Input stage, sound supply noise

[spyrusthevirus]

Hello folks.
I'm making a simple clipper pedal based on this rather smart idea http://www.diystompboxes.com/smfforum/index.php?topic=89363.0 . I'm using different topologies and signal chain, but the idea behind it all is the same vacuum tube diode.
I haven't worked with a guitar(don't even own one), so a few electrical things are a bit fuzzy (oh, the puns) for me. While any usual sound source, like a computer, was ok to plug in my dirty breadboxed version, the guitar created tons of noise in what I suspect to be some kind of parasitic positive feedback. My dirty version did not have a buffer put in though, so that might just answer it. The chain was like this: Opamp acting as an inverting 2 pole filter -> clipping stage ->speakers.
But what is the best strategy to handle a guitar as an input anyway? That thing has coils and magnets, picks up any shit that electromagnetically floats around the room and all that sounds like trouble in my engineering mind. Is a simple opapmp buffer ok? Could I combine the buffer stage with a pre-gain stage, or should I use yet another opamp?

I also suspect the tube had a little to do with the noise. In that wire jungle that was my breadboard anything could pick up coupled signals from stuff next to it, and a heater with 300mA going through it propably played some part in the whole thing. My power supply was not clean (switcher passed through an lm317) and my ICs were running from 9V batteries. Messing up the ground also comes to mind.
So to be sure that the tubes don't cause any noise and ripple, I intend to give them their own power transformer, and sufficiently large capacitor (plus, linear regulator, ofc). The question is, how large? What levels of ripple would be acceptable for this application?
The signal processing part will have it's own transformer and pleantly of voltage headroom, making sure that no clipping is caused by the opamps. This might be an overkill since those clip at about 0.4V at 4-5V heater voltage but the cost is negligible.
Lastly, how do I handle grounds in this seperate power supply scenario? Just use the IC ground for the signal ground as usual and don't really care about the tube supply?
Is it customarry to use 3 terminal(ie, grounded) power chords for pedals, or should I go for the two terminal one.

I'm pretty tube and guitar naive, but I know a thing or two about electronics. I need to use conventions I know little about (like, what's the usual signal level of a guitar output? of an effect chain?) to make this "compatible" with your average guitarist.
Just so you know, the preliminary test sounded pretty sweet.

[PRR]

Welcome.

> guitar ... has coils and magnets, picks up any sh*t that electromagnetically floats around the room and all that sounds like trouble in my engineering mind.

The pickups are dipoles. Hum-buckers are small dipoles, but even single-wound pickups are dipoles. Sensitivity to stuff in the room falls-off with distance very quickly.

Yes, they do pickup stuff. But with good magnets and strings, and staying *away* from power systems, it is acceptable (obviously).

> what is the best strategy to handle a guitar as an input

Over 100K (better over 470K) resistance, under 200pFd capacitance.

They were evolved/designed to feed a cable and a vacuum tube grid. A few hundred pFd total (mostly in the cable) is part of the tuning (treble peak). Low resistance would swamp this peak. Tubes easily offer 1Meg inputs and pickups/cables are designed to need something in this area.

Merlin's plan has that, already in there.

> created tons of noise

When you want to *fix* something, don't say "noise". Identify the noise. Hiss, humm, squeals, Mexican radio, crackles, cackles, burps.....

> that wire jungle that was my breadboard anything could pick up

Part of the art of breadboarding is learning what "noise" is sure to happen with a certain technique of breadboarding, Hay-wire on cutting-board sitting under fluorescent lamp is sure to buzz. Working a mile from a 50,000 Watt AM radio station, expect a lot of sports-talk.

> suspect the tube had a little to do with the noise.

Divide and conquer. Take the tubes out. With Bright and Fat pots at minimum R, the circuit should pass signal, considerable gain (possible distortion), but ideally no garbage.

> pleantly of voltage headroom, making sure that no clipping is caused by the opamps.

How are they gonna clip? Input and output buffers are unity gain, and guitar signal is always less than 9V peak-peak. The tube-clip stages will (with tubes in and hot) clip in-the-tubes at 0.5V-1.0V peak, far short of the 3V peak of the opamp with 9V supply.

> to make this "compatible" with your average guitarist.

I think Merlin already considered this question?

While you may be able to make it "better", for today I think you should build it EXACTLY as show in his excellent essay. Then find a guitar and see where (in your opinion) it falls short.

[spyrusthevirus]

thank you for your input.
You're right, I should describe the noise. It was rather loud (louder than the signal after passing the clipping point at the pot), containing a low frequency (although I'm almost certain it was higher than 50hz) hum, with a few low-mid harmonics and some higher frequency thing that I believe was kind of a reapeating click. The clicking's texture was not very sharp and it repeated at what seemed to be almost constant tempo, something in the 90-130 bpm range.
This is the best description I can think of, if it makes any sense.

There is another thing I forgot to mention, which might be important, and seemed interesting anyway. When my friend holded the guitar and touched the strings, the noise increased dramatically. If the guitar was left alone, the noise was def. there, but was heard a lot less.

And what about the seperate psu's for the tube and opamps? anything in particular about supressing any kind of tube noise?

I re-breadboarded the thing again, making it very tidy this time, to try a few other things out. I will consider the rest carefully and get back if something comes up.

I know merlins design works and it does have a nice sound(breadboarded and tried out with various test signals but not a real instrument.). But I know enough about sound sculpting (I'm very interested in synths) and electronics to want to tackle making my own design. Some more trial and error will be involved, but the process is more fun to me this way.



Thanks for the help.