I have already replaced originals 2SC2275 and 2SA985 with TTC004B and TTA004B respectively.
Same as I did for the right channel.
Just exploring an idea--- don't actually try just yet: Would slack in the cables allow you to connect right channel output devices to left channel driver PCB? Or vice versa?
It may be possible to come up with something, but the original cables are very short, I would have to improvise.
Swap a few suspicious components from right channel to the left channel would be easier.
Swap a few suspicious components from right channel to the left channel would be easier.
Was just a thought, don't bother.
I'm going to assume Q723 and Q725 are OK. I'm suspecting the "bias spreader" network (Q715, Q717, and surrounding parts) and trying to think of a way to test without going crazy.
I'm going to assume Q723 and Q725 are OK. I'm suspecting the "bias spreader" network (Q715, Q717, and surrounding parts) and trying to think of a way to test without going crazy.
Hi Bokey,
The failure you noted above we spotted and addressed back in post 8. mgrisoli faces the difficulty of making stable voltage measurements while his DBT is cycling about every 2 seconds.
With power off, I suggest making resistance measurements in the bias spreader circuit in the good channel and comparing with same points in the bad channel. If that doesn't reveal anything, I can suggest a powered test but it would require some test preparation.
The failure you noted above we spotted and addressed back in post 8. mgrisoli faces the difficulty of making stable voltage measurements while his DBT is cycling about every 2 seconds.
With power off, I suggest making resistance measurements in the bias spreader circuit in the good channel and comparing with same points in the bad channel. If that doesn't reveal anything, I can suggest a powered test but it would require some test preparation.
The resistance comparison measurements should reveal something. I'd lift the transistors off the faulty board to check them thoroughly, one by one. Definitely would check all 5 diodes.
@mgrisoli
start populating the amp schematics with voltage values. It'd be so much easier to have a full picture. I am sure the root cause would be much easier to spot.
I think the DC servo is trying to correct a large DC offset at the speaker output, hence that large voltage spike every 1.5 seconds.
Lift the connector to LEVEL IND.Asembly while testing. I'd also look at disconnecting the regulator PCB while testing/fault-finding.
I am not sure if you have variac... If you do, lift one side of R721 and R751 while slowly increasing the voltage. This may reveal the issue more easily.
R747 might be open circuit & Q709 might be faulty...
@mgrisoli
start populating the amp schematics with voltage values. It'd be so much easier to have a full picture. I am sure the root cause would be much easier to spot.
I think the DC servo is trying to correct a large DC offset at the speaker output, hence that large voltage spike every 1.5 seconds.
Lift the connector to LEVEL IND.Asembly while testing. I'd also look at disconnecting the regulator PCB while testing/fault-finding.
I am not sure if you have variac... If you do, lift one side of R721 and R751 while slowly increasing the voltage. This may reveal the issue more easily.
R747 might be open circuit & Q709 might be faulty...
There are 4 outputs per channel so it is most likely one channel is good.
The shorted channels drivers will need replacement as well. It could have killed something in the bias circuit as well but it is unlikely.
The shorted channels drivers will need replacement as well. It could have killed something in the bias circuit as well but it is unlikely.
Let me review where I believe we are. mgrisoli, please correct me if I misstate any details and please elaborate where you wish.
The problem is in the left channel; I believe the right channel is working, but I don't know this explicitly.
In the left channel, when the interface cable to the output devices is unplugged, DBT cycling stops, protection releases, and the driver board behaves sufficiently to bias its output to 0v (Post 19) and present a reasonable voltage at servo error amp (Post 17). Note that driver board bias and servo paths remain intact despite disconnected output devices.
The output transistors have +/- 57V present, but do not pass enough current to provoke fault protection or overload the DBT. Apparently, there are no collector to emitter shorts. There might be base-emitter damage somewhere amongst the 4 devices, but this seems unlikely, I think.
Given the above, my suspicion is that the "base spreader" network (Q715, Q717, and associated parts) is misbehaving by applying excessive drive voltage and provoking rail-to-rail current when connected to the output devices. Comparing resistances with the other channel might reveal a problem. Chief suspects would be the two thermistors and their paralleled resistors, and the bias pot. (I would note the Ref Des, but I can't read them with any confidence. The pot should be adjusted to minimum resistance for initial powered testing.)
I believe it would be possible to test the powered spreader circuit independently of the output devices, but such a test is frustrated because Q723 and Q725 are missing collector bias when the interface cable is removed. The following would be a bold experiment, but with the cable removed, I believe it would be safe to install jumpers from Q723 and Q725 collectors to their respective supply rails. Q719 and Q721 should clamp any excessive drive to Q723 and Q725. If resistance checks don't reveal a problem, this procedure might be considered. There is some risk; circuit analysis is strictly mental. Adjusting the bias pot should vary the voltage between the emitters of Q723 and Q725.
Comments welcome.
The problem is in the left channel; I believe the right channel is working, but I don't know this explicitly.
In the left channel, when the interface cable to the output devices is unplugged, DBT cycling stops, protection releases, and the driver board behaves sufficiently to bias its output to 0v (Post 19) and present a reasonable voltage at servo error amp (Post 17). Note that driver board bias and servo paths remain intact despite disconnected output devices.
The output transistors have +/- 57V present, but do not pass enough current to provoke fault protection or overload the DBT. Apparently, there are no collector to emitter shorts. There might be base-emitter damage somewhere amongst the 4 devices, but this seems unlikely, I think.
Given the above, my suspicion is that the "base spreader" network (Q715, Q717, and associated parts) is misbehaving by applying excessive drive voltage and provoking rail-to-rail current when connected to the output devices. Comparing resistances with the other channel might reveal a problem. Chief suspects would be the two thermistors and their paralleled resistors, and the bias pot. (I would note the Ref Des, but I can't read them with any confidence. The pot should be adjusted to minimum resistance for initial powered testing.)
I believe it would be possible to test the powered spreader circuit independently of the output devices, but such a test is frustrated because Q723 and Q725 are missing collector bias when the interface cable is removed. The following would be a bold experiment, but with the cable removed, I believe it would be safe to install jumpers from Q723 and Q725 collectors to their respective supply rails. Q719 and Q721 should clamp any excessive drive to Q723 and Q725. If resistance checks don't reveal a problem, this procedure might be considered. There is some risk; circuit analysis is strictly mental. Adjusting the bias pot should vary the voltage between the emitters of Q723 and Q725.
Comments welcome.
@BSST
I left this amp in the thinking corner for a few days while I was busy at work, I came back to it today and decided to cross some Ts.
I found that the collector trace of Q727 was lifted and with a bad solder joint.
Once fixed it the amp stopped oscillating. I was able to regulate the idle current and listen to some music.
A few additional findings that I presume have no relation with the Q727 problem:
Left channel shows some noise in the oscilloscope. A 400mhz sine wave line shows thicker than the right channel. In close-up there is a 60hz noise coming through.
I havent done anything yet to troubleshoot the issue above. If you want to keep helping me a couple more miles I would really appreciate that, however your assistance so far has been incredibly helpful.
I left this amp in the thinking corner for a few days while I was busy at work, I came back to it today and decided to cross some Ts.
I found that the collector trace of Q727 was lifted and with a bad solder joint.
Once fixed it the amp stopped oscillating. I was able to regulate the idle current and listen to some music.
A few additional findings that I presume have no relation with the Q727 problem:
Left channel shows some noise in the oscilloscope. A 400mhz sine wave line shows thicker than the right channel. In close-up there is a 60hz noise coming through.
I havent done anything yet to troubleshoot the issue above. If you want to keep helping me a couple more miles I would really appreciate that, however your assistance so far has been incredibly helpful.
Last edited:
I'm glad to hear power amp seems to be working, since they are often vexing to debug.
I'm happy to help brainstorm any continuing issues.
Best,
Steve
I'm happy to help brainstorm any continuing issues.
Best,
Steve
Update...
Correction, the noise is in the right channel, not the left.
I traced the audio path starting from the speaker terminals.
I disconnected the level indicator circuit, the noise remains.
It comes through the relays, power transistor PCB and in the driver PCB comes out from FET Q702.
Previous to that FET it is just the input board, signal is weak and noisy at that stage, but I think I can assume the FET is bad.
Other hypothesis would be a power supply ripple, but then both channels should be affected, not just the right channel.
Now I am researching replacements for the dual 2SK150A.
Correction, the noise is in the right channel, not the left.
I traced the audio path starting from the speaker terminals.
I disconnected the level indicator circuit, the noise remains.
It comes through the relays, power transistor PCB and in the driver PCB comes out from FET Q702.
Previous to that FET it is just the input board, signal is weak and noisy at that stage, but I think I can assume the FET is bad.
Other hypothesis would be a power supply ripple, but then both channels should be affected, not just the right channel.
Now I am researching replacements for the dual 2SK150A.
I had such noise from defective FET a few times. May be hard to source a non fake replacement.
https://www.diyaudio.com/community/threads/2sk150a-gr-replacement-or-substitute-suggestions.267214/
Best may be to adapt SMD parts.
https://www.diyaudio.com/community/threads/2sk150a-gr-replacement-or-substitute-suggestions.267214/
Best may be to adapt SMD parts.
Last edited:
@Turbowatch2
Thank you for the suggestion. I ended up buying some 2SK117s.
All that to conclude the problem is not with the FETs.
I am still chasing down the noise on the right channel. Frequency of the noise is 1.1Mhz and level is about 200mV peak to peak on the speaker output.
I have replaced all transistors in the audio path except two, I ordered them today.
Thank you for the suggestion. I ended up buying some 2SK117s.
All that to conclude the problem is not with the FETs.
I am still chasing down the noise on the right channel. Frequency of the noise is 1.1Mhz and level is about 200mV peak to peak on the speaker output.
I have replaced all transistors in the audio path except two, I ordered them today.
This is only a wild guess, but I suggest checking the Zobel network: R790 and C742, 10 Ohms, 0.047 uF. An Ohmmeter will allow check of R790. The cap has reactance of about 3.1 ohms at 1.1MHz, so if you measure 200mV at the output, there should AC volts dropping across both R790 and C720 if they are both viable.
As this is a very old amp, caps should always be considered "not new".
If you got ripple on one side only, this can be the result of a dry cap.
As long as one side of a stereo amp is working, fault finding should be possible, just by comparing the same points, left and right.
Did you check the voltages at C705 /C706? (707/708) Is there a lot of dirt on the DC? If this cap is bad it could induce noise (and make the FET look bad) only on one channel.
If you got ripple on one side only, this can be the result of a dry cap.
As long as one side of a stereo amp is working, fault finding should be possible, just by comparing the same points, left and right.
Did you check the voltages at C705 /C706? (707/708) Is there a lot of dirt on the DC? If this cap is bad it could induce noise (and make the FET look bad) only on one channel.
What ??? These old Japanese designs are WAY ahead of the rest. Just about all the better present DIYA designs are based on these classic (Japanese) designs.
One of mine is similar to the Nikko. It eats 3886's for lunch .
Yeah... I used to dream about Nikko amps when I was a teenager... could never afford one... great amps.
