Tonepad MXR envelope filter issue

[idy]

For what it's worth here are my voltages.  My working A-gua filter shows voltages like supercolio. Mine works, very pronounced filter, not as much regeneration (peak or resonance) as some others...and like I said above at some point I added a mosfet booster with pregain knob to the input. But the attack, threshold and emphasis pots all do their things.  Took them several times, they don't vary much. Can't say I understand them much.

All three supply pins have around 4.1v. ( The power in reads 9.3v at the jack. After the 3k resistor I see about 4.1v.) I did not fiddle with the knobs or the direction switch, But mine seem similar. I am thinking supercolio is confusing "VS" and his 9v input though: on the tonepad schematic "VS" is used as the voltage after the 3k resistor. Don't schematics sometimes call for VS and VD? Source and Drain? In any case on this PCB the voltage marked VS will be exactly what appears on all three power pins. The schematic shows VS several places but does not show the power to the ICs.

IC1 (4069)
1: 2.11
2: 2.01
3: 2.11
4: 2.25
5: 2.11
6: 2.2
7: 0
8: 4.18
9: 0.01
10: 2.21
11: 2.12
12: 2.12
13: 2.11
14: 4.2

IC2 (4069)
1: 2.15
2: 2.0
3: 2.0
4: 2.16
5: 2.16
6: 1.59
7: 0
8: .34
9: 2.80
10: 4.02
11: 1.16
12: 1.09
13: 2.09
14: 4.18

CD 3 (4066)
1: 0.06
2: 0.01
3: 0
4: 0
5: 0
6: .35
7: 0
8: 2.08
9: 2.19
10: 2.08
11: 2.18
12: 0.35
13: 0.03
14: 4.14

Thank you Mark Hammer for starting to explain how this thing works. I have read some other things you ahve written and want to add a decay knob to mine next (C9 and R17.)
So the oscillator has its duty cycle shaped by the envelope voltage. How does this then "switch" the filter open and closed? I was thinking the two switches in the 4066 were like the LDR/LED pairs in other filters. How does a different duty cycle affect them? That's just another way to vary the power being delivered, like a lamp dimmer?

[supercolio]

For what it's worth here are my voltages---

Thank you! It seems IC1 and IC2 are looking pretty good, but IC3 has differing voltages (esp. pins 3-5). I wonder why 

Also I mislabeled "Vs" when I posted my voltages; I meant I got 9.3v before the 3k resistor, and ~4v after it, as pins 14 show it.

I did some scoping and this is what "TG" point looks like on IC3's pin 12, with and without reverse sweep switch. Looks pretty good?

http://imgur.com/a/zUq43     ( I didn't know how to scale images, they came up huge, so I linked the album instead).

The effect sounds like it's not being triggered, no quacking. Could the problem lie somewhere around the 4066? I can post pictures of the pcb if someone wishes so, but it looks fine to me.

[duck_arse]

I don't want to labour a point made many times before, but, there is a rule for using cmos, which say "ALL inputs MUST be tied to a logic level, ALWAYS".

and in this pcb and circuit, there is a pair of unused transmission gates in the 4066, which have unconnected control inputs. best practice would have one spare connected parallel to each of the used, but just pulling the two unused "TG's" either high or low would satisfy all. and the similar applies to the IC1b inverter, which could well have its pin9 'C' connected always to 'G', to satisfy all. (the voltage reading on the pins for these two unused will be meaningless until the control lines are taken care of. [and even then.])

your scope picture looks like something, and its inverse. I can't remember what it is supposed to look like tho. does the width of the pulses change with envelope? are you getting a sweeping envelope follow at D2/C9/R17?

and photos of your board, yes.

[supercolio]

I don't want to labour a point made many times before, but, there is a rule for using cmos, which say "ALL inputs MUST be tied to a logic level, ALWAYS".

and in this pcb and circuit, there is a pair of unused transmission gates in the 4066, which have unconnected control inputs. best practice would have one spare connected parallel to each of the used, but just pulling the two unused "TG's" either high or low would satisfy all. and the similar applies to the IC1b inverter, which could well have its pin9 'C' connected always to 'G', to satisfy all. (the voltage reading on the pins for these two unused will be meaningless until the control lines are taken care of. [and even then.])

your scope picture looks like something, and its inverse. I can't remember what it is supposed to look like tho. does the width of the pulses change with envelope? are you getting a sweeping envelope follow at D2/C9/R17?

and photos of your board, yes.

Next chance I have to work on this is next weekend, so I'll post pictures then. The width changes by tweaking potentiometers. I have to check D2/C9/R17 next weekend, didn't think of that. Thank you all for your help so far!

[Mark Hammer]

Thank you Mark Hammer for starting to explain how this thing works. I have read some other things you ahve written and want to add a decay knob to mine next (C9 and R17.)
So the oscillator has its duty cycle shaped by the envelope voltage. How does this then "switch" the filter open and closed? I was thinking the two switches in the 4066 were like the LDR/LED pairs in other filters. How does a different duty cycle affect them? That's just another way to vary the power being delivered, like a lamp dimmer?

You're welcome.
Virtually all filters will involve a capacitor and a resistance.  The resistance is going to determine how quickly (or slowly) the cap in question is going to charge up.  If the cap value is large/high, it will take longeer to charge it up.  But if the resistance limiting how much current is fed to the cap is large/high, it also takes longer to charge up the cap, even if the cap value is small/low.  The rolloff frequency will be lower, the more slowly the cap charges.

In the case of the EF, and a bunch of other circuits that use duty-cycle of a switch as their "control element", it's a bit like turning an LED light on and off at very high speed.  If the overall percentage of time that the lightswitch is "on" is high, then the light will be brighter, even though it spends some time off.  If the percentage of time is low, the average brightness will be lower.

Turning the CMOS switch "on" completes a path made up of a 47k fixed resistor in series with the switch.  Varying the duty-cycle of the switches makes the resistor+switch combo behave like a  resistance whose  average resistance varies, just like a dual-gang pot.  If you were to replace the two switches with, say, a dual-ganged 100k pot, you would be able to mimic the sound of an EF.  Naturally, what allows this magic to happen is the fact that it is occurring at a high frequency - not CPU/RF high, but higher than the amp and speakers will reproduce - so we tend not to hear any aliasing or  "zipper" sounds like we would if the clock was workng at, say, 8khz.

There are a couple of MXR phasers that use PWM-controlled switches to substitute for FETs, that would normally be used as variable resistances.

Make sense?

[idy]

Crystal clear, thanks. but it's weird to think of audio being PWMed instead of power for a light bulb. How fast is the oscillator?

[StephenGiles]

Faster than we can hear

[Mark Hammer]

Incidentally, I see that Analog Mike Piera has a limited edition release of MXR-EF clones, using NOS PCBs, with our collective mods, including the hi-lo range and variable emphasis/Q.

[supercolio]

I don't want to labour a point made many times before, but, there is a rule for using cmos, which say "ALL inputs MUST be tied to a logic level, ALWAYS".

and in this pcb and circuit, there is a pair of unused transmission gates in the 4066, which have unconnected control inputs. best practice would have one spare connected parallel to each of the used, but just pulling the two unused "TG's" either high or low would satisfy all. and the similar applies to the IC1b inverter, which could well have its pin9 'C' connected always to 'G', to satisfy all. (the voltage reading on the pins for these two unused will be meaningless until the control lines are taken care of. [and even then.])

your scope picture looks like something, and its inverse. I can't remember what it is supposed to look like tho. does the width of the pulses change with envelope? are you getting a sweeping envelope follow at D2/C9/R17?

and photos of your board, yes.

Hi! sorry it took so long to answer, having all sorts of stuff going on at the moment.

Here's a photo (it ain't pretty!) http://imgur.com/a/TFJJC
If someone wants more precise photos I'll post them.

I'm getting envelope follow at D2, at least the scope picks up my playing and attack control adjusts how fast the voltage changes (so I got that going for me..)

On 4066 I tied pins 1, 3, 5 and 13 (SIG A in/out, SIG B in/out, CONTROL A and CONTROL B (http://www.ti.com/lit/ds/symlink/cd4066b.pdf)) to ground. I'm not sure if that made it, but now the pedal sounds more like an envelope control, adjusting the attack of the signal (not like it should, though). There is no treble whatsoever and the output is low.

I'm not sure where the problem lies. Circuit oscillates and it reacts to my playing. I might check voltages tomorrow. Sigh...

[idy]

Oh oh oh you are getting closer. And this is a very nice sounding filter when it works!
The envelops section is working, its the filter that isn't cooperating.
You should audio probe the points around IC1 pins 1-6.

Will reread the thread to see what has/hasn't been modded. What is that jumpery thing, ah a repaired 9v in...

[idy]

By the way, did you ever try it with a booster between guitar and filter. Drive it a little.

[idy]

Also thinking the differences in our IC3 voltages have to do with flipping the "up/down" toggle.

[supercolio]

By the way, did you ever try it with a booster between guitar and filter. Drive it a little.

Hey! I tried some boosting with mxr micro amp with no success. Audio probing is next, right after xmas!

I should upload some sound samples, in case someone might figure out what's going on. For the record, this project is in no hurry at the moment, I'll keep posting updates once in a while. I hope we'll figure this out so that someone else could find this thread useful someday!