Ampeg B2

[Euthymia]

Gotta love those IRFP(9)240 based Crampegs; they just keep coming in.

One on the bench now is a fuse-blower (or lightbulb limiter lighter, whichever you prefer).

Only condition under which it won't light the bulb is if I depopulate either the high or low side (or both) of MOSFET's.

Soon as I have at least one in on each side, poof.

Whichever MOSFET I put in the high side heats up faster.

The MOSFET's test okay to the extent of my ability to test them with my Fluke set to "diode." They are not shorted.

I'm getting 2.5V at TP9 and TP10 (47R gate resistors), and the rails are down to 11V, but is that because of the light bulb doing its thing?

What to test next?

[Enzo]

By vacating one side at a time, you have neatly demonstrated that the problem is both sides are turned on at once.

So depopulate the MOSFETs. NOW what do the rest of the TPs read?

Verify all the gate resistors at 47 and not open, also verufy all the ballast resistors as .47 and not open since the MOSFETs are out.

if they are all turned on, look to the driver stage. Q4,13 are the drivers. They OK? 33 ohm resistors there OK? Both getting rail through D29,31?

is the bias string OK? Q5,6. If the amp was blown up, I would be replacing those four xstrs anyway. MPSA06 were 6 cents each last time I bought a hundred. How much labor should I put into testing the old ones? The drivers cost more than 6 cents, but are not expensive. When the outputs short, stress can fall on the drivers. MAke sure the four 13v gate limiting zeners are ok, or better yet replace them. How many cents are zeners? Speaking of limiters, how are the limit xstrs Q3,8, and all the stuff associated with them? When something like this blows, I usually verify EVERY resistor on the thing. Any semi exposed to the fault current path gets replaced. Remember the two bias xstrs have a collection of resistors around them.

Q2,17 are probably OK, but check them, and verify their TP levels. is the +/-16VDC for them OK?

Make sure the spkr relay has not welded itself together as happened here recently to someone.

[Euthymia]

Ampeg B2

By vacating one side at a time, you have neatly demonstrated that the problem is both sides are turned on at once.

Okay.

So depopulate the MOSFETs. NOW what do the rest of the TPs read?

9 and 10 read 13.5 and -13.3, respectively.

7 and 8 read 5.5 and -5.5.

Verify all the gate resistors at 47 and not open, also verufy all the ballast resistors as .47 and not open since the MOSFETs are out.

All good.

if they are all turned on, look to the driver stage. Q4,13 are the drivers. They OK?

Will do.

33 ohm resistors there OK? Both getting rail through D29,31?

Yes.

is the bias string OK? Q5,6.

Q5,6, and R27, 28, 29, 30 test good.

R78 and 79 are also good.

If the amp was blown up, I would be replacing those four xstrs anyway. MPSA06 were 6 cents each last time I bought a hundred. How much labor should I put into testing the old ones?

I have this tres nifty Hickock semiconductor tester, so it's easier than even just doing a DMM diode test.

The drivers cost more than 6 cents, but are not expensive. When the outputs short, stress can fall on the drivers. MAke sure the four 13v gate limiting zeners are ok, or better yet replace them. How many cents are zeners?

The issue is not so much the cost of one zener as it is the cost of obtaining and stocking them. It's an hour round trip to the closest source of zeners. Same with the 2N5415 and 2N3440. If I could pick them off a tree for $5 each, sure it would be smarter than testing them, but I either have to drive an hour to get them or place an order and pay shipping, etc.

Speaking of limiters, how are the limit xstrs Q3,8, and all the stuff associated with them? When something like this blows, I usually verify EVERY resistor on the thing. Any semi exposed to the fault current path gets replaced.

I'll test Q3 and 8. Thing is, the fuse seems to have actually protected the circuit fairly well. None of the MOSFET's are shorted. I usually associate "blown output stage" with the death by shorting of at least one of the megaMOSFET's.

Remember the two bias xstrs have a collection of resistors around them.

They all test good.

Q2,17 are probably OK, but check them, and verify their TP levels. is the +/-16VDC for them OK?

They are seeing -53 on the collector of Q2 and -56 on the collector of Q17. Is that a clue?

Make sure the spkr relay has not welded itself together as happened here recently to someone.

It seems to be working correctly. I'm tuned in to the treachery of the B2's speaker relay.

Thanks for the help, as usual. Any ideas from what I've told you?

[Euthymia]

Moron this B2

Q4 and Q13 test good out of circuit.

I decided to probe the junction of R22 and R23 for DC and got -7V.

Going upstream to TP6 (pin 7 of opamp IC4) shows -14. According to the chart, there could be "some offset" at that point. But how much is too much?

[Enzo]

TP7,8 sound right - 36v across three 4.7k nominally so about 12v each. close enough.

I have xstr testers too, but they don't put the real world loads on things nor real world voltages. Like checking a 450v cap with an ohm meter, it is only part of the story. I like to replace xstrs exposed to failure currents because who knows what stress they faced, and are they weakened and ready to fail later?

Once I buy 100 of something, they last a while. Jameco gets $2.80 for 100 MPSA06 for example and 3 cents for a zener. And while ordering parts a coiple at a time eats you up on shipping, a bulk order spreads it out. Mouser has no minimum, and Jameco has none if ordered online. I keep a full set of zeners in stock, so 15 and 16v zeners are always here. 13 was an odd voltage, but not any more. Mouser wants 59 cents for the 3440 and 34 cents for the 5415. They do on the other hand want $7 for 100 MPSA06. But zeners are under a dime, and a nickel for 100 lot. I usually figure $6-7 shipping for a light bag of semis.

We mentioned all the four resistors in the bias string, but is the pot OK too? Does it vary the TP9,10 voltages? It should.

Q2,17 sound OK. TP on the bases sounds OK and rail on the collectors is right.

IC 4 does NOT sound right. SOme offset usually does not refer to pretty much the whole rail voltage. I'd sub it. If you have no 5532, any dual op amp will work at least for testing and stabilizing the amp.

I could be wrong, but your 13v TP9,10 sound high to me, especially when the TP list calls for more like 5v. Which leads me back to the bias string as suspect.

And don't mix up the 240s and 9240s... Hey, I've done it.

[Euthymia]

3 paws stuck in the tar

I have xstr testers too, but they don't put the real world loads on things nor real world voltages. Like checking a 450v cap with an ohm meter, it is only part of the story. I like to replace xstrs exposed to failure currents because who knows what stress they faced, and are they weakened and ready to fail later?

Not that I disagree, but how often have you seen a transistor that tests okay on the DMM diode, yet fails when actually put in the circuit? My answer to that question is never, not once. I know it's theoretically possible, and I usually do as you do and replace jellybean stuff like MPSA06's and 2N3904's, but I've never actually had a transistor test good in the tester and not work in the amp.

For the things to test "good," including realistic gain, with the Hickock, it would have to be a pretty darned exotic failure mode.

Once I buy 100 of something, they last a while.

My fear is that if I ordered a bunch of these, then I'd never get another B2, 3 or 4 in for repair....hey wait a minute, that would be so totally worth it!

We mentioned all the four resistors in the bias string, but is the pot OK too? Does it vary the TP9,10 voltages? It should.

The pot tests good out of the circuit. How much should it vary that voltage?

IC 4 does NOT sound right. Some offset usually does not refer to pretty much the whole rail voltage. I'd sub it.

Removed, socketed, and subbed. -13V.

I could be wrong, but your 13v TP9,10 sound high to me, especially when the TP list calls for more like 5v. Which leads me back to the bias string as suspect.

So what can I do now to further investigate the bias string?

And don't mix up the 240s and 9240s... Hey, I've done it.

Thanks for the warning!

[Enzo]

I've never actually had a transistor test good in the tester and not work in the amp.

OK, it's not like its a thing with me, when they test good they generally are good. then again who among us hasn't replaced everything that seemed bad in an amp only to have it blow back up a short time later? I figure replacing the things exposed is cheap insurance. If you are confident with it, then thats OK with me.

hey wait a minute, that would be so totally worth it!

Y'see? Actually I don't recall just what therad, but somewhere around here I was talking about subbing xstrs with someone, and they are extremely flexible little things, so a drawer full of 2N4401 for example takes the place of not only that type, but also the 2N390x, and a host of others. I like that MPSA06 as a general purpose beast. it was my fave little xstr long before I saw a SS Ampeg. At 80v and 500ma it can do many things. I use it for little relay drivers in stuff I invent a lot. I'd use it in place of a 2N4401 for that matter, except I stock them too. I stock a lot more little xstrs than most folks need to. But again, it ain't a thing for me, just a suggestion.
Something liek 15 and 16v zeners really are common, a lot moreso than MPSA06.

The pot tests good out of the circuit. How much should it vary that voltage?

That's a great question, I don't know. I'd find out what it does and see if it makes sense. I just expect some action. My assumption - and I am willing to eb proven wrong, so take my advice with a grain of salt - is that there won't be much in the way of gate current on teh MOSFETs, so the voltages there ought to trim up fairly close.

Removed, socketed, and subbed. -13V.

Damn. Oh well. Looking at the IC4, it looked like it ought to sit fairly in the middle. But it does sample the output buss. With the MOSFETs out is there still some DC on the output bus? Where is it coming from if so. Pop IC4 out of its socket so its offset drive won't confuse things. And if it has 13v on its output, what are the input pins reading at the empty socket? Somethign is driving it to close to rail, wonder what. Even if the part itself is OK, it is not OK to have the output of it sitting where it does. I guess it is trying in its own way to compensate for the output not being what it should be.

So what can I do now to further investigate the bias string?

DOn't know, just my spider sense tingling when I look at the bias ckt.

WEre it in front of me now, I'd be in zen mode at this point and maybe hard pressed to explain why I look where I look. Figure out why IC4 is doing what it is doing.

[Steve Conner]

If this is the Ampeg circuit I think it is, and IC4 is the chip I think it is, its output will just peg to either rail until you have the whole power amp working again. It's part of the power amp's negative feedback loop.

[Enzo]

Makes sense I guess, it just looked like it ought to be stable with zero on the out.

[Euthymia]

Oh dear

Were it in front of me now, I'd be in zen mode at this point and maybe hard pressed to explain why I look where I look.

Maybe if I did sort of a Trial of Billy Jack vision quest with starvation, dehydration, and rattlesnake venom I'd gain the level of intuitive insight needed to fix it so an CrAmpeg B2 bass head doesn't lunch its main fuse as soon as it's turned on.

I'll leave it on the shelf for a while to see if the path comes to me in my sleep, then I guess it'll go back to the client and from there to the dumpster.

[Enzo]

If it needs a reference on that feedback from the output, just ground the thing. That should at least stop the DC from wandering, at least in the short term.

[Steve Conner]

The reason why the output of that chip pegs to a rail (and will wander even if you ground the amp's output) is that it's set up as an integrator with almost infinite gain at DC, and that like all op-amps it has a slight DC offset error at its input.

If you want to kick it out of the way to help troubleshoot the rest of the power amp, you could just pop the chip out of the socket that you handily installed and jumper the socket hole where the output pin used to be to ground.

If you're anything like me, it'll cause a load of mental pain trying to fix this thing, but when you do it'll give you a kick to know you won