Sunn Amplifier Power Amp Puzzle

Hi All,

I was working on the Power Amp board of my Sunn Beta lead to find out why the -15vdc supply was only reading -13vdc.

Wondering whether IC-1 might be partially shorted and dragging down the voltage, I pulled it from its socket (see attached schematic) so I could power on the amp for 2 seconds and take a voltage measurement. To me, judging by the schematic, this seemed like a harmless enough move (note: I'm not an advanced tech, maybe a step above a novice).

Well, the second I turned the amp on to take the measurement, I got a horrendous hum from the speaker. I powered down and found Q19 blown.

My question is . . . why? What did removing IC-1 do to the circuit that caused the output transistor to short out? To me, IC-1 seems to be taking inputs for monitoring purposes only.

Thoughts?

Any thoughts/input/ideas appreciated,
Bud

Bud_J said:

My question is . . . why? What did removing IC-1 do to the circuit that caused the output transistor to short out? To me, IC-1 seems to be taking inputs for monitoring purposes only.

Click to expand...

You apparently only looked at the upper 1/2 of IC-1.
The lower 1/2 of IC-1 at the bottom of the page is an integrator in a feedback loop that maintains the DC bias at zero volts at the output.
Without that the bias goes open loop and the output likely slammed to one of the output rails which probably blew the transistor due to excess current.
You're lucky it didn't also blow your speaker.

Where does the -15v come from?

crutschow said:

You apparently only looked at the upper 1/2 of IC-1.
. . .
Where does the -15v come from?

Click to expand...

Thank you, Crutschow!

You're absolutely right. I did initially overlook the 2nd half of IC-1, but being kind of a newbie I still didn't understand why it blew the transistor. In analyzing it after, I didn't see it as part of the feedback loop. I'm learning a lot by trying to fix this amp. I put in a sacrificial output transistor and it blew right away on re-power-up, as I half expected it might. I had tested all the nearby diodes and transistors and had found no shorts. But the bias transistor looked a bit funky testing it in circuit, so I will pull that later and test it out of circuit.

Interestingly, I found Q15 with two cracked socket pins while looking things over. The socket had been loose, I noticed, in earlier tests, but I wonder if the over-voltage finally cracked those pins or if it's just coincidence. Oddly, the transistor itself reads fine -- hfe is within range, etc.

As for the -15vdc supply -- it's generated in the Q11 circuit with its associated Zener. I can "trick" the voltage up to -15 with three 1n400x diodes in series with the Zener and the amp behaves much better, Q11 and its power resistor run much cooler. Nice to know I can do that, but my main reason for working on this old amp it to learn electronics. So I removed the 3 1n400x diodes I had put there and went searching for the real problem a few days ago. That was when I committed the folly of removing IC and now causing problems on my PA board. I think I'll have the PA board back to where it was soon, but I'm still curious about what might be dragging down the -15vdc supply. Disconnecting the preamp board has no effect, so it's nothing there. That's why my eye landed on IC-1 (well, HALF of it!)

I have a LM1458 I can put it its place once I get the PA board back to where it was. Then I can test that -15 supply again and see if it's where it's supposed to be. I also wonder if the cracked socket of Q15 was causing it to read low?

But first I have to trace out why Q19 keeps blowing. I'll update as I go along. In the meantime I'm open to all thoughts/ideas/suggestions. I'm trying to learn!

Peace,
Bud

Bud_J said:

...........In analyzing it after, I didn't see it as part of the feedback loop.

Click to expand...

It's strictly DC negative feedback to stabilize the DC output voltage.
It supplements the AC/DC feedback going to the base of Q3.
Any deviation of the output voltage from zero, causes the op amp integrator output voltage to change the voltage at the amp input in a direction such as to bring the output voltage back to zero.

crutschow said:

It's strictly DC negative feedback to stabilize the DC output voltage.
It supplements the AC/DC feedback going to the base of Q3.
Any deviation of the output voltage from zero, causes the op amp integrator output voltage to change the voltage at the amp input in a direction such as to bring the output voltage back to zero.

Click to expand...

Thank you, Crutschow, for that explanation. I guess that is the meaning of the 0.v notation at both the output and input of that half of the op-amp. I had been wondering about that.

I must admit I get pretty disheartened when I blow something up, but I do always seem to learn something from it.

I'm going to go pull that bias transistor now and test it out of the circuit . . . along with a few other components that are suspect. I'll post an update afterward. Thank you for your help. I like learning!

Peace,
Bud

UPDATE:

I was hoping to find a bad diode or transistor (or something), but did not. I pulled CR4, CR10, Q2, Q3, Q6, Q13, Q15, Q16, and all the power transistors from the circuit. They all checked out fine. Couldn't find any bad caps or out-of-value resistors.

Since Q19 blew the 2nd time exactly as it blew the first time with the op-amp out of circuit, I'm thinking that when I put the op-amp back in the circuit (before I found Q19 was shorted) maybe this blew the op-amp?

I could put in a new op-amp, but I'd hate to blow Q19 again. I have several sacrificial output transistors, so that's not my concern. I'm more concerned that blowing it will take out other stuff. So how does one go about this? Is it possible to power-up this board without the power transistors in place to test voltages on the op-amp? What is best practice here? Or are techs sometimes forced to do a "smoke test" in certain situations?

Thoughts?

Thanks all,
Bud

Bud_J said:

Since Q19 blew the 2nd time exactly as it blew the first time with the op-amp out of circuit, I'm thinking that when I put the op-amp back in the circuit (before I found Q19 was shorted) maybe this blew the op-amp?

Click to expand...

It's possible that the op amp was damaged since there's no particular protection against input overvoltage.

Bud_J said:

Is it possible to power-up this board without the power transistors in place to test voltages on the op-amp? What is best practice here?

Click to expand...

Powering up a circuit without all the components is problematic, as you found out.
The way such failed electronics are often tested is with a Variac variable AC transformer.
That way you can slowly bring up the AC power while watching the circuit voltages and current, and hopefully find what is wrong before blowing something.

Do you know anyone who has one?

crutschow said:

The way such failed electronics are often tested is with a Variac variable AC transformer.
. . .
Do you know anyone who has one?

Click to expand...

Thank you, Crutschow. I don't know anybody with a variac, so I just picked one up off eBay. Not a brand name, but I don't expect to need it a lot, so I imagine it should serve my purposes. I got a 20A model, so at least I know it will handle anything I've been plugging into my household outlets. I guess I'm in a holding pattern till it arrives, if I want to do things properly. In the meantime maybe I'll find and build a simple circuit that tests op-amps. If I prove the op-amp to be blown, I'd feel much better about throwing power at this PA board again -- especially since I've tested every other semiconductor on the board.

I've been wanting to get a variac for awhile now. I guess I finally gave myself a reason to ;-)

I'll post back when I have more results.

I really appreciate all your input!

Peace,
Bud

Bud_J said:

I'll find and build a simple circuit that tests op-amps

Click to expand...

All you need is a simple follower circuit with the (-) connected to the output.
You then put a varying voltage into the (+) input.
The output should then follow the input over the op amp common-mode range, if it's working properly.
You can even do this test with just one power supply and ground.

crutschow said:

All you need is a simple follower circuit with the (-) connected to the output.

Click to expand...

Thank you, Crutschow. I breadboarded this one: https://www.diystompboxes.com/smfforum/index.php?topic=63394.0

I built it with a known working LM1458, then tested the op-amp that was on the PA board when the output transistor shorted. It tested perfectly with a 9V source. The numbers were dead-on.

At this point, the only known failed component when Q19 shorted the 2nd time is Q15, which I later found had two cracked socket pins. I've since soldered this directly to the board, but I don't think it was the problem, esp. since the 2nd failure of Q19 was exactly the same as the first, making me think something in that DC-feedback circuit is still not right. As I await the variac, I'll study the schematic and try to see what else might've been taken out. A bit stymied at this point . . .

Then it seems we await the arrival of your Variac.

crutschow said:

Then it seems we await the arrival of your Variac.

Click to expand...

ETA of 7/18. One thing I can do in the meantime is build a dummy load. I was lucky to have not lost my speaker. No sense pushing my luck. Will post back when my new toy arrives ;-)

Bud_J said:

ETA of 7/18. One thing I can do in the meantime is build a dummy load. I was lucky to have not lost my speaker. No sense pushing my luck. Will post back when my new toy arrives ;-)

Click to expand...

Actually it does no harm to test the amp without a load.
Tube amps shouldn't be operated without a load, but transistor amps can be.

crutschow said:

Actually it does no harm to test the amp without a load.
Tube amps shouldn't be operated without a load, but transistor amps can be.

Click to expand...

Thank you, Crutschow. I had wondered about this. In the early stages of testing, I found that if I did not connect the speaker, then the voltages would not come up. This, despite the fact the schematic says that all measurements should be taken without a load (or an input).

Then, later when I was testing things (this was all back when I was first checking the +/- 15vdc supplies), I seem to remember that the voltages *did* come up without a load. I should have paid closer attention to what the exact scenarios were. Could be that none of the voltages came up till I tricked the -15vdc supply (which I've since un-done). It's also possible I had something plugged into the input but not the output when the voltages re-appeared. I can't remember the timeline. I just remember being upset that I had to connect my speaker.

Of course now the amp is dead, so I can't test this. Once I get the PA board going again, I'll look more closely at this. Perhaps there's something on the inputs and/or outputs that's a safety of some sort, and maybe requires *something* to be connected? Equally possible the amp has other issues with it that I've yet to uncover . . .

I can think of no reason to require a load be connected before the amp comes up.
It should operate equally well with no-load or a load.

crutschow said:

I can think of no reason to require a load be connected before the amp comes up.
It should operate equally well with no-load or a load.

Click to expand...

Can you think of any fault condition that would cause that symptom? When it was happening, I also tried it with a very small dummy load (10w ceramic resistor @ 4-ohms) just to see. The amp would turn on, but the +/- 40vdc and the +/- 15vdc supplies would only read a few volts. But when I shut it off and connected the actual speaker then turned it back on, then all voltages returned. I had done this many times in a row because I was mystified by it and really didn't want to run the amp with a speaker.

Then, after awhile (wish I could remember exactly when) the voltages would come up whether a load was connected or not.

Perhaps the Q15 with the cracked socket leads was causing a weird issue? Needing x amount of current draw to overcome the weak connection (grasping at straws here!).

I would have taken better notes at the time, except I thought it was supposed to work that way. I saw it only as an annoyance rather than one of the amp's problems.

Thanks for all the input. I'm learning things just by talking it out.

Peace,
Bud

Bud_J said:

Of course now the amp is dead, so I can't test this. Once I get the PA board going again, I'll look more closely at this. Perhaps there's something on the inputs and/or outputs that's a safety of some sort, and maybe requires *something* to be connected? Equally possible the amp has other issues with it that I've yet to uncover . . .

Click to expand...

No, there's no such 'feature', and there would be no use whatsoever in such a feature.

I would suggest the reason is because you've blown the amp, and one of the faults is causing this particular action.

As someone who's repaired amps professionally for the last 40+ years you don't go round changing just one transistor

As the amps are DC coupled, one failure almost always causes multiple other device to fail, and often the original fault is intermittent - so you replace all the faulty parts you can find, and it blows again shortly afterwards. In your case the intermittent possibility isn't likely, as you appear to have caused the fault by removing a critical IC.

Nigel Goodwin said:

As someone who's repaired amps professionally for the last 40+ years you don't go round changing just one transistor

Click to expand...

Thank you, Nigel. There's definitely another faulty component because Q19 blew a 2nd time after I changed it. It's tricky for me to find because I can't find anything obviously shorted as Q19 was. Nor have I found anything open. So I'm awaiting my variac so I can hopefully at least confirm it's still something in the DC feedback circuit that is not working right.

As far as the transistors . . . I've been pulling and testing them. If the junctions read properly on my diode test setting, and the hfe is in spec with the data sheet when I put the transistor in the transistor tester, then is it safe to call that transistor "good?" Or do techs change even these ones out as a matter of course?

Thoughts?

Thanks for all your input,
Bud

Bud_J said:

As far as the transistors . . . I've been pulling and testing them. If the junctions read properly on my diode test setting, and the hfe is in spec with the data sheet when I put the transistor in the transistor tester, then is it safe to call that transistor "good?" Or do techs change even these ones out as a matter of course?

Click to expand...

Generally I used to change all transistors in the DC path, this results in very few returned repairs.

the circuit looks not too complex for simulation (especially when there're check points with -values listed ← enables fine tuning the components/discreets) ? much cheaper than playing around with real board

i found the IC1 1/2 implemented rising/falling signal/slope detector curious as the second time i see such ... as is **broken link removed** is also and this one here frequency band oriented but here one only seems to work at signal going away from SG e.g. ± 1-st & 3-rd quarter-cycle (i don't want to get down to philosophy of such) - looks like something implemented fast* without much thought to it . . . ?? is such* and the rest of the amp . . .