Mackie PPM1008 Powered Mixer

[lowell]

For any Mackie repairmen out there. I'm not posting a schematic as I'm not sure LOUD Tech would be ok w/ that.

Channels 2 and 3 have a bad hum to them. The hum goes away when the level is turned down. I signal traced it and the hum does not seem to appear before the output of opamp U74B. There is no hum on either of this opamp's inputs or before this point in the circuit. I replaced that opamp and it's still humming. Both +15v/-15v rails seem good at ~14.78v +/-. Any thoughts on where to go w/ this?

[Martin Pedder]

I don't have the circuits for this, but as nobody else has chipped in, I'll do what i can. You are welcome to email the schematic to mpedder@lineone.net

I assume that the other channels are all OK.

First, I would scope the rails to make sure that they are clean. You don't need much ripple on some mic amps. I would expect this to affect all channels however, unless there are duff caps on the two with problems. My hunch is that this will not be the case.

If you do have power rail ripple, the most likely thing is a failed diode in the bridge. That is assuming a conventional power supply. Remember I am working blind.

Hum is often picked up by induction. We have all been thrown by a lead picking up hum from a mains transformer. Look to see if the hum varies as you move the board or move a metal plate near the board. I have been fooled by the transformer in my soldering iron on more than one occasion.

Hum can sometimes be caused by a high frequency oscillation out of audible range and demodulated to something you can hear like an AM radio. Check that none of the chips are hooting. A few cm of wire on the end of a scope probe can act as an aerial to help locate the source of an oscillation.

[Martin Pedder]

Hi again. Can you elaborate on the problem. eg,
What affects the hum? EQ, Insert etc.
Is it on channels 1 and 4 at all?
When you say "The hum goes away when the level is turned down", does this mean the respective channel level controls?
Does "I signal traced it" mean that you have an oscilloscope?
If the hum is on the channel level control, it presumably means that it is also on the Efx, Mon1 and Mon2 busses too. eg, if you listen to the Mon1 out, can you hear the hum and adjust it on channel 2's Mon control (assuming all other Mon controls are down).

I have an idea where this may be going, but we need to be very logical in our approach. Please give me a guide as to the level of your electronics knowledge and the test equipment available. I don't want to talk down to you, but don't want to talk over your head either.

The more tests you can do and the more detail you can give me, the easier this will be.

[Enzo]

Hi MArtin. Lowell has been with us a while, has almost 2000 posts here. He's working in a shop. WHat he generally needs is help troubleshooting.

[Martin Pedder]

There is a block diagram in the Owners Manual on page 29. I have attached it here:
PPM1008 Block Diagram.pdf
While we narrow down exactly where the problem lies, this is easier than the schematics (which I now have).

You do not need any speakers attached to the mixer's speaker outputs for these tests, but a simple powered speaker (like a powered computer speaker or hi-fi amp will do nicely for our tests..

Testing before the insert point
Let's conclusively eliminate the preamps on the first 4 channels first. If you use a mono jack plug connected to an amp/speaker, you can listen at the insert point on each channel. As the line input is muted when no input is connected, you should just hear the preamp hiss at the insert point. The gain switch should make the hiss increase and decrease. Now feed tone (or even music) into each line input and check that all of the first 4 channels have the same gain (ie, they are all at about the same level) at the the insert point. The level pot and the EQ controls will have no affect listening at the insert point as they are both post insert. The compressor however is before the insert. If you increase the level of your tone, you should see the overload LED light and the compressor (threshold) pot should affect the level. What you are doing here is proving that all 4 input stages are behaving the same. Assuming they are, we move on the testing after the insert.

Testing after the insert point
We don't need to listen for this test. Assuming all the inputs were OK, set all the EQ controls (Ch 1-4) to the centre position. Set all MON and EFX pots to zero. See if there is any level difference between the 4 channels. Set the Main Master graphic to mid point on each slider, the "FX to Main" to zero, and the Main Master level to mid point. Feed tone to channel 1, adjust the channel level control to read something sensible on the meters. Then feed tone to channel 2, 3 and 4 in turn and adjust their level pots to get the same level on the meters. Did any level controls finish at a different position to the others?

Finally, for now; Use your amp/speaker to listen on the Main output sockets and check a few things for me.
Does the BREAK button kill the hum?
Does the LEVEL pot on Ch 2 and 3 affect the hum and can it eliminate it completely?
Does the COMP pot on Ch 2 and 3 affect the hum?
Does the hum sound like a mains hum?
Is it a smooth hum or a raspy buzz sound?
If you turn up the FX TO MAIN pot, does the hum get affected by FX pots on ch 2 and 3?
Now listen to the MON1 output. Is there hum there? Do the MON1 pots on ch 2 and 3 affect it? (level has to be up for FX/Mon2 to work).

When we have the answers, we can progress to component level. Do you have a scope? If not, can you rig up a probe made of a capacitor (anything from 1uF to 470uF will do) into an amp with a volume control. We can use this to listen at various points in the circuit. Something like this will do.

Attach the negative end of the capacitor to the lead. The screen (or black croc clip) will go to the mixer chassis. we will probe with the + end of the capacitor and adjust the amp's volume as required.

[lowell]

Tried to upload the hum sample but cannot figure out what type of video to upload. m4v doesn't work. the hum sounds lower than 50hz. It's NOT 60hz. Here are the answers to your questions Martin. Thanks for your help.

Testing before the insert point
Chnl 1 - good
chnl 2 - peak light does NOT go on, and signal is weak
chnl 3 - peak light goes on but signal is weak
chnl 4 - good

Testing after the insert point
same results as above but results are displayed on the meter

Finally, for now; Use your amp/speaker to listen on the Main output sockets and check a few things for me.
Does the BREAK button kill the hum?
Yes

Does the LEVEL pot on Ch 2 and 3 affect the hum and can it eliminate it completely?
yes and both level pots completely kill hum when turned down.

Does the COMP pot on Ch 2 and 3 affect the hum?
yes

Does the hum sound like a mains hum?
sounds lowere than 50hz, tried to record it but couldn't upload it, it's raspy in quality

If you turn up the FX TO MAIN pot, does the hum get affected by FX pots on ch 2 and 3?
no it does not

Now listen to the MON1 output. Is there hum there? Do the MON1 pots on ch 2 and 3 affect it? (level has to be up for FX/Mon2 to work).
no there is no hum

[Martin Pedder]

Just off to bed 00:40 here and up at 06:00 for a very long day with 450 miles driving.

I am intrigued with this, but we should be able to suss it out now. I will get back to you, but not tomorrow I suspect, probably on Thursday.

Of course, if anyone else wants to chip in first; Go for it.

Martin

[lowell]

Thanks Martin.

[Martin Pedder]

Testing before the insert point
Chnl 1 - good
chnl 2 - peak light does NOT go on, and signal is weak
chnl 3 - peak light goes on but signal is weak
chnl 4 - good

Excellent. This means that the problem is before the insert. This eliminates the EQ section and fader as they are both post insert.

Does the LEVEL pot on Ch 2 and 3 affect the hum and can it eliminate it completely?
yes and both level pots completely kill hum when turned down.

This test, combined with the one above, ruled out a suspicion I had that there might be a bus or summing mixer problem.

Now listen to the MON1 output. Is there hum there? Do the MON1 pots on ch 2 and 3 affect it? (level has to be up for FX/Mon2 to work).
no there is no hum

This puzzles me. I would expect to get the hum here too. Referring to the Block Diagram I posted a few days ago; If the hum is present at the insert, then when the insert is unplugged, it gets through the EQ and fader onto the main L+R busses. It surely must also be able to get on to the Mon1 bus. To hear this, you should plug an amp into the Mon1 Send. The Mon1 knob on Ch2 or 3 would need to be turned up and so would the Monitor 1 Master Level knob.

Does the COMP pot on Ch 2 and 3 affect the hum?
yes

The compressor is the last thing before the insert point. It is even after the overload LED. The fact that both problem channels are low level as well as humming would imply that the problems are related. Either we have two channels with similar faults or there is something common to both channels.

There are chips like U1, U63 and U69 that are shared by channels 1 and 2. Similarly channels 3 and 4 share the equivalent chips U68, U80 and U81. Most of the electronics for channel 2 has nothing to do with channel 3, except for one thing. The compressor uses a Quad VCA chip SSM2146 called U3. This chip is shared by all of the first 4 channels and so becomes our prime suspect. It lives between the Comp pots of Ch 2 and Ch 3.
The datasheet is here:
SSM2164S datasheet(1/12 Pages) AD | Low Cost Quad Voltage Controlled Amplifier

Before we go there however, I want to check some easier things:

Does the phantom 48V switch affect the hum?

Does the gain switch on the problem channels affect the hum?

Does shorting the mic input by putting a wire between pins 2 and 3 of the Mic XLR on the problem channels affect the hum?

Can you check the outputs of all these op-amps with a DC volt meter, measuring to ground. We should expect close to zero volts in every case.
Ch1 (for reference)
U5-1, U5-7, U7-1, U63-1
Ch2
U71-1, U71-7, U7-7, U63,7
Ch3
U76-1, U76-7, U77-1, U80-1
Let me know if any of these are not zero.

Can you measure the DC Voltage to ground of the bases of the transistors in the pre-amp. The easiest place to do this is on the diodes just below the Low EQ pot on channels 2 and 3. Measure:
Ch1 (for reference)
where D6 meets D7
where D8 meets D9
Ch2
where D35 meets D34
where D37 meets D36
Ch3
where D43 meets D42
where D45 meets D44
Let me know if anything seems unusual on the faulty channels.

Assuming we find no problems with any of the above, we need to look at the circuitry around the compressor and it's VCA chip. Some of these may not be zero, but compare what you get on Ch 1 with Ch2 and Ch3.
Ch1 (for reference)
U69-1, U1-1, U3-5
Ch2
U69-7, U1-7, U3-4
Ch3
U81-1, U68-1, U3-12
Let me know if anything seems unusual on the faulty channels.

Martin

[lowell]

Sorry Martin you are correct there IS hum out of the monitor 1 send jack. My bad.

Before we go there however, I want to check some easier things:

Does the phantom 48V switch affect the hum?
Yes it reduces the hum when engaged. In fact after engaging it and then unengaging there is now a TON of hiss on top of the hum on Channel 2.

Does the gain switch on the problem channels affect the hum?
No

Does shorting the mic input by putting a wire between pins 2 and 3 of the Mic XLR on the problem channels affect the hum?
Yes it greatly reduces it but doesn't get rid of it.

Can you check the outputs of all these op-amps with a DC volt meter, measuring to ground. We should expect close to zero volts in every case.
Ch1 (for reference)
U5-1 1.6v
U5-7 1.6v
U7-1 0v
U63-1 0v

Ch2
U71-1 1.8v
U71-7 13.7v
U7-7 11.9v
U63-7 -11.7v

Ch3
U76-1 1.5v
U76-7 1.5v
U77-1 0v
U80-1 14v

Let me know if any of these are not zero.
So it seems there are some bad ICs in channel 2 and 3. Namely U71, U7, U63, U80.

Can you measure the DC Voltage to ground of the bases of the transistors in the pre-amp. The easiest place to do this is on the diodes just below the Low EQ pot on channels 2 and 3. Measure:
Ch1 (for reference)
where D6 meets D7 10mv
where D8 meets D9 10mv
Ch2
where D35 meets D34 345mv
where D37 meets D36 8.4v
Ch3
where D43 meets D42 10mv
where D45 meets D44 10mv

Let me know if anything seems unusual on the faulty channels
So channel 2 transistors seem off too.

[Martin Pedder]

Does the phantom 48V switch affect the hum?
Yes it reduces the hum when engaged. In fact after engaging it and then unengaging there is now a TON of hiss on top of the hum on Channel 2.

The 4 diodes D34, D35, D36, D37 are there to protect the preamp that follows from surges that go beyond the power rails (in this case +/- 15V). Switching 48V phantom on and off will produce such surges. Check each diode is being a diode. Use a multimeter on a 1 kOhm or 2 KOhm scale. You would typically see 300 to 500 Ohms with the probes one way across the diode and a high impedance when you swap the probes round. Some meters have a diode tester that can measure the voltage drop of a diode. This is 0.6 to 0.7V. I have seen diodes like this fail in mixers. A leaky diode in channel 2 could be the cause of the problem, or an open circuit diode could have let a damaging spike through to the transistors or op-amp.

Does the gain switch on the problem channels affect the hum?
No

I think the hum you hear is on the power rails and getting through only because some of the chips outputs are nearly at one rail or the other. If the hum had changed in frequency, we might have had an oscillation problem.

Does shorting the mic input by putting a wire between pins 2 and 3 of the Mic XLR on the problem channels affect the hum?
Yes it greatly reduces it but doesn't get rid of it.

Again this supports the theory that it is not an oscillation causing the hum. I can see the hum reducing on Ch2, but I am not sure I would have expected to see a difference on Ch3. No matter, it is not important.

Can you check the outputs of all these op-amps with a DC volt meter, measuring to ground. We should expect close to zero volts in every case.
Ch1 (for reference)
U5-1 1.6v
U5-7 1.6v
U7-1 0v
U63-1 0v

Ch2
U71-1 1.8v
U71-7 13.7v
U7-7 11.9v
U63-7 -11.7v

Ch3
U76-1 1.5v
U76-7 1.5v
U77-1 0v
U80-1 14v

Let me know if any of these are not zero.
So it seems there are some bad ICs in channel 2 and 3. Namely U71, U7, U63, U80.

Not necessarily. Interestingly, although you appeared to have similar symptoms on the 2 faulty channels, it looks like they actually have quite different faults. Let's look at what the results tell us.
Ch2
U71-7 is clearly wrong. The question is whether it is U71 itself or something feeding it causing the 13.7V. We look at the transistors that feed it lower down this page. as with any circuit with a feedback loop, it is hard to tell what is cause and what is effect.

U7 however is probably OK. It is just amplifying the difference between U71-1 and U71-7. You would expect U7-7 therefore to show a big output as it has a big differential voltage on pins 5 and 6.

U63 is also probably OK. If you follow the circuit from U7-7, it feeds U63-6 which is the inverting input. The 11.9V from U7 is correctly being inverted to -11.7V. OK, this voltage should not be there but the chip is doing its job with the signals it is getting.

Ch3
I agree U80-1 should not be 14V, especially when U77-1 was zero. It could be getting an incorrect input from U3-12 but it is most likely to be U80 itself that is faulty. U80 needs changing. Are these conventional (through-the-board) or surface mount ICs? If it isn't an SMD, you could fit a socket in case it need changing again.

Can you measure the DC Voltage to ground of the bases of the transistors in the pre-amp. The easiest place to do this is on the diodes just below the Low EQ pot on channels 2 and 3. Measure:
Ch1 (for reference)
where D6 meets D7 10mv
where D8 meets D9 10mv
Ch2
where D35 meets D34 345mv
where D37 meets D36 8.4v
Ch3
where D43 meets D42 10mv
where D45 meets D44 10mv

Let me know if anything seems unusual on the faulty channels
So channel 2 transistors seem off too.

Ch2
I don't think that the 345mV is significant. The 8.4V is though.
Measure the voltage on U71-5 (=U71-3). This is derived from a potential divider made of R564 and R565 from the +/- 15V rails. It should be about -12V. If it isn't, compare the same point on channel 1 U5-5 (=U5-3).
Measure the voltages on U71-2, and U71-6. These are the collector voltages from Q29 and Q30. The op-amp is looking at the difference between that and the -12V reference.

I suspect that it is U71 that has failed. It is also possible that Q30 has died causing U71-6 to have the wrong voltage on it. Check Q30 for two diodes B-E and B-C when switched off. Let me know if not sure how to do this.

This circuit of Ch2 shows your measured voltages in red and the correct or my calculated voltages in green.

[lowell]

Martin,
This is very exciting to me. Thanks for the help and being so concise. I'm learning alot about DC offset in cascading opamp topology here.

Okay so...

All opamps are SMD. No biggie. I'll replace U80 in Chnl 3.

Measure the voltage on U71-5 (=U71-3). This is derived from a potential divider made of R564 and R565 from the +/- 15V rails. It should be about -12V. If it isn't, compare the same point on channel 1 U5-5 (=U5-3).

This is indeed spot on at -12v.

Measure the voltages on U71-2, and U71-6. These are the collector voltages from Q29 and Q30. The op-amp is looking at the difference between that and the -12V reference.

Q29 C: -11.8v
Q30 C: -12.34v

The difference between -12v reference and Q30's -12.34v is certainly not the measured 13.7v. It seems U71 is bad. ?

I suspect that it is U71 that has failed. It is also possible that Q30 has died causing U71-6 to have the wrong voltage on it. Check Q30 for two diodes B-E and B-C when switched off.

These junctions test good with my diode function on my meter.

[lowell]

Martin I swapped U80 and still the hum persists.

You said:

Ch3
I agree U80-1 should not be 14V, especially when U77-1 was zero. It could be getting an incorrect input from U3-12 but it is most likely to be U80 itself that is faulty. U80 needs changing. Are these conventional (through-the-board) or surface mount ICs? If it isn't an SMD, you could fit a socket in case it need changing again.

So maybe we need to investigate U3-c?

[Martin Pedder]

Martin,
This is very exciting to me. Thanks for the help and being so concise. I'm learning alot about DC offset in cascading opamp topology here.

This forum is all about cascading knowledge. I am happy to pass on anything I understand better than you and I am sure the reverse is also true. I am well aware that there are many people here whose have far better electronic knowledge than me. That it why I joined, so I can pick some of those brains. To anyone reading this thread, "Please jump in, especially if I am missing anything or getting it wrong".

Okay so...

All opamps are SMD. No biggie. I'll replace U80 in Chnl 3.

Measure the voltage on U71-5 (=U71-3). This is derived from a potential divider made of R564 and R565 from the +/- 15V rails. It should be about -12V. If it isn't, compare the same point on channel 1 U5-5 (=U5-3).

This is indeed spot on at -12v.

Measure the voltages on U71-2, and U71-6. These are the collector voltages from Q29 and Q30. The op-amp is looking at the difference between that and the -12V reference.

Q29 C: -11.8v
Q30 C: -12.34v

The difference between -12v reference and Q30's -12.34v is certainly not the measured 13.7v. It seems U71 is bad. ?

Not quite. Let's cover a little theory first.

Op-amps do look at (compare) the difference between the two inputs. When being used as a "comparator" (ie when there is no negative feedback), if the + input is higher than the - input, the output will swing fully positive. If it is lower, then the output swings fully negative. They usually can't quite get to rail voltage, so when U71-7 is at 13.7V, that is about as close to the 15V rail as it can ever get. Comparators are therefore a bit like switches. The output is fully on or fully off.

What a comparator is actually doing is amplifying the difference between the inputs. Most op-amps have an open loop gain (ie no negative feedback) of 100,000 to 1,000,000. In other words, the tiniest difference between the input pins will cause the output to swing fully positive or fully negative.

They are used in this way in LED bargraph meters. If the signal is above a certain level, the red LED goes on. Each LED has its own comparator comparing the signal to a reference voltage for the LED. Here is a simple example. You can see the resistor ladder producing a different reference voltage for each comparator.


For amplifying audio signals, a designer wants to be able to swap one faulty chip, valve or transistor with a replacement and have the circuit still work. As all chips, valves or transistors will have slightly different gains, this could mean us having to recalibrate every time a component was replaced. What designers usually do is to choose a chip, valve or transistor that naturally has far more gain than actually required, and then control that gain with negative feedback. That way, if you are trying to double an input signal (gain of 2), you feed most of the output back to the input, out of phase, usually with a pair of resistors acting as a potential divider. That way, if the replacement chip, valve or transistor has a higher or lower gain than the original, it doesn't matter as the amplifiers overall gain is being controlled by the resistors.

With op-amps, the feedback is usually quite easy to spot. Look for any signal path from the output back to the inverting "-" input. Here, the signal goes in to the non-inverting "+" input so the output is in phase with the input. All of the output is also fed back to the inverting input so there is no gain. This is a unity gain buffer amplifier.


Here, you can see a potential divider feeding 1/10th of the output back to the inverting input. This amp has a gain of 10.

See https://www.circuitlab.com/circuit/m...ing-amplifier/ for the full explanation.

Op-amps are so versatile. There is loads of info on the net. This "Handbook of Operational Amplifier Applications" looks like a good start: http://www.ti.com/lit/an/sboa092a/sboa092a.pdf

Back to our circuit. The signal coming out of transistors Q29 and Q30 will be swinging around the -12V area as they are referenced to the -15V rail with resistors R561 and R562. So, the op-amps U71-A and U71-B are going to amplify those signals but using a -12V reference to give an output somewhat closer to zero. In practice, the signal seems to be centred around 1.6V on a healthy channel. They could have tweaked the reference resistors to make it closer to zero, but they just needed to ensure that the amplified signal could never swing anywhere near the power rails, and so cause distortion. They also needed some voltage there for the transistors to turn on. Notice there is no feed from the positive rail to run those transistors. They use the 1.6V that comes out of the chip as their power source. I am not sure what the signal amplitude will be at U-71-1 and U71-7, but it won't be more than a volt or two so safely away from hitting a rail.

What Mackie have done here is to use the transistors in the feedback loop of the op-amp. See how the signal from U71-1 gooes through R566, through Q29 and back to the inverting input on U71-2. I will be honest here. I do not fully understand how this works or why they have done it.

Anybody help?

What I do understand is the gain switch. As the two halves of the input signal are out of phase, the are simply shunting part of the signal from the hot leg to the cold leg by lowering the value of R601 by putting R563 across it.

I suspect that it is U71 that has failed. It is also possible that Q30 has died causing U71-6 to have the wrong voltage on it. Check Q30 for two diodes B-E and B-C when switched off.

These junctions test good with my diode function on my meter.

Here is the circuit again with your new voltages in it (in smaller red letters). Can you check the 0.9V and 9.0V I showed. I assumed these transistors are both turned on so the emitter voltages would be 0.6V higher than the bases. Best to check though.



The most likely candidates for causing our incorrect voltages on Q30 and U71-B are those components themselves. With Q30 metering OK, then U71 becomes prime candidate. There are a few other parts that could be causing the problem and worth checking first. Consider D36 for example. If that was leaky, that could explain the 8.4V on Q30-Base. Also, if phantom was turned on and C380 was breaking down, this could put 48V on to Q30-Base. This is worth checking. Assuming the phantom is actually 48V (and it often isn't) it is very close to the 50V rating of those caps (C379 & C380). If C380 had failed, it might have killed Q30. On the drawing, it shows the input sockets section in its own box. If they are on a separate board, you might be able to unplug it to see if the voltages on the chip comes good.

Martin I swapped U80 and still the hum persists.

You said:

Ch3
I agree U80-1 should not be 14V, especially when U77-1 was zero. It could be getting an incorrect input from U3-12 but it is most likely to be U80 itself that is faulty. U80 needs changing. Are these conventional (through-the-board) or surface mount ICs? If it isn't an SMD, you could fit a socket in case it need changing again.

So maybe we need to investigate U3-c?

Yes. Please can you take DC measurements around U80-A, U3-C, U68-A, U81-A? We will expect to see DC voltages on U68-A and U81-A as they aren't handling audio. U81-A is a precision rectifier that detects how loud the signal is. You will find it on page 88 of the "Handbook of Operational Amplifier Applications". It produces a varying DC voltage, proportional to the signal level, and this is smoothed by C436. This voltage is compared on U68-A with a reference voltage from the Threshold pot. The result is a DC control signal that controls the gain of the VCA (Voltage Controlled Amplifier) made of U3-C in the feedback loop of U80-A. The result is to reduce the gain of the signal whenever a loud peak of the music occurs above a level set by the threshold.

If you have a scope or even an analog voltmeter, (one with a needle), you could look at the VCA control signal on U68-1. It is much easier to see varying voltages on an analogue meter than to watch changing digits on a digital meter. If U3-C has failed, some of the DC control voltage might be getting out of its audio pins.

It could be worth plugging a stereo jack plug (with no lead) into the insert socket, just to completely isolate it from what follows.

If you still have DC on U80-A and have already changed the chip, it must either be coming from U3-C or U77-A. There is no capacitor blocking DC from U77-1, but you might see a voltage drop across R620. Again, if the compressor is on a separate board as the diagram implies, you might be able to unplug "COMP_IN_3" to see if the DC is before or after that point. If you can't then lifting one end of R620 would also isolate the two sections.

Martin

[lowell]

Martin,
Waiting for parts for Channel 2 so I'm just gonna focus on Channel 3.

Yes. Please can you take DC measurements around U80-A, U3-C, U68-A, U81-A? We will expect to see DC voltages on U68-A and U81-A as they aren't handling audio. U81-A is a precision rectifier that detects how loud the signal is. You will find it on page 88 of the "Handbook of Operational Amplifier Applications". It produces a varying DC voltage, proportional to the signal level, and this is smoothed by C436. This voltage is compared on U68-A with a reference voltage from the Threshold pot. The result is a DC control signal that controls the gain of the VCA (Voltage Controlled Amplifier) made of U3-C in the feedback loop of U80-A. The result is to reduce the gain of the signal whenever a loud peak of the music occurs above a level set by the threshold.

U80A
1: 14v
2: -280mv
3: 0v

U3C
10: 7.4v
11: 3.5v
12: -280mv

U68A
1: -3.5v
2: 6.4v
3: 0v with compressor pot down/400mv with compressor pot maxed up

U81A
1: 14v
2: 13.4v
3: 14v

It could be worth plugging a stereo jack plug (with no lead) into the insert socket, just to completely isolate it from what follows.

If you still have DC on U80-A and have already changed the chip, it must either be coming from U3-C or U77-A. There is no capacitor blocking DC from U77-1, but you might see a voltage drop across R620. Again, if the compressor is on a separate board as the diagram implies, you might be able to unplug "COMP_IN_3" to see if the DC is before or after that point. If you can't then lifting one end of R620 would also isolate the two sections.

There's 0v on U77A pin1 so it must be coming from the compressor circuit - U3-C?