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

[orangetones]

Judging by your voltage readings you have the ICs in correctly.  Pin one is in the square hole for each IC correct?

[captntasty]

Yes, the IC's are all in correctly.  I did find a solder bridge.  Here are the new voltages compared to frequency centrals:

V=12.08v (the voltage went down a few 100ths - ?)

LM13700

1 - 1.025        0.91
2 - 0.002        .017
3 - 5.98          6.03
4 - 5.98          6.03
5 - 7.13          0.11   *
6 - 0.00          0.00
7 - 7.13          0.11   *
8 - 5.93          0.00   *
9 - 6.01          0.00   *
10 - 7.19        0.11   *
11 - 11.99      12.08
12 - 7.19        0.11   *
13 - 5.99        6.03
14 - 5.99        6.03
15 - 5.99        .025   *
16 - 1.025      0.89   *

TL062

1 - 5.99         6.04
2 - 5.99         6.04
3 - 5.97         5.75
4 - 0.00         0.00
5 - 5.78         5.98
6 - 5.96         6.04
7 - 5.95         6.04
8 - 11.99       12.08

LM324

1 - 0.001        0.25   
2 - 0.009        0.04
3 - 0.003        .005
4 - 11.99        12.08
5 - 0.197        0.12     
6 - 0.655        0.52     
7 - 0.655        0.52     
8 - 5.96          6.04
9 - 6.00          6.04
10 - 5.99        5.98
11 - 0.00        0.00
12 - 5.99        5.98
13 - 6.00        6.04
14 - 5.99        6.06

So the LM324 and TL062 seem in order.  The LM13600 is still not correct.  I'll keep looking...

[orangetones]

Have you read over the datasheets for the LM13700 versus the LM13600?

Found this:
http://dropmix.xs4all.nl/pipermail/synth-diy/2008-August/008908.html

[captntasty]

Interesting...  I think I'm going to put this one aside for the time being.  The way I have the modes switched by SPST switches independently is having a negative effect.  I will be ordering the correct SP3T switch and I'll also get some LM13700's.

[frequencycentral]

Interesting...  I think I'm going to put this one aside for the time being.  The way I have the modes switched by SPST switches independently is having a negative effect.  I will be ordering the correct SP3T switch and I'll also get some LM13700's.

Seems like you've made some progress - well done! The 324 and 062 do look good now. Maybe you have another solder bridge? It would certainly help having the right switch for less confusion. Boogdish used a 13600 in his build - and the circuit I based (stole?) my design on used a 13600 too. So, I think the 13600d should work. I'll do some checking today, after I've finished the snowman!


EDIT: If you order a 2P3T switch you can hook up LEDs to the other pole to give a visual intication of which mode you are in. Unnessecary but cool.

[Boogdish]

Capntasty, like Rick said, mine uses a LM13600D just as yours does.  I also used  SPST switches for selecting modes.

Have you tried measuring the voltage coming out of the envelope follower as you play?  I'd make sure that it's putting out a good amount of voltage first so you can isolate your problem.  If you don't have the extra hands to play the guitar or clips for your test leads to do a normal voltage check of this, you can try running the output of the buffer on the envelope follower into an LED, if the LED lights up when you play, you're putting out enough voltage.

Is your resonance knob affecting the sound at all?

This might be a bugger to debug, since Rick, you and I all used different layouts and I used different control voltages.

Also, Rick, I was being light-hearted when I said you stole the design, I hope that came across okay on the internet.  To quote The Simpsons "Everything's stolen now adays, why the fax machines just a waffle iron with a phone attached to it."

[captntasty]

I'll take another whack at it... where are the test points for measuring voltage on the envelope follower?

edit: has anybody else used orangetones layout?

[frequencycentral]

.....you can try running the output of the buffer on the envelope follower into an LED, if the LED lights up when you play, you're putting out enough voltage.

That's a really good idea, and what I did when I was breadboarding the circuit. Has a look at the circuit I based (stole?) my design on - that has a LED attached, you could connect it like that. It would confirm that the follower part of the circuit is working.

Also, Rick, I was being light-hearted when I said you stole the design, I hope that came across okay on the internet.  To quote The Simpsons "Everything's stolen now adays, why the fax machines just a waffle iron with a phone attached to it."

Yup - no problemo - I have no qualms about 'borrowing' or tipping my hat to influences. The filter is also very similar to the state variable filter on the LM13700 data sheet. I like the Simpsons quote!

I'll take another whack at it... where are the test points for measuring voltage on the envelope follower?

You should hook up a DMM pin 1/16 of the 13600, that should change with the transient of any note you play. It should settle to a constant voltage with no note played.

[captntasty]

So there is no voltage change when a signal is applied on pins 1/16...  what does this tell me?

[frequencycentral]

So there is no voltage change when a signal is applied on pins 1/16...  what does this tell me?

The issue must be with the envelope follower, either the signal not reaching it or not getting through it. Check the connections and component values around the three opamps that make up the follower.

[Mark Hammer]

A naive question: If the LM13600 is governed by the current applied to pins 1/16, then are you supposed to see envelope-associated voltage changes?

[frequencycentral]

A naive question: If the LM13600 is governed by the current applied to pins 1/16, then are you supposed to see envelope-associated voltage changes?

Equally naive answer: Uh I had assumed so but obviously not now you mention, as the 10K resistor changes the voltage output of the follower into a current.........so maybe it would be better to measure the voltage the other side of that resistor. I just checked now, and that obviously does change from almost nothing to heavily +ve when a note is played. Thanks Mark.

[puretube]

A naive question: If the LM13600 is governed by the current applied to pins 1/16, then are you supposed to see envelope-associated voltage changes?

As much as the voltage across 2 seriesdiodes(-drops) changes with current-changes... (= ~ a few millivolts around the ~1.4V mark...)

[Cliff Schecht]

A naive question: If the LM13600 is governed by the current applied to pins 1/16, then are you supposed to see envelope-associated voltage changes?

Yes. I haven't looked at the schematic of this pedal to see if it's single or dual rail power but essentially, you can monitor the voltage at the Iabc pin to determine where your gain is. When working with these OTA's, you can easily figure out the current feeding the pin by looking at the voltage on the pin and dividing by the resistance. Full off (or in this case, the lowest filter cutoff) is heard when the voltage seen at the Iabc pin is equal to the voltage that is being applied to it, and full on is when you are feeding the device the max amount of current you designed your OTA circuit for (and the accompanying voltage that is "converted" through the current limiting resistor). It's the same as designing with a differential amplifier, you just have to limit the amount of current that your connected emitters see and from there, it's easy to determine things like gain, dynamic range, etc..

[chicago_mike]

Will a CA3280 work in this?? I have a lot fo those, but no lm13600/13700's.

[frequencycentral]

Will a CA3280 work in this?? I have a lot fo those, but no lm13600/13700's.

Hmm - I'm not sure. I have no experience with working with that device. A quick look at the data sheet suggests that it should be functionally similar to 13600/13700 - but the pinout is very different. I also had trouble identifying where the buffers are. My guess is that it should work as a replacement with maybe some tweaking needed. The CA3280 has been discussed over at electro-music.com, and is featured in a few designs over there. Pesonally, I would breadboard it first. It is apparently superior to the 13600/13700.

[orangetones]

I finally breadboarded this up.  Have some issues though.  Maybe someone here can point me in the right direction.  I get a very loud popping when the filter engages.  Loud pop/click as the sound comes in, then it sounds fine.   But this is of course not fine as everytime you pick a note, you get this loud popping surge!  Anything I should be looking for that might contribute to this?

I am using a 13700N, LM324 and NE5532(for the Vref and Output Buffer).  All stock values for everything else except 50k for the 47k pot and 500k for the 470k pot.  I am using some log pots, but that shouldn't really change anything sound wise, just control wise.

Thoughts on this?

[Cliff Schecht]

I finally breadboarded this up.  Have some issues though.  Maybe someone here can point me in the right direction.  I get a very loud popping when the filter engages.  Loud pop/click as the sound comes in, then it sounds fine.   But this is of course not fine as everytime you pick a note, you get this loud popping surge!  Anything I should be looking for that might contribute to this?

I am using a 13700N, LM324 and NE5532(for the Vref and Output Buffer).  All stock values for everything else except 50k for the 47k pot and 500k for the 470k pot.  I am using some log pots, but that shouldn't really change anything sound wise, just control wise.

Thoughts on this?

Your problem is that the signal is rising too fast. To your LM13700 has a really fast slew rate (50 V/us) and a quick dynamic like a guitar pluck (where the attack is perceivably instant) will look like an impulse function (the delta dirac function for you differential equations fans). This function is centered around 0 (0 V DC in our case), is infinitely thin (no frequency content) but is infinitely tall (LOUD!). In the case of your OTA circuit, this impulse function is a quick changing DC signal that is at the amplitude set by your envelope follower circuit. What do you get when you run a DC impulse into an audio circuit (or a speaker)...? A loud "THUMP!". The same effect happens when you touch a battery to a speaker really fast, you hear a real life example of the impulse function. To alleviate this, you need some sort of slew limiting circuit. This can be done a few different ways but to be honest, I've never designed or used one.

Deriving the definition of slew rate is easy and may lend towards some circuit ideas, however. Since slew rate is I/C, first derive the units of current and capacitance. Current is defined as the rate of charge per second (Q/t) and capacitance is the cap charge divided by the voltage across the cap (Q/V). Dividing these gives us our units, volts per second (usually given in V/us). So what you need is a circuit that limits the amount of current a capacitor sees.. I would think that a voltage to current converter feeding an integrator will you do what you need.. I'll see if I can come up with a circuit tonight (I'm obviously bored ).

[orangetones]

Hey, Thanks for the reply Cliff. 

When you said "to my lm13700" are you referring to the 13700N? or the lm13700 in general?

Why would this work for Rick, but not me?  Presumably he used the same chip.

[Cliff Schecht]

Every LM13700. I assume that the averaging capacitor in your envelope detector is too small. Making it larger (the cap going to ground after the diode) and will make it so that the envelope signal can't change as fast. I need a schematic to go any further, but I know your problem because it's a more general problem with audio stuff.