SOLVED Median compressor (sanity check)

[mybud]

You'll forgive me for asking, I trust, but this is either the most subtle compression I've heard or it isn't functioning as intended.

First off, I started by soldering the 5457s in place (biased to around 1.4V as per @andrewsrea's build reports) and used 733s as opamps. When only the level and blend controls seemed to function, I replaced these with sockets (which of course I should've done in the first place ) and tried a slew of JFETs (5457, 5458, 5459, J201, J113, and so on). I also fudged the attack control to get close to 25k (50k in parallel with 51k resistor) for lack of a C25k pot.

Problems (?): The compression (ratio) knob has little effect on the signal level, if any. It slightly changes the tone but seems to do little else. Likewise, the attack and release knobs do little to affect the signal, it seems. Currently I'm using J113s and 072s but swapping out the JFETs and rebiasing each time doesn't change anything much.

I'm wondering if I've lost my mind or whether the limiting effect is very subtle.

Apart from no obvious compression, the pedal does nice things to the tone (kind of focuses it a bit more clearly) and brings in a lot of top end on the tone control but I'm not convinced it's behaving as it ought to.

Any advice gladly received. TIA.

 

[jwin615]

Try a little boost in front of it. It should become more obvious *if* it's functioning.
Full disclosure, no experience with this one but my Keely pro can be super transparent.
You could probe IC3 pin 2 with compress cranked. Should hear the very compressed signal there.
Also listen to the noise floor. If your playing isn't louder but the noise is, it's working as intended

 

[jwin615]

I think your resistor on the attack pot should be across lug 1 and 2. Lug 3 is floating(as opposed to being tied to lug 2) in the schem.
It's early but morning brain fog tends to think you're not getting the proper effect as is. I think you could bridge pins 2 and 3 as well. Maybe try a temporary jumper to validate/decline my sanity.

 

[mybud]

You could probe IC3 pin 2 with compress cranked. Should hear the very compressed signal there.

Thanks, @jwin615. There's no signal from IC3 at all. Let me try your pot suggestion (sly wink). I think you're onto something, fog regardless

 

[jwin615]

Oof. Meant pin 1, though should have something at pin 2...
If there's no signal there, need to start tracking back through the circuit.
How about IC2 pin 1 and 7?

 

[mybud]

Yes, getting signal there. Let me check again with the updated pot in place.

 

[mybud]

No, still nothing on IC3.1. At least now I know it's not working as intended . Will keep slogging away, thanks for your help.

 

[DeadAirMD]

I know that at least one of those TL072 is fake. The one on the bottom right, Dr. Shy. The others also look suspicious "CHN"/China. Not sure if that's causing the issue or not. Actually looked again, it's both of the bottom ones. I think someone here was keeping a tab on fake TL072 and that was definitely one of them.

I built this compressor and it definitely works great. See if you can replace the IC chips with real TL072's.

 

[andrewsrea]

No, still nothing on IC3.1. At least now I know it's not working as intended . Will keep slogging away, thanks for your help.

This is a subtle effect compared to a Dynacomp type compressor, but you should definitely hear squish when the ration is at noon.

Given that you are hearing the guitar with the effect engaged, that indicates the signal path and Vref are working in IC 1, IC 2, IC 4 and IC 3.2. The issue is in the AC to DC conversion and or the gating function (FETs allowing signal to be released to ground between IC 2.2 and IC 2.1). So, please post the pin voltages of IC 2.1 & IC 3.1.

In the meanwhile, send a 1Khz sine wave from a signal generator (you can use a phone app signal generator and an 1/8"x 4 conductor male plug to 1/4" male plug or RCA and use a 1/4" plug adapter). Using a oscilloscope (or a DVM set at AC volts in a zero to 20v range). Follow the signal From IC 2.1 pins 3 & 1 to IC 3.1 pins 2 & 1. If you see signal, then check for corresponding DC voltage at the juncture of D1 + D2 + R12. If yes, measure the DC voltage at the gate of Q1 with no sine wave, then with sine wave. You should see the Q1 gate voltage increase. If it does, perform the same check on Q2. If it is working, then it could be your FETS bias are beyond the range of the trimmer.

 

[DeadAirMD]

Thread 'PSA - more bad TL072s'

Aug 16, 2023

If you’re troubleshooting a build that involves TL072s, don’t assume the chips are good even if purchased from reputable sources. I built a @szukalski Blue Sheep yesterday that just would not work. I narrowed it down to the ICs. Popped in known good NE5532s and it sounds awesome. The TL072s I was using were either from Digikey or SBP, so it’s def not a dodgy seller. Anyway, carry on building something! Hopefully this note will help somebody’s sanity.

EDIT: I traced this tube of chips back to a Digikey order.

• Laundryroom David
• Replies: 29
• Forum: Troubleshooting

• 

The 18MDSHY or what I've been calling "Dr. Shy" is located in this thread. @Cybercow has been compiling a list of fake TL072's and that's 100% on 3/4 of them. The 4th chip looks suspect as well. I would definitely bet since your voltage are wrong, that it's the fake chips causing your compressor not to work. I know you've built many pedals and have gotten all of them to work, so you definitely know what you're doing in terms of circuit construction. My bet is on the chips.

 

[mybud]

This is a subtle effect compared to a Dynacomp type compressor, but you should definitely hear squish when the ration is at noon.

Given that you are hearing the guitar with the effect engaged, that indicates the signal path and Vref are working in IC 1, IC 2, IC 4 and IC 3.2. The issue is in the AC to DC conversion and or the gating function (FETs allowing signal to be released to ground between IC 2.2 and IC 2.1). So, please post the pin voltages of IC 2.1 & IC 3.1.

Thanks so much. Will do so tomorrow.

 

[mybud]

Replaced the MDSHY chips.

Q1 biased to 1.4V
All controls at noon
IC 2.1 = 7.8V
IC 3.1 = 4.5V

I notice that the blend knob cuts out when fully CW (if that indicates anything amiss). I assume component tolerances but could easily be wrong.

I know you've built many pedals and have gotten all of them to work, so you definitely know what you're doing in terms of circuit construction.

Much as I appreciate @DeadAirMD's confidence in my abilities, let me do due diligence and double-check component values and soldering so that such issues are eliminated. I've ordered two more of these in any case but these take about six weeks to be processed through local customs. Lots of time to perfect this recalcitrant board...and to try @andrewsrea's suggestions.

Thanks to you all for responding and all the best for the new ear to come.

 

[andrewsrea]

Replaced the MDSHY chips.

Q1 biased to 1.4V
All controls at noon
IC 2.1 = 7.8V
IC 3.1 = 4.5V

I notice that the blend knob cuts out when fully CW (if that indicates anything amiss). I assume component tolerances but could easily be wrong.



Much as I appreciate @DeadAirMD's confidence in my abilities, let me do due diligence and double-check component values and soldering so that such issues are eliminated. I've ordered two more of these in any case but these take about six weeks to be processed through local customs. Lots of time to perfect this recalcitrant board...and to try @andrewsrea's suggestions.

Thanks to you all for responding and all the best for the new ear to come.

If you can, provide those IC's DC voltages in reference to ground, by each pin (1-8). They should be something like this:
IC 2.1 & 3.1
Pin 1, 2 & 3 = 4.5V
Pin 4 = 0v

Note you have 7.8V on any of those pins, that IC could be defective, or The Compress control lug #2 isn't connecting to ground.

Also, to absolutely hear the compressor, turn the Blend control fully counter clockwise.

 

[mybud]

Replaced all 072s with 833s.

Voltages follow:

IC1 IC2 IC3 IC4


1 4.5 4.3 4.3 4.3


2 4.5 4 4.3 4.3


3 4.2 4 4.3 4.3


4 0 0 0 0


5 4.3 4.2 4.3 4.3


6 4.3 4.2 4.3 4.3


7 4.2 4.2 4.3 4.3


8 8.7 8.7 8.7 8.7

Also, to absolutely hear the compressor, turn the Blend control fully counter clockwise.

Not to be a smartass, @andrewsrea, but the Wampler Ego 76 site says the opposite: 'Blend: Blend the signal between clean and compressed, clockwise will increase the amount of compressed signal present.' This is borne out by my experience: IOW, turning blend clockwise reduces the apparent signal level.

Truly unsure of this one. Tempted to mark it solved and move on and build a better version around the end of January when the new boards arrive. Thanks again, y'all.

 

[andrewsrea]

@mybud the voltages look good.

No harm taken from your statement ergarding the Wampler Ego 76. I recommend you consult the schematic for the blend control for the Median Compressor, vs. the Wampler model (they could be different).

On the schematic, the control operates as a voltage divider (somewhat like a guitar volume control), by sending either the unaffected signal or the effect signal to ground, via the Vref. Lug 1 = 'effect', Lug 2 = the sweeper contact to ground and Lug 3 = 'original signal.' Thus, turning control fully clockwise connects Lug 2 & 3, which sends all of the original signal to ground and only allows the effect signal to pass to IC 4.1.

The Blend is at a slightly quieter signal than fully counterclockwise (sending the effect signal to ground via Lug 1&2), as some of it gets leaked through R24 to ground. Notice R24 is 1/10th the resistance of R25 in the passive mixer stage, which i believe is why this occurs. The signal is also loudest in the non-extreme settings, as is the nature of a variable voltage divider.

I just fired mine up to verify. This might be why you are not hearing the compression.

 

[mybud]

Brilliant explanation, @andrewsrea. Thanks for your patience as ever.

 

[comradehoser]

Truly unsure of this one. Tempted to mark it solved and move on and build a better version around the end of January when the new boards arrive. Thanks again, y'all.

I feel your pain.

I'll let you know how my go around #2 goes as well.

 

[mybud]

I feel your pain.

I'll let you know how my go around #2 goes as well.

I think it’s working but I’d still like to try it with the correct attack pot (not my fudged version). This just adds another variable and more room for error. I’ve checked out some YouTube demos and mine seems really subtle by comparison. By all means, LMK how your revised version goes. Might be an idea to socket the JFETs in case; I know I will Time will tell whether congrats or commiserations are the order of the day

 

[comradehoser]

I think it’s working but I’d still like to try it with the correct attack pot (not my fudged version). This just adds another variable and more room for error. I’ve checked out some YouTube demos and mine seems really subtle by comparison. By all means, LMK how your revised version goes. Might be an idea to socket the JFETs in case; I know I will Time will tell whether congrats or commiserations are the order of the day

Dude, you know I am going to socket every dang thing this time.

 

[mybud]

socket every dang thing

Socket to me, brother