Need assistance troubleshooting a DOD 440 Envelope Filter (Vactrol) Clone

[bushidov]

Hi All,

So I made a clone of the DOD 440 Envelope Filter based off of the General Guitar Gadgets one (http://www.generalguitargadgets.com/effects-projects/filters-envelope/dod-440-ef/)

Earlier, I tried making one based off the Beavis Audio site, but discovered (as did the folks at PedalPCB), that they have an incorrect schematic, due to the Vactrol being used is a 4-pin one, when to make a DOD 440, it is a 5 pin one.

Attached is the schematic I made off the GGG one.


I assembled the PCB and socketed the Vactrol, as I've heard rumors that XVive ones may be a problem (apparently, there is a current post on the forum indicating as much).

It "envelope filters"... kind of.

Here's the sound it makes.

Here's the pot adjustments:

This is what is happening with my XVive VTL5C4/2 vactrol. Some folks indicated that this "distorted wobble" I am hearing is ripple. If that is the case, how do I fix that? Get another vactrol? Roll my own? If roll my own, what LEDs and LDRs and where do I get them? Or is this another problem all together. I checked my capacitor and resistor values and checked orientation for my diodes. I did check in EAGLE Cad that I didn't have any air-wires missed or traces that accidentally intersected, but it is possible that my schematic is flawed.

However, what I think is incorrect is the XVive Vactrol. I didn't get it from a sketchy ebay site, but rather from Small Bear. They've been really reliable for not getting crap, so I'd like to hope the Vactrol is ok, but it might just be bad?

Any help?

[Mark Hammer]

Wobble/ripple will generally not be a product of the vactrol/LDR itself, except in some very rare circumstances with some relatively rare LDRs.  For the most part, vactrols/LDRs are used precisely because they don't exhibit, or even smooth over, ripple.  The ripple comes from the actual follower stage, 99% of the time.

The 440 is set up for fastest attack and response time.  Consider changing C5 from 1uf to something larger, like 3u3 to even 10uf.  If 10uf seems too slow a decay, you can speed up decay time by wiring in a pot or 3-posiion toggle to place a resistance to ground of 22k-1meg in parallel with C5.

And note that some of the ripple is not because of imperfections in the rectifier circuit, but because the guitar itself provides a very uneven amplitude as plucked strings decay.  Perceptually, it sounds like a smooth fade-out to us, but physically it has a lot of bumps, which is why ripple is always worst as the strum dies out.

I made a 440 clone a while back, and used a square LED I pulled from a junked cassette deck or something, that allowed LDRs to be attached on each side face and receive the same illumination.  It's a nice pedal when you get it working right, with a very vocal sound.  Hope you can get that.

[bushidov]

So, is it supposed to be making that sound that is in my YouTube samples? I'll try changing out the values today.

[rankot]

Wobble/ripple will generally not be a product of the vactrol/LDR itself, except in some very rare circumstances with some relatively rare LDRs.  For the most part, vactrols/LDRs are used precisely because they don't exhibit, or even smooth over, ripple.  The ripple comes from the actual follower stage, 99% of the time.

Mark, I believe that using vactrol or LDR with really short release (fall) time can cause this as well, because the difference between 30ms and more than 1 second is huge. So he can probably try to compensate slow fall time with what you suggested - bigger smoothing capacitor, or even changing R10 value (maybe put a 100k pot there, and try what happens).

Erik, you can try to put 2M2 instead of 430k at R13, when I simulate this circuit it works much better with this bigger value.

[Mark Hammer]

Agreed.  Given that LDRs with those very short recovery times are uncommon, that's what I meant by "some very rare circumstances".  Indeed, many of the more common applications of LDRs/vactrols often use units intended to be less responsive at faster rates, almost as a way of automatically adjusting sweep width as sweep rate is increased.  The prototypic case of the need for an LDR with a super-fast recovery time would be a peak limiter.  Many compressors can do fine with sluggish LDRs because they yield the slower gain/level-recovery times suited to long sustain.  tBut if the goal is to simply prevent transients from blowing speakers or sending the VU into the red, then you want something that keeps a lid on level but gets out of the way quickly.  Envelope filters want something in between what phasers and compressors want.  A little bit of slow-to-wake-up is good, and makes the filter sound more vocal.  One can always devise a way to introduce slower attack, but it can complicate the circuit, or alternatively - if done simply - reduce the sensitivity.  An LDR/vactrol with a wee bit of lag does the heavy lifting for you.

[Kipper4]

Seems like the envelope is not responding well or the vactrol is super slow to decay.
Sometimes it's a compromise as to how gain on the envelope detector. And the decay of the charge pump envelope detector.

It's a weird filter this one.
It's not often you see a positive feedback filter in stompboxs.

The level pot has a few things going on it contributes to the dc biasing on the vactrols led and also acts as a decay for the envelope detector.
Clever in terms of parts economy and knob footprint.

Is your C5 really 1uf?
Are the diodes orientated right? D1 D2?
Does the voltage change on the level pots wiper when you turn it?
Have you metered the vactrols resistances? Vtl5c42 assuming that's what you used. Pins3-4 and 4-5.
The data sheet here says (page 15)
http://denethor.wlu.ca/pc300/optoisolators/analogoptoisolatorintroduction.pdf

400M dark ,fast response. Yours is very sluggish to my mind. Or the level pot is leaking.

All references your diagram.


Looks like the dual vactrol is controlling 3 different filtering actions if I'm looking at it right.

Perhaps post some ic voltages too.
Hope this works out.

[Mark Hammer]

Okay, so looking at the various schematics I have for the EF440, it is starting to look like you have an excessive envelope signal.  Your schematic shows feedback resistor R1 as 2M2, while other posted schematics for the EF440 (including the one I built my own enjoyable unit from) show it as 470k.  After finally listening to your YTs of the issue, my first thought was "Man, there is absolutely NO rise time on that, and it sweeps ridiculously high!".  So, here's what I'll suggest before you go ripping anything apart.  Tack on a 560k-680k fixed resistor in parallel with R1, and see if it makes the difference I'm thinking it will.  That will bring the combined parallel resistance down to 445-519k.

[bushidov]

So, I don't have and 560K or 680K resistors on hand. I have a ton of 470K and 1M, but nothing in the middle. Thankfully, I have a good desoldering gun at home, so no worries there.

So, here is the video recording of the 470K resistor in place of R1.

Much improvement! It still has this ripply, kind of square-wavey distortion going on that I am not sure about. Maybe that's how it is supposed to sound? I am used to my clone of a DOD FX25 that I made and modded, which is based around an OTA LM13700 chip. It just sounds like a quacky auto-wah, but clean as a whistle.

So, Mark, either you nailed it, or you got me a ton closer. But now this leads me to another question then. Why do General Guitar Gadgets, Beavis Audio and PedalPCB have their schematics claim R1 has a value of 2.2M?

[Mark Hammer]

The Beavis Audio drawing (at least the one I have; it may have been corrected since) not only has an error in the vactrol but yes shows 2M2.  The 2003 GGG drawing I have shows 470k and so does another 2003 drawing I have, which I think might be RG's drawing.  I don't know what may have transpired between JD's 2003 and 2012 drawings.

I wish I could compare yours directly to my own, but having a look through "the collection" just now, it appears mine is either on long-term loan to someone that I've forgotten about, or else was sold to someone.  Either way, it has been at least 18 months since I've been able to use it, however I do remember it as sounding pleasingly smooth.  Could have been something in the homemade vactrol.

This person here posted a YT of his build, using the GGG schematic.  It sounds in the ballpark of yours, with the feedback-resistor change, including the envelope ripple, although he fiddles with the Range/tuning control a bit, such that the filter sweep can get higher than yours.  If you're in the experimenting mood, do consider upping the value of the time-constant "averaging" cap from 1uf.  The perennial tradeoff with simple half-wave rectifiers like the 440 uses is that using a larger averaging cap gets you a slightly longer attack time, but can result in a protracted decay time; you get "smoother" at the cost of responsiveness.

Once you get it working to your satisfaction, there is a mod for varying the resonance that David DiFrancesco of DOD published.  I scanned and posted it, but can't remember where.  I can dig up the original and give you the specs if you're game for it.

[bushidov]

Hi Mark,

Well, I did bump that cap up to 4.7uF and that helped it a lot too. I'm reading your notes on the GGG site:
http://www.generalguitargadgets.com/effects-projects/filters-envelope/dod-440-ef/ and it does mention on the schematic that they changed it from 470K to 2.2M for "Great Response on Range Control", however, I am not seeing it. I like your 470K better.

I am interested in your bass mod for it though. I am going to draw that out tomorrow, just to make sure I understood your instructions correctly.

You 'da man!

[bartimaeus]

I think something might be odd with the envelope follower circuit on your pedal. I built my own from scratch with a smallbear xvive vactrol and get a much cleaner sound with less ripple. Try using 1uf Elec for C1, 470r for R2, and 1M for R1 and see if that helps. I also used 20k resistors for both sides of the level pot, not sure if that matters but maybe it lets some extra current through R14?

You may also want to try a tl022 or other opamps

[Mark Hammer]

THis isn't another one of those autowahs that works best with a 1458, is it?