Noise control in modulation and other pedals

[Mark Hammer]

Before the Signetics NE570/571 became a frequent fixture in modulation pedals as a way of keeping internal noise under control, a number of pedals used a gate approach.  You can see an example of this in action here in the Boss CE-1. http://www.mif.pg.gda.pl/homepages/tom/files/bossCE1.jpg  The same basic circuit idea isd also used in the A/DA Flanger, the Ibanez SC10 chorus, and other pedals.

Referring to the CE-1 schematic, IC3 is the section in question.  It takes its input from the first gain stage, before anything else is done to the signal.  The several diodes on its output provide a half-wave rectified "envelope" signal.  Follow the signal along from there, and you can see that it goes to the gate of Q12, a FET.  That FET, in tandem with R48 (2k2) and R50 (100k), forms an attenuator circuit.  When Q12 is off R48 and R50 acts like a 102.2k pot turned down juuuuuuuuuust a hair.  When Q12 is turned on, its resistance drops dramatically, and the signal is essentially turned wayyyyy down by the attenuating action resulting from the low-resistance path to ground.

The "trick" used here is that the FET is turned off when there IS signal, and turned on when the signal drops enough.  R66/67 and D13 feed a default voltage in that the envelope essentially counteracts, resulting in the sort of "reverse gate".

The nice thing about this is that, in this and similar applications, there is always a dry signal present, so turning down the processed signal level removes any noise (click, hiss, clock whine, etc.) stemming from the wet signal path, while leaving the dry path undisturbed.  Since the dry signal is always there, the gating action on the wet signal goes largely unnoticed, even though the same gating applied to the combined signal would be irritating.  It might be possible to design another gate from scratch, but the nice thing about this one is that the attenuating action can be made to occur slowly.  Again, because the clean signal is always present, masking everything else as the signal level starts to decline, we tend not to hear this as a brief moment where the hiss/noise becomes audible then suddenly silent.

Although the manufacturers using this sidechain/wet-gate approach never seem to include any adjustment of it, the sensitivity of the gate CAN be adjusted to be more and less responsive.  In the CE-1 example, increasing or decreasing R73 (470k) would increase or decrease, respectively, the gain of that stage and its sensitivity.  Similarly, tinkering with the values of C42 and R68 would get you different decay times to suit the properties of the pedal and your taste/playing.

I raise this because people have commented on ticking or other noise in assorted modulation pedals over the years.  Sometimes the solution to that is simply to improve the decoupling of chips, using the same PCB and a few extra components on the copper side.  Other times, the problem doesn't go away quite so easily, and it seems to me this little add-on could be the basis of productive noise reduction.  The nice thing is that, as a sidechain, it can truly function as a separate "daughter board", and be patched in at the relevant points in the problem circuit.  Does it provide better noise control than a 570?  My guess is no, but the nice thing is that it is adaptable to almost any dry+wet modulation circuit (and some wet-only types too, as well as some hissy fuzzes) and does not require a complete redesign the way that implementing a compander chip would.  Since it only involves a single op-amp, a FET, and a couple of passive components, it is the sort of thing that could easily fit on a small daughter board and find its place nestled in amongst the pots.  The tricky or nuisance part would be the wires running to the main board (V+/Vref/Gnd, input signal, FET drain),

[Ben N]

Mark,
That idea of a sidechain noise reduction daughterboard is a great one. Would work right into some of RG's modular fx concepts down the years, or would make a great board/project for Tonepad/GGG/Small Bear.
That mind of yours just keeps on churning! (Musta been pretty bored in shul, eh?  )
Ben

[Joe Kramer]

Similarly, tinkering with the values of C42 and R68 [of the Boss CE-1] would get you different decay times to suit the properties of the pedal and your taste/playing.

Interesting coincidence, as today I happen to be doing a check-up of my friend's CE-1 before he goes off on tour to Japan.  He complained that when he uses a high-gain effect before the CE-1, it makes a sort of momentary whooshing sound and then falls quiet.  I explained to him the CE-1's noise gate and told him I'd take a look at trying to speed-up the gating effect.  But now that I'm looking at the PCB, C42 is a .47 electrolytic, whereas the schematic indicates ".047."  Since the schemo does show C42 as a polarized cap, I'd guess that the cap on the board is correct and the schemo is wrong.  Otherwise, a .047 film cap would have easily fit.  In any case, I'm going to try a few different values, starting around .1, and see what happens. . . .

EDIT: I've tried this mod.  The smallest you can go with C42 before you hear "chattering" is .1uF, but this isn't really acceptable because the decay is too abrupt, and is especially noticeable with the vibrato setting, where all you hear is wet signal.  Using .22uF is okay, but while the decay is less abrupt, it's still too extreme.  And of course going any bigger retains the same "whooshing" problem as the stock value.  I think I'm going to leave it stock.

BTW, your daughterboard idea could also include the gating of only higher frequencies, say 1KHz and up, so to avoid the perception of abrupt decay.  I guess that would entail roughly splitting the wet signal into two bands, low/mid and mid/hi, and then applying the gate only to the mid/hi band.  Fancy but still pretty simple I'd think.

Regards,
Joe

[cpm]

BUMP!

Mark, you should make a compendium from all your post and snippets scattered along the interwebs, related to BBD designs.
i am satarting with these and usually my searches always en on a text of yours.

[Auke Haarsma]

good read and thanks for the bump!