Discrete opamp trouble shooting.

[elskardio]

Are the opamp's transistors easy to identify?

 

[Johnny1234]

From photo, looks like xsistor #s are scrubbed off. There are only 8 transistors in the circuit. Surely they can be tested EBC and a quick beta check. One will turn up bad.

 

[pucho812]

From photo, looks like xsistor #s are scrubbed off. There are only 8 transistors in the circuit. Surely they can be tested EBC and a quick beta check. One will turn up bad.

I did that. All tested ok.
Tested with both a dmm doing the diode check and my peak electronics transistor tester.
Numbers are hard to see on some but they are all the same either a bc456 or bc556

 

[Johnny1234]

I would take voltage measurements from emitter to ground between working and not working unit and see what is wrong. resistors are rarely a problem, but I have seen it before. Solder joints? Are the pin gripping the mother board tightly. Is the b+ and or b- showing up to full value. Just some dumb thoughts! .......DC voltages

 

[pucho812]

So when the discrete is working. I can measure +4 dBu at the output of the opamp
I don’t think the resistors are an issue as they all measure the same resistor value in the same spots. I can’t see the tiny ceramic as the culprit for my issues yet all 446 and 556 transistors test as working.
Of the 4 not working correctly. I get -1dBu at output when in circuit

 

[Johnny1234]

I would take voltage measurements from emitter to ground between working and not working unit and see what is wrong. resistors are rarely a problem, but I have seen it before. Solder joints? Are the pin gripping the mother board tightly.

 

[Johnny1234]

Just thought of some thing else. there seems to be no coupling caps on the board. It may be dc coupled. So your EBC readings and beta measurement may be misleading. I know a pain in the >>> Did you check the transistors out of circuit?

 

[Johnny1234]

Open junction looks correct because another junction is turning on

 

[pucho812]

Just thought of some thing else. there seems to be no coupling caps on the board. It may be dc coupled. So your EBC readings and beta measurement may be misleading. I know a pain in the >>> Did you check the transistors out of circuit?

Yes. Dc coupled.
Yes I measured each transistor out of circuit.

Open junction looks correct because another junction is turning on

 

[Johnny1234]

Without reviewing all previous posts, have you checked all the supporting circuitry.

 

[Johnny1234]

The Peak meter can be wrong as beta goes down as the current goes up. The Peak meter is probably not supplying a high enough current to be accurate for the circuit.

 

[dmeek]

Please post a photo of the foil side.

 

[ccaudle]

I measured each transistor out of circuit.

Assuming that the protection circuitry disconnects the output with relays or similar, and doesn't just shut off the power supply, the most sure method is measure the voltages across each transistor junction, base to collector, base to emitter, emitter to collector. One of the transistors will likely show up as not biased on correctly, or will have drifted so far that the input pair is no longer matched closely enough to work at DC. Feedback should get the output back to reasonable if one of an input pair has just drifted a little bit, but something like excessive bias current creating DC offset base-to-base in the input pair is outside of the feedback loop and cannot be corrected, you will always get DC input offset multiplied by closed loop gain at the minimum.

 

[Johnny1234]

well put ccaudle.

 

[pucho812]

Without reviewing all previous posts, have you checked all the supporting circuitry.

Yes. I can bring you up to speed if you like?

 

[pucho812]

Will have more on this tomorrow. As I got to get it finished.
So far if I move known good discrete sections around they work in other spots which just confirms what I have been chasing on this, that the discrete sections that are deemed faulty are misbehaving. Not much of an update, but it's nearly impossible to address these in circuit due to layout. any more desoldering I risk damage to the motherboard. luckily it's only a 2 layer.

 

[Johnny1234]

You need to do a ccaudle says, get a sheet of paper and take down and record all the measurements suggested on a good unit and compare with a defective unit.

 

[pucho812]

You need to do a ccaudle says, get a sheet of paper and take down and record all the measurements suggested on a good unit and compare with a defective unit.

yes I am on it.

 

[pucho812]

i have to be doing something really stupid. I also can’t continue to desolder and resolder the discrete opamps in and out for testing.
I went over the discrete opamps again, yesterday.
All resistors check out and continuity from solder tabs to next component follow the traces exactly. The one ceramic cap to ground is working.
Transistors test out as working.

I now have 2 which when in circuit have dc on their input pin.
I now have 2 which seem to work all be it not correctly. These two show me level on input and output that is about 3dB off from the working unit.
Nothing like chasing your own tail.

Known good discrete in the same positions behave as they should. But I proved that the other day.
The hard part here is that there is not room to work on this things.

 

[ccaudle]

I also can’t continue to desolder and resolder the discrete opamps in and out for testing.

Would be useful to build up a socket where you can plug in an op-amp module to test before installing back into the full system.