peavey 300ch

[woodyc]

I thought this was a curious failure. (Its an old monitor amp.)

Here's the schematic and the condition when it came in...

http://music-electronics-forum.com/a...0ch_module.pdf

U1 burnt, all leads open
CR4 short
CR11 short
Q6 bc short
R12 open
Q3 bce short, base lead burnt off
R15 open
Q7 bc short
Q8 bce short
R17 open
R23 open
R25 open

All the fuses were fine. The opamp looked pretty funny...



I think it only took one failed part to do all this, namely the bc short in Q7?
If anybody has another opinion about the cause, I'd love to hear it.

[Jazz P Bass]

I'll bet the unit was powered on a number of times in the failed state.

[Enzo]

cascade failure, yes. But it is a losing battle trying to decide which part started it. Like trying to determine which rock started an avalanche. And the problem is, even if we can nail it down to a certain part that went first, then what? Unless the same system fails over and over everywhere, we can't decide there is some design weakness.

If I had to come up with a story for it, it would probably go like this...

Q8 fails shorted. That means Q7 tries to carry all the current, and it too fails. The heavy current through those causes R17 to burn open. That leaves R15 as the current path and it too burns up. This drags a lot of current through limiter Q6, burning it and R12 up. All this is causing excess currents all over, so Q9 burns out R25. R23 has no choice but to burn out now. Over current now burns up plus side driver Q3. SHorted Q3 and Q7 leave 90v across CR8,10,11. CR11 fails short. And shorted CR11 puts -45v on pin 1 of the IC U1. Blows the top off it. The little pile of slag that used to be U1 puts that -45v with nothing to limit current onto zener CR4, so it burns.

Works for me.

Here is a technique I believe in. Working copies. I will either print a scratch copy of my schematic, or shove the original into a clear plastic "page protector" sleeve. Never write on your schematics unless it is information you will always want to have. I use a yellow china marker on the pastic sheet, or a highlighter on my scratch copies. Now any part I have to replace, I color in yellow. (Just how you indicate it doesn't matter) As I move along, the failure current path is revealed. There will be a big yellow path from V+ to V- or from one of those to ground, whatever. The value of this is what it reveals.

Look at Q3, R14, R15, Q7. Nice straight path through there. You didn't find R14 open, but surely the same current went through it that went through the others. You have your current path with a gap in the middle at one part. I'd replace it. Q6 burnt up, but the same current went through CR14. I;d replace that. It may test OK, but it surely was stressed. I think another filure current path was Q7, CR11, U1, but just to be safe, I;d spend the couple extra cents and replace CR8 CR10, in case they got hammered by the dead Q3. I think U1 killed CR4, but once it shorted, it increased stress on its dropping resistor R20. (a little below CR11) Current doesn't skip over parts, so what we look for is a path through parts and see if we are overlooking a part that checks OK but was in the failure path.

Cascade failures are not uncommon in solid state amps.

And the fuses? The two fuses on the board are for the +/-16v rails and were not involved. That leaves the 5A mains fuse. Folks wonder why the fuse didn't blow when some resistor burns to a crisp. Think in terms of power. That 5A fuse at 120v represents 600 watts of power. A half watt resistor or even a 10 watt resistor will burn up WAAAAAAY before the amp draws 600 watts. You had a short on the V- side, but until one of the V+ side outputs shorted, we don;t have a rail to rail short, and that is about what it takes to draw 600 watts, and take the fuse.

[g1]

we don;t have a rail to rail short, and that is about what it takes to draw 600 watts, and take the fuse.

That reminds me of the old "crowbar" primitive protection schemes, where the protection "sensors" would trigger an SCR. The SCR was from the supply rail to ground so when it turned on it would cause the fuse to blow or circuit breaker to pop. Simple but effective.

[Enzo]

They still use crowbars. Peavey has them across most of their power amp outputs.

[woodyc]

...but just to be safe, I;d spend the couple extra cents and replace CR8 CR10, in case they got hammered by the dead Q3.

Yeah I agree. I replaced those.

The only reason I think it was Q7 rather than Q8 is that they are on the same rail and Q7 has the lower voltage rating. In this particular amp, the drivers were MJ243/253, which are only rated for 100V. At first glance it might look like the negative rail is only -45V, but it actually swings to -80 (it would go to -90 if not for Q6/R12/R16.) So there isn't all that much headroom for glitches.

I love it when wacky amps like this come in, especially when its one I haven't seen before.

[Jazz P Bass]

I love it when wacky amps like this come in, especially when its one I haven't seen before.

The 300CH design is basically a Peavey Output Section.
Tried & true, the basic design is used on a lot of Peavey's equipment.
Master it & that many other amps are easier to repair.

[Enzo]

None of the parts exceed their voltage, so I tend not to like that agrument. Current is what burns up these parts. Glitches are still limited to the power supply at hand.

[woodyc]

None of the parts exceed their voltage, so I tend not to like that agrument. Current is what burns up these parts. Glitches are still limited to the power supply at hand.

Respectfully disagree. Overvoltages can easily damage semiconductors junctions and the big currents come after. And the danger generally comes in on the power line or generator the amp is plugged into, not the supply. My point is just that with the particular parts in the amp I fixed, a driver could be taken out by an overvoltage 1/3 the size of what it would take to damage a power transistor.

[Enzo]

Yes, but it is all academic. You have your scenario, and I have mine. I imagine there are other reasonable scenarios. I have no reason to say yours is wrong, I am just happier with mine. With cascade failures like this, we will never know. It is an academic exercise, which is interesting, but not definitive.

The nature of a transistor is not that voltage ratings are brick walls. 100v OK, 101v it blows up. So I'd agree that it might be easier to over-voltage a 100v part than a 250v one, but we still have to get to that over voltage. A huge transient would have to get past the filter caps and all the rest of the circuit and still hit that part hard enough to kill it. That is a big transient.

Will too high a voltage damage a transistor? Yes, of course. It is just my feeling - right or wrong - that in these circuits, powerline glitches are not what is killing the output stages. I suspect an excess current did it. But you may be totally correct also.

[woodyc]

Yes, but it is all academic. You have your scenario, and I have mine. I imagine there are other reasonable scenarios. I have no reason to say yours is wrong, I am just happier with mine. With cascade failures like this, we will never know. It is an academic exercise, which is interesting, but not definitive...

Wow Enzo, thanks for cutting me some slack. Never saw that coming. In gratitude, I'm going to agree with you too. If somebody hooks up too many speakers to this amp, the output transistors are going to overheat. I know that didn't happen in this case, but its an equally plausible failure mode and it would almost certainly take out a power transistor first.

[Enzo]

It is tempting to put togethre a this then that cause and effect for something. But so often it is more complicated than that. Amps can be compromised but still seemingly function for weeks, months, years. Things no one would ever suspect damage an amp, but not enough to kill it, but like a cracked wheel on your car, it can soldier on for a long time then finally croak. We experience turning the corner into the driveway and our wheel falls off and wonder why the driveway killed our wheel. But in reality it was that pothole we took at 40mph last winter that killed the thing, it just took this long for the crack to finally set all the way through. Just so amplifiers. Bad cords can damage amps. "Oh sure, we had a bad cord, but that was LAST YEAR..." Yeah, but still can happen. Driving a speaker hard, and the plug comes halfway out of the cab, shorting the amp output. Plugging a speaker cord into the amp during setup on a dry winter day can subject the outputs to thousands of volts in static charge. Er, discharge I suppose. That spark you get at a doorknob on a winter day can be 10kv or more. Talk about overvolting a transistor.

And thunderstorms. DOn;t laugh. A LOUD nearby thunderclap sends a strong pressure wave through the air. It hits your speaker cones, and they move the voice coil, generating a voltage spike. The voltage generated need not be huge, probably isn;t. But when an amp is running, the outputs are strung between V+ and V-. SO slamming a positive voltage into that output brings the one side of the transistor closer to V+. But when power is off, now the supply rails are at zero, and the same positive pulse on the putput now appears at the transistor as a reverse voltage. In other words in a powered situation we might have +50v on the collector and zero at the speaker lead - emitter. Push +20v into the emitter and now there is only +30v across the part. COllector 20v positive with respect to emitter. Now with power off, zero on collector, the same +20v on the output puts the emitter 20v more positive than the collector, reverse voltage on the part.

Stuff like that happens.

Not only that, but long speaker wires act as loops, and the current of a nearby lightning strike can induce a current in the wires. Lightning doesn't have to hit us electrically, it can just act like a giant transformer with our wires. Lighting hits your wires and the whole booth gets burned out, but just strike outside and we can kill the amp but leave the mixers and stuff intact.

And those things don;t have to act instantly. I have to assure people it was nothing they did, most times. I all the time hear things like: "I changed channels with my boost bedal still on, and the amp blew. They should tell us you can't change channels with the boost pedal on."

I actually am interested in this stuff, not intending to be argumentative.

[Jazz P Bass]

A bit off topic, but not.
Back in the day I had a run in with fake Toshiba output transistors.(before I knew that they where or could be fakes out there)
The receiver was a Luxman.
All was well until the tape recorder (source material) hit the end of the tape & 'clicked' off.
That spike sent the transistors into flames.
Tried another pair of transistors (same fake ones) & just for grins & giggles I made the tape click off.
Flames again.
F this.
Installed OnSemi transistors & never looked back.