My studio Mic compressor build

[11-90-an]

But TRIGGER pot is not working as intended

define "not working as intended"... 
does it change anything?

also... I'm not sure, but try checking this part...


make sure that R78 has a connection to -VE and lug 3 of the TRIGGER pot has continuity to lug 3 of the DEPTH pot... while your at it, check TRIGGER pot value too...

[jfrabat]

define "not working as intended"... 
does it change anything?

Sorry, I mean not working AT ALL.  It does absolutely nothing.

also... I'm not sure, but try checking this part...

Check values you mean?

make sure that R78 has a connection to -VE and lug 3 of the TRIGGER pot has continuity to lug 3 of the DEPTH pot... while your at it, check TRIGGER pot value too...

R78 shows 7.03V going to R79 and 1.5mV going to Pin 2 IC6.

WAIT, WAIT...  You are using an old schematic!  OK, never mind.  Resistor values were changed to match the original schematic. 

What you call R78 is now R70.  R70 is reading -15V on one side, and -12.18V on the other.
TRIGGER pot value is correct
Lug 1 of the TRIGGER pot has continuity to lug 1 on the DEPTH pot (had to switch them around for proper rotation).

Voltages for IC5c are above (I think your diagram says IC6c, but that is now IC5c).

[Rob Strand]

Pin 7: 14.74V

Pin 7 IC5 high means the unit thinks there's no signal.   Trigger won't do anything.

You need to check beyond Pin 7 IC5

You can check,
- Q7 emitter,  should be close to 15V;  it should be about 0.6V above Pin 7 IC5
- pin 14 IC6,  should roughly be same as pin 7 IC5.
- Wiper of PR9,  Depth pot
  Between 0V and voltage on pin 14 IC6, depending on PR9 position

So I suspect you are probably getting that far already since you quote 7.11V on the depth pot.

Next is  Q9.   That's probably where the problem is.

Check,
- When the depth pot is at 7.11V you should see about +0.15V on the base of Q9.
- You should see +0.65V to +0.7V on the emitter of Q9
- Measure the voltage *across* R90 (12k).   
  That needs to be 0V in order for the gate to be closed.

I think the problem is going to end up that R87 (1k) is too low.   That means Q9 cannot shut-off and the voltage across R90 cannot get to zero volts.   And that means the gate can never close.

Maybe try increasing R87 maybe 2k2.      Increasing R89 to 1.5k might also help.

[11-90-an]

WAIT, WAIT...  You are using an old schematic!

There's a new one? 
I'm using the one you posted on reply #13 in this thread:

EDIT: Here is the schematic:

If I may ask... what schem are you guys referring too..? 

[Rob Strand]

Maybe try increasing R87 maybe 2k2.      Increasing R89 to 1.5k might also help.

While those might work, perhaps decreasing R88 might be a better way of tuning when the gate closes.

That whole circuit around Q9 is a bit touchy.   

I checked the noise-gate design and I don't get it.  It looks like the noise-gate has gain. 
Does it sound like it has gain when you turn off and on the noise-gate bypass?
Don't worry about this   I realized I forgot to allow for the diodes on pin 15, 14, 13 inside the LM13600.

If I may ask... what schem are you guys referring too..?

I use that one and I also use the ETI article PDF file that Slowpoke posted in the old thread.  I find the original article easier to read, the Eagle schematic is a little bit blurry and I can't make out the designators sometimes.

[Rob Strand]

Here's the thread with the schematic and also the discussion on the errata,
https://www.diystompboxes.com/smfforum/index.php?topic=125167.msg1190505#msg1190505

Here's the link to the ETI article from that thread,
https://web.archive.org/web/20150408034353/http://www.spontis.se/pdf/eti_dicompgate.pdf

[jfrabat]

WAIT, WAIT...  You are using an old schematic!

There's a new one? 
I'm using the one you posted on reply #13 in this thread:

EDIT: Here is the schematic:

If I may ask... what schem are you guys referring too..? 

That's the correct one. Except 70 looks like 78 because of the fuzziness.  But the numbering is the same as the schematics in the article that Rob mentioned (except where it is wrong!).

You can check,
- Q7 emitter,  should be close to 15V;  it should be about 0.6V above Pin 7 IC5
- pin 14 IC6,  should roughly be same as pin 7 IC5.
- Wiper of PR9,  Depth pot
  Between 0V and voltage on pin 14 IC6, depending on PR9 position

IC5 Pin 7: 14.74V
IC6 Pin 14: 14.27V
Q7e: 15.27V
PR9 middle: 7.02V (after all the readings, including below, I tried moving it from side to side, I can go from 0 to 14.20V)

So these seem to check out...

Next is  Q9.   That's probably where the problem is.

Check,
- When the depth pot is at 7.11V you should see about +0.15V on the base of Q9.
- You should see +0.65V to +0.7V on the emitter of Q9
- Measure the voltage *across* R90 (12k).   
  That needs to be 0V in order for the gate to be closed.

PR9 middle: 7.02V
Q9b: 0.2V
Q9e: 0.731V
R90 (across): -0.234V

Q9 seems to check out, but there is an issue with R90...  What could be causing this?

EDIT: Another thing I noticed is that I am supposed to set RV2 until IC2 Pin 7 is precisely 0V.  But all I get is between 29 and 35mV as far as range (I cannot get it to 0V exactly).  I know this has to do with the gate...  Could that be an issue?

[Rob Strand]

PR9 middle: 7.02V
Q9b: 0.2V
Q9e: 0.731V
R90 (across): -0.234V

Q9 seems to check out, but there is an issue with R90...  What could be causing this?

EDIT: Another thing I noticed is that I am supposed to set RV2 until IC2 Pin 7 is precisely 0V.  But all I get is between 29 and 35mV as far as range (I cannot get it to 0V exactly).  I know this has to do with the gate...  Could that be an issue?

The voltage in Q9b is close to what *I* expected based on how I see the circuit as is.  That doesn't mean it works or is correct. I actually think there *is* a problem.   That's why you can't get the gate to close.

The reason you can't get the voltage across R90 to zero is because the voltage across the base and emitter of Q9 isn't low enough.
So at the moment we have Vbe =  0.731 - 0.2 = 0.531V.    We need to reduce that a 'bit'.   How much doesn't matter.   What's important is the voltage across R90.     (BTW, when you measure the voltage across R90 you need to put the + probe on the Q9 side to get a positive voltage measurement.).

So there's a number of ways to do it.     Reducing  R88 (270k) to 180k or 150k might be a good start.   However see details below.

I suggest this method:
- Set the Depth pot RV6 mid-way.     (The voltage on the Depth pot wiper will roughly be 7V.   Exact value not important.)
- Measure the voltage across R90
- Tweak (ie. reduce) the value of R88 until you get 0V across R90.

At this point I think we should try to get the voltage across R90 to zero when RV6 is set mid-way.    If you can't get zero (no mattter what R88 is set to) then 10mV to 50mV is OK but remember we want to get this low voltage when RV6 is mid-way and that's why we have to change R88.

So if you get that far, see how the Noise gate works.   If it still doesn't seem right   try adjusting RV6.       If you don't seem to be able to get it work work properly with RV6 then try increasing R88 up or down by two resistor steps (by two steps I means if it's 150k then you would got to 220k or 100k).   Try to work out if bigger or smaller R88 sounds better.


EDIT:

Oh more important.   The voltage across R90 is a way to measure what is going on.   When the voltage across R90 is zero the gate should close.   The goal is to get the gate to close when RV6 is mid-way by adjusting R88.    Measuring the R90 voltage allows you to find the point where the gate closes without overdoing it.

[jfrabat]

I screwed something up...  I changed the pot, but now, regardless of what pot I put in, I am getting 14.2V (don't remember the EXACT number, but it was above 14V) across R90.  Also, Q9b is at that voltage.  I have NO IDEA what I did, but screwed it up I did!

By the way, I measured the second board, with the 270K resistor in place, and it is at 9mV.  That pot is still missing some components, but they are basically LEDs and XLR connectors.  But, then again, I could not get IC2 Pin 7 to 0 because RV2 is not changing the voltage.

I am just giving up for tonight; I am tired (went to bed past 2am last night fixing my wife's PC) and tomorrow is a long day at work.  I will retake this tomorrow night!

[Rob Strand]

I screwed something up...  I changed the pot, but now, regardless of what pot I put in, I am getting 14.2V (don't remember the EXACT number, but it was above 14V) across R90.  Also, Q9b is at that voltage.  I have NO IDEA what I did, but screwed it up I did!

Maybe you fried Q9.   Measure all the voltages on Q9 with no signal.  That will give an ideal what going on there.

By the way, I measured the second board, with the 270K resistor in place, and it is at 9mV.  That pot is still missing some components, but they are basically LEDs and XLR connectors.  But, then again, I could not get IC2 Pin 7 to 0 because RV2 is not changing the voltage.

I guess the next question is does the voltage on IC2 change at all with the RV2 adjustment?   What range from one end of RV2 to the other?

I am just giving up for tonight; I am tired (went to bed past 2am last night fixing my wife's PC) and tomorrow is a long day at work.  I will retake this tomorrow night!

Maybe you need to take a break!

[11-90-an]

I am just giving up for tonight; I am tired (went to bed past 2am last night fixing my wife's PC) and tomorrow is a long day at work.  I will retake this tomorrow night!

Maybe you need to take a break!

Yup, yup. Don't overwork yourself... 

[Rob Strand]

I was looking at the Noise gate LED circuit to see if I could get an idea what voltages should come out of the Depth pot wiper.
It seems that in order to *turn off* the "Gate open" LED you pretty much have to have the Depth pot RV6 on full.   The LED circuit doesn't really to allow to Depth pot to be backed off (It' probably could be tweaked but we will leave it.)

Anyway after noticing that I read over the ETI article.   

I found two pieces of information that hint at how to set-up the noise gate:
- p 49:   The noise gate supposed to be set-up for 60dB attenuation.    From my calculations that means
  the voltage across R90 should be around 15mV to 16mV when the gate is closed.
- p52 says to set RV6 to fully clockwise.

So the update to tweaking the R88 value is as follows:
- Set the Depth pot RV6 to full.     (The voltage on the Depth pot wiper will roughly be +13V.   Exact value not important.)
- Measure the voltage across R90
- Tweak (ie. reduce) the value of R88 until you get 15mV to 16mV across R90.

[jfrabat]

Also, Q9b is at that voltage

I was actually measuring Q8b...  That is how tired I was!

Maybe you fried Q9.   Measure all the voltages on Q9 with no signal.  That will give an ideal what going on there.

Board 1:
Q9e: 0.723V
Q9b: 0.641V
Q9c: 0.675V

Q8e: 15.34V
Q8b: 14.67V
Q8c: 6.75V

I am getting 14.27V across R90

Board 2:
Q9e: 0.735V
Q9b: 0.203V
Q9c: -13.65V

Q8e: 15.35V
Q8b: 14.73V
Q8c: 15.34V

I am getting 186mV (and drops) across R90

I guess the next question is does the voltage on IC2 change at all with the RV2 adjustment?   What range from one end of RV2 to the other?

In board 1 (old one), right now, no change in IC2 Pin 7 with RV2.  It used to be a about 10mV before I screwed up.  Pin 7 measures 31mV.  On board 2, it also not moving.  Pin 7 is at 3.7mV.

So the update to tweaking the R88 value is as follows:
- Set the Depth pot RV6 to full.     (The voltage on the Depth pot wiper will roughly be +13V.   Exact value not important.)
- Measure the voltage across R90
- Tweak (ie. reduce) the value of R88 until you get 15mV to 16mV across R90.

Board 1:

Depth Pot: 14.19V
R90 (across): 14.28V

Board 1:

Depth Pot: 13.77V
R90 (across): ~30mV (about 3mV of oscillation)

[Rob Strand]

Board 1:
Q9e: 0.723V
Q9b: 0.641V
Q9c: 0.675V

Q8e: 15.34V
Q8b: 14.67V
Q8c: 6.75V

I am getting 14.27V across R90

Well, the high R90 voltage is because of the Q9c voltage .

So is Q9 fried?   The Q9 voltages don't look right for Q9c. 
You could try to test a short between e and c of Q9 but it might be easier to just replace Q9.
With 1k between b and e on Q9 in-circuit checks won't make much sense anyway.
You could check Q9 when you remove it just to prove the point it is fried.


As for Q8.   Well  I don't know.   That part of the circuit is running on a thin line.   It looks like someone adjusted parts by hand
and the circuit might not work for low gain Q8's.

Board 2:
Q9e: 0.735V
Q9b: 0.203V
Q9c: -13.65V

Q8e: 15.35V
Q8b: 14.73V
Q8c: 15.34V

I am getting 186mV (and drops) across R90

So the Q9 voltages make more sense.   High R90 voltage means R88 needs to be decreased until the R90 voltage drop to 15mV or so.

In board 1 (old one), right now, no change in IC2 Pin 7 with RV2.  It used to be a about 10mV before I screwed up.  Pin 7 measures 31mV.  On board 2, it also not moving.  Pin 7 is at 3.7mV.

Now I look at it, RV2 might not adjust pin 7 unless you have a signal.    I know the test procedure says otherwise.
If the compressor works don't worry about this for now. 

Board 1:

Depth Pot: 14.19V
R90 (across): 14.28V

Board 1:

Depth Pot: 13.77V
R90 (across): ~30mV (about 3mV of oscillation)

So board 1 we know has a problem.

I'm assuming the second test should be Board 2.
I think board 2 is working.  Voltage across R90 is a little high.  So maybe try lowering R88.

I'm bit confused because just a bit above board 2 got 186mV.

[jfrabat]

I'm bit confused because just a bit above board 2 got 186mV.

Depth pot setting changed from ~7V to 14V in the tests.

[Rob Strand]

Depth pot setting changed from ~7V to 14V in the tests.

OK cool, that explains it!

[jfrabat]

Came home for lunch, and had a quick stab at this.

So is Q9 fried?   The Q9 voltages don't look right for Q9c. 
You could try to test a short between e and c of Q9 but it might be easier to just replace Q9.
With 1k between b and e on Q9 in-circuit checks won't make much sense anyway.
You could check Q9 when you remove it just to prove the point it is fried.

Yup.  Fried Q9.  Did not bother testing it; just replaced it.  We are back to normal voltages. 

I replaced R88 to 120K.

With DEPTH at 6V, I get 43mV across R90.  With DEPTH at 14.2V, I get 0.1mV.  Should I go up to 150K?

[Rob Strand]

Fried Q9.  Did not bother testing it; just replaced it.  We are back to normal voltages. 

OK good.  Easy fixed.

I replaced R88 to 120K.

With DEPTH at 6V, I get 43mV across R90.  With DEPTH at 14.2V, I get 0.1mV.  Should I go up to 150K?

Yes, just edge it up.   The Q9 change might have affecting things anyway.

While you have 120k in there, try to use the noise gate (with say a guitar) with the depth at 14.2V, then at 6V and see if it is working.   But also *how* it is working.  Play with the attack and release. 

The main thing to realize is the lower R88 values give you more attenuation when the gate is closed.    However, if you have too much attenuation it might seem sluggish getting opening the gate even when the attack is fast.  Or it might seem to shut of unnaturally.

The difference is behaviour/sound is largely covered by the  6V to 14.2V Depth pot adjustment.   Can you even tell the difference when unit is as 6V and at 14.2V?  Don't worry if the noise-gate LEDs don't work properly.   If you can find a point you like record the voltage across R90 and the Depth pot wiper voltage.

Then change R88 and try it paying with it again.   See if you can tell any difference in the behaviour.

I have a feeling if set R88 to have get 15mV across R90 it will be good enough.


[Editted]

[jfrabat]

(WARNING: Star Wars reference coming!)

I feel like Wedge Antilles in A new Hope...  "Almost there....  Almost there... "

OK, the gate is working with the 120K resistor.  Well, sort of.  With a guitar plugged in, the TRIGGER only works until about 9 o'clock.  After that, gate closes no matter what.  So I just need it a little less aggressive.  So I take it R88 fixes that, right?   So go to 150K then and retry?

DEPTH works as advertised, by the way.  And I also took the opportunity to fix IC2 Pin 7 to 0V (it oscillates between +1.5 and -1.5mV with a signal) and the click.

[Rob Strand]

I feel like Wedge Antilles in A new Hope...  "Almost there....  Almost there... "

Tell you the truth after the 3rd release (in time) I don't know what happens in Star Wars  .

OK, the gate is working with the 120K resistor.  Well, sort of.  With a guitar plugged in, the TRIGGER only works until about 9 o'clock.  After that, gate closes no matter what.  So I just need it a little less aggressive.  So I take it R88 fixes that, right?   So go to 150K then and retry?

Yes the aggressiveness is determined by R88.   You might even try 180k.     You want it so when the Depth (RV6) is full (14V) it's about as aggressive as you need.   Then you might need to back-off RV6 to soften it. The spec is -60dB  but that's fairly aggressive anyway.   -40dB max would mean setting the R90 voltage to about 150mV.       

DEPTH works as advertised, by the way.

If you have to back-off Depth (RV6) too far to make it usable then maybe increase R88 further.   As you found, for sound it's OK to back-off RV6 but I think it stuffs-up the noise-gate LEDs so you would want RV6 to bet setting between half and full, preferably between 3/4 and full.


As for the Threshold adjustment.  Don't worry about where it is positioned ATM.  The position vs when the Gate opens depends on the signal level, and the parts that set the sensitivity are not around the Depth control.

And I also took the opportunity to fix IC2 Pin 7 to 0V (it oscillates between +1.5 and -1.5mV with a signal) and the click.

Good, I thought you might need a signal.  Your signal generator looks pretty distorted so I wasn't sure how representative the zero would be using it.