Okay, I rest my case. It is difficult to trace a fault with messages, it will take you forever. Now if there is no issue other than the two off-sets that are different and probably only inconvenient, then I suggest a simple solution. Decouple the amp from the speakers with say 2200 uF capacitor and you are done. You don't even need the protection relay thingy.
The driver board opamp appears to be fine, I swapped it with the opamp from the donor amp and no change. I also did the same with the opamp of the protection board.
I agree but the goal is not just to repair the amp but to learn something in the process.
I welcome and appreciate all input.
A signal tracer is an invaluable tool. Try knocking one together. You hook it up to an earpiece and listen following the sound through your amp by probing on different components. BTW to get a zero off-set on any commercial amplifier is not that common. Assume you talk about 20mV that is about 10 micro watt into the load. The background noise is probably higher. Maybe you are splitting hairs. Theoretically you can maybe have 0V off-set. Can your meter read that voltage reliably?
You can make a signal tracer with a single op-amp and variable gain control hooked to an old cell phone earpiece. Once you have one you won't believe how you could have lived without one.
So you're observing 8mV more drop across one of the emitter resistors. 8mV/0.3 ohm = ~ 26mA. So this an estimate of the mystery current. If you still can't resolve this symptom, I suggest lifting one end of L850. Then measure the voltage across R853 (12 ohm). If the mystery 26mA current flows through R853, you'd see about 320mV. This should give some insight. Be sure to recheck the drops across the emitter resistors, as opening L850 may change the picture. If you don't find significant voltage drop across R853, the mystery current is flowing along some other path. Try to find where it's going.
Before I pull the outputs, I took @BSST suggestion and disconnected one end of L850 and measured across R853 and after 5 minutes or so, it settled in between 50 to 60 mV, same for the left channel after taking L750 out of circuit and measuring across R753.
Removing the outputs is no easy chore with this amp, I going to study the schematic and see if I can de-solder a few wires which will be much less work.
Removing the outputs is no easy chore with this amp, I going to study the schematic and see if I can de-solder a few wires which will be much less work.
Ok, so there's 60mV/12 ohm = about 5mA through R853. That's not a huge current, though still not clear where it's going. What are the corresponding voltage drops across the emitter resistors? Has the disparity in voltages across the resistors reduced? Or is there still about 20mA going somewhere?
I'm a strong advocate of in-circuit diagnosis measurement before resorting to "exploratory surgery." 😉
I'm a strong advocate of in-circuit diagnosis measurement before resorting to "exploratory surgery." 😉
With L850 (R channel) & L750 (L channel) out of circuit, I get...
Right channel........15 mV & 11 mV (max)
Left channel...........17 mV & 13.4 mV (max)
With L850 & L750 in circuit, I now get...
Right channel........20 mV & 15 mV
Left channel...........20 mV & 15 mV
I had to see how it looked on the scope so I hooked up dummy loads and injected a 1K sine wave and slowly increased the volume until at the 12 o'clock position. Sine waves looked good, left channel had slightly more amplitude and discovered a spot in the potentiometer from 10 to 11 o'clock where the sine wave amplitude did not increase. I also did a quick run through from 20-20K Hz and the amplitude remained constant.
Took some temperature readings and Right channel heatsinks were 3-4 degrees warmer than Left channel. Right channel DC Offset 28 mV, Left channel, 20 mV
I realize this is fairly rudimentary but it's a start.
Before I remove the outputs, I'll look into getting a signal tracer as @Nico Ras suggested and also going to go through the schematic and note actual measurements along side those that are annotated just to be a little more systematic.
Right channel........15 mV & 11 mV (max)
Left channel...........17 mV & 13.4 mV (max)
With L850 & L750 in circuit, I now get...
Right channel........20 mV & 15 mV
Left channel...........20 mV & 15 mV
I had to see how it looked on the scope so I hooked up dummy loads and injected a 1K sine wave and slowly increased the volume until at the 12 o'clock position. Sine waves looked good, left channel had slightly more amplitude and discovered a spot in the potentiometer from 10 to 11 o'clock where the sine wave amplitude did not increase. I also did a quick run through from 20-20K Hz and the amplitude remained constant.
Took some temperature readings and Right channel heatsinks were 3-4 degrees warmer than Left channel. Right channel DC Offset 28 mV, Left channel, 20 mV
I realize this is fairly rudimentary but it's a start.
Before I remove the outputs, I'll look into getting a signal tracer as @Nico Ras suggested and also going to go through the schematic and note actual measurements along side those that are annotated just to be a little more systematic.
Last edited:
The resistor measures to spec out of circuit, I also tested the odd capacitor adjacent to it and it's fine as well.
I went through the schematic and measured all voltages.
Power supply voltages are good.
Driver board voltages are all within 2% except or Q710 & Q810, schematic shows -0.4 V at the bases, my measurements are -0.7 V
Protection board shows the largest deviation from spec, Q901, Q902, Q921, Q922 should be +.9 V or -.9 V, I get readings of +1.3V, -1.3V & +1.24V, -1.24V.
Q910 & Q930 are off quite a bit more.
I don't know if any of this has any effect on the bias issue this amp exhibits.
Next step would be to pull the output transistors (again)
Power supply voltages are good.
Driver board voltages are all within 2% except or Q710 & Q810, schematic shows -0.4 V at the bases, my measurements are -0.7 V
Protection board shows the largest deviation from spec, Q901, Q902, Q921, Q922 should be +.9 V or -.9 V, I get readings of +1.3V, -1.3V & +1.24V, -1.24V.
Q910 & Q930 are off quite a bit more.
I don't know if any of this has any effect on the bias issue this amp exhibits.
Next step would be to pull the output transistors (again)
I advise against removing the transistors. Troubleshoot the circuit and symptoms, and allow measurements to lead the way. I predict the transistors will check ok out of circuit.
You reported:
Protection board shows the largest deviation from spec, Q901, Q902, Q921, Q922 should be +.9 V or -.9 V, I get readings of +1.3V, -1.3V & +1.24V, -1.24V.
Remember in the era this amp was designed and characterized, they may have used passive meters; the meter’s load resistance may have led to a reading of 0.9V but the DVM presents a lighter load and the diode drops less voltage. Or so I conject.
I need to attend to other matters but will offer more thoughts in an hour or so.
Thanks.
You reported:
Protection board shows the largest deviation from spec, Q901, Q902, Q921, Q922 should be +.9 V or -.9 V, I get readings of +1.3V, -1.3V & +1.24V, -1.24V.
Remember in the era this amp was designed and characterized, they may have used passive meters; the meter’s load resistance may have led to a reading of 0.9V but the DVM presents a lighter load and the diode drops less voltage. Or so I conject.
I need to attend to other matters but will offer more thoughts in an hour or so.
Thanks.
Interesting idea about the meter, I have an old analog meter I restored and calibrated, I'll give that a try.
The suggestion for pulling the output transistors was to test the circuit without them, but because doing so is particularly difficult in this amp, I would prefer not to.
The suggestion for pulling the output transistors was to test the circuit without them, but because doing so is particularly difficult in this amp, I would prefer not to.
Last edited:
To test if the protection circuit is somehow contributing to the problem, you might disconnect the wire(s) from the Darlington bases to protection board (i.e. X901, etc).
To review a bit, am I correct that present concern is the too low, mismatched voltage drops across the 0.3 ohm emitter resistors? I believe at one point one channel (left?) was looking ok, but now both are misbehaving in similar fashion?
It's still curious to me that the collector-emitter voltage across Q710, Q810 appears to be too low. I would expect to see about 2.4V. The protection might be involved, as it acts to shunt drive to the Darlington bases.
Edit: If you disable protection as I suggested, be very careful with the bias pot. Start with pot set for minimum current and advance slowly.
To review a bit, am I correct that present concern is the too low, mismatched voltage drops across the 0.3 ohm emitter resistors? I believe at one point one channel (left?) was looking ok, but now both are misbehaving in similar fashion?
It's still curious to me that the collector-emitter voltage across Q710, Q810 appears to be too low. I would expect to see about 2.4V. The protection might be involved, as it acts to shunt drive to the Darlington bases.
Edit: If you disable protection as I suggested, be very careful with the bias pot. Start with pot set for minimum current and advance slowly.
Last edited:
BSST, you are probably right something other than the main amplifier connected somewhere else supplying power to it. Even an indicator light for that matter.
Referring to the data in post #70, there's about 4mV difference in the drops across the emitter resistors. But be conservative and assume only 3mV. 3mV/0.3 ohm = ~10mA. So that 10mA is flowing somewhere. If through R753 (12 ohm), that's about 120mV drop. But output voltage is about 20mV, and ostensibly there's light loading on that path, since no load is connected. So where's that current flowing? Maybe that's a clue. Just to emphasize, I can't find an explanation that's consistent with the data, so those are quandaries to be explored and resolved.
With power off, you might explore resistance to ground at R753. Maybe lift an end of R753 to isolate from output transistors. Perhaps drive that node from a signal generator and trace with your scope?
With power off, you might explore resistance to ground at R753. Maybe lift an end of R753 to isolate from output transistors. Perhaps drive that node from a signal generator and trace with your scope?
Last edited: