Release yourself - more experiments with filters

[Mark Hammer]

As noted in an earlier thread this week ( http://www.elixant.com/~stompbox/smfforum/index.php?topic=41437.0 ), I've been experimenting with altering the value of the resistance in parallel with the averaging capacitors in envelope controlled filters.  I had success with the Bass Balls, and pleased to now report that the MXR Envelope controlled filter (the A-Gua layout at Tonepad) also lends itself very nicely to this remarkably painless mod.

The capacitor to ground, in tandem with the small resistance (usually <1k) leading up to it set the attack or onset/rise time of the filter as it sweeps.  The cap also sets the decay time, with larger caps taking longer to discharge.  Until recently, I've been foolishly recommending people to switch to larger or smaller caps to vary the decay time.  How blind I have been.

Typically, the cap discharges through whatever paths are available to it, but one can accelerate that discharge time by providing the cap its own specialized path to ground via a resistor/resistance to ground in parallel with the cap.  Theoretically, you can start off with a larger value cap than called for and provide a parallel resistance to ground that varies from fairly high (where the cap bleeds off at a rate not much different than having no resistor) to a rate much faster than you'd expect for a cap that size.  All of this is, of course, mere child's play and old hat to those with a taste for analog synthesis, but it was one of those things I had to finally get around to doing to know its value.  Dare I say it, I think varying the release time has more impact on the feel of the pedal than varying the attack. 

No, actually, let me revise that and say that it is easier to achieve meaningful variations in feel by altering the release time than it is to achieve variations in feel from changing attack.  Once you've increased the attack time a bit, you start to get diminishing returns by virtue of the way the series resistance reduces envelope drive at the same time as it increases attack time marginally.  In contrast, you can vary release time over a wide range without impacting negatively on performance.

How do you do it on the A-Gua?  The key component is R17, a 1M resistor tied to pin 14 on IC1.  The suggested implementation is to remove it from the board.  Next, solder a 47k resistor (1/8W would probably work best) to one of the lugs of a 1M pot and secure it with a piece of shrink wrap to reduce strain and risk of fracture.  Now solder one of the leads from the board to the free end of the 47k resistor, and solder a lead from the other R17 pad to an adjacent lug on the pot.  You'll now have an A-Gua that will go from the normal sluggish settling down of the filter to whip fast.  While you're at it, you might want to change C9 (.05uf) to a .1uf value.  This will give you release times from much faster to much slower.  In combination with a very effective attack time control, this makes for a filter that can achieve a wide variety of feels.

On the Dr. Quack, Nurse Quacky, and of course Dr. Q, there is no existing bleed resistor, but you can add one easily.  I suspect the same 47k+1meg arrangement can be used.  This capacity to expedite release and filter settling may well be an excellent addition to the Quack and Q for the more trebly filter mode.  My experience with them is that the Bass mode settles down a lot faster than the treble mode, with the treble mode often taking too damn long to be useful.  You would place it in parallel with the 10uf or 22uf cap to ground (whichever you have).

You will note that the Meatball already has such a control.

If you have a Boss TW-1 Touch-Wah, it already HAS such a resistor...but you can still make it variable.  It is normally a 68k resistor in parallel with what looks like a 1uf cap (fuzzy schem, sorry).

The Rattle-Crow has a 100k resistor in parallel with a 2.2uf cap.  That's the one.  The DOD 440 has a 1uf cap and no resistor.  You can add one.  The Funny Cat has a similar 1uf cap with no obvious bleed resistor.

There are a number of pedals that have a slightly different arrangement in which the cap is charged up by being tied to a DC source, and discharged (well, "counter-charged" really)  by whatever the output of the envelope follower contributes.  Think of the Dynacomp/Ross compressor in which there is normally a 10uf cap to ground and a 150k resistor from V+ to the cap.  The variable recovery mod for the compressor involves reducing that resistance so that the cap gets charged up faster.  150k = slowest recovery/recharge, and 10k = fastest recovery/recharge.

Pedals in this category include the Mutron, the DOD FX-25, the Maestro/Oberheim Sample and Hold (FSH-1).  Here you can vary the resistance between the voltage source and the cap, just as you do in the Ross/Dynacomp.  The modded version of the FSH-1 over at GGG already has such a control shown (thanks!!).

I think the nice thing about adding a release control is that you can start with a bigger value cap than the schematic calls for, in order to assure the *possibility* of nice long ripple-rejecting decay, and just edge the control down to get a release time that balances speed with ripple rejection.

Try it, you'll like it.

[zeta55]

Intresting reading Mark. Gonna try this mod on my MXR filter when I get some time alone with my soldering station.