For fans of the Sunn Coliseum Lead thread

[Euthymia]

Not to worry if you thought it was all over. Just got a phone call from the guy who picked it up.

"I plugged it in at our practice space, and all that came out of it was a loud hum, then smoke, so I turned it off."

Guess I'll be going in search of another "aha."

Of course, after I got it working this last time, I kept it on the bench, running full-tilt into my 1000W dummy load for at least 15 minutes, just to make sure that the repair was a success.

And of course, when I asked him about the cabinet he was plugging it into, he said that as soon as smoke came out, he plugged his other amp into it and finished the practice session.

[Euthymia]

Okay, the Sunn is back in the shop.

Plugged it in through the lightbulb limiter and got a dim bulb. So far, so good.

Connected the dummy load, and got a bright bulb. Not so good.

Here's a photo of the famous driver board:



Whatever happened took out half of the output transistors, 2 on the high side and 1 on the low side.

[Enzo]

Who knows.

One thing I'd offer. When you plugged it in with the bulb and got a dim bulb, the next step is NOT to add a load. The next step is apply your voltmeter to the output jacks and look for DC there. if there is DC, then don't connect the load, all that will do is draw a lot of current.

What impedance cab did he use, and is the amp rated for it?

[Euthymia]

Before I report further, I'll say that I have it back up and running and making sound, albeit with only one output transistor on each string plugged in. Gotta order the other replacement transistors.

To get it working, I first replaced everything burnt (R213, R214, R220, R221) or exploded (Q205). Then tested all transistors (Q206 and Q207 were shot).

I also replaced CR207-210 for good measure. CR207 was open.

So where we are now is the amp works fine and doesn't burn up, but I am not getting expected behavior from the bias adjustment: it varies the voltage between the bases of Q206 and Q207 just fine, from over 3V to under 1V, but I'm not seeing much change in global current draw as a result.

After it's all back together, I'm going to have the guy bring his speaker cabinet here to the shop to play through before he takes it away.

[dwhutchens]

I seem to remember something about powering on or off (don't remember which) on some solid state amps years ago would cause a loud pop that would take out the output transistors. Leaving the volume up was a no-no. I blew up a PA amp like this once. I think it was a Peavey. But then again my memory aint what it used to be.
David

[Enzo]

Cr207 is part of the current limiters, and it is there to help prevent blowups.

Q204,205 can be disabled for testing, but they must be there for operating the amp in the real world.

The output current is sampled by the voltage drop across R9 (which is also the current path for Q1 and Q207 as well as Q2). When it gets large enough, it will turn on Q204, which through CR207 will shunt any further drive away from Q207. The negative side works the same in mirror fashion.

SInce CR207 burnt, I'd replace Q204 as well, and if Q205 shorted, I'd replace CR210 - which you did.

Over by the output, make sure L1 is OK, and also check R14 for opens.

[Euthymia]

SInce CR207 burnt, I'd replace Q204 as well....

Sorry, I did do that, neglected to put it on the list.

Over by the output, make sure L1 is OK, and also check R14 for opens.

I get a DC short across L1 with an ohmmeter. R14 measures 21 ohms. Anything else I should do to test L1?

Poking around further, I find that couple of the emitter resistors on the + string opened up since the last time I checked them. They have now been replaced.

Still can't get the bias pot to affect the global current draw at idle.

[Euthymia]

The question is now "how do I find a replacement for L1?"

I removed it for testing on my impedance bridge, and noticed that the thin wire that goes through the center of the ferrite is broken. The big copper coil on the outside is intact, and I guess that's what's testing zero ohms, but the straight-through center conductor is broken.

The schematic calls it "8.8mHy coil," but I have no idea what other specs it might need for replacement, like current capacity, etc.

What effect would a problem with L1 have?

[Enzo]

Usually the coil itself is all there is. Didn't know af any secondary stuff inside it...Sometimes the parallel resistor is in the middle of the thing. All I was concerned with was whether or not the coil wire was broken. Sometimes they break from vibration or crack their solder. I never worry about the resistor openeing, just measuring across it lets me know if the coil is there or not. If you get the 21 ohms, then the coil is open. Or removed, of course.

If the coil opens, then the resistor has to handle all the output current. And if the coil break repeatedly touches and untouches, there can be inductive kicks.

[Euthymia]

Usually the coil itself is all there is. Didn't know af any secondary stuff inside it...Sometimes the parallel resistor is in the middle of the thing.

All I was concerned with was whether or not the coil wire was broken. Sometimes they break from vibration or crack their solder. I never worry about the resistor openeing, just measuring across it lets me know if the coil is there or not. If you get the 21 ohms, then the coil is open. Or removed, of course.

If the coil opens, then the resistor has to handle all the output current. And if the coil break repeatedly touches and untouches, there can be inductive kicks.

The resistor is a standard sand rectangle in parallel with the coil.

The coil is large gauge copper wound around a ferrite, with a thinner wire soldered in parallel, apparently running through the center of the ferrite. This thinner wire has broken off too close to re-solder.

I need to source another L1 and have no idea what specs to look for except the 8.8mHy.

Also still don't know why the bias didn't affect the global current draw.

[Enzo]

Usually the inductor is simply a coil. If they added a ferrite bead in parallel, it might have been for some sort of very high freq attenuation, but I don't think it adds measurably to the inductance when compared to the coil. No reason you can't sub most any similar sized FB in there.

If you are concerned over the heavy wire coil, I don't think yuo need to be, but if you want to replace it, just get a good idea what gauge the wire is that it is wound from and chase it down that way. You don't then need current specs. Alternatively, you know what the power output is into a specified load, so you can determine current through the load - and thus the coil - using Ohm's law. Then double it or more for safety margin.

Parts Express has a great selection of crossover parts, including coils. Heavy ones even.

MArtin Sound has some too. DOn't know that you will find 8.8 or whatever, I think they tend to come in whole digit steps. But since you have a way to measure, I have done this in t he past. If I need 3.6, I buy a 4 and unwind some turns until it reads what I want.


The bias thing? The bias adjustment basically determines the voltage difference between bases of Q206, Q207. WHich of course is the same as the voltage across CR203-206. Monitor that voltage and turn the control a little. If nothing changes, look for an open pot or R211. An open there would make the amp run too hot.

If both the pot AND a diode there opened, then Q206,207 would be free to slam against the rails and turn both sides on hard. That would at least blow a fuse if not some outputs.

[Euthymia]

Usually the inductor is simply a coil.

Here are a couple of photos of the removed part:

Parts Express has a great selection of crossover parts, including coils. Heavy ones even.

Ah, great idea, Parts Express caters to speaker builders.

The bias thing? The bias adjustment basically determines the voltage difference between bases of Q206, Q207.

The bias pot does vary the voltage between the bases of Q206 and Q207, but doing that does not affect the amount of current the amp draws from the mains. Therefore, I can't set the bias via the Enzo method of monitoring the draw from the wall while turning the pot.

[Enzo]

Oh, I see. I don't think that is a ferrite bead, I think it is a resistor. Many coil makers use a resistor as a coil form. The coil doesn't care if there is some parallel resistance, any more than your speaker would care if you put a 400 ohm resistor in parellel with it.

SNip the lead off the other end and slide it out. See if it doesn't have color bands. or for that matter, get a meter lead on the good end and stick your other probe on the broken off stub. Measure the resistance through the part with the coil no longer in parallel. A wire through a bead would still have zero resistance. A resistor would have some... resistance. The value is not critical. It is even possible they used a "real" coil form - a resistor body without the resistive element inside - an open. SCratch at it, it looks brown to me, not ferrite color.

And you could easily duplicate that coil by hand. Not all that many turns, and you can find a form of proper size.

[Euthymia]

Oh, I see. I don't think that is a ferrite bead, I think it is a resistor. Many coil makers use a resistor as a coil form. The coil doesn't care if there is some parallel resistance, any more than your speaker would care if you put a 400 ohm resistor in parellel with it.

SNip the lead off the other end and slide it out. See if it doesn't have color bands. or for that matter, get a meter lead on the good end and stick your other probe on the broken off stub. Measure the resistance through the part with the coil no longer in parallel. A wire through a bead would still have zero resistance. A resistor would have some... resistance. The value is not critical. It is even possible they used a "real" coil form - a resistor body without the resistive element inside - an open. SCratch at it, it looks brown to me, not ferrite color.

Slid it out, and it still looks like a ferrite to me. The wires coming out of the ends show infinite resistance. Scraped it, and it behaves like a ferrite. I think the photo showed some superficial brown crud on the ends.

It looked like they went to a lot of trouble to solder those wires to the coil wire, but that may have just been to hold it in place.

The coil itself is fine, it's just that ferrite/resistor/form thing that has its wires busted off. I guess I can put it back in.

Any advice on how to adjust the bias? R225 varies the voltage between the bases of Q206 and Q207 as it should, but there is no corresponding change in how much current the amp draws from the wall.

[Enzo]

If it is ferrite, I stand corrected. FInd one of that dimension and slip it in if you like. But if it is open through the center, then it sounds more like a coil form to me. A ferrite bead is around a wire. This sounds more like a coil forming center slug, and the wires are there for mounting and support. In that case a drop of glue would work as well as a wire end.

What voltage range is that differential between those bases? If it doesn't alter the mains draw, then either something is funnny or the bias range is not enough to get both sides conducting. What is the mains draw then too?