Input of the amp is shorted? Maybe it is something picked up.
Please make short on the RCA input (via wire or male RCA connector). After this if you have measured on the AC setting something then you have oscillation and this should be sorted out...
Please make short on the RCA input (via wire or male RCA connector). After this if you have measured on the AC setting something then you have oscillation and this should be sorted out...
Please measure now what you have on the base of the Q29 and Q30, if D8, D9 and C17 are shorted there shouldn't be any DC voltage. Also on the emitter resistor (R57 and R58) of the Q29 and Q30 shouldn't flow current and shouldn't be any voltage.
Is C17 still shorted?
I don't understand how you got this 177mV between this two bases. Did you change bipolar capacitor C27 (100uF/25V), it is placed in the Offset voltage circuit detector, maybe some current leak through it to the GND.
S2 switch is in what position?
I don't understand how you got this 177mV between this two bases. Did you change bipolar capacitor C27 (100uF/25V), it is placed in the Offset voltage circuit detector, maybe some current leak through it to the GND.
S2 switch is in what position?
A couple of speculative musings:
Some symptoms appeared to improve when the amp input was shorted. Any chance C1, C2 are open? With the innards of the amp open for testing, is there abnormal feedback suseptability to the amp input? Was the input always open during prior tests?
There are apparently AC signals present with very little DC bias in the output end of the amp. Are they definitely internally generated, not external EMI? I note that with C17 shorted, there's lots of crossover distortion/dead zone in the feedback loop. The resulting amp output might be very noisy and erratic with no signal present. An interesting experiment would connect a resistor (22k ?) between the output (or driver) and either supply rail to provide a light bias current. I wouldn't be surprised if noisy waveforms cleaned up as a result.
Some symptoms appeared to improve when the amp input was shorted. Any chance C1, C2 are open? With the innards of the amp open for testing, is there abnormal feedback suseptability to the amp input? Was the input always open during prior tests?
There are apparently AC signals present with very little DC bias in the output end of the amp. Are they definitely internally generated, not external EMI? I note that with C17 shorted, there's lots of crossover distortion/dead zone in the feedback loop. The resulting amp output might be very noisy and erratic with no signal present. An interesting experiment would connect a resistor (22k ?) between the output (or driver) and either supply rail to provide a light bias current. I wouldn't be surprised if noisy waveforms cleaned up as a result.
That's not going to cause any issue such as we have with the bias.
At face value it can only be an issue with the pre driver stage. It should look like this:
I'm going to say that you need to remove one diode from each of those series chains of three. Lets say D11 and D12. Lets remove that from suspicion.
With it in the above state the output voltage line should basically follow whatever voltage is on the shorted C17. Although it should be zero volts of course (no DC offset) if you happened to have +20 volts or -9 volts on C17 the output node should follow and there should always be zero bias current flowing.
If something were feeding current into the output node then that current would show in only one half of the stage depending on polarity. You would not see volt drop across both 33 ohm or each upper and lower set of 0.47 ohm for example.
Removed D11 and D12
The voltage to Pin23 (ground) of C17 is 75mv
Voltage across R89/R90 is 0mv
Voltage across R65 is -0.8v
Voltage across R62 is +0.8v
Voltage to earth of R65 is -0.85v at Q32 connection and 0v at the output
Voltage to earth of R62 is +0.85v at Q31 connection and 0v at the output
I also replaced C1/C2 in case they were leaky
The voltage to Pin23 (ground) of C17 is 75mv
Voltage across R89/R90 is 0mv
Voltage across R65 is -0.8v
Voltage across R62 is +0.8v
Voltage to earth of R65 is -0.85v at Q32 connection and 0v at the output
Voltage to earth of R62 is +0.85v at Q31 connection and 0v at the output
I also replaced C1/C2 in case they were leaky
Wow, we seem to be back with the same picture.
Would you confirm amp input is still shorted. Would you additionally report voltages at bases of Q29 Q31 Q30 Q32 and TP16? TP16 should be the same as the same as the amp output, but I’ve been surprised that there hasn’t been at least a small DC offset present. Perhaps the relay is open? Would you report TP16 with mV resolution? Additionally, would you probe these points with your scope for AC activity?
If things remain similarly unchanged, would you humor me and connect a 22k resistor between TP16 and a supply rail. Is there any interesting change?
Would you confirm amp input is still shorted. Would you additionally report voltages at bases of Q29 Q31 Q30 Q32 and TP16? TP16 should be the same as the same as the amp output, but I’ve been surprised that there hasn’t been at least a small DC offset present. Perhaps the relay is open? Would you report TP16 with mV resolution? Additionally, would you probe these points with your scope for AC activity?
If things remain similarly unchanged, would you humor me and connect a 22k resistor between TP16 and a supply rail. Is there any interesting change?
Amp input is shorted
Base voltages are Q29 = 200mv/ Q30 =-22mv/Q31= -1.1/Q32 =1.2 and TP 16= 2.37mv
all measured with 5.5 digit siglent DVM
and it is my pleasure to measure anything you request..🙂
so AC Voltages on bases Q29=6.4v/Q30= 6.4v/ Q31=60mv/Q32=56mv and TP 16 =0.5v
measured with scope and DVM
I will do the 22k in tomorrow and report back
thanks again
Base voltages are Q29 = 200mv/ Q30 =-22mv/Q31= -1.1/Q32 =1.2 and TP 16= 2.37mv
all measured with 5.5 digit siglent DVM
and it is my pleasure to measure anything you request..🙂
so AC Voltages on bases Q29=6.4v/Q30= 6.4v/ Q31=60mv/Q32=56mv and TP 16 =0.5v
measured with scope and DVM
I will do the 22k in tomorrow and report back
thanks again
Last thoughts tonight:
I recall that you were having trouble getting a frequency measurement of the oscillation. If your scope is digital, you might try a single sweep capture so you can disregard sync stability challenges.
Having seen this data, I doubt my requested 22K experiment has any value.
I should have also asked for AC voltage at the (shorted) C17 node for a more complete picture, but the peculiar combo of low DC biases that don't make sense and large AC swings sure suggests oscillation.
Without the current gain of the output transistors and their low impedance drive, I suspect the Zobel network (C24, R89,R90) will be a difficult load for the drivers. Given the absence of the output devices, you might try lifting C24 and see if that helps stop the bias loop oscillation.
When you eventually install output devices, I'd keep input shorted until everything else is resolved and input susceptibility is the only uncertainty remaining.
I recall that you were having trouble getting a frequency measurement of the oscillation. If your scope is digital, you might try a single sweep capture so you can disregard sync stability challenges.
Having seen this data, I doubt my requested 22K experiment has any value.
I should have also asked for AC voltage at the (shorted) C17 node for a more complete picture, but the peculiar combo of low DC biases that don't make sense and large AC swings sure suggests oscillation.
Without the current gain of the output transistors and their low impedance drive, I suspect the Zobel network (C24, R89,R90) will be a difficult load for the drivers. Given the absence of the output devices, you might try lifting C24 and see if that helps stop the bias loop oscillation.
When you eventually install output devices, I'd keep input shorted until everything else is resolved and input susceptibility is the only uncertainty remaining.
We have to find out where that 0.8 volts is coming from across those resistors.
You could lift the 22 ohm resistors R59 and R60 and do as we did with the others and tie both to ground. That absolutely has to give 0.00 across the 33 ohms. I don't think there would be any issue doing that although it does leave the first driver pair as 'outputs' with 270 ohm emitter resistors. No signal and nowhere for anything to go and it should not be a problem.
You could lift the 22 ohm resistors R59 and R60 and do as we did with the others and tie both to ground. That absolutely has to give 0.00 across the 33 ohms. I don't think there would be any issue doing that although it does leave the first driver pair as 'outputs' with 270 ohm emitter resistors. No signal and nowhere for anything to go and it should not be a problem.