Envelope filter construction

[duck_arse]

I'm pretty sure Rob Strand has done a pull down and fix up of the MXR envelope filter, at least the tonepad circuit copy. and I'm pretty sure he covered that 3 invertor section and how to make it work right - but the search is far too daunting a task for me to take on.

[Mark Hammer]

I suspect the high values are to do with the fact that it's using CMOS invertors in their roughly-linear region as op-amps. Consequently, reducing the values might not work so well.

It's an odd design choice on MXR's part, but I suppose the temptation of six amps in one package for cheap was too much!

Aside from the mixing of the three SVF outputs, the add-an-extra-invertor to flip the PWM waveform and reverse the sweep is very cute.

One of the things to remember about guitar signal is thatit has the most harmonic content during the pick attack.  Filters normally sweep up, such that they start out essentially blocking harmonic content, allowing more in/through asthe filter sweeps up.  But since there is less harmonic content in the signal, shortly after the pick attack, the sweep feels less aggressive.  In contrast, when the sweep goes down, instead of up, the filter starts out emphasizing all that initial harmonic content. By the time the filter reaches the bottom of the sweep, theres not a lotofharmonic contentleft to filter out.

The result is that downward sweep can sound much more aggressive - in an unpleasant way - and jarringly fast.  It isn't really faster, but because the tonal contrast between start and end points of the sweep is so great, it FEELS faster, and sometimes too fast.

In a perfect world, thestart point of a downward sweep would not simply be the end point of an upward sweep.  One can tweak the sweep so the downward starts from a lower point, but that starts to complicate matters and alter the circuit.  The easier solution is simply to lengthen the "attack" time and reduce the threshold.  At least that's what I do to make it sound "normal".  It's still a very clever mod, thanks to Stephen Giles.Itjust needs some taming.

[PRR]

> I really do not like non-standard components, which are like 22 Mohm. there are many such components in the part of the detector. in my opinion, these are simple inverter amplifiers

They are also R-C time-constants. So you need bigger caps. And then you may need bigger amplifiers (not 3-cent CMOS inverters). It's all inter-related.