"Funky MF (Multimode Filter)" - LM13700 Multimode Envelope Filter

[frequencycentral]

Frequency: It always helps to understand the circuits you design with, but it's obvious that you aren't all that lost. Understanding op-amp design is really not that hard once you understand a few limitations, and even those LM13700 OTA's aren't hard once you understand how a transconductance amplifier works. TBH, everything in your schematic is pretty dang standard and so explaining the circuit isn't as hard as you think. If you want, I can break it down for you in more detail than you'll care to know .

That would be awesome if you could actually Cliff - I can 'do stuff' with electronics but it's quite scary to me that I know how much I don't know - if that makes sense. I design with pre-existing 'blocks' and tweak on the breadboard. There is very little theory underpinning what I do. Scary huh? I probably understand tubes more than I do opamps.

Rick, I am wondering which of the opamps in the schematic you used the tl062 and the lm324 for?  Is it possible that having the TL062 performing the function of a specific part of the circuit has something to do with the problem?

In my pedal the output buffer and the unity gain buffer are the TL062.

BUT

When I breadboarded it I used two LM324's. I only used the TL062 in the final build.

Thinking it through though, it was only in the final build that I replaced the resistor between the unity gain buffer and pins 1/16 of the 13700 with a 10K. It had previously been 22K I think. So maybe a TL062 gives out less current than a LM324? And therefore needs a smaller current limiting resistor? The 22K that I replaced was holding the filter too 'open'. So maybe you should try a few different values in that position.

Having played about with this pedal for a couple of months, I have to say I just love the sound of it, some great stuff can be coaxed out of it. It is THE FUNK. I knew I would like it before I built it - I've been putting guitars through synths for years, and have a very similar Oberheim SEM multimode filter in one of my modular synths. But 'coaxed' is the operative word. It's a very 'touchy' device, and needs to be set up just right to get the good sounds. That means not only how all the knobs are set, but also which guitar you use, which pickup (bridge or neck), how the guitar's volume and tone are set, and which effects come before it. I would not dare to tweak in a sound on the fly in a live-on-stage situation. The only way I would consider using it live would be to have it on its own loop with a dedicated distortion before it, and set both up beforehand and not touch either apart from to stomp it in and out on the a/b switcher. Maybe it's the nature of envelope filters - a guitar really being an 'imperfect source' to derive gate and control voltages from.

I had a brainwave re the envelope follower this morning. I'm thinking of maybe dropping a couple of clipping diodes in just before the range control. I have this idea that the follower works better with a clipped waveform, so adding the clipping diodes would do this, particularly if the gain stage before the range control was increased from x100 to x1000 by replacing the 10K to vref with a 1K. It would be kind of like having a Distortion+ before the follower. I don't know if it will work out but I'll try it. I've try it when I get some time - I have two tube builds to finish this week.

[orangetones]

Well folks, I have tried a larger pot for the attack pot.  No luck.  Most of the travel does nothing.  The 50k is right.  Tried increasing that 10k resistor to pin 1/16 of the lm13700.  No luck there either. 

I love the sound of this filter once engaged.  It has a really good feel.  I just wish we could solve this problem!  How can this possibly be!  Rick, there must be some magic in yours!

Cliff or anyone else out there, would you breadboard this and see if you can get it to work properly for guitar?

Something I thought I would add is that once you get past that initial volume jump/pop, the filter if fine, but then if you don;t play, and the envelope closes so so speak, you have to pluck to get back in.  Mind you, if you pluck the strings and open the envelope, you can hear a slight hissing in the background and you can actually hear the envelope closing.  when it finally closes, you get a lower thump sound.  Not as loud as the pop when you engage.

Damn!  This thing is frustrating me to no end, but it sounds soo good once engaged!

Rick, you used all lm324's to breadboard.  Hmmm... so there isn't anything there then I guess...

Anyone want to breadboard and team debug?  I unfortunately don't know enough about envelope followers to debug this myself.

Would it help to post sound clips?

[Cliff Schecht]

Well folks, I have tried a larger pot for the attack pot.  No luck.  Most of the travel does nothing.  The 50k is right.  Tried increasing that 10k resistor to pin 1/16 of the lm13700.  No luck there either. 

I love the sound of this filter once engaged.  It has a really good feel.  I just wish we could solve this problem!  How can this possibly be!  Rick, there must be some magic in yours!

Cliff or anyone else out there, would you breadboard this and see if you can get it to work properly for guitar?

Something I thought I would add is that once you get past that initial volume jump/pop, the filter if fine, but then if you don;t play, and the envelope closes so so speak, you have to pluck to get back in.  Mind you, if you pluck the strings and open the envelope, you can hear a slight hissing in the background and you can actually hear the envelope closing.  when it finally closes, you get a lower thump sound.  Not as loud as the pop when you engage.

Damn!  This thing is frustrating me to no end, but it sounds soo good once engaged!

Rick, you used all lm324's to breadboard.  Hmmm... so there isn't anything there then I guess...

Anyone want to breadboard and team debug?  I unfortunately don't know enough about envelope followers to debug this myself.

Would it help to post sound clips?

Are you using active pickups or sending a LOT of signal to the filter pedal (maybe with a preamp?) then the envelope follower would be overly sensitive and rise too fast (causing that impulse I talked about earlier and hence, a pop). Op amp choice isn't very critical in this circuit to be honest, the only reason to use a TL0XX series FET is for the ultra high impedance that the LM13700 filter sections feed. The really high impedance presents no loading of the OTA based filters and is a common approach.

Another problem could be that 2.2 uF cap. Is it really 2.2 uF? Maybe take a picture of your build as it is on the breadboard. Also, sound clips never hurt either .

[orangetones]

I am playing a stock American Standard Strat.

I have a 2.2uF yes.  I tried a larger cap.  4.7 I think or 10uF.  Have to try that again.  I am pretty sure that is not it.

I will record some soun clips.

An a pic of the breadboard...  sure.  I am not a very neat breadboarder.  But I assure you I have built it as per the schemo.

Sound clips in a bit.  Maybe in the next couple of hours.

[frequencycentral]

What about double checking that all the vref points and ground points are actually correct? It's easy to connect a vref point to ground by mistake. For example, if you were to connect the 100k in the 220K/100k voltage divider (at pins 1/16 of the 13700) to vref instead of ground it would screw things up.

[orangetones]

I will check that, but I am pretty sure of myself here.  I am going to upload some photos/sound clips in a sec.

[slacker]

I`d implement the voltage divider at the input-side of the buffer... (but keep the 10k where it is!)

and remember, that any voltage applied to the 10k must exceed 2 diode-drops, before any current flows into pins 1/16...   

I think this is the problem causing the thump, I simmed the envelope follower and the 200k/100k voltage divider are pulling the voltage down to 3-400mv. This means the filter will be completely closed and will suddenly jump open once the voltage output of the buffer reaches a certain point. Moving the divider like Puretube suggests solves this.

[frequencycentral]

I`d implement the voltage divider at the input-side of the buffer... (but keep the 10k where it is!)

and remember, that any voltage applied to the 10k must exceed 2 diode-drops, before any current flows into pins 1/16...   

I think this is the problem causing the thump, I simmed the envelope follower and the 200k/100k voltage divider are pulling the voltage down to 3-400mv. This means the filter will be completely closed and will suddenly jump open once the voltage output of the buffer reaches a certain point. Moving the divider like Puretube suggests solves this.

That's interesting Ian (and Ton) - thanks for the input. Though I'm a little perplexed as to why mine doesn't thunk!

[slacker]

Not sure, it could be down to the opamp used. Your envelope follower might be putting out a higher voltage at rest so the divider won't pull the voltage down as far.
You could measure the voltage at the output of the buffer and the control pins of the LM13700 and see how they compare with orangetones'.

[frequencycentral]

Not sure, it could be down to the opamp used. Your envelope follower might be putting out a higher voltage at rest so the divider won't pull the voltage down as far.
You could measure the voltage at the output of the buffer and the control pins of the LM13700 and see how they compare with orangetones'.

My pins 1 and 16: 1.025v
Orangetones      : 0.522v

My buffer output: 0.655v
Orangetones     : 0.055v

[slacker]

Looks like that's the problem then, your control pin voltages are Ok, but oranagetones' are too low.

[orangetones]

So moving the voltage divider to the input of which buffer are we taking about?  the unity gain one?  So I basically put the voltage divider on the + input of that buffer where the 2.2uF is connetced?

Pardon my ignorance here guys, I am still learning all of this.  I really do appreciate the help.

[slacker]

Yeah that's it.

[orangetones]

Moved the voltage divider.  Now the envelope is not being held shut, however, having the voltage divider there seems to affect the controls (specifically the attack and release).  I had to change out the release pot for a 50k, and even then much of it is unusable, as the filter stays open and won't close for most of it's travel.  Have to try a lower pot.

Slacker, is there any way other than moving that divider to the input to get the envelope not to be held shut?  The functionality of the envelope follower is not as good as before.

I do have to play around with this a bit more, but I can say that the loud popping is gone.

Thoughts on the issue of moving this voltage divider and the other issues that it creates?

[Cliff Schecht]

I`d implement the voltage divider at the input-side of the buffer... (but keep the 10k where it is!)

and remember, that any voltage applied to the 10k must exceed 2 diode-drops, before any current flows into pins 1/16...   

I think this is the problem causing the thump, I simmed the envelope follower and the 200k/100k voltage divider are pulling the voltage down to 3-400mv. This means the filter will be completely closed and will suddenly jump open once the voltage output of the buffer reaches a certain point. Moving the divider like Puretube suggests solves this.

You could move the bias resistors, but I don't see why the 10k resistor isn't just eliminated in the circuit, it really doesn't provide any useful service here (well, aside from current limiting!). Instead, remove the 10k resistor and place in a 100k resistor in the same spot, remove the other 100k resistor (the one going to ground) and leave the 220k resistor where it is. Essentially, you instead have a 100k resistor in series from the op amp output to the Iabc pins, and the 220k goes to VCC from the Iabc pincs (this acts as a voltage divider AND pulls the Iabc pin up to the correct voltage). In the case of what I describe in this post,  it's at about 1.8 V which keeps the filter on and hopefully pop free. The dynamic range of the filter (for a given envelope peak value, that is) can be increased by decreasing the 100k and 220k resistor values (forces more current to charge that cap to Vref) but if you're happy with the range, then stick with these values. If you guys want a schematic, let me know.

Frequency, you should give your components designators (R1, C13, Q3, IC3, etc..) so I can actually describe the changes I'm implying!

[frequencycentral]

Frequency, you should give your components designators (R1, C13, Q3, IC3, etc..) so I can actually describe the changes I'm implying!

Yes I should - lesson learned!

Orangetones: looks to me that tweaking the values of the voltage divider (in its original position) is the answer. Try changing the 100K for 250K pot, and tweak the pot until you get good sounds but no thunk - your voltages should be closer to mine then anyway. Measure the resistance across the pot and replace it with a fixed resistor.

Or maybe even replace the 220K/100k with a 1M pot, lug 1 @ +ve, lug 2 @ pins 1/16, lug 3 at ground, and tweak for the correct ratio soundwise and voltage wise. Measure and replace with fixed resistors.

Yours is so close to working correctly. The key must be in getting that voltage divider ratio corrent so your voltages are closer to mine at pins 1 and 16. I've gone over and over the schematic for errors, and taken my unit apart to check values. I am stymied by the variation between yours and mine, but ultimately the solution will be there for future reference. I have no doubt you will conquer it today.

[orangetones]

Well folks, I tried the 100k straight off the buffer like Cliff suggested, with the 220k going to vcc.  No dice.  Held it to much open.

Put in a pot to use as a voltage divider in place of the 100k/220k voltage divider.  Can set the voltage just right so that I get Rick's values.  But I still get a light thud as the filter closes.  As well, when I set pins 1 and 16 for 1.025 volts, I only get 0.385v on the output of the unity gain buffer.  So I guess I really didn;t match Ricks voltage there.

The other pots when the filter is held just open enough to avoid the popping (by using the voltage divider pot) jst don't seem to function very well.  The envelop filter is some how not working as nicely as when it was popping to open.

Strange...  I am starting to think that this thing may be more trouble than it's worth.

I would still be interested to have someone else breadboard this and see if they can get it to work stock with their components.  Cliff? Slacker? anyone?

[liquids]

It may take me a while, but I'm dying for a synthy filter, and this looks feasible...count me in.

Only, tell me...I'm a big fan of reverse sweep.  For after I get it functioning as is (if!), tell me - what would be needed for a reversed sweep if at all possible--which I find more synthy than the typical sweep.   

[frequencycentral]

It may take me a while, but I'm dying for a synthy filter, and this looks feasible...count me in.

Only, tell me...I'm a big fan of reverse sweep.  For after I get it functioning as is (if!), tell me - what would be needed for a reversed sweep if at all possible--which I find more synthy than the typical sweep.   

You just need an inverted envelope to do that, a single extra opamp would do the trick, plus a toggle switch to select normal/inverted.

[orangetones]

Liquids, have you thought of trying the MXR Envelope filter from Tonepad?  There are mods for it to do different ranges and reverse sweep.  It is a pretty versatile filter.