noisy filter/distortion pedal

[blackieNYC]

After deciding I was hearing only breadboard-based noise, I perfed and boxed up this pedal.  Its the Madbean Karate Shop followed by a distortion plus with a couple mods.  For a fixed-wah distortion sound. There is a lot of pink/brown noise and also buzz.  The guitar sound is very good otherwise.  Voltages on the op amps look good.
The noise changes with the filter frequency pot, but if the distortion pot (1 meg) is turned down, and the output volume is turned up to compensate to equal loudness, the noise is almost gone.  So, even though the dist circuit is after the filter circuit, the distortion seems to be affecting the filter. Maybe.
Forget the Karate Shop power supply - the power supply used is in my Distortion Plus drawing. And I've added a 100nf cap to the 100uf cap on the 9v supply.
Should I separate the two circuits 9volt distribution somehow?  Anything wrong with my distortion circuit?  Is the karate shop just noisy?  Should I sub for the 741?

[antonis]

Try to put 10k between 1M & 1nF ..

[blackieNYC]

Thanks.  I'll try that.  It would seem to make a 15kHz rolloff with the 1nf - how do you think it will help?

[Mark Hammer]

Hi Alan,
My Karate Chop adopts a simpler approach.  I simply strap a back-to-back pair, with a treble-taming cap, across the feedback path of IC1A.  I use a 250K Gain pot to get the amplitude needed to produce  decent clipping, and a SPDT toggle that either connects the diode pair and cap across the Gain pot OR lifts their connection and subs an 82k fixed resistor across the pot to bring the maximum gain down to something closer to normal.

Granted, sticking a Dist+ after the Chop affords more opportunity for tone shaping, and also provides harder clipping, but what I suggested is simpler and requires no daughter board.

Alternatively, use a DPDT toggle to run a back-to-back pair to ground from the junction of R8 and C6, and simultaneously lift a fixed resistor in parallel with the suggested 250k pot, or lift the diodes and drop the max gain with a fixed resistor.

Again, a simple change to the basic Karate Chop.

[ElectricDruid]

It would seem to make a 15kHz rolloff with the 1nf - how do you think it will help?

That's it, basically. It'll make a 15KHz rolloff and help kill any out of band noise that would otherwise subsequently get amplified. The rule I was always taught for amplifiers is that you only amplify the bandwidth of interest. Anything else is just increasing noise you don't want.

Where's this filter circuit you're talking about exactly? I see a gain circuit with tone shaping, but I wouldn't exactly describe it as a filter circuit, and then it's followed by clipping diodes and a volume control.

I would imagine the noise disappears when you turn the gain down and the volume up because you turned the gain down. Same signal, less amplified noise.

And yeah, the 741 is pretty noisy, so if you can't fix it any other way and you want it quieter, a better op-amp will probably help, but it will turn it into a different thing ("modern" instead of "vintage", with everything that implies).

[blackieNYC]

By filter I mean the karate shop wah thing. I don't know why it's not called "chop". It precedes the dist+.

[duck_arse]

C9 10uF on the chop appears to be drawn backwards. it surely couldn't hurt to have an R//C supply filter - at least - between the two circuits, maybe even an R/C for each.

[iainpunk]

the UA741 is inherently noisy. not super noisy, but still, you might want to look in to other single opamps to use here.
i like using the CA3130 or CA3160 for such circuits, but they can be quite expensive as far as chips go and they benefit from a big compensation cap between pins 1 and 8, i use 2.2nf.
another good (probably better) IC for this purpose is the OP07, but it also wants a compensation cap, this one only needs 30pf between pins 1 and 8.

cheers, Iain

[blackieNYC]

It's still pink-noisy.
I tried a 5534 instead of the 741CN - no change.  That's the only other type of single op amp I have.
I added a 10k series resistor between the 1M and the 1nf on the input of the distortion plus.
I tried to isolate power by giving the karate chop a 100ohm series resistor from the 9v supply with a 220uf cap - no change.
I tried adding a grounded 220uf cap to the bias supply - no change.
I found that if I ground the input to the distortion plus (the output of the wah) the noise is gone. So the distortion plus isn't noisy on its own.
I've added a 10nf cap across the 50K gain pot of the karate chop, which reduces noise but does weaken the high end of the wah range a little.

[blackieNYC]

Bump

[iainpunk]

i have this gut feeling it might be the Jfet buffer source to ground resistor that is to high value, but i can't back it up with rationale...
seems like its time to bust out an audio probe and look for the origin of where the noise originates...

cheers

[blackieNYC]

I reduced the value of that source resistor to 6.8K - no change.  The audio probe shows a very low noise at the output of the filter circuit/input of the distortion+.   
I can't believe I housed this in an enclosure and I'm trying to fix noise now.  It's the Karate Shop thing.  It has just enough noise that the high-gain distortion+ circuit that follows turns the noise into an ocean.  At this point I'm curious about adding crossover distortion in the form of two Ge diodes in series.  I've read that it is possible to get noise-gating with this.  I tried this on the very input to the distortion+, but that didn't work.  If I pull the 10K on the output of the 741, is it possible I can gate out the noise by inserting two anti-parallel diodes there?

[iainpunk]

I've read that it is possible to get noise-gating with this.  I tried this on the very input to the distortion+, but that didn't work.  If I pull the 10K on the output of the 741, is it possible I can gate out the noise by inserting two anti-parallel diodes there?

i'm a connoisseur of "crossover diodes" / ''dirty gate" style. it can be an awesome anti-noise tool, but only if the noise level is atleast an order of magnitude smaller than the signal level, if its an ''ocean of noise'' you describe, it can be that it will just pass over the gate without much hassle.
Si diodes are better at reducing noise that Ge diodes are due to their inherent leakage. this does mean however, that you need just under 6db more gain to ''open'' the gate when using silicon diodes.
if you add one of them dirty gates, place it before the clipping diodes, if you place it after, it might be a silencer.

cheers, Iain

[blackieNYC]

Tried it with Si diodes - good enough!