My studio Mic compressor build

[jfrabat]

I don't think it's silly either.    Having a real box means you can use it with other "real" equipment.

Thanks, Rob!

If you want to check in-circuit measurements you could also measure the circuit around IC6b and IC6c.  It's virtually identical.

Lost me there...

One obvious thing we haven't done is simply replace IC5!   I mean it's easy to do and maybe save a lot of headaches.

Just tried it.  Did not measure voltagees, but LED did not change (regardles of pot positions).

Another thing to try is to solder a small cap, say 10pF,  across pin 1 and pin 2 of IC5a.  If you don't have 10pF anything upto 33pF.   And if you don't have 10pF to 33pF, even 100pF is OK for a test.

Well, i screwed up trying this one...  Put the 10pF cap in the socket and inserted the IC back in (no issues doing that).  But apparently, putting the IC backwards is bad if you fire up the circuit...   Pfffft....  smoke...  I hate these design flaws in these circuits! 

So, a THIRD IC went in.  Hope I did not fry anything else!  Compressor seems to be working fine (at least LEDs are working fine!  Did not plug it to an amp).

[11-90-an]

Lost me there...

I think Rob means that the voltages in those 2 op-amps are identical, or somewhat the same... i think 

THIRD IC went in.

You surely must have a large stock of these ICs, don't you...

[jfrabat]

I think Rob means that the voltages in those 2 op-amps are identical, or somewhat the same... i think 

Ah, that makes sense...

You surely must have a large stock of these ICs, don't you...

Still got about 4 or 5 left...  When you burn as many as I do, you better have some stock!   

[Rob Strand]

Well, i screwed up trying this one...  Put the 10pF cap in the socket and inserted the IC back in (no issues doing that).  But apparently, putting the IC backwards is bad if you fire up the circuit...   Pfffft....  smoke...  I hate these design flaws in these circuits! 

Oh well, that's life.

On a few boards I had to fix for friends they put the sockets in reverse like that.   I marked pin 1 with liquid paper because I can't stand boards asking for ICs to go in the wrong way  .    It's going to happen!

[jfrabat]

Socket is oriented correctly; the error was 100% the interface between the chair and the circuit.

So, what do I try next? 

[Rob Strand]

So, what do I try next? 

So adding the small cap did nothing?
and replacing IC5 did nothing?

One thing we haven't explicitly checked is that pin 3 IC5 is connecting to ground.    The voltages look like it is grounded but maybe it's actually not connecting.   Make sure it gets to the top of actual IC pin, not just the socket.

Probably good idea to do the same check on the power pins.  I'm pretty sure these are OK.   


If you don't find any of the above connection problems, at this point I'd probably do the following:

- Solder a 100k resistor to IC5 pin 2

- wire the other side of the 100k to +15V
- measure the voltages:
  IC5 pin 14         expect - 7.1V 
  IC5 pin 1           expect -5.2V
  D3 cathode       expect -2.4V
  D4 anode          expect -4.7V

- now move the connection on the 100k from +15V to -15V
- re-measure the voltages
  IC5 pin 14        expect -7.1V
  IC5 pin 1          expect +7.7V
  D3 cathode      expect +7.1V
  D4 anode         expect 0V


All the above is with no signal.

[Edit: fixed first set of expected voltages]

[jfrabat]

So, what do I try next? 

So adding the small cap did nothing?
and replacing IC5 did nothing?

OK, I did replace the IC and install the cap.  Here are the voltages with those changes:

IC5 Pin 14:
No Signal: -12.94V
Signal: -13.4V

IC5 Pin 13:
No Signal: -4.25V
Signal: -4.82V

IC5 Pin 10:
No Signal: -11.76V
Signal: -11.751V

IC5 Pin 9:
No Signal: -12.84V
Signal: -13.26V

IC5 Pin 1:
No Signal: +2.19V
Signal: -2.00V

IC5 Pin 2:
No Signal: +1.4mV
Signal: +1.4mV

IC5 Pin 3:
No Signal: 0V
Signal: 0V

D3 Cathode:
No Signal: +5.59V
Signal: +3.69V

D3 Anode:
No Signal: +2.65V
Signal: -1.98V

D4 Cathode:
No Signal: +2.65V
Signal: -1.98V

D4 Anode:
No Signal: -5.21V
Signal: -8.51V

IC5 Pin 12:
No Signal: -6.02V
Signal: -8.51V

R64/C8:
No Signal: +1.4mV
Signal: +1.4mV

IC5 Pin 4: +15.23V
IC5 Pin 11: -14.85V

All measurements taken directly at the leads (not the solder or the sockets).

- Solder a 100k resistor to IC5 pin 2
- wire the other side of the 100k to +15V

Done.  Actual measurement is 96.3K.  I actually put it in the socket from pin 2 to pin 4.

- measure the voltages:
  IC5 pin 14         expect - 7.1V 
  IC5 pin 1           expect -5.2V
  D3 cathode       expect -2.4V
  D4 anode          expect -4.7V

  IC5 pin 14         expect - 7.1V        Actual: -12.97V
  IC5 pin 1           expect -5.2V        Actual: +1.270V (fluctuates about 0.02V)
  D3 cathode       expect -2.4V        Actual: +4.62V
  D4 anode          expect -4.7V        Actual: -6.07V

- now move the connection on the 100k from +15V to -15V
- re-measure the voltages
  IC5 pin 14        expect -7.1V
  IC5 pin 1          expect +7.7V
  D3 cathode      expect +7.1V
  D4 anode         expect 0V

  IC5 pin 14        expect -7.1V         Actual: -12.11V
  IC5 pin 1          expect +7.7V        Actual: +3.47V
  D3 cathode      expect +7.1V        Actual: +5.50V
  D4 anode         expect 0V             Actual: -2.18V

For all the measurements, I left the ceramic cap in between pins 1 and 2.

Obviously something is way off...

EDIT: Re-did the values.  It was called to my attention that not connecting the meter negative lead to GND gives inaccurate readings...

[11-90-an]

I think there's a micro-short somewhere...

how many pcbs of this compressor do you have, Felipe?
I assume 3 since you got them from OshPark...

Maybe populate another one?

[Rob Strand]

I think there's a micro-short somewhere...

how many pcbs of this compressor do you have, Felipe?
I assume 3 since you got them from OshPark...

Maybe populate another one?

Maybe I can go even further.   I have  been simulating faults to see what can give the behaviour we are seeing.

For the normal circuit  + no signal  + *without* my added 100k test:

If I connect pin 12 IC5d  to -VE through a 100k resistor I get very similar behaviour.    The 100k is not the 100k's I mentioned about.  It's a mysterious 100k which is on the PCB *somewhere*, an unintentional short.

So what might be happening is pin 12 IC5C is shorting to a 100k resistor on the PCB and the other end of that 100k resistor is connected to -VE (or to an opamp output which is at -VE).

So maybe set your DMM to continuity and look for shorts from pin 12 IC 5d!  The short might not be to an actual 100k, the short might just look like it.  The idea is to find an unintended connection.

[Rob Strand]

  IC5 pin 14         expect - 7.1V        Actual: -12.97V
  IC5 pin 1           expect -5.2V        Actual: +1.270V (fluctuates about 0.02V)
  D3 cathode       expect -2.4V        Actual: +4.62V
  D4 anode          expect -4.7V        Actual: -6.07V


  IC5 pin 14        expect -7.1V         Actual: -12.11V
  IC5 pin 1          expect +7.7V        Actual: +3.47V
  D3 cathode      expect +7.1V        Actual: +5.50V
  D4 anode         expect 0V             Actual: -2.18V

When I add to 100k and simulate the fault on this case:
normal circuit  + no signal  + *with* my added 100k test

I don't get the same voltages as the ones you just measured.


So maybe in the right track but not exactly.

[jfrabat]

I think there's a micro-short somewhere...

how many pcbs of this compressor do you have, Felipe?
I assume 3 since you got them from OshPark...

Maybe populate another one?

Actually 5; they sent me 2, but because they sent me 2 (one was damaged during fab), they sent me 3 extra...  I think I have enough components to do so.  I just need to flip the polarity of the 2 IC's that I know are reversed again.

  IC5 pin 14         expect - 7.1V        Actual: -12.97V
  IC5 pin 1           expect -5.2V        Actual: +1.270V (fluctuates about 0.02V)
  D3 cathode       expect -2.4V        Actual: +4.62V
  D4 anode          expect -4.7V        Actual: -6.07V


  IC5 pin 14        expect -7.1V         Actual: -12.11V
  IC5 pin 1          expect +7.7V        Actual: +3.47V
  D3 cathode      expect +7.1V        Actual: +5.50V
  D4 anode         expect 0V             Actual: -2.18V

When I add I simulate the fault on this case:
normal circuit  + no signal  + *with* my added 100k test

I don't get the same voltages as the ones you just measured.


So maybe in the right track but not exactly.

You just went a couple of levels above my head!

[jfrabat]

You know what, I have the boards, and I have all the components.  I even have an extra enclosure...  And I also wanted to have 2 of these.  And the power supply should be able to drive about 3 of these!  So, I will populate another board and make another compressor.  Then I can compare voltages, and it SHOULD be easier to find the issue, right?

[Rob Strand]

You just went a couple of levels above my head!

I made a typo which didn't help.

What I do is enter the circuit into a circuit simulator program (LTspice) then deliberately add faults which aren't on the original circuit.  That's done by adding extra components.

Using that method  I can see what faults match the voltages you measure on the actual circuit and "guess" where the problem might be.

When I add a "fault" which is 100k from IC5 pin 12 to -VE I see similar voltages to you no-signal measurements.    So shorts on pin 12 IC5 might be somewhere to start looking.    When I add the same fault and *also* add the 100k to +VE or -VE in the additional tests you didn't I don't quite get the right answer.   That means we are close to finding the fault but it's not exactly like adding 100k from pin 12 IC5 to -VE.

[Rob Strand]

When i look at the waveforms you posted a long time back they look pretty weird.  I can't seem to match those waveforms.

[jfrabat]

When i look at the waveforms you posted a long time back they look pretty weird.  I can't seem to match those waveforms.

Just to be sure, I should ground the scope to the GND for the wavefor reading, right?  I will retake the signal tonight.

[Rob Strand]

Just to be sure, I should ground the scope to the GND for the wavefor reading, right?  I will retake the signal tonight.

Yes.

I'm not sure if it's going to help narrow down the problem though.

Perhaps if you removed IC5 from the socket then measured following the DC voltages:
- IC5 pins 1, 2, 3
- IC5 pins 12, 13, 14
- D3 cathode
- IC5 pin 9

All the voltages should be zero.

If you find a non-zero voltage that means that point has a short to something else.

For each point with a non-zero voltage, follow that tracks going to that point on the PCB.
Measure the voltage on tracks that run next to the track with the non-zero voltage and see if it has the same voltage.
If you find a match turn off the unit then use the continuity tester on your DMM to confirm a short (or high-resistance link).

[jfrabat]

Maybe populate another one?

Started work on that tonight.  Got all the diodes and most of the resistors in.

Just to be sure, I should ground the scope to the GND for the wavefor reading, right?  I will retake the signal tonight.

Yes.

I'm not sure if it's going to help narrow down the problem though.

Perhaps if you removed IC5 from the socket then measured following the DC voltages:
- IC5 pins 1, 2, 3
- IC5 pins 12, 13, 14
- D3 cathode
- IC5 pin 9

All the voltages should be zero.

If you find a non-zero voltage that means that point has a short to something else.

For each point with a non-zero voltage, follow that tracks going to that point on the PCB.
Measure the voltage on tracks that run next to the track with the non-zero voltage and see if it has the same voltage.
If you find a match turn off the unit then use the continuity tester on your DMM to confirm a short (or high-resistance link).

OK, took out IC5.  First thing I notice is that now DEPTH actually influences the LED indicator.

Also, no signal is used in these measurements.

- IC5 pin 1: -2.3 mV
- IC5 pin 2: 1.3 mV
- IC5 pin 3: 0V
- IC5 pin 12: -0.2 mV
- IC5 pin 13: -1.6 mV
- IC5 pin 14: -3.2 mV
- D3 cathode: 0V
- IC5 pin 9: -6.4 mV

None are particularly far off from 0... That's good, right?  I mean bad for finding the issue, but it should be like this, right?  Or should I back trace that -6.4 mV, for example?

[11-90-an]

yea that back-trace would be useful...

forgive me for my ignorance, but is the depth pot supposed to control the LED indicator?

[Rob Strand]

Also, no signal is used in these measurements.

None are particularly far off from 0... That's good, right?  I mean bad for finding the issue, but it should be like this, right?  Or should I back trace that -6.4 mV, for example?

I'd say those measurements show it is unlikely there is a short to something else.

Using the DMM continuity tester, check there is no shorts to ground on any of these pins,
- IC5 pins 1, 2, 3
- IC5 pins 12, 13, 14
- D3 cathode
- IC5 pin 9

An also check this pin *is* connecting to ground,
- IC5 pins 3


If that all checks out,  we have
- IC5 OK
- resistor values look correct
- no shorts
- continuity of parts back to the IC check out OK so no open circuits either.

Everything checks out OK.   But we can clearly see it doesn't work.

So when that happens I'm thinking oscillations.   

So things we can do,

1) I have a feeling you need to re-do the test with the 100pF cap soldered between IC5 pin 1 and pin 2.

2) Another thing you can try is soldering a 100nF cap across the power pins of IC5, pins 11 and 4.    If that's no good maybe one 100nF from pin 11 to ground then another from pin 4 to ground.

Do the changes and check the DC voltage on pin 14 IC 5, maybe pins  1, 12, 13  also.   No signal is fine.

If no luck maybe look pins 1 and pin 14 with the oscilloscope to see if you can find any oscillations.

It's starting to be a tough one to work out.

[Rob Strand]

I'd say those measurements show it is unlikely there is a short to something else.

I was thinking it is still possible there is a short to some negative voltage.

When IC5 is in it creates some negative voltages but then IC5 is removed then those voltages are gone.
So we need IC5 out to see them but when we pull IC5 out they are gone!

So that got me thinking  maybe check these points
- IC5 pins 1, 2, 3
- IC5 pins 12, 13, 14
- D3 cathode
- IC5 pin 9

are not shorting to other pins on IC5.   I'd be really looking at pins 5, 6,7 if IC5 first.