anyway, there is continuity.

your rails might be OK, but are they actually getting to teh output xstrs?

Yes.

we need to take voltage readings with respect to the output bus...What voltage is at the bases on the pos side? SHould be about .5 at Q2, 1 at Q1, and about 1.5 at Q207.

Referenced to the output bus:
Q2 .78V
Q1 1.4V
Q207 2V

refernced to the neg rail, so we then expect the .5v at Q6 to then be across R218, and teh base of Q5 about 1v across R217 .

Referenced to the negative rail:
Q6 1.3V
Q5 1.3V

But Q206 is right side up so to speak, so its base ought to be more or less 1.5v negative referred to the output bus.

Referenced to the output bus:
Q206 -.75

I you want, you can lift CR207, 210 to disable the limiters, and make sure they are not involved.

Instead, since they are socketed, I pulled Q204 and Q205, but it had no effect.

So the bases of Q206,207 opught to have about 3v between them. In practice, it will be lower, more like 2-2.5v.

Ah, here's where we go really astray. The measured voltage between them is about 86VDC, depending on the setting of the bias pot. It's 85.8-86.8.

I bet that's a big clue, but damned if I can follow it.

What could be wrong with Q203?...If not shorted, then look for an open CR201,202 or R210. Maybe shorted Q201, or shorted C202.

All of the above components test good according to my semiconductor analyzer (transistors out of circuit) or DMM (resistors and diodes in-circuit).

BTW, your message is the first time I've ever caught that with a quasi-complementary, you test output transistor voltages with respect to the output bus and/or negative rail. That is gold right there. As you can imagine, I've always found the voltages I've measured at output transistors to be insanely off from what I thought they should be. I've yet to work on a transistor amp that was not quasi-complementary or complementary.

[Don Moose]

Since one of the things you're looking for here are opens, measure impedances pin-to-pin around Q206/7 and the components mentioned after that.

Sockets do sometimes fail to make good contact - especially the (flat)leaf-contact types liek cheap IC sockets. If they're tin-lead plated contacts, consider a bit of cleaning (DeOxit?).

Hope this helps!

[Enzo]

Referenced to the negative rail:
Q6 1.3V
Q5 1.3V

Well that's odd. Same volts at both places? Is Q5 shorted?

If the voltage between the base of Q206 and Q207 is 80v, then the bias string is open. Specifically the part where the four diodes are.

Power on, and measure the voltage across each of those diodes. NOT the diode test voltage, but just the actual voltage in the circuit across each. it SHOULD be about half a volt. If one has like 80v across it, it is open. If any have zero across them, they are either shorted, or the circuit board traces are open somewhere. Also measure voltage across the bias pot and across R211 470 ohms. Neither should be over a couple volts. Possible open resistor or pot AND a diode. Or all the parts are OK, but a trace is cracked on the board.

DO continuity check from cathode of CR206 to the end of R211 and to the base of Q206. ANd do a check from teh anode of CR203 to the collector of Q203.

[Euthymia]

Jesus wept

Well that's odd. Same volts at both places? Is Q5 shorted?

No, not according to my semiconductor analyzer. R217 is getting REALLY hot during my brief tests, though.

If the voltage between the base of Q206 and Q207 is 80v, then the bias string is open. Specifically the part where the four diodes are.

Gotcha.

Power on, and measure the voltage across each of those diodes. NOT the diode test voltage, but just the actual voltage in the circuit across each. it SHOULD be about half a volt.

The voltage across each is about .53V. The anode of CR203 has about 40V on it.

measure voltage across the bias pot and across R211 470 ohms. Neither should be over a couple volts.

Voltage across R211 is 2.2V.

Across the bias pot, 7V with the wiper all the way. Resistance measurement across the pot with the wiper all the way is about 800K. With the wiper run down to about 250R, the voltage is about 1V.

Possible open resistor or pot AND a diode. Or all the parts are OK, but a trace is cracked on the board.

DO continuity check from cathode of CR206 to the end of R211 and to the base of Q206. ANd do a check from teh anode of CR203 to the collector of Q203.

Of course, all test OK.

[Enzo]

Um, if there is half a volt across each diode in CR203-206, then how can there be more than 2 volts across R225 - the bias pot?

Also, the pot measures 800k? It is a 500 ohm pot. There is a clue for you right there. When you measure 800k, reverse your meter leads and see if you don't get a reading of "negative resistance." Pull the pot and measure it out of circuit.

There should not be 40v at CR203. However if the bias string is open, it will shoot towards the + rail, just as the lower end will shoot towards the - rail. Remember the rubber band.

OK, what is the voltage at each diode CR203-206? If you have 40 at the top and .53 acrossd each one, at the bottom of CR206 there ought to be about 37.88v. And then ther should also be the same 37.88 volts on the base of Q206 - after all they are wired together.

Something is open, a part, a trace, something. And with 7v across the pot and another 2v across the resistor, that makes 9v right there across the four diodes. So either there is 9v across them or four times .53. But not both.

[52 Bill]

I've pulled both trimpots, and tested them across their ranges with an analog meter. They worked from end to end.

Is it possible you switched the bias and offset pots when you replaced them? The offset pot is a 1M.

[Enzo]

Now there is a good call, and a 500 ohm in the offset position would do no favors either, in fact it would likely cause serious skew.

That still doesn't explain 80v across four diodes, but it might be the rigth direction.

Just shorting across the bias pot puts the amp in the coolest setting, which would be fine for testing.

[Euthymia]

the pot measures 800k? It is a 500 ohm pot.

Sorry, sorry, typo. I meant 800R. Perhaps it got roasted.

Something is open, a part, a trace, something. And with 7v across the pot and another 2v across the resistor, that makes 9v right there across the four diodes. So either there is 9v across them or four times .53. But not both.

I'll look specifically for something open.

Just shorting across the bias pot puts the amp in the coolest setting, which would be fine for testing.

I'll jumper it and try the voltages again.

[Euthymia]

How did SS ever get such a reputation for being difficult to service?

As I had been fearing might happen, all of this powering up and checking out-of-spec voltages finally resulted in a big POOF, smoking Q204-207 and R218. Also one of the output transistors on the plus side.

I replaced the earlier suspect, the bias pot. I replaced Q204-207 and R218.

So here are some of our sample voltages as they now appear with the bias pot jumpered, Q204 and 205 removed, only one output transistor on each string installed and the whole thing going through a lightbulb limiter.

The rails are at 45V and -45V.

There is about 24VDC on the speaker jack with no load, and about 3.5 with a load.

Along the plus string, referenced to ground Q207 has 29V at its base, Q2 has 28V, and Q1 29V.

Along the neg string, across R218 we have .13V, base of Q5 is .3V, and Q206 base is 70V from the negative rail.

(Q205 takes this iteration's Wildly Out-Of-Spec Voltage award, in a surprise victory over the plus string)

Last iteration's winner, the voltage difference between the bases of Q206 and Q207, is now down to a relatively normal 1.8V.

The voltage across the bias diode string is -1.8V, with the drop evenly divided amongst the 4 diodes.

[Enzo]

If there is 24VDC on the output not loaded, there is no reason to slap a load on it, all that will do is stress things.


If R218 burnt, then Q5 is immediately suspect.

OK, base of Q206 is 70v pos from neg rail, whuch at 45v rail means it sits at about +25. That is fairly consistent with the +29 or whatever on Q207. So the whole bias string is slammed up towards +. So either Q203 is goofy or there is trouble down in R215,216 or C205 is shorted. What is at the base of Q203? IS C205 OK? what voltage is at each end of R215, 216?