Making a generic envelope follower

[soggybag]

Here's a schem for a generic envelope follower. I Made this by adding the GGG buffer in front of the envelope from the Dr. Q. I think I used the Dr. Q drawing from Jack Orman's site so it probably has some his mods in it.

I breadboarded this and it seems to be working well. At this time I've just been lighting an LED by playing notes ont eh guitar. It seems to be working pretty well.

I'm hoping this could be a simple envelope follower that could be added to any circuit that might need one. My own plans are to add to the Rate control of the Frobnicator.

I'm looking for suggestions or thoughts for improvements. I'm a novice I'm sure I'm not seeing something that could be done.

[Mark Hammer]

The one I'd like to see a board for is Harry Bisell's envelope follower that he designed expressly for guitar (well, for his guitar synth) and published in EDN.  I've misplaced the link, so someone with better research skills (or the link itself) will have to help out.

The one you've drawn (and thanks) from Jack's chip-adaptable Dr. Quack is certaily simple and cheap.  Harry's is just plain responsive.  I can vouch for it because I've played it. He nails the "holy grail" of envelope followers - ultra-fast response time with little detectable ripple.  If you can imagine and play a dynamic difference, this thing will pick it up.  Usually, if you want a fast attack, you end up with more envelope ripple, and if you try and smooth out the ripple you have to forfeit the fast attack.  Harry's idea uses a round-robin combo of what are essentially 3 envelope followers combined into one.  Sort of a big build.  In fact more components than a great many effects pedals themselves.  But if the goal is a nondedicated envelope follower that can be assigned to different things, this is the shiznaz.

[Dave_B]

The one I'd like to see a board for is Harry Bisell's envelope follower that he designed expressly for guitar (well, for his guitar synth) and published in EDN.  I've misplaced the link, so someone with better research skills (or the link itself) will have to help out.

Mark, is it this one? http://go.cadwire.net/?3572,3,1 (click on the links for the schem)

[Mark Hammer]

That's the one.  A real keeper.

[Paul Perry (Frostwave)]

How difficult would it be to implement Harry's envelope follower idea in a PIC micro?

[StephenGiles]

A very good weekend to you Mark, this is the link I have to Harry Bissell's full envelope circuit and there seems to be a PCB, but I think it needs a bit of unravelling! Do try my Spacedrum adapted Envelope Generator which is great with the Bi Filter Follower filter, albeit a little fuzzy, for reasons better known by greater minds than mine!

[R.G.]

How difficult would it be to implement Harry's envelope follower idea in a PIC micro?

Difficult. The analog portion of sensing the peaks is not something that the PIC does well - yet. The digital part is trivial, of course. A 16F629 in an 8 pin DIP would do nicely, and is smaller than an oscillator and a 4017.

The one I'd like to see a board for is Harry Bisell's envelope follower

I did one of those. It turns out not to be a nicely behaved circuit for effects. The board is big, even trying to stuff it down. There are four dual opamps, three transistors, and two other chips, either 14 pin or one 14, one 8. You can replace the 14 and 14/8 with an 8 pin PIC, but you still have enough junk in the peak/holds to make an entire phaser. This is one circuit that begs to be made into an SMD module that you could graft onto another board. I've done that as well, but most people are afraid of SMD.

[StephenGiles]

RG that sounds a bit pessimistic, I would expect to use a considerable number of chips if the end product warranted the effort, and Mark seemed very impressed with it.
Stephen

[R.G.]

To me, circuits have a feeling attached when I do layout. That's my own personal interpretation, of course.

The three-peaks envelope follower does not produce a satisfying layout, either of the several times I've done it. It's somehow resistant to tight, efficient layout, and that irritates me. I've messed with the circuit, recast it with other parts, etc. I've even looked for oddball ICs that could be pressed into service for it to do the layout better, with no luck.

I have no doubt at all that it's a superior envelope follower. The simulation runs show really good results.

There are a few semi-natural-laws with effects, a result of the history and what guitarists expect. You can violate one or two of them, but things get unwieldy after that. One of those is a smallish box. An envelope follower that takes up as much circuitry as the effect itself is a problem, it feels like IMHO.

So, yeah, I'm pessimistic about the practical aspects of layout and inclusion in a box, but not at all about the performance. That's why I suggested encapsulating it in an SMD self-standing circuit.

[soggybag]

I feel there must be simple effective envelope that can be used in many circuits. Something that can be made with a single op-amp and a few more parts. A complex circuit is not as useful, and causes problems for amateurs like myself. I had the idea of coming up with something that could be grafted onto any circuit where you felt an envelope might enhace or create new sounds.

[StephenGiles]

Ah RG that's the perfectionist in you springing forth! I'm just happy if I get something to work, and as 95% of what I make is on veroboard (that's gone quiet hasn't it!) .......need I say more? I must confess that when I gave the Bissell Envelope Follower a try on breadboard I couldn't get it to work but I certainly intend to try again.
Stephen

[Dave_B]

How difficult would it be to implement Harry's envelope follower idea in a PIC micro?

Difficult. The analog portion of sensing the peaks is not something that the PIC does well - yet.

Is there a reason for that?  In my mind you're just measuring the first peak off the A/D converter.  From there it would follow the waveform until it reached the noisefloor, whatever that might be.  As it rides the waveform down, you'd be smoothing the waveform in software  rather than relying on caps, so it seems like it could be pulled off. 

/scared to attach my homemade PIC programmer to the serial port so I may never know...   

[StephenGiles]

I'm not convinced we want to follow the waveform because that is probably asking for ripple. Surely it is better to read the initial peak of the note, then have the decay decided by separate circuitry independent of the decay of the guitar note, which is just what this does - although there should be a 10k resister out of pin 3 of the 13600 to the wiper of the start frequency pot.

Stephen

[Transmogrifox]

How difficult would it be to implement Harry's envelope follower idea in a PIC micro?

I am actually not very skeptical of that at all.  The first derivative of the waveform always crosses 0 at each peak on the waveform.  The second derivative tells you if it's concave up or concave down (so you can determine if it's a maximum peak or minimum).  The "first difference" is very easy to implement, and you just set a test bit to determine when x[n] - x[n-1] is reasonably close to 0, then you store the sample at that point in time in a registor, or even a RAM location until
1) Something with a higher amplitude comes along, then store the new maximum
2) you reset it to 0, then take the next series of maximums into the buffer and hold until you reset it again.

You could do this with 3 registers or memory locations in a round robin fashion--you're just choosing a different time to reset the register (or memory location) at different times.

In the end of it, you can add the outputs all together, drop it into the A/D converter and throw a simple RC filter on the output of the PIC chip like on the output of Harry's circuit to avoid having to pull any crafty programming to implement a digital filter in between the rest of the processing.

[soggybag]

I think I'm with Stephen, an envelope generator might be a better choice. A PIC has great potential but you need special equipment to create and program the PIC, not to mention the know how.

If there were an envelope generator that could be made from a dual op-amp this would be perfect.

[StephenGiles]

Why do you want to use a just one dual opamp soggybag?
Stephen

[soggybag]

I got started on the idea when I was playing with the Ring Fronnicator, I thought it might be interesting to add an envelope to control the frequency. This could be interesting in the Ringing mode. I made up the Dr Q envelope on a breadboard. I used an LED with an LDR across the frequnecy pot. It worked pretty well after I added the buffer.

After working on this I began to think that a simple envelope follower might be a good circuit block to have around to add to anything. And it would be especially handy if it was compact. A dual op-amp and some other parts makes a nice easy to work with package.

The schematic above is what I came up with. It's sort of a kluge of the op-amp buffer and the envelope from the Dr Q. The control is simple, one knob determining how strong the effect. It works with an LED/LDR which is pretty easy to fit into most circuits.

Having only one control makes it easy to add to a project. If you have an effect with two knobs, you don't want to add three more. But you might want to add an envelope to your Phase 45 or easyvibe with a knob to determine how much the envelope is effecting the sound. If the whole thing is just a dual op-amp, a few extra parts and one knob it could be added to a lot of projects that are already boxed up with minimal trouble.

[StephenGiles]

Ah that explains it.
Stephen

[Paul Perry (Frostwave)]

It's easy to get sucked in, with envelope follower designs.... I find myself working wiht an "ideal" guitar signal in mind, rather than the horibble spiky mess full of crossovers that you actually see on a scope. I think that is why there are so many disappointments in envelope follower design. I have a PIC based product (the Sonic Alienator wiht variable sample and bit number, also distortion algorithms) that works on a PIC. (But, someone else did the programming!). So I'm sure it is possible to use a PIC.
Maybe Steve or someone would be happy to sell an envelope follower PIC as a product, with a couple of bucks royalty to the programmer??

[smccusker]

Why couldnt you have the envelope detector as a separate box with a CV out, like analogue synths?
then just mod your pedals so that they could take a CV in, and have some sortof switching to switch between which values are controlled by the envelope of your signal.