Marantz PM400 - diagnosing cause of noise in one channel

[Since1984]

Hi all,
I've got a Marantz Pm400 on my bench, nice amplifier. It works, one channel playing fine, the other has weird noise which almost gets dangerous to speakers - it's something like a cracking/almost a bit like fireworks, sometimes gets lower, then suddenly comes back again, never stops. Still, that channel plays music with normal volume, too. No overheating, it's not a hum or buzz (it's volume independent), it's not the power supply (as said, other channel is fine). Changed most electrolytics in the defective channel, no difference. Decided to start isolating the problem, so I disconnected the power amp section from the pre amp section and grounded it's input. As expected, still the same noise problem in the defective channel, while other good channel now dead silent, as it should be. Grabbed my oscilloscope and started to trace the noise and also compare with the other good channel at some strategical points.

Here I add an excerpt of the schematic of the defective channel (power amp part), where I marked green dots on places where there was no noise present and red dots, where it was present. The noise looks like some random "spikes" around some millivolts, quite intense, of varying frequency and amplitude, it's not white noise or something. It's more a crackling.



From what I understand, after doing some research, this will probably be caused by some transistor in one of the stages of the power amp which is failing and developed this noise, right? If it was a resistor or a capacitor, would be more like white noise and more consistent, lower amplitude, no?
Also, doesn't seem to be a broken solder joint or intermittent "short circuit", as I tapped around that section with a wooden tool quite thoroughly and nothing really changed.
Haven't done the freeze spray trick yet.

Now I'm reaching out to you guys, as I don't understand very well the design of the output stage, it's a little complex, I know the first stage is the diff amp and I know what the last stage does, obviously, but can't identify well the rest (don't know where the VAS is, for example), so not sure if maybe with some help understanding the circuit better, finding the culprit would be easier?

As you can see from the red dots, I'm pretty sure the noise appears somewhere after the stage of Q707 and Q709, as it is not present at the base of Q705 nor Q707, but is present at base of Q713 and Q711. It's also not present at the common point for differential stage (at Zener), but it is present directly at emitter of inverting diff amp transistor and its base, which is where the feedback arrives from the output power transistors (which obviously carries the noise), which is also why it is present at the emitters of QN01 and QN03. Am I thinking correctly here? Or could the noise go the other way round, could it be the output transistors and the feedback connection injects the noise signal in the whole circuit over inverting input of diff amp stage?

What would be best to do next? Please, some opinions on this.

a) Apply freeze spray (I only have a non-invertible compressed air can, I hope it does the job) to the transistors which are suspects, to see if something changes

b) Disconnect R723 and 725 (output transistor resistors), so that no feedback goes to inverting input of diff amp and see if that changes noise? Or would that make the amp instable or other problems during this brief testing?

c) Disconnect R709 (resistor to base of inverting diff amp transistor) Q703, so that feedback doesn't get there and start isolating this way one transistor after the other, working from left to right, interrupting signal temporarily to test?

d) is it possible to measure (in terms of voltage, ohms or current) a transistor which has developed this kind of noise problems? or, as it is still partially working (it stil puts the music through), measurements would all be normal?

e) lastly, should I simply start substituting transistors in pairs, if necessary, starting with Q709 and Q707, then QN01 and 03, then Q703, hoping to eliminate the problem by luck, so not to loose much time in tracing it? this is a paid task, but not per hour, obviously.


Thanks in advance. And sorry for low quality of schematic, it's the only service manual I can find.

 

[AE Bob]

It is most likely a noisy transistor in the differential pair (Q701 and Q703) at the input of the power amp. I would just replace them to start with. It is a very common issue in many units.You can spray the pair with freeze spray and I bet it will either cause the issue to get worse, or to almost go away. I would NEVER disconnect the feedback through R709, as the amp will surely go unstable without any feedback. You can sometimes find a transistor that is noisy by its leakage, but it is not likely possible with a multimeter.

I bet you a cup of coffee it is that differential pair Q701 and or Q703. You can try spraying them with freeze, and see if it effects the noise. I would also replace the same pair in the other channel as well.

 

[Since1984]

Thanks for your suggestion, AE Bob and your warning about the feedback, I had a feeling that it could be a bad idea, even only for a quick test. I will do that, first try the freeze trick, but probably just replace both of them! It's easy and in line with what I've read and come across when researching. I suppose that (just for temporary testing purposes) it won't hurt if I substitute both transistors with not-so-ideal "equivalent" replacements? Then, with the test done and problem identified I would order correct replacements.

 

[Since1984]

Again, thanks for the hint. It was spot on, I'd now happily buy you a cup of coffee, wasn't it for the distance ...

It took me quite some time, though, as (weird enough) there was only one defective transistor, Q701. And what is even more odd: in the first measurements I did, no noise was visible at its collector. But I hadn't measured its base. So, at the time, as I was measuring noise at the base of Q703, I (wrongly) deducted the defective transistor would be Q703 and exchanged it with the one from the other good channel. It didn't solve the problem, obviously. The noise only got there coming from the feedback loop. It was only when I decided to short the base of Q701 to ground (before, I had always shorted in front of the 470ohm resistor, now I shorted behind it) that I noticed that the noise went away, so I took the short out and hook my probe onto the base of the transistor (note that resistor was floating free, then there's 33k to GND), and there was the noise! I assumed it wouldn't come from the resistors, would be really unlikely, as there not of the old carbon composite kind, so I decided to swap Q701 and Q702, and bingo! Now the noise went to the right channel instead of the left! It's Q701. It has to be said, btw, that none of the four transistors of the differential pairs of the two channels are the ones indicated in the schematic, I don't know why. They are all the (infamous) 2SA750, known to become noisy (it's on the list in that one really useful sticky thread here on this forum). I can't understand why Maratnz would have decided to go for these instead of the 2sa1127 originally indicated in schematic. I guess this could be the result of some old repair done on this amp? I mean, some solder points look like they are not from the same time as others. Maybe the amp had some major failure once and someone decided to substitute all 4 diff pair transistors with A750, which was a really bad choice, if that's true...

So, I now have to look for the best equivalent transistor and order, say, 20 of them, so that I can match them? I've never done this before, how can this be done easily? I don't have a hfe meter. I'll search about this on the Internet, but any hint is welcome.
One last question: do I have to adjust the two bias pots afterwards, even with matched diff pair transistors? I mean, it seems well adjusted right now, it has about -11 mV DC, just like requested limits in SM. Or am I mixing things up and this has nothing to do? Thanks!

It is most likely a noisy transistor in the differential pair (Q701 and Q703) at the input of the power amp. I would just replace them to start with. It is a very common issue in many units.You can spray the pair with freeze spray and I bet it will either cause the issue to get worse, or to almost go away. I would NEVER disconnect the feedback through R709, as the amp will surely go unstable without any feedback. You can sometimes find a transistor that is noisy by its leakage, but it is not likely possible with a multimeter.

I bet you a cup of coffee it is that differential pair Q701 and or Q703. You can try spraying them with freeze, and see if it effects the noise. I would also replace the same pair in the other channel as well.

 

[Since1984]

For what I can find, the 2SA992 seems to be a good equivalent. BC560 could work and is easier to find here, but I suppose it would be risky, as amp seems to have rails at about +-43 V and the BC560 datasheet states maximum VCE at 45 V...
I had a search about hfe matching (unfortunately my DMM doesn't have that function) and seems there are some test rigs to do this, not too easy, but I wonder: if I buy them together, hopefully all from the same lot, won't hfe be sufficiently similar? Or do they HAVE to be hand matched for a diff pair? I've read some varying opinions about this...

 

[AE Bob]

I would use 2SA970, as that is the transistor I stock for those (120 volt, 100 ma, 300 mw, 100 mhz , >200 hfe Low Noise Audio transistor). I find if I spot check the transistors from the same batch, the hfe comes out very close on all of them (I usually buy them 100 at a time ,to get the best price, as I use many). I doubt that Marantz matched the transistors originally. You will likely need to reset the offset, but the idle current shouldn't change (although I would check it anyway).

 

[Since1984]

Thanks Bob, you've been most helpful with this! I've read about the 2sa970 also, and if you have had the experience of them coming with very similar hfe in the same batch, then this would make it a lot easier for me, as I won't have to build some fancy test rig... I think I'll go with that transistor you're suggesting then. I'll report back on here once I've got them and done the repair.

 

[Since1984]

Repair done, testing done, it definitely was a defective transistor, to go safe I changed all 2sa750 for the recommended 2sa970 and the result is great, amplifier working fine without any noise! Thanks again for the help!

 

[Since1984]

Unfortunately this unit is back to my bench - not exactly for the same problem (a bad connection in Phono section), BUT: I detected while testing, that the unit still shows random "pops" (almost sounding like gun shots) during operation, but only on the right channel.
I managed to isolate the problem, again, it's the power section. This had the majority of small signal transistors replaced by me, as you can read above, except for the last drivers and power output transistors. Drivers are something like D667 and D666, I've never come across them listed on any place on the net as potentially being problematic, so I don't believe it could be a transistor? Or could it still be one of them? Unless it was one of the output power transistors? Is that even possible? Unfortunately I can't easily see which are mounted (they are below a metal plate which is hard to get off), as many times the service manuals of Marantz don't corrspond to actual transistors installed.

I decided to check most of the passive components in that channel and all measure fine, also no bad solder, as far as I can detect. No bad connections to speakers, too. No speaker relay mounted in this unit. So, I'm lost. I noticed a carbon composite resistor in the output and temporarily took it out but no improvement. Electrolytic caps have been replaced before by me (no mistakes made, triple checked). Any ideas? Thanks!