ARCHIVED FORUM -- March 2012 to February 2022READ ONLY FORUM
This is the second Archived Forum which was active between 1st March 2012 and 23rd February 2022
A few weeks ago I replaced all the electrolytics (except the big 10000uF ones and the ones on board 2). I also replaced the output transistors because I had some lying around and the original ones presumably have been running hot for quite some time. Obviously I replaced the trimmers and adjusted according to manual.
The speaker ran just fine for about two weeks. We went away for a couple of days, during wich time I unplugged the speakers. When we got home I plugged them back into mains and we had them running for maybe an hour at moderate volume. The it just switched off and the LED went orange.
Now, it's not because of heat that's for sure. Unless of course the thermistor is broken in some way. So I guess it's drawing excessive current for some reason. It's a bit difficult to measure things since the protection circuit puts the amp in standby as soon as it turns on. When plugged into the mains it goes into standby as normal with the red LED. But as soon as it detects a signal and turns on it immediately switches off and turns orange.
Any ideas on what might be the culprit? Where do I start looking?
Thanks!/ Johan
I did some poking around and found TR25 to be a dead short. Still have to check the negative side, and see if I can find the cause of the failure. I suppose that if it's only that transistor it could just be that it was a bad one, or maybe bad thermal coupling but that seems unlikely.
Anyone have any ideas?
Did you check the soldering around TR23 and TR28?
Offsettrimmer replaced /adjusted?
I check every time the mainboard with a loupe. But in this case it can be a trace which is broken caused by replacing TR25-27 and TR31-33?
I plugged it in after replacing TR25. It now powers up like it should but I've not supplied a signal or hooked the amp up to the speaker.
Beobuddy: Did you check the soldering around TR23 and TR28?
No I have not.
Beobuddy: Offsettrimmer replaced /adjusted?
Yes.
Beobuddy: I check every time the mainboard with a loupe. But in this case it can be a trace which is broken caused by replacing TR25-27 and TR31-33?
Ok, I've found a weirdness. When measuring the idle current across R103/R104 it increases. And it does so pretty quickly. Adjusting the trimmer brings it back down, but it increases again.
But I should also add that when the speaker first stopped working the amp was not really warm to the touch, so I'm not sure what's going on here.
/ Johan
Update: Short version, there were fireworks and smoke.
Long version, as I described earlier the amp sort of worked, but the idle current was increasing in a worrying manner. In order to be able to get some useful measurements I plugged it in very briefly a few times. Then the inevitable obviously happened. There were sparks. The narrow trace going from the +68V rail to the collector of TR23 was burnt off in one place. Obviously there was a lot of current running through TR23. Now, the transistor itself measures fine (so does TR28) so I can only assume that there's something at the base of TR23 that is turning it on full blast. I've been poking around a bit to try and find the fault but I'm really stuck right now.
Suggestions anyone? I could really use some support, if nothing else moral support. :-)
Johan: Update: Short version, there were fireworks and smoke. Long version, as I described earlier the amp sort of worked, but the idle current was increasing in a worrying manner. In order to be able to get some useful measurements I plugged it in very briefly a few times. Then the inevitable obviously happened. There were sparks. The narrow trace going from the +68V rail to the collector of TR23 was burnt off in one place. Obviously there was a lot of current running through TR23. Now, the transistor itself measures fine (so does TR28) so I can only assume that there's something at the base of TR23 that is turning it on full blast. I've been poking around a bit to try and find the fault but I'm really stuck right now. Suggestions anyone? I could really use some support, if nothing else moral support. :-) / Johan
It sure sucks when you try to improve something and in the end it gets worse (this is why despite being for certain in the minority, I don't touch my gear, even the old BM8000 until something actually breaks... I'm sure I'll have my own light show soon). Moral support for you as hopefully it will get resolved soon. The Pentas are terrific. Just keep remembering that beautiful sound as you get through it.
It sure does. Thanks! :-)
Did you replace the idle current trimmers or did you just adjust them ?Do you still have the original output stage transistors ?
Martin
That's shitty! Sorry to read about your progress.....
Did you replace both trimmers?
Check R91 and the section around TR10-TR11. If the offset goes wrong, same thing happen and causes malfuction.
Úse the light bulb method if you don't have a variac.
Check thorougly on broken traces. Sometimes traces can break next to a solderjoint.
@Martin: I did replace both trimmers. I do have the original transistors.
@Beobuddy: I will check the suggested areas. You mean that this could be the cause of the drifting idle current?
I suppose it's silly to think that solely turning TR23 on fully would blow a trace in under a second... There is probably a short somewhere in the negative side, no? I will go through and test all output transistors once more and check once more for broken traces. I have a few more transistors spare and the original ones.
Can't do anything til tonight though.
Thanks a lot for your input. Much appreciated!
Sadly I must report that TR23-27 and TR29-32 measure ok. So do R91 and TR10 and TR11.
I also could not find any broken traces around any of these components. I will continue the search for broken traces.
What should I look at next?
Cheers!
I guess my real question is: what am I looking for?
To recap: The amp stopped working with an orange light. I found TR25 (NPN) to be shorted. After replacing TR25 I found that the idle current was drifting fairly quickly. Now, I suppose that this could have been the case before the amp blew and thus the cause of it blowing, but I don't know that as I didn't see it drift when I adjusted the trimmer the first time (maybe I was just being too eager to get it done.. not sure). While having the drifting idle current I plugged in and turned on the amp a few times briefly to try and get some measurements. Then it blew with a burnt trace from +68V to TR23's collector.
What may have caused that flow of current? My guess is a short somewhere on the bottom of the schematics (negative, voltage-wise).
What may cause the drifting idle current, as the suggested culprits seem to measure fine.
Before I was focusing my attention to the positive side, but now I'm more thinking the fault is on the negative side. Am I right in doing so?
Thanks!
That's right. Don't know how to put it exactly. But you need the negative part to limit the current through TR23. But beside that, if the offset goes up (did you check?) the section around TR10-TR11 isn't working properly and will influence the idle current also.
Beobuddy: That's right. Don't know how to put it exactly. But you need the negative part to limit the current through TR23. But beside that, if the offset goes up (did you check?) the section around TR10-TR11 isn't working properly and will influence the idle current also.
Right, so I guess the search continues then.
I did measure TR10 and TR11 and they seem fine. I will check the passives around there too, hopefully today.
Ok, update time. I had no time to work on the amp this weekend, but had a couple of hours today.
I looked over the board with a loupe but could not find any broken traces. I reflowed pretty much all joints that looked even vaguely suspicious.
I measured all the output transistors again, after which I soldered them back in (but not mounting them on the sink).
I'm happy to report that I'm back to about where I was before the fireworks. The amp starts up and the voltage over R103/R104 increases steadily and pretty quickly.
I also measured the offset on the output of the amp, and it is stable at about 60 mV, so that shouldn't be the issue, right?
I checked the passives around TR10-11 and they're all fine.
Now, when I plug the amp into the mains it goes into standby. I can turn it on and the light goes green. If I then pull the mains and plug it in after not so long, the light goes orange. If I wait longer the normal behavior is back again. Sounds cap related, no? Well, I mean that there is a cap charged up and it discharges and somehow engages the protection circuit comparator. Does this sound at all reasonable?
Then of course I'm still struggling with the increasing voltage over R103/R104.. no idea what to do next?
Ok, so I really need some guidance now.
Today I soldered in the original output transistors (all the sink-mounted ones TR24-27, 29-32, with no observable difference in behavior, so I can only draw the conclusion that the fault is not at those transistors.
I've had the pcb free from the chassis for some time for easier access and consequently have not had the transistors sink-mounted. I thought that it'd be a good idea to actually mount them properly and get a feeling for what role heat is playing in this thing. Huge difference, is the answer.
The voltage across R103/104 is still increasing, but an order of magnitude slower. Before it was about a mV a second, and now about 100 uV/s.
Further, cooling down the sink drastically decreases the voltage, but of course it climbs again (with increasing temperature?)
Now, this may all be obvious to you (and kind of is to me too, hence the test) but what do you make of it? Is it bad thermal contact, or is it some failed component somewhere else?
I kind of would expect the voltage to climb with increasing temperature and then reach an equilibrium at some point where it stays, no? If so, is there an explanation for the initial failure? Was it just bad luck or is it the increasing idle current?
Could it be that something is oscillating? If so, where is it likely to be? I put my scope to the base of TR24 and TR29 but the voltage is stable as a rock (although slightly higher than in the schematics, about 1,3 volts as opposed to 1,1 in the manual.)
Could oscillation be the cause?
A long shot.
Since you changed the caps, there is a possibility (although very small) that one of the new once are faulty. Try to eliminate that possibility.
The was also a discussion lately about high & low ESR caps, can that be the reason here?
These are just wild guesses.
//Bo.A long list...
Thanks for the suggestion.
I guess it's worth checking them at this point. Although my other speaker is running fine, and has been for many months, and I used the same caps in that one.
I suppose one of them could be faulty, but since almost all of the electrolytic caps are filter caps to ground I guess it would manifest itself as increased ripple, something I have not seen. But checking C39 carefully might be interesting (previously only checked that it wasn't shorted.)
Ok, I replaced C39 (even though I couldn't find any fault on it) with a different type of cap. No difference at all.
Any other suggestions? Anyone?
Is my assumption sound, that failed caps elsewhere are unlikely to cause these symptoms, at least without showing up as ripple? Or is there no ripple without load? But then, why would I have this problem without load...?
Further suggestions are most welcome.
Johan:I also measured the offset on the output of the amp, and it is stable at about 60 mV, so that shouldn't be the issue, right?
I measured on one of mine and DC-offset started at Zero and stabilized at 7 mV after some minutes. Your 60 mV seems high.Why don't you adjust it to Zero?
Johan:1,3 volts as opposed to 1,1
Have you investigated this furhter?
I adjusted my amp, measured an got these values.
TR29 1.162VTR24 1.178VTR23 1.748VTR28 70.6VIdle current 29mVDC offset 0,2mV
Don't know if this is of any help but i think your values are a little to much off.
i can add to that :
Collector current throughTR23 is 9,9mA. If I remember correctly, indepent from the setting of R95 (current limiter)
Another thing; increasing R95 (higher value), gives a higher idle current.
And about IC1 (shown as PCB5) this IC1 limits the voltage accros the TR28 B-E. It's mounted on the heatsink to get a stable environment (current)
When the current through TR28 goes up, it will be limited by IC1 as it lowers the B-E voltage from TR28.
Sure you didn't get any of the new caps wrong polarized?
Tack Bo!
I've been away for a few days in the countryside without Internet connection (Midsommar, you know), that's why I haven't replied, sorry.
I did adjust the offset to near zero. My point was just that it was stable and not drifting.
Thanks for the measurements. I will get in there and compare to see what we get.
I'm pretty darn sure I got all caps in the right way, but I will check again.
Thanks Beobuddy!
Right. I will get in there and measure.
Not sure I'll have time today, but soon. I'll be back.
Cheers
Any progress?
BO: Any progress?
Yes! (I completely forgot to update you guys. Sorry!)
I have them up and running and have had for months now.
What happened was that I adjusted the idle current a bit low and monitored it for a long time, probably and hour or so. It did rise but slower and slower and then seemed to settle and stay quite stable. I adjusted it a bit more and then decided to leave it. It has been working fine ever since. The amp does not get warmer than the other one.
It probably wasn't the best way to go (not finding the actual culprit I mean) and I was a bit weary at the beginning but it works fine so...
Anyway, thanks for all the help! And again, I'm sorry I forgot to update.. I really think one should do that!
Happy ending