What sorcery is this? - Weird interaction between guitar and pedal - update

[rockhorst]

I just had the most surreal gear testing experience. I'll give away the clue: we suspect that what's causing the problem is a faulty volume pot on a Les Paul *EDIT it's NOT the pot!*

We were going to test different caps (socketed) in a finished medium to highish gain pedal to adjust bass content. Signal chain was essentially:
Guitar => a buffered bypass pedal (mainly tried an Ibanez CS9 compressor and some others) => the DIY gain pedal => amp

With the CS9 (or other buffered pedal) engaged, all was fine. Omitting the buffered pedal, going straight into the gain pedal was also fine. With the buffered pedal on bypass and the gain pedal engaged, the weird stuff started to happen.
When hitting pinch harmonics, especially between the 5th and 12th fret, we got what is best described a volume swells. Really digging into the notes would cut of the signal, swelling in as the vibrations died out. This also happened with hard hitting chords, but not as audible: it was especially evident on the squeels. Kicking the compressor back on again, everything appeared fine.mwe did this dance for an hour, checking everything. We suspected faulty cables, broken solder joints, thought the CS9 was crap. It turned out that it only happened with one particular Les Pauls bridge pickup selected. Other, similar output guitars gave no problems.
The guitar has had a recent pickup change and a NOS capacitor for the tone. We took out the cap, but that didn't change the problem.

Now, I have yet to find a way to make sense of this. Anybody? What's so special about a (potentially) faulty pot in combination with a disengaged buffered bypass pedal?

[rockhorst]

Slightly edited the post above, because it turns out not to be the pot. It was replaced and the problem persisted and we've found another LP that also has the weird swell effect. My best guess is that a really hot signal spike somehow overloads some part of the signal chain. Maybe there's two traces or stray wires barely touching somewhere? We're going to go over and change all variables the upcoming week (cables, guitars, amps, etc) to see what we can find, but looks like there really is some spooky action in the guitar => buffered bypass => DIY gain pedal. I'm very confused.

If anybody has any ideas, please share.

As a reference, this is the schematic of the DIY gain pedal.

[Morocotopo]

Maybe the high frequencies in a pick harmonic triggers some parasitic oscillation in the high gain circuit, that dies as the note dies out? Just a idea. Would be visible with an oscilloscope.

[Quackzed]

if the bridge pickup has chrome covers that are grounded to the pickup base plate, AND the bridge pickup is wired out of phase(intentionally or not), the chrome cover will be flipped from ground to + and because the strings are at ground , tuching the string to the pickup cover directly or with sweaty fingers will 'short' the signal to ground.
it may be that the technique of pinch harmonics is such that your inadvertently making a connection through your thumb from the string to the pickup cover and grounding or shorting the signal(cover) to ground(string).
easy to test, play a big open chord and with a finger push the high e string down till it touches the pickup cover. if it mutes the sound when the string contacts the pickup cover then thats your issue.
i have an sg that has 'out of phase switch for the bridge, when out of phase is selected for the bridge and i touch any string to the cover it 'kills' the signal.
i left it that way but at times of heavy picking i sometimes accidently get a 'mute' via the cover shorting to string ground...
to remedy this you can either dissconnect the cover from the pickup base plate ground or flip the polarity of the pickup wires so its 'in phase'. though if both pickups are accidentally wired out of phase such that the middle pickup selector position sounds normal the way its wired you may need to flip both pickups wires for both to be in phase with each other and also not short out if making contact with the pickup cover/covers.

[rockhorst]

@Quackzed
I got an Oscilloscope recently, still need some probes, haven't played around with it yet. How would I spot something like that on the scope?

@Morocotopo
Will try that tomorrow! Interestingly enough, I couldn't get the notes to die out with pinch harmonics which we attributed to 'sissy' playing. But it might've been a hint.

Thanks for the suggestions, I'll get to it. I just got off the phone with the guitarist. He switched leads and amps and the problem persisted. He raised the pickups of another LP and that one started doing the same thing. So that would support the idea that it either has something to do with a very hot guitar signal, or he is grounding his signal during the pinch harmonics. I was also thinking maybe it had something to do with the way I wired the clipping diodes and their threshold. Seems unlikely though because I that stage would be well over the threshold voltage already.

Other suggestions?

[pinkjimiphoton]

sounds to me like you have too much signal, peak to peak, hitting the first stage of the effect, effectively pushing it beyond it's cutoff point for audio..with the comp on, it doesn't get the signal peaks, with it off, it's overloading the first stage. check the biasing on it.

i do stuff like that intentionally, sometimes i will overload a fuzzface with a compressor full blast to get that kind of effect. as the envelope begins to decay, it drops in amplitude enough where the transistor swells back in.

it can be a cool effect, if you want it.

if you have dirty fingers pickups in the les paul, those are REALLY hot... probably 3-5 times the output of "normal" pickups.

i would check the input of the effect, and the blocking caps in the first stage..a leaking cap can make a transistor or tube saturate by overloading it with DC sometimes.

just thinking on my feet here..

[rockhorst]

Thanks Jimi. I'm thinking the same thing but I can't get my head around the tandem with the non true bypass compressor. Guitar straight into the gain pedal doesn't seem to give any problems (but I'll double check that when I get my hands on the unit tomorrow). I can see how the comp would smooth things out when activated. What I don't get is why the comp off, in buffered bypass, would give problems as opposed to true bypass. Building the unit I hand picked the JFETs and their bias resistors, but that will be my next stop after I check Morocotopo's suggestion.

[pinkjimiphoton]

well, the GOOD thing with these head-scratchers tho, is it will get sorted out, and it will help a whole lot of us. keep at it bro, you'll get it!!
thinking the buffered compressor may be causing an impedance mismatch maybe?

man...stumped..

[Quackzed]

if thats the reason, it might be that those pickups have a high freq peak that when un buffered, gets rolled off enough to not flood the next stage too much, but when buffered those high frequencies are full tilt and do flood the stage... 

[rockhorst]

I haven't been able to play around and test the setup described previously (the other guy was a bit busy), so I did some of my own experiments. Sure enough, it's quite easy to make everything crap out by overloading it with signal. I took a humbucker loaded PRS, into a minibooster on full tilt and into a dual drive pedal (either two other mini boosters or an Xotic BB+). Putting a buffer in front of it makes it even easier. I haven't been able yet to have stuff crap out on pinch harmonics, those keep singing. But bang an open E power chord or hard hit some lead lines and you get the aformentioned volume swells. At no time did I get a completely dead signal, the level just drops significantly.

Now, I'm assuming that this is similar to what's happening in the original problem (with maybe some specific pickup resonance thrown in the mix). I was wondering, since all the pedals are on 9V, that would be the maximum signal output you can get from any chain of pedals. I would've thought that if you keep slamming boost into boost into OD, you'd just get an extremely distorted mess without much definition. Instead, some input stage seems to get overloaded/swamped and reacts by shutting off (JFET) or reducing the level (opamp?). How does that happen? Is it possible to turn of a properly biased JFET by putting a very hot signal on it's gate?
Another thought would be that this testing is being done with tube amps at bedroom levels with their preamps turned down. Maybe it's not so much of a problem live when the tube amp is turned up and has more room to breath/swing.

Thoughts?

[Lurco]

Too big an input capacitor? R2=2.2M with C1=220n gives 0.3Hz fc. Kinda like a motorboating effect?

[rockhorst]

Good suggestion Lurco. Since the guy using the pedal is big on 'tight bass' (so we're already filtering a lot of bass out), it can't hurt to up that value a bit and see what happens.

[Lurco]

Good suggestion Lurco. Since the guy using the pedal is big on 'tight bass' (so we're already filtering a lot of bass out), it can't hurt to up that value a bit and see what happens.

Don`t UP it, but DOWN it.

[rockhorst]

Yes I know, my bad.

[rockhorst]

So I finally had the chance to test the actual pedal at the guys' place. Still his pinch harmonic technique is way more aggresive than mine, I can't get them to the level that the signal overloads the drive pedal, but it can be done. Testing with my PRS instead of his Gibson even made it easier (because those PRS pickups are very high output). Playing on the PRS the signal also started crapping out on normal notes.

Short version: nothing of the above solves the problem. What I've tried and didn't work:

-Lowering the effective input cap by putting another cap in series with the effect, which makes a high pass filter with the pop-resistor
-Putting an adjustable low pass filter in front of it (simple cap with potmeter setup). This did help, but by the time the attenuation becomes effective, the cut off frequency is already in the audible range
-Bypassed the clipping LED section entirely (no difference)
-Tried another buffered pedal in front of it (TC electronic delay, switchable between true and buffered bypass). Same result: true bypass no problems, buffered bypass troublesome
-Put the guitar signal on a scope. Since the guitar signal is so weak, it's kinda hard to get a clear reading on the scope using an FFT, but I could cleary see the usual frequencies in the audible range put no weird spikes beyond say even 10k.
-It's a lot harder to have the signal crap out with IC based overdrives as compared to this triple FET stage type thing.

So, I'm here again fishing for suggestions. It's definitely got to do with some sort of signal overload, too much going in. Next test is going to be the output of the separate FET stages, maybe the effect itself over amplifies something (though it really isn't as high gain as say a muff or dr. boogie). In the meantime I'm going to mod the pedal to exact Plexi Drive input specs, as it was derived from that, hoping that it will solve something.

Any other ideas? I'm almost out of them

[rockhorst]

Eureka! And the price for the winning suggestion goes to Lurco! I tested a lower input capacitor yesterdag (10n in conjunction with 2.2M resistor) but that just wasn't low enough. Today I convinced myself that it had to be the bass content so I left the 2M2 pop resistor in place, put in a 10n input cap and changed the following resistor from 2M2 to 820k, which seems good enough especially since the pedal is placed directly after a buffer. The effective -3dB frequency is now about 20Hz, which seems to filter out enough junk to make the pedal stable.

Cheers everyone!

[Quackzed]

if the gitar is active, needs a battery, it might be a good idea to put a voltage divider --IN--- the guitar itself.
any decent ( or indecent/crappy) buffer or 'active' battery powered guitar will easily be able to put out a signal that will not be effected by a voltage divider in the guitar.
by a voltage divider i mean a 100k in series before the output jack of the guitar, and a same resistor 100k resistor to ground just before the output jack of the guitar.
this will cut the output of the guitar in half and prevent any problems (at least heavily reduce them ) with overloading the 'first effect' after the guitar.
so, if the guitar takes a battery. pop a voltage divider into the guitar.


wire to output jack of guitar--------100k----/----output jack signal tab
                                                                 /
                                                                1
                                                                0
                                                                0
                                                                 k
                                                                /
                                                              ground tab of out jack  

100k may be unneccisarilly big, 10k's would probably be fine, but 100k's will definately be fine- for an ACTIVE circuit(with battery) ie... buffered.
this will cut the guitars overly loud outpuit in half. and therefore prevent it from hitting anything its plugged into too hard.