I don't understand what you're telling me. What circuit node is showing a short to ground?
Hi, just thinking out loud, probably incorrectly, as the 10k resistor goes to ground.
I apologize for the confusion.
Just saw post 181. This may be encouraging.
Try exercising the bias spreader from min to max and report available voltages present across C26. Restore to minimum spread when done.
Next experiment is to vary input voltage at input at R57 and confirm that we can push the driver stages from +rail to -rail. If you have a convenient adjustable voltage source, use it the drive input positive and driver stages should swing to the +rail; then apply negative input voltage, and driver should swing to -rail.
Try exercising the bias spreader from min to max and report available voltages present across C26. Restore to minimum spread when done.
Next experiment is to vary input voltage at input at R57 and confirm that we can push the driver stages from +rail to -rail. If you have a convenient adjustable voltage source, use it the drive input positive and driver stages should swing to the +rail; then apply negative input voltage, and driver should swing to -rail.
The 10k to ground was to ensure a known bias at the input of the power stages, since we have removed both Q18 and Q6. Absent the 10k, the output at L1 isn't very predictable.
You didn't mislead. I believe that earlier you were testing bias spreader and it looked good. In its most recent configuration, the front end is open loop, i.e. very large gain, and drift to a rail might be normal. So let's try to confirm front end:
Try tacking a 22k resistor between either end of C26 and R56. This sets front end gain to (1+22k/270) = ~82. Then check DCV to ground at C26. If front end is OK, the measured voltage should be near ground; there may be a few volts offset because the DC servo is disabled by the shorted C21, C22. If this test is as hoped, then the test of 100k between R57 and +/- 16V should be a valid exercise.
Let me know what you see. Good luck.
Try tacking a 22k resistor between either end of C26 and R56. This sets front end gain to (1+22k/270) = ~82. Then check DCV to ground at C26. If front end is OK, the measured voltage should be near ground; there may be a few volts offset because the DC servo is disabled by the shorted C21, C22. If this test is as hoped, then the test of 100k between R57 and +/- 16V should be a valid exercise.
Let me know what you see. Good luck.
Well, nothing makes sense:
Assuming C28 is still +22.8V re ground, the voltage at feedback (R56) should be = 22.8V * 270/(22k+270) = 0.276V. But the amp is not delivering feedback voltage that corresponds to the voltage divider, as you measure 0.1V at Q30 base.
Positive input at amp measures 0.18V, but it should be near 0V. The feedback base should be more positive than the input base, but it isn't.
Let's remove still more parts so the defect area is further restricted. Please remove the temporary 22k resistor tacked at R56. We should be left with only R56 providing bias to Q28 and Q30; input bias to Q29 and Q31 comes via R58 and R57. Voltage presented to the spreader stages isn't well defined, but it should be safe to let it go wherever it wishes. Would you apply power and report voltages at collectors of Q23 and Q38, and b,c,e of Q30, Q31, Q26, Q28. I hope these measurements will help to explain the mysteries noted above.
Thank you.
Assuming C28 is still +22.8V re ground, the voltage at feedback (R56) should be = 22.8V * 270/(22k+270) = 0.276V. But the amp is not delivering feedback voltage that corresponds to the voltage divider, as you measure 0.1V at Q30 base.
Positive input at amp measures 0.18V, but it should be near 0V. The feedback base should be more positive than the input base, but it isn't.
Let's remove still more parts so the defect area is further restricted. Please remove the temporary 22k resistor tacked at R56. We should be left with only R56 providing bias to Q28 and Q30; input bias to Q29 and Q31 comes via R58 and R57. Voltage presented to the spreader stages isn't well defined, but it should be safe to let it go wherever it wishes. Would you apply power and report voltages at collectors of Q23 and Q38, and b,c,e of Q30, Q31, Q26, Q28. I hope these measurements will help to explain the mysteries noted above.
Thank you.
Last edited:
I'm sorry for creating confusion. The top half of post 195 outlined some of the incongruities that need to be resolved. C26 and C28 share a common connection, and my earlier attempt was to get the that node to bias to near 0V. But it became more evident that the problem may lie within the input stages.
So the lower half sets up an experiment to troubleshoot the input stage transistors. We'll just ignore all the other stages for now.
So the lower half sets up an experiment to troubleshoot the input stage transistors. We'll just ignore all the other stages for now.
Hi Peter,
The 0.2V at the bases Q28 and Q31 remains mysterious. Would you explore along its bias path? Given the 0.2V at the bases, you should find +0.2V at R57 and about 9mV across R58. I'm suspicious that R57 might be open or have defective circuit traces, so confirming about 9mV across R58 holds keen importance.
Would you also report Q35 and voltages across R45, R46, R69, R71 ?
Thank you.
The 0.2V at the bases Q28 and Q31 remains mysterious. Would you explore along its bias path? Given the 0.2V at the bases, you should find +0.2V at R57 and about 9mV across R58. I'm suspicious that R57 might be open or have defective circuit traces, so confirming about 9mV across R58 holds keen importance.
Would you also report Q35 and voltages across R45, R46, R69, R71 ?
Thank you.