R13, R14 have nearly rail-to-rail voltage applied and are 3W, so it's normal for them to be toasty.
I haven't yet found R7. Can you steer me in the right direction?
I'm obsessing a bit about the bias spreader, especially since you seem to have had a fried R53, and it's a sure-fire way to blow the outputs if it's not working properly. So, would you remove both Q18 and Q6; previous tests suggest this should allow the 60V rails to rise. Then see what presents across C26 and voltage at C26 with respect to gound.
I haven't yet found R7. Can you steer me in the right direction?
I'm obsessing a bit about the bias spreader, especially since you seem to have had a fried R53, and it's a sure-fire way to blow the outputs if it's not working properly. So, would you remove both Q18 and Q6; previous tests suggest this should allow the 60V rails to rise. Then see what presents across C26 and voltage at C26 with respect to gound.
Just to make sure I'm not ignoring a problem/clue, which resistors are getting hot? I would expect the two added current-limit resistors to heat quickly, as they have full 60V applied at power-on. However, as the 60V rails rise, the volts across the current-limit resistors should drop and their dissipated power should abate. Is this what you observe? Are there other resistors concerning? If we aren't causing damage, I'd like to have 60V rails elevated to facilitate testing of the bias spreader.
If the above reassures you to proceed, I'd like to get some measured voltages in the bias spreader to find its problem. With voltmeter's negative lead referenced to Q32 emitter, I'd like measurements of:
Q32 base and collector
Q8 base
Q10 base, emitter, collector
Junction of R65 and R53 (set pot for maximum resistance)
Top of R66
Bottom R49
Output
+/-80V rails
+/-60V rails
Measuring these voltages may be problematic if C26 voltage re ground is drifting. I surmise the C26 voltage re ground may be moving relatively slowly? If so, I suggest letting the C26 voltage follow its own course to see if it settles to a stable voltage. Another idea is to connect a 1k resistor between amp Output and the Q32 emitter reference; I think this should allow amplifier bias to servo the reference to near 0V. Let me know of any issues.
Believe it or not, I think we're making progress.
If the above reassures you to proceed, I'd like to get some measured voltages in the bias spreader to find its problem. With voltmeter's negative lead referenced to Q32 emitter, I'd like measurements of:
Q32 base and collector
Q8 base
Q10 base, emitter, collector
Junction of R65 and R53 (set pot for maximum resistance)
Top of R66
Bottom R49
Output
+/-80V rails
+/-60V rails
Measuring these voltages may be problematic if C26 voltage re ground is drifting. I surmise the C26 voltage re ground may be moving relatively slowly? If so, I suggest letting the C26 voltage follow its own course to see if it settles to a stable voltage. Another idea is to connect a 1k resistor between amp Output and the Q32 emitter reference; I think this should allow amplifier bias to servo the reference to near 0V. Let me know of any issues.
Believe it or not, I think we're making progress.
Last edited:
So, put a fan on the board and let it sit for a few minutes.
44.7V is the highest it got across C3.
R26 goes from ~+60V to -60.9V before stabilizing.
Q9 & Q11 got quite warm. Everything else felt fine with a finger test. Those resistors weren't too warm with the fan on them.
Q32 base and collector: 0.7/1.6
Q8 base: 0.3
Q10 base, emitter, collector: 1.2/1.6/0.9
Junction of R65 and R53 (set pot for maximum resistance):1.17
Top of R66: 1.18
Bottom R49: 0.32
Output
+/-80V rails: 85.8V
44.7V is the highest it got across C3.
R26 goes from ~+60V to -60.9V before stabilizing.
Q9 & Q11 got quite warm. Everything else felt fine with a finger test. Those resistors weren't too warm with the fan on them.
Q32 base and collector: 0.7/1.6
Q8 base: 0.3
Q10 base, emitter, collector: 1.2/1.6/0.9
Junction of R65 and R53 (set pot for maximum resistance):1.17
Top of R66: 1.18
Bottom R49: 0.32
Output
+/-80V rails: 85.8V
We are poking around in a problematic area, but I can't yet sort out what's a definite defect. Data is still confusing.
Some mysterious symptoms:
Q32 voltages suggest it's active and might be controlling the shunt spreader voltage. But it's not clear how Q32 gets its base current: the voltage span between the bases of Q8 and Q10 are 0.9V, seemingly not enough to get the pair to conduct current; scrutinizing the voltages across R65 and R53 suggests R53 might be open--- (that's my best guess at this point) but then how does any base current get delivered to Q32?
So these are questions to be resolved. To that end, would you report voltage at Q32 emitter re ground? Explore the voltages across R65 and R53; being 270 ohm and 200 ohm, they should be roughly similar. And is there any voltage across R67? Should be 0V given the low spreader voltage. Any follow your instincts re any other voltages that don'y make sense.
Thanks.
A lot of these voltages are confusing, hence the need to confirm that they are established by nearby components and not being clamped by other circuits.
Some mysterious symptoms:
Q32 voltages suggest it's active and might be controlling the shunt spreader voltage. But it's not clear how Q32 gets its base current: the voltage span between the bases of Q8 and Q10 are 0.9V, seemingly not enough to get the pair to conduct current; scrutinizing the voltages across R65 and R53 suggests R53 might be open--- (that's my best guess at this point) but then how does any base current get delivered to Q32?
So these are questions to be resolved. To that end, would you report voltage at Q32 emitter re ground? Explore the voltages across R65 and R53; being 270 ohm and 200 ohm, they should be roughly similar. And is there any voltage across R67? Should be 0V given the low spreader voltage. Any follow your instincts re any other voltages that don'y make sense.
Thanks.
Well, that should be easier to troubleshoot!
Let's take a fresh look at the whole picture, especially since you were reporting about -5 at the output.
Is this the present state of the amp?
Q18 is removed?
Current limit resistors are still present, so 60V rails are at a reduced voltages?
Bottom of C26 is stuck around -60V?
All voltages are settled and stable enough to make voltage readings to present a stable picture?
If so, would you report voltages, with respect tp ground:
C26 bottom and top
Q6 base, emitter, collector(i.e. -60V rail)
+60V rail
Output
Feedback node, eg at base Q30 or equivalent
Input, eg. at base Q31 or equivalent
I'm sure there'll be followup exploration, but hope this will point the way.
Thanks.
Let's take a fresh look at the whole picture, especially since you were reporting about -5 at the output.
Is this the present state of the amp?
Q18 is removed?
Current limit resistors are still present, so 60V rails are at a reduced voltages?
Bottom of C26 is stuck around -60V?
All voltages are settled and stable enough to make voltage readings to present a stable picture?
If so, would you report voltages, with respect tp ground:
C26 bottom and top
Q6 base, emitter, collector(i.e. -60V rail)
+60V rail
Output
Feedback node, eg at base Q30 or equivalent
Input, eg. at base Q31 or equivalent
I'm sure there'll be followup exploration, but hope this will point the way.
Thanks.
Q18 is removed? YES, as is Q6
Current limit resistors are still present, so 60V rails are at a reduced voltages? YES
Bottom of C26 is stuck around -60V? NO, +60 - -60V
All voltages are settled and stable enough to make voltage readings to present a stable picture? YES
If so, would you report voltages, with respect tp ground:
C26 bottom and top: -60/-61.6
Q6 base, emitter, collector(i.e. -60V rail)
+60V rail : 0, -79, -61.6 (Q6 not present)
Output: 0.0V
Feedback node, eg at base Q30 or equivalent: 0.59V
Input, eg. at base Q31 or equivalent: 0.3V
Current limit resistors are still present, so 60V rails are at a reduced voltages? YES
Bottom of C26 is stuck around -60V? NO, +60 - -60V
All voltages are settled and stable enough to make voltage readings to present a stable picture? YES
If so, would you report voltages, with respect tp ground:
C26 bottom and top: -60/-61.6
Q6 base, emitter, collector(i.e. -60V rail)
+60V rail : 0, -79, -61.6 (Q6 not present)
Output: 0.0V
Feedback node, eg at base Q30 or equivalent: 0.59V
Input, eg. at base Q31 or equivalent: 0.3V
Digging around a bit, I had a check on the D7 & D25 diodes.
D7 & D25 act differently!
D7 - If I put the red onto anode and black on cathode, my DMM says 600(0.6V, I assume). If I do the opposite, it rises past 2000 then goes blank, as expected.
D25 - doesn't! It's over 2000 in BOTH DIRECTIONS! This is in circuit, but this means we found a defective component! Now, does this fit the behavior? Does it make sense?
D7 & D25 act differently!
D7 - If I put the red onto anode and black on cathode, my DMM says 600(0.6V, I assume). If I do the opposite, it rises past 2000 then goes blank, as expected.
D25 - doesn't! It's over 2000 in BOTH DIRECTIONS! This is in circuit, but this means we found a defective component! Now, does this fit the behavior? Does it make sense?
Last edited:
D25 sure sounds suspicious. You're probably suing diode test mode; I'm curious about the diode check behavior and resistance reading out-of-circuit.
If I interpret your data correctly from post 70, the C26 voltage starts at +60V and then transitions to -60V and sticks there. So whatever leakage current a suspect D25 is presenting, it's not enough to short the bias regulator to the +60 rail. So the state of D25 is sill curious.
Assuming this is the case, there's still too little voltage across C26/ Q32, i.e. there's still something failed in the bias spreader. With the bias spreader stuck at -60V and the output stages disconnected (Q6 and Q18 absent), this is a good configuration for debugging.
First confirm there's 0.6V or more across R63; this assures there's enough supply current to power the bias spreader. Then, repeating from an earlier test, with voltmeter's negative lead referenced to Q32 emitter, would you measurement:
Q32 base and collector
Q10 base, emitter, collector
Junction of R65 and R53 (set pot for maximum resistance)
Top of R66
Bottom R49
Thanks!
If I interpret your data correctly from post 70, the C26 voltage starts at +60V and then transitions to -60V and sticks there. So whatever leakage current a suspect D25 is presenting, it's not enough to short the bias regulator to the +60 rail. So the state of D25 is sill curious.
From the schematic, I would expect the "top" voltage to be more positive than the "bottom." Did you intend C26 top = -60V and C26 bottom = -61.6V ??
Assuming this is the case, there's still too little voltage across C26/ Q32, i.e. there's still something failed in the bias spreader. With the bias spreader stuck at -60V and the output stages disconnected (Q6 and Q18 absent), this is a good configuration for debugging.
First confirm there's 0.6V or more across R63; this assures there's enough supply current to power the bias spreader. Then, repeating from an earlier test, with voltmeter's negative lead referenced to Q32 emitter, would you measurement:
Q32 base and collector
Q10 base, emitter, collector
Junction of R65 and R53 (set pot for maximum resistance)
Top of R66
Bottom R49
Thanks!
Maybe it's a good idea to lift at least one end of D25 so that it can be checked out-of-circuit, since there are so many possible pitfalls to in-circuit tests. Restore or replace as appropriate.
Would you also measure the voltage across R63 and the voltages at top and bottom of C26 re ground? I'm trying to reassure myself that there's adequate operating current available to the bias spreader to produce believable voltages; at present, the spreader voltages don't hang together in a sensible picture, even when I try to imagine possible failed components.
Thanks.
Would you also measure the voltage across R63 and the voltages at top and bottom of C26 re ground? I'm trying to reassure myself that there's adequate operating current available to the bias spreader to produce believable voltages; at present, the spreader voltages don't hang together in a sensible picture, even when I try to imagine possible failed components.
Thanks.
I'm sorry, but I don't follow what you're telling me. The diode should behave out of circuit--- i.e. about 0.6V when forward biased, open in the other direction. Otherwise, it's failed and needs replacement. Would you clarify?
If you applied diode test to the circuit without the diode, you'd probably observe open in both directions, but all sorts of other symptoms are possible because of semiconductor junctions, and caps with possible stored charge.
Assuming the D25 diode issue is resolved, we can continue with troubleshooting the bias spreader. My apologies, because I don't know I managed to ask about R63 voltage. My intent was to ask voltage across R76, to confirm current flow in Q38 and Q11 an thus current delivery to the bias spreader circuit. But 0.8V across R63 suggests there maybe very little drop across R76.
Would you report voltage across R76?
Thanks again.
If you applied diode test to the circuit without the diode, you'd probably observe open in both directions, but all sorts of other symptoms are possible because of semiconductor junctions, and caps with possible stored charge.
Assuming the D25 diode issue is resolved, we can continue with troubleshooting the bias spreader. My apologies, because I don't know I managed to ask about R63 voltage. My intent was to ask voltage across R76, to confirm current flow in Q38 and Q11 an thus current delivery to the bias spreader circuit. But 0.8V across R63 suggests there maybe very little drop across R76.
Would you report voltage across R76?
Thanks again.
Great!
Assuming that C26 remains stuck on the negative rail, that should make for valid measurements in the bias spreader.
So again as in earlier test, with voltmeter's negative lead referenced to Q32 emitter, would you measure:
Q32 base and collector
Q10 base, emitter, collector
Q8 base
Junction of R65 and R53 (set pot for maximum resistance)
Top of R66
Bottom R49
Thanks!
So again,
Assuming that C26 remains stuck on the negative rail, that should make for valid measurements in the bias spreader.
So again as in earlier test, with voltmeter's negative lead referenced to Q32 emitter, would you measure:
Q32 base and collector
Q10 base, emitter, collector
Q8 base
Junction of R65 and R53 (set pot for maximum resistance)
Top of R66
Bottom R49
Thanks!
So again,