Distortion noise reduction

[Mark Black]

Hi all -

I'm in the mood the last few days to investigate noise reduction schemes for knocking down the hiss on a favorite distortion pedal (old big-box Rat). I would visualize some sensing circuitry at the input controlling a downward expander on the output, and would like to build it in rather than have yet another box in my crowded pedalboard. Looking for a way to cut down on my pedal tap-dancing...

So - any ideas or pointers to existing designs (especially clever low-parts-count ones)or informative articles would be much appreciated.

Thanks!
Mark

[Ray Ivers]

Mark,

There's always the 'diode-gate' circuit - two 1N34 diodes back-to-back, in series with the signal; you'll probably also want to include a resistor in parallel with the diodes so it doesn't gate to silence, but rather to a reduced signal level. 1N34's are germanium and will gate at about .6V peak-peak, whereas 1N914's or similar silicon ones will do it at about 1.2V P-P, IMO too high for a pedal-level output.

Will it work as well as a Hush unit or other good noise-suppression device? Absolutely not - but you can't beat that parts count!

Ray

[Mark Black]

Wow Ray - that has to be the king of the minimal parts count! Thanks for that idea, and I might have to try it just to see what it sounds like.

Well I'm sure this would involve a lot more parts, but I'm now kicking around an even more grandiose idea: creating something that would shift the output from the distortion circuit to straight-through clean based on input level so notes could tail off nicely. An input-triggered gradual bypass if you will, like a sort of auto-balance control. Could this be (gasp! ) an actual new idea or is it just old news I'm not aware of?

[Ray Ivers]

Mark,

I'm now kicking around an even more grandiose idea: creating something that would shift the output from the distortion circuit to straight-through clean based on input level so notes could tail off nicely. An input-triggered gradual bypass if you will, like a sort of auto-balance control. Could this be (gasp! ) an actual new idea or is it just old news I'm not aware of?

Well, I haven't heard of it! Don't be coming up with this stuff, now - haven't you heard that all guitar-amp circuits are just copies of earlier designs?

Unlike compression - which I feel has to be of a certain minimum "quality" in terms of dynamics tracking, recovery time, 'breathing', etc. or it can sound bad, noise reduction can be fairly crude and still work OK (although the best NR units aren't crude at all, IMO).

You could do the dirty/clean transition as you suggested simply by reducing the gain of the overdrive channel, using one LDR, with its control LED illuminated by a simple positive half-wave rectifier (resistor/diode/cap/low-voltage Zener) powered from the signal voltage, in conjunction with a similar supply off the heater winding, and the resistor connected anywhere in the circuit that reducing resistance reduced gain (grid resistor, plate resistor, NFB, etc.). In the presence of signal, the Zener-ed signal DC would hold the LED cathode voltage up near to the anode voltage, and it wouldn't conduct (and the resistor would show several megohms of resistance - IMO it's important that it not affect the sound at all), and then after the signal dropped below the Zener threshold the LED would come on, and the resistor value would drop to a pretty low value (all depending on the LDR used - they vary quite a bit in minimum 'on' resistance, response time, etc., as I'm sure you're aware). Anyway, this is just a bare-bones idea; you might need other components and some tweaking to get it right, but it should work.

Ray

[R.G.]

Compand around the pedal, or just around the distortion part.

Or, dig in really deeply on the noise of the pedal. You mentioned that the pedal was old. Most old pedal designs were not done in view of the damage that even small transients can do to transistors and opamps. For instance, reverse breaking a bipolar transistor's base-emitter junction even once will cause a measurable increase in its noise. That 5-7V can easily come from caps in the circuit that get charged up during operation and then discharge reverse through that junction at turnoff, so the breakage happens once per power cycle. A similar thing happens in bipolar opamps which don't have input protection. I wrote a lot of this up in the article "When Good Op Amps go Bad" at GEO - http://www.geofex.com.

You might rebuild the pedal, changing out the input buffer transistor(s) for new ones, changing the resistors to metal film to get rid of carbon's excess noise, and replacing the opamp. If that doesn't help enough, then dig into companding and noise gating.

[Mark Black]

Thanks Ray and R.G. for the ideas.

I really hadn't considered trying to improve the existing noise floor - just figured it was a really noisy design. Could be interesting to quantify which components make a difference based on scope observation.

If/when I come up with something worthwhile I'll post back. In the meantime other priorities prevail, so off to the back-burner for now... (like so many other projects)

Thanks again!

Mark