Vexing SM-400 (old TO-3 version) power amp problem

[asfi]

Hello all, this is my first post despite lurking here for a few months now, so obligatory capsule introduction: I'm an old hand with electronics though pretty new to solid-state amps. I've got a reasonably solid grasp of transistor theory but getting a blown-up power amp back into the saddle is still hit-or-miss for me.

So I'm begging the forum's collective indulgence for help with my homework: an old SWR SM-400 that's on my bench today.

It came to me in terrible, terrible shape. Both amplifier modules blown, filter caps blown and half the chassis covered in remnants of corrosive fluff from the caps. Repairing that damage was time-consuming but rewarding: this beast at least looks like a bass amp once again, the power supply's alive again, and all preamp functions have been restored.

I've rebuilt one of the power amp modules and now I've got trouble. I'm powering the chassis with a series 100W bulb for current limiting, and the rails are at +/- 43.3V. Output transistors are MJ21194G's purchased recently from Digi-Key (it seemed much less risky than buying 2SD424's on eBay). I replaced all of the other semiconductors as well, plus new electrolytics and several new resistors. I started with very nearly a bare PCB!

Now, the measurements:

The output is pegged just 0.5V below the +ve rail, and yet none of the transistors seem to be blown. Referring to the attached schematic, Q14-C, Q7-B, Q4-B, Q4-E, and Q3-C are all at 43.3 V.

Q9-B/Q10-B are 42.8V, as is Q9-E and Q11-C. Ditto for Q5-B and Q6-B.

So the high side drivers seem to be fully on, bottom side drivers definitely off, and I'm baffled why.

Any suggestions?

[Jazz P Bass]

Pull Q14 & see what you get at the base of Q7.
That 390pf cap is a suspect also.
Q14 is a PNP transistor.
If the base & the emitter are both high then the collector should be off.

[52 Bill]

If you've completely checked every possible thing and the limiter still lights up, you may need to take a leap of faith with this amp. On a couple of repairs of SWR amps, I found that the amps won't power up correctly with the light bulb limiter in the circuit. I attributed it to the start up bias circuit. If you are sure that there are no shorted devices and that all of the parts that you replaced are correct and in the right places, then you might have to go for it and power up without the limiter in place. I know that it puts a lot of parts and time at risk, but it may just be the limiter that's causing your problem. If you do go for it make sure to power up without a load and check immediately for dc on the output.

Do you have a Variac? Maybe you can bring it up slowly with the Variac and see what happens.

[woodyc]

The balance in that amp is controlled by the diff amp, i.e. if Q2 or the 27K was open, you'd get an imbalance where Q7 is on and Q8 is off.

But thats just one possibility. If you measure all the voltages, its usually obvious where the problem is.

[Steve Conner]

If you substituted transistors, do the new ones have the same pinout as the old ones?

What voltage do you get on Q3 base? Q1 base? Q2 base? Q1/Q2 emitters? the 18k tail resistor? between collector and emitter of Q4?

Check the 270 ohm resistor that supplies negative voltage to the bottom of the input stage.

Is the limiter bulb dim or bright?

[asfi]

Q14 is conducting. Q13 is saturated, a fraction of a volt across it.

With Q14 pulled, Q7-B is initially negative as the PSU filter caps charge, but stabilizes at roughly 0V. This is as I would expect, without any current path.

FWIW, I replaced all the ceramic caps as part of the rebuild. The old ones had an ugly discoloured look, probably just soot from combusting resistors but so cheap and easy to replace.....

[MarkusBass]

I rebuild this amp after both power amps were set on fire.

If you remove the output transistors, forget about the short circuit protection and the muting circuit, you end up with 6 transistors and few diodes - this is very simple amp. I suggest to remove the output transistors and check what is wrong with the rest of the amp.

Mark

[Mark Black]

All good troubleshooting info so far.

I would second Steve Connor's suggestion to check the 270 ohm resistor on the negative rail (labeled as a 1W in the schematic). I would further suggest to check the 2.7K 1/2W resistor which is tied to the 2nd input transistor collector. I have found both of these drifted quite high on these amps when they exhibit instability.

The later revision schematic #1011 has these doubled in wattage, and I believe even that may be too conservative. Things are fine with just the DC voltages, but when the amp is driven to full output the AC component impressed across those resistors causes much higher dissipation needs.

If memory serves I ended up with 2W for the 2.7K and 5W for the 270 ohm to make things happier.

Best of luck!

[asfi]

HEAD-BENCH!

Even though I'd replaced nearly all of the chassis wiring, I'd missed a broken signal ground connection, and that meant Q1-B was basically floating. Get that 68k resistor tied to GND like it's supposed to and.....success! At the moment I'm putting nearly 100W into a dummy load and all is well.

Thanks, everyone!

[Jazz P Bass]

With Q14 pulled, Q7-B is initially negative as the PSU filter caps charge, but stabilizes at roughly 0V. This is as I would expect, without any current path.

I would verify the transistor pinout of Q14.(and that it is indeed a PNP)
Like I said, that is the current path to Q7.
None other.
And being a PNP, with the constant voltages on the base & the emitter, the collector has to be off.
No?

[asfi]

Hahahaha, my "it's working now" post was premature. As part of the rebuild I replaced the 270 ohm resistor with a 3W part, but replaced the 2k7 one with the original 1/2 rating. Well, the latter already looks pretty burnt, so a 2W upgrade it is.

Thanks once again.....

[Mark Black]

Glad to hear it's nearly mended.

Measure the AC voltage & compute the wattage across the 270 ohm at full output - if I remember correctly I ended up at 3.7W or so, hence the 5W replacement. Also measure across the 2.7K (don't remember what I got across the that one and can't find my notes - aargh!). It might end up over 2W or too close for comfort - and I honestly can't remember whether it was a 2W or 3W part I replaced it with...

[asfi]

Ah, OK. 5W and 3W it is, then....so easy to do. Looking at the newer schematic I see that the 180 ohm resistor's also underpowered at 1W, so that's getting an upgrade too. Good thing it's a pretty roomy PCB (unlike, say, the HA3500).

[asfi]

If you remove the output transistors, forget about the short circuit protection and the muting circuit, you end up with 6 transistors and few diodes - this is very simple amp. I suggest to remove the output transistors and check what is wrong with the rest of the amp.

Mark

Not that I strictly need to know at this point, but in the interests of getting over my noobishness, is the short circuit protection and the muting circuit the portion involving Q5 and Q6?

[52 Bill]

Not that I strictly need to know at this point, but in the interests of getting over my noobishness, is the short circuit protection and the muting circuit the portion involving Q5 and Q6?

Q5 and Q6 are the protection circuit and Q13 and Q14 are the "mute" circuit.