Thumb Sucker Compressor "scrapy" sound

[Matopotato]

Hi,
I just "finished" building a Thumb Sucker Compressor. Sounds very cool, more impressed than I had expected.
But..
1. There is an element of distorted scratchy sound. More with humbuckers than single coils. It is not distortion as you would like from a distortion pedal. More scratchy. It is all in the box so I doubt there is some loose cable or joint (famous last...). I am not very used to the controls yet, but it seems to vary with very low setting on release but maybe not. More so with (higher) ratio above 12 and (lower) threshold below 10-12...
(2. The attack and release knobs. Checked connection and that seems to be all good at least in that vicinity of the board. But they do not change things much. Is it because they are that subtle? Or do they only come in play with other settings? (CORRECTION: Asked this here as well: https://forum.pedalpcb.com/threads/thumb-sucker.6769/ post #6 I think. sorry for double posting))

 

[Matopotato]

Did you run it at 12V? Just wondering if that makes any difference with respect to the Threshold?

Just wondering if the extra power affects that.

Not yet. Will try but it takes some tweaking first...

 

[Matopotato]

So I did the most unscientific test, Friday late night, but hey...
I ran with 18, 12 and 9 volts. There might be all kinds of objections, but my take on all this is to run at 18V, learn how to dial in the pedal and look for subtle but good. I am closing in on the first...
Oh, almost forgot, now that I am beginning to accept this build and pedal, it deserves a nice decal print. Compressor to me is an Atlas Copco yellow smooth-cornered machine from the 70's so I plan on using that somehow...

Here are my settings:


Single coil, A Fender Strat

https://soundcloud.com/user-558594533%2Fsinglecoil-18-12-9volts

Humbucker, LP Classic

https://soundcloud.com/user-558594533%2Fhumbucker18-12-9volts

 

[Matopotato]

I took out R10, put sockets in.
Tested with 18V.
1k gives less distortion when on 18V vs 9V
2k sort of bring back some distortion.
So I tried with 470R and the distrortion is basically gone.
But it sounds as "less compression pedal" being the cost.
Is that an accurate observation? (@Chuck D. Bones ?)

 

[Matopotato]

I think I can slightly almost sense something happening with ATTACK if I tamper a lot with RELEASE. But (in my case) not enough to have the knob in the first place.
Low RELEASE seems to be lifting the sound more than natural as it rings out. So it acts like an after-boost, sort of as if the pedal forgot do do its work gradually, got reminded and rushes it. Almost an oscillating thing. Not so musical IMHO.
The scraping distortion is still there, and I can appreciate that extreme settings and low strings will produce it, but to have to swap resistors etc while on noon settings feels as if I am not getting the right effect out of the pedal compared to what ought to be possible and what I was expecting/hoping for. I am no compressor wizard, let alone any decent player, but feels as if something is missing still.
I plan to put in a 2k trimpot for the R10 and try to dial in some unicorn sweet spot of high compression effect and low scrapiness, but my hopes are not very high. It feels as if something is wrong in the build, but since the extremes cause the negative effect mostly, there might not be any more advice to get.
I might try to randomly step through the "non-clean" side of the path, but I wont know what to look for, so right now this pedal stands out as the first among some 40+ that does not make it to any board... Sad...
I will try to build a different compressor kit that emulates Diamond Compressor. Hope that works better. Else I will just have to buy some real deal compressor, because despite my problems, there was enough effect with the Strat at least to make me want to pursuit some better version of it.

 

[Matopotato]

I got a 2k trimpot for R10 today. With 18V, Ratio on 3 o'clock and Threshold on 9, I had to dial it down to about 340 ohms. And at that point I am not able to hear as much of compression as with 1k or recommended 2k.
So I tend to think that there is something else wrong, not just pushing the boundaries (which I do not feel that I am doing anymore).
Any ideas of how to trace the compressed part of the circuit, or how to test the LM13700 would be much appreciated.

 

[Chuck D. Bones]

I do not recommend less than 1K for R10. The LM13700 is spec'ed for Iabc (current into pin 1 or 16) 1mA or less. If you go above that, performance is undefined. My Thumbsucker does not do what yours is doing. I'm using an LM13700 I bought on Amazon, of all places.

One interesting thing I found out on my Thumbsucker breadboard is that C5 and C6 should be 1uF film capacitors. Aluminum electrolytic is too leaky. Even tantalum is unacceptable at longer RELEASE settings. The swell you hear during the note decay is normal. The high freq content tapers off more quickly than the fundamental as the guitar signal decays. When the compressor turns the gain up as the note decays, there will be more and more fundamental content. This is particularly noticeable at high RATIO settings. Mine sounds best with ATTACK & RELEASE below 10:00, RATIO below 2:00, THRESHOLD above noon.

 

[Matopotato]

I do not recommend less than 1K for R10. The LM13700 is spec'ed for Iabc (current into pin 1 or 16) 1mA or less. If you go above that, performance is undefined. My Thumbsucker does not do what yours is doing. I'm using an LM13700 I bought on Amazon, of all places.

One interesting thing I found out on my Thumbsucker breadboard is that C5 and C6 should be 1uF film capacitors. Aluminum electrolytic is too leaky. Even tantalum is unacceptable at longer RELEASE settings. The swell you hear during the note decay is normal. The high freq content tapers off more quickly than the fundamental as the guitar signal decays. When the compressor turns the gain up as the note decays, there will be more and more fundamental content. This is particularly noticeable at high RATIO settings. Mine sounds best with ATTACK & RELEASE below 10:00, RATIO below 2:00, THRESHOLD above noon.

Thanks for coming back to me. Felt a bit lonely there for a bit...
I will keep it at 1k for R10. Makes sense. I assume this is for 18V.
I got the LM13700 from Das Musikding as part of the kit. I will see if there is a way to test it to see that it is OK or not, but not quite sure how.

C5 and C6 are 1uF electrolytes, and have polarity marks on the pcb. I might get a chance to swap them out later on for "film box" MKT (?), would that be something to try?

Any hints on how to "trace" the compression side for anomalies? And/Or the LM13700...

Thanks

 

[Chuck D. Bones]

No need to quote me message in your replies.

Power supply voltage doesn't matter. 1K is the max recommended value for 9V, 12V or 18V.
Das Musikding does not have the best reputation for parts quality.
Testing the LM13700 is not easy and I would use a scope and sig gen to do it. I'll see if I can come up with a simple test circuit and get back to you. Might take a while. Do you have a breadboard?

Does turning RELEASE down below noon make the scrapy noise go away?

I do not have any other troubleshooting suggestions at this time.

 

[Matopotato]

Sorry,...

OK, I felt a clear difference as to my issue being lot less prominent on 18V compared with 9V. I also assume that 18 would raise the headroom (from what I read) and possibly result in less clipping problem in my (previously) more extreme settings.

Yes I have a breadboard (with a Dist + on it to try out Jack Orman's Saturation variations and some asymmetrical clipping, but it is easy enough to disassemble again). No hurry at all. If you find the time I am grateful.

Will try RELEASE down later on, but with my previous settings of RATIO at 3 and THRESHOLD at 9, it had no impact on the scrapy stuff IMHO. (I thought scrapy was a word in English... Might have been wrong all along.)

Your recommended settings are fine, not much scraping there, although a little bit less "compression effect" than what I was trying to find.
I am not very subtle with my playing and gear, more all in, for better or for worse. So I liked the first feel of the compressor but I suppose I set it quite "hard". With more subtle settings I struggle as of now to hear and feel much difference from clean tone. I guess there is a learning curve to compression and I built the kit because I was sceptic if compression would be to my "taste".
But the first impression with this build was "Wow, my tone just got more lively and fatter" and I liked that! Then scrapy...

 

[Chuck D. Bones]

I have no problem with headroom at 9V as long as I keep THRESHOLD above noon. I don't think the scrapy sound is from overdriving IC1.1 because the scrapy sound doesn't happen during the attack, it happens during the release. Make sure Q1 is tight in the socket. I don't recommend socketing transistors in those SIP strips because the leads are too small for the socket.

 

[Matopotato]

Yes, I have noticed sometimes how SIP sockets are not tight enough.
What alternatives are there/do you use?

EDIT: I actually soldered it into another SIP and stacked them to keep things tighter.

 

[Chuck D. Bones]

I solder transistors. Sometimes I'll socket ICs. I used to socket germanium transistors, but these days I audition them on a breadboard then solder 'em in. Sockets are just one more thing to go wrong.

 

[Matopotato]

I did change the 1uF electrolytes, C5 and C6 to Film Caps. I has been a while since I tested, but spontaneous feeling was that the scrapy sound got less scrapy, but more present. I mean even with previously unaffected settings, it was now audible. It could well also be in my mind only, and the knobs were not on so I could not tell for sure what positions I had dialed in.
Since I had unboxed it, I tried to take some shots, although I am not very good at that. Still it might be helpful if someone has sharper eyes and more experience than I do.

I am still trying to figure out how to possibly test the LM13700 on a breadboard. I googled a bit, but couldn't quite understand how to set it up.

 

[Chuck D. Bones]

I think your LM13700 is working fine. Something I noticed when I monitored IC2-1 on a scope was a tendency toward oscillation. That might be what we're hearing in your first sound clip. I ended up installing a 47pF cap in parallel with D1 and that fixed it. I have some other mods in mind, but they're not ready for prime-time just yet.

There's a lot of flux on the board. I recommend cleaning it off because when exposed to humidity, flux becomes conductive and can upset the circuitry.

 

[Matopotato]

Thanks, will try.
I am not adding any flux myself, must be part of the solder then. Thought that was conservatively included. Will figure out something to clean with and make it a habit.
Thanks for checking and I assume that the solders looked reasonably ok apart from solder.
47pF can be ceramic I suppose.

 

[Matopotato]

It was not easy getting to the D1, but I kind of managed to have contact and strum a string, but with the occasional touching the wrong thing. Pop and gradual come back. I first thought that was effect of the cap, but got wiser.
In the end I could not tell much difference. I tried 47pF, 82pF (felt a tiny bit better, but not sure) and 120pF (no difference). Seems some hiss or similar goes away though, but with the pedal in the open it is sensitive to the surroundings. I used 18V, but that has been better than 9V for me from the start, so I doubt 9V would make it better.
I really hoped this was the fix, but could not get it to happen. I used slightly pushed settings to tell if the noise goes away.
Good point that it might not actually be distortion per se, but possibly something else like oscillation. I'd be eager to try any of tour other ideas and appreciate your patience in this.

 

[Matopotato]

My friend came over with his Engineer's Thumb and it did not have the scrapy noise. With RATIO max and THRESHOLD min there was some noise as expected, but it disappeared quite soon.
We tested my Thumb Sucker with 47, 82 and 120pF in parallel with D1 and it seemed to get better occasionally but once the caps settled in it was back to bad. Not all that easy to test even being 2, but no solution unfortunarely.

 

[Matopotato]

We should have swapped LM13700's just to rule that out during the comparison, but weren't too clever about it...

@Chuck D. Bones : "I have some other mods in mind, but they're not ready for prime-time just yet."
Did you manage to try out any other of those mods you were thinking of?

 

[Matopotato]

Hi again,
It's been sitting around for a bit while other projects were completed.
Now I managed to re-flow every solder. Flux used to make it flow nicely. Felt that everything went well.
Disappointingly the scrapy sound remains and worse than I had remembered.
I already swapped places with the TL072s and I also tested with a 2N5087 for the 2N3906. Even removed the transistor all together. No change .(Wonder what that transistor is contributing with...)
So besides swapping each resistor for a new one, each cap for a new one etc etc, there is only the LM13700 remaining that is easy enough to test.
So that'll be the last thing to do. As mentioned above, my friend has an LM13700 in his Engineer's Thumb and his sounds as it should and could be expected, so the test should be easy enough.
If that is not the cause, then I will most likely give up on the Thumb Sucker project.

 

[Matopotato]

<snip> I have some other mods in mind, but they're not ready for prime-time just yet. <snip>

Any of those ideas that you could share? Thanks.