Supreme Compressor (Engineer's Thumb) Unwanted Distortion

[Wook22]

Hello all.

I'd really appreciate some help with my build please, if possible. I have followed the troubleshooting steps in the FAQs but nothing seems to have had any effect. This is my first build with a PCB rather than a veroboard which I though would be more straightforward. That was certainly the kiss of death!

The kit is the Supreme Compressor from Effect Pedal Kits. It is based on the Engineer's Thumb, with 4 control knobs.

A lot of fuzz-like distortion is being produced when the input signal is higher, i.e. when a string is played hard. Sounds ok when played gently. Exhibits the behaviour with both humbuckers and single coils. Can be reduced by playing with the vol and ratio pots but not removed entirely. Sounds as though when trying to compress it is distorting instead. Bypass works fine.

Schematic is shown below.

No modifications to circuit or parts subs.

Battery voltage 9.02V

Voltage at the circuit board end of the red battery lead = 8.88V
Voltage at the circuit board end of the black battery lead = 0V

For the following, vol, attack and threshold turned fully clockwise, ratio turned fully anticlockwise.

Q1
C = 1.08V
B = 4.80V
E = 4.41V

U1
5.23V
4.45V
3.98V
0V
4.17V
4.43V
4.43V
8.83V

U2
1.08V
30mV
4.42V
4.42V
4.44V
0
0
0
100mV falling
0
8.82V
0
0
0
0
0

U3
4.46V
4.45V
3.98V
0
4.21V
4.45V
4.86V
8.83V

D1
A = 4.21V
K = 4.45V

D2
A = 4.46V
K = 4.46V

D3
A = 4.84V
K = 4.45V

Grateful for any assistance!

[merlinb]

Sounds like you should add the two LEDs at the input, which seem to be missing from the kit schematic (D6/D7 in my circuit). They're used for pre-clipping, to stop the rest of the circuit being driven to overload, which is not pleasant.

[Wook22]

Thank you Merlin, I shall certainly give that a try.

Seems like an odd omission if it does turn out to be the issue!

[matopotato]

I built a "Thumb Sucker" from PedalPCB. It has 5 knobs and a bright switch, but else very much the same.
I have the problem when I go to Ratio passed noon or 3 o'clock, and Threshold turned down to noon to 9.
My friend built an Engineer's Thumb (different kit) and he can set to 3 and 9 without distortion. Or oscillation. Bad sound anyway.
I have tried several things and traced with a probe. Got many good tips in PedalPCB forum (https://forum.pedalpcb.com/threads/thumb-sucker-compressor-scrapy-sound.10516/) but none helped.
Eventually I did re-solder every joint and every wire. Swapped around the 8-pin ICs and I think I even rotated in a third one.
So although the help I got said it would not be the LM13700, that is what remains for me to try.
I'll see my friend over the weekend, So I plan to swap LM13700 with his Engineer's Thumb to see if that is the issue. If not, I am at a loss and might give up on it for good. But really would like it to work...
Not sure if this helps you in any way, but with all the help I have gotten in this forum I feel I must pay back whatever I can when I can.

[Wook22]

Thanks. Good to know I'm not alone if nothing else!

My issue seems pretty similar to the one in your posted sound file, although min is more of a "crackle" than a "scrape" I would say...

Currently having a bit of a play with the pre-clipping LEDs. So far it's having an effect (at least I think it is) but it still doesn't take much effort to get the distortion back, particularly on the lower strings.

Desoldering/re-soldering may be on my list of next steps, but I think I'll wait until you've had a play with an alternative LM13700 first!

[matopotato]

Thanks. Good to know I'm not alone if nothing else!

My issue seems pretty similar to the one in your posted sound file, although min is more of a "crackle" than a "scrape" I would say...

Currently having a bit of a play with the pre-clipping LEDs. So far it's having an effect (at least I think it is) but it still doesn't take much effort to get the distortion back, particularly on the lower strings.

Desoldering/re-soldering may be on my list of next steps, but I think I'll wait until you've had a play with an alternative LM13700 first!

Cool, I'll report back.
Have a scan trough the tips I got in the other thread.
I first verified that my components could handle 18V then tried that. And it reduced the problem but did not fix it.

[Keppy]

A lot of fuzz-like distortion is being produced when the input signal is higher, i.e. when a string is played hard.

Based on that description, my first thought is rectifier distortion. This would be possible if you had diodes D4/5 or electrolytic caps C4/5 reversed or otherwise installed incorrectly, or if the resistors on those components aren't right. That type of problem wouldn't show up in your stable voltage readings, but would be immediately obvious if you scoped or audio-probed the C5/opamp junction and saw/heard half-wave distortion.

[Eb7+9]

Sounds like you should add the two LEDs at the input, which seem to be missing from the kit schematic (D6/D7 in my circuit). They're used for pre-clipping, to stop the rest of the circuit being driven to overload

what (other) comp/limiter design needs diode/LED clamps at the input to work cleanly ? ... that's a red flag right there

and, what are LED's going to do to single-coil pickup signals ? 
... even played hard, pretty much nothing or not much anyway

I mentioned too many times already the author has no real understanding of what can go wrong with his invention
I provided many hints for his wizardship to take a second look and have a chance to fix this ...

thanks for all the insults so far bud

---

I mentioned very clearly from the get-go that some people would randomly run into the problem posted here
which I'll explain later once we get confirmation of a potential deviation/correction in performance
(something that's easily recreated on the simulator)

https://www.diystompboxes.com/smfforum/index.php?topic=105610.msg1203643#msg1203643

I even suggested where to look for a solution ... but, I'm an asshole right ?!

---

well, I knew eventually someone would pipe up about it ...

so, I'm glad to say matopotato and wook22 are finally at the threshold of a big discovery here
(and it doesn't just apply to this circuit)

I will offer one suggestion, try out a bunch of 13700 IC's to compare ... and hopefully you're only using one of the two OTA
two OTA's makes the problem worse, as I also pointed out

like, if you have several of them tell us how many IC's you tried and how many made the circuit work clean or not ...
leaving everything else the same, and the front LED's out of circuit

---

eventually, the next question will be the following : whether or not there is a fix that precludes having to "hand select" the offending part

let's leave that for part B of my reply even though I showed everybody where to look too many times already

[Wook22]

Based on that description, my first thought is rectifier distortion. This would be possible if you had diodes D4/5 or electrolytic caps C4/5 reversed or otherwise installed incorrectly, or if the resistors on those components aren't right. That type of problem wouldn't show up in your stable voltage readings, but would be immediately obvious if you scoped or audio-probed the C5/opamp junction and saw/heard half-wave distortion.

Thanks Keppy. I believe D4 and D5 equate to D1 and D2 on my layout.

I have added images of the layout from the plans and my board. R2 and C1 are on the veroboard I am using to play around with clipping diodes and I have swapped out R1 to be the same value as Merlin's plans to see if it makes a difference.  I think I've "idiot-checked" it to the point that I'm happy that components are all correctly oriented, but would be very happy to be proven wrong!




Building an audio probe is on the to-do list, I now suspect it should have been further up!

[Wook22]

let's leave that for part B of my reply even though I showed everybody where to look too many times already

I appear to have inadvertently walked into an on-going conversation.

I'm not sure I have much to contribute other than trialling any suggestions and reporting my findings and as my electronics knowledge is fairly limited!

[matopotato]

Yes, I felt the same. I have learned also to try and listen to all suggestions, but also with my limited knowledge it takes a while and some study hours to check all out. Here I am a newbie to the functional schemas as well as other stuff.

I was a bit puzzled on a "higher level": if (in my case) the Thumb Sucker was made into a PCB, and sold to several people and most have no issues ( I seem to be the distortion exception rather than norm) then how can the design/schematic be the main culprit?
I would think my bad soldering, orientation mistakes, or my own goofs first. Then unlucky bad component.
Anyway, will try LM13700 as planned, but with less hope now and see.

Probe: quite easy to build. Just google or aske me again.
Very useful tool to help follow the signal. You need a looper or tone generator to send in a signal.
That and a decent DMM to check "beep-connectivity"

[Wook22]

Thanks again Matopotato.

There seem to be positive reviews for the kit I have too, therefore my first instinct was that I'd ballsed something up, which may still be the case!

I have a plan ready for a probe, just need to get on with it...

[merlinb]

I appear to have inadvertently walked into an on-going conversation.

Ignore the troll

[matopotato]

Sorry, totally missed it  :
@Wook22: Very welcome to the forum!
   

[Wook22]

Thank you!

So, I have concluded that whilst the addition of the LEDs to the front end does make a difference, it hasn't got rid of the issue. This leads me to 3 possibilities:

1. Builder error (most likely)
2. Component fault.
3. An issue with the PCB or design (least likely).

If no one can spot anything obvious from the image I posted (component orientation for example) then I think the next step is the sound probe route to see if I can zone in on where the problem may lie.

[matopotato]

Managed to get an old oscilloscope and ditto tone generator. And we tested around a bit this evening. Really cool to see the effect and also with guitar. It gets very obvious when the distortion happens. Remains to figure out where.
One thing I noticed though is that the Ratio pot does not deliver expected values. Could be because of its context, but also a bad pot. A1M that read 19k in full CW feels a bit off. So now I gave to things to try.
Which beats zero things by far.

[DrAlx]

One thing I noticed though is that the Ratio pot does not deliver expected values.

Sounds like you are trying to measure the pot resistance while it is in the circuit. You cant do that in this case. You should disconnect it (or at least one of the two used pins) from the circuit in order to measure it. Pot is probably OK.

[matopotato]

One thing I noticed though is that the Ratio pot does not deliver expected values.

Sounds like you are trying to measure the pot resistance while it is in the circuit. You cant do that in this case. You should disconnect it (or at least one of the two used pins) from the circuit in order to measure it. Pot is probably OK.

Ok, thanks.
Yes wasn't sure. Circuit is not powered and no signal fed either. The other 4 pots check out.
But your point is what I was debating with myself as well.

[merlinb]

U1
5.23V
4.45V
3.98V
0V
4.17V
4.43V
4.43V
8.83V

U1 pin-1 seems abnormally high. Could your output cap C3 be leaky? Do you measure any DC across the VOL pot?

R6 should be 220R, it looks like you have installed 10k.

[DrAlx]

U1 pin 7 = 4.43V   which is the bias voltage. Looks correct.

U1 pin 3 and U3 pin 3 are both at 3.98V.

Looking at schematic I would think they should be at bias voltage also (they connect to the bias voltage via the threshold pot).
I assume measured voltages are reported  OK so what am I missing here?
Could it be measurement error because of multimeter impedance?

Good spot on the build error BTW.