Envelope filter construction

[POTL]

Hi all.
How about brainstorming?

I thought about creating an envelope filter. All analog circuits are roughly the same. Something more popular and often cloned (Dod 25, Mutron iii), something not so popular (MXR) or not used in commercial devices.

In general, the envelope filter is structurally (but not sonically) a cousin of tremolo, phaser, compressor, etc.
Yes, yes, because it essentially has an audio part and variable resistors that can be controlled by a detector or LFO. We live in a time when analog components are becoming a thing of the past (Jfets, Germanium, Opticals, OTA etc.) and devices are moving to new components, DSPs or even plugins.

Conservatives or just lazy people (like me) who didn't study programming and likes to dig into the components continue to build or design Analog devices, while others are programming with might and main bypassing the limitations that we have in the analog world

I thought about the analog envelope filter and the components from which we can assemble it.
My standard approach is to study how something works, then study what it can be assembled from, and in the end choose the least problematic option.

ok, let's go.

1) OTA - mad professor swaw, dod, simple, convenient, but has a number of disadvantages, such as noise and sound, which is not a role model, since OTAs have never been famous for being a good audio amplifier.

2) Jfets - I haven't come across commercial schemes with them, but there is nothing difficult in adapting the scheme for them. The cons we know are clipping, the need for matching, and they're as rare as unicorns.

3) VCA - does not have the disadvantages of OTA or Jfets, but it is very expensive, let's leave them for compressors and noise reduction. I have not seen commercial projects.

4) LDR, in many countries of Europe or the USA it is rare and expensive, requires matching, is gradually becoming a thing of the past, but is still relevant for projects.

5) cmos/pmw - not an example of good sound, may have clipping (but not as bad as Jfets), are produced and cost little, do not require matching.

It would seem that take option 5 and rejoice, but the schemes known to me (MXR) require a bunch of non-standard components.

And I thought, because you can take the standard Mu tron, Q-tron circuit and replace LDR with cmos. we only need to set up the detector. You can forget about the search for rare components and, in theory, get a good sound. I know successful projects in phasers circuits, maybe there will be a good result here too. someone tried it? any ideas?

[ElectricDruid]

Personally, I'd say that (3), (4), and (5) are all valid courses of action. Here's why:

3) VCAs aren't that expensive. Ok, THAT series stuff is, or some parts by Analog Devices. But you can get an AS2164 clone for cheap, or an AS3360. Or even a V3320 or AS3320 VCF. Why even build the filter yourself if you can get one on an IC?

4) Again, LDRs are still around. Xvive clones do the job, not rare or very expensive. Ok, the matching is pretty poor, but it always was. In many ways, that's part of the "character" of vactrol circuits.

5) PWM Sound is "not good"? Where did you hear that? The 4016 and 4066 are still around and there are a ton of circuits that used/still use them, so there's plenty of options for "research". The sound quality depends on how high you get your PWM frequency (high enough that it doesn't interfere with the audio, hopefully!) but with PWM filters you've got a head start, since they naturally reject a lot of the high-frequency noise that you might otherwise see (in a PWM compressor, for example).

Yes, you could absolutely take a Mutron-style SVF circuit and adapt it for PWM switching using CMOS. That's pretty simple to do. I seem to remember someone here doing exactly that. They were driving it with a PIC rather than an envelope generator, if In remember correctly, but it was a nice sounding PWM filter. Apologies to whoever it was, because I can't remember the rest of the details. Do please jump in and correct me.

[POTL]

Hello.
as always glad to see your answer!
V/AS3320 I will go to study this chip
Ok google show me... 

[Mark Hammer]

Omitted from the list were BJTs, as seen in the Dr. Q and Bass Balls.  The rectifier stage could use a little work, in both instances, but the staggered bandpass filters in the Bass Balls present some interesting possibilities.  Note the similarities between the Bass Balls and Craig Anderton's Bi-Filter Follower.

One of the things I'm curious about is the use of separate lowpass and hipass filters to achieve wider passbands, rather than the traditional peak-with-sloped-sides.  So, imagine a 2-pole lowpass with high resonance and a one-pole tunable hipass, such that the low part of the pass band can be "trimmed back" to the user's taste, providing thicker and thinner-sounding sweeps.

My longtime favorite filter was the original MXR unit, because of the variable attack, which allowed for better matching to the feel of the song. There is a temptation to want to add variable attack, but the MXR had the advantage of buffering the envelope.  In many instances, the simple addition of a series pot to vary the attack time undermines sensitivity, so I avoid it in other applications beyond the MXR.  Variable decay is easier to achieve without interfering with anything.

Finally, don't ignore mixing clean with filter, and the phase of the clean as well.  I did this to a Dr, Quack for a bass-playing buddy and he loves it.  Don't underestimate the value of making a filter sweep sound subtler.

[Mark Hammer]

Here's another angle....
Consider the EHX Blurst.  It's a modulated filter.  You can adjust the LFO shape, and there is an expression pedal that can be assigned to any of three parameters, but ultimately, it's a single filter that sweeps up and down.

Okay,now consider the various "talking" pedals. These generally involve countersweeping bandpass filters to mimic movement of speech formants.

What if we were to combine one BP modulated by an LFO, and a second one, modulated by picking strength?  Sometimes, if the envelope-controlled filter was sweeping up, and the LFO modulated section was sweeping down, you could get unpredictable "talking".  Not Hatsune Miku, but somewhere in that neck of the woods.

Ideally, this includes a mixer and balance control to adjust level of each filter..  Make it fully envelope, and you have a Dr. Q.  Make it fully LFO and you have a simplified Blurst.  The "magic" comes with combinations.

[Kipper4]

SVF?
Easily built and modulated with 3 outputs,


https://www.diystompboxes.com/smfforum/index.php?topic=119314.20

[idy]

The MXR is very good sounding and practical. The PWM method seems artificial but the result is as "natural" as anything.

Meatball/Meatsphere is a good rabbit hole to go down. I add the choice to blend in pre/post effects loop dry signal (buffered) and invert the filtered sound: with SVF the HP and LP are inverted. Inverted is not necessarily worse, but it is really different (when you mix with dry.)

I also liked the Snow White Auto Wah.

[Kipper4]

   ^ That reminds me of Ray Rings mini mutron (aka cream wah)

[PRR]

5) cmos/pmw - not an example of good sound, may have clipping (but not as bad as Jfets), are produced and cost little, do not require matching.

OTOH, the EMT-sold PWM (and I think CMOS?) dynamic range processors developed by Blesser were stunningly good, but COSTLY. For a while they were on front of the best disc-cutters. Despite high cost to refurb and the rise of digital "copies" they are still sought-after.

If you are just going to make one, and value your time, it will be much cheaper to just go buy one than to explore every branch on the decision tree.

[garcho]

AS3320 I will go to study this chip. Ok google show me...

I just built a voltage controlled low pass filter pedal with an envelope detector using that IC. It sounds great. With the envelope detector and the resonance up, it's super funky. Bwow to the max. But there are a lot of sound possibilities beyond that, too. I used +/- 9VDC via LT1054 and it works just fine. Tom put up a ton of super useful and helpful info and schematics at the Electric Druid website. And of course, the datasheet. Start there, not google   

[Mark Hammer]

AS3320 I will go to study this chip. Ok google show me...

I just built a voltage controlled low pass filter pedal with an envelope detector using that IC. It sounds great. With the envelope detector and the resonance up, it's super funky. Bwow to the max. But there are a lot of sound possibilities beyond that, too. I used +/- 9VDC via LT1054 and it works just fine. Tom put up a ton of super useful and helpful info and schematics at the Electric Druid website. And of course, the datasheet. Start there, not google   

Did you use it as a 4-pole lowpass?

[POTL]

Omitted from the list were BJTs, as seen in the Dr. Q and Bass Balls.  The rectifier stage could use a little work, in both instances, but the staggered bandpass filters in the Bass Balls present some interesting possibilities.  Note the similarities between the Bass Balls and Craig Anderton's Bi-Filter Follower.

One of the things I'm curious about is the use of separate lowpass and hipass filters to achieve wider passbands, rather than the traditional peak-with-sloped-sides.  So, imagine a 2-pole lowpass with high resonance and a one-pole tunable hipass, such that the low part of the pass band can be "trimmed back" to the user's taste, providing thicker and thinner-sounding sweeps.

My longtime favorite filter was the original MXR unit, because of the variable attack, which allowed for better matching to the feel of the song. There is a temptation to want to add variable attack, but the MXR had the advantage of buffering the envelope.  In many instances, the simple addition of a series pot to vary the attack time undermines sensitivity, so I avoid it in other applications beyond the MXR.  Variable decay is easier to achieve without interfering with anything.

Finally, don't ignore mixing clean with filter, and the phase of the clean as well.  I did this to a Dr, Quack for a bass-playing buddy and he loves it.  Don't underestimate the value of making a filter sweep sound subtler.

Hi Mark.
Thank you for your answer.
I actually forgot about BJT, but I never considered Bass Balls and Dr.Q to be a scruffy thing either. If I considered MXR and BOSS pedals to be 99% fundamental and serious projects, then EHX products look like the work of a mad genius, on the one hand they are incredible things, on the other hand something simplified and untidy. Design decisions (schemes) for the names of effects and so on are also very controversial. the most controversial company of the big ones.

I'm sorry, I digress.
tell me more about the modified attack, I do not fully understand the MXR scheme. By the way, we can put the control circuit in order, it has a lot of components with non-standard values, like 22MΩ. Perhaps the chain can be reduced by 10 times?

[POTL]

ok i cant buy 3320 chip in my country. so I'm going back to my originals. LDR - I can buy them everywhere and cheaply. PMW is a similar situation.

[ElectricDruid]

You can't get 4016 or 4066 cheaply anywhere? Where are you?!? Those seem pretty basic, but maybe now they're too vintage to be easily available, I dunno?

I'll send you a 3320 to play with.

[POTL]

You can't get 4016 or 4066 cheaply anywhere? Where are you?!? Those seem pretty basic, but maybe now they're too vintage to be easily available, I dunno?

I'll send you a 3320 to play with.

4066 and 4016 cheap and available, just like photocells. Thanks for the offer to send me 3320. I think it will not be so easy to do in the current conditions, to Russia, as far as I know, not many companies carry out the movement of parcels, as well as difficulties with financial transfers. but thanks for the suggestion!

[garcho]

Did you use it as a 4-pole lowpass?

Yep. I basically just made the datasheet circuit with the output buffer/resonance compensation a la Sequential Pro-One, thanks to the info Tom posted on his site here, about CEM IC filter designs. Earlier in the thread I linked to a different article than I intended.

[POTL]

thank you for your responses let's try to talk more about the MXR envelope.

as I said, 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 and we can reduce the value of the resistors by 10/100 times, right?
1) R8, R6 = 100K, 510K
2) R24, R15 = 220K, 100K
3) R3,R9,C1 = 100K, 10K, 10uF In general, the scheme is clear, but I can lose sight of something.
in the diagram, I don't understand the fragment after R28 and before the TG control point. Who can explain to me?

[idy]

I have not seen "that" mxr filter: it looks like low mid high are three outputs of SVF you can mix, neat.

[ElectricDruid]

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.

[POTL]

OK I will look deeper into the issue the final solution is to adapt the detector in the mu tron ​​circuit to replace the optics. there is an idea to build several identical layouts with different elements being variable resistors and compare them.