Geofex's Current Controlled Amplifier Application 4 - CD40106 LFO

[john_h]

I recently stumbled on R.G. Keen's gem of an article at Geofex
http://www.geofex.com/Article_Folders/VCA%20Applications.pdf

It really has me thinking about some neat applications for the Stereo Trem-Pan circuit.

I am digging into the LM13700 datasheet to try to get a grasp on this IC as I have absolutely no previous experience with it, but I would love it if you folks could help me a bit. I thought it could be fun to implement another one of R.G's ideas, being the pseudo-random LFO into this circuit. I am waiting for some missing components to arrive before I can breadboard everything, but Im curious as to what you all think about the challenges I might face with this project or some things to keep in mind before I dig in as an admitted novice with OTAs. 

*editing this post to ask for specific clarification on some things as opposed to just asking broadly "is this going to work"*

1) I am wondering weather to use the non-inverting input or the inverting input on the second (-1) op-amp feeding the lctl on the first CCA. Would the unused input simply be sent to 0v? Also regarding biasing, my plan is to use a simple old TL072, any suggestions for the values of those 2 unmarked resistors?

2) In the pseudo-random LFO article, it is mentioned that "while the waveform does occasionally get as large as the full supply voltage and as small as 0V, that it spends most of its time near the middle of the power supply." This would mean I would have to adjust the dc level down to reach the required 0-9v swing, correct?

3) In the CCA schematic, there is an un-identified capacitor before the output. Is there a reason for this? Will a suitable cap be be chosen depending on the application?

4) Im not sure exactly why I am having trouble with this, but for the 4.5v sources, would I need to create a simple voltage divider to supply those points with the needed 4.5v?

Im sure Ill think of other things as I delve further into this, but in the meantime, I really appreciate any and all help that can be afforded.


{EDIT- cleaned URL link --PRR}

[ElectricDruid]

(I don't quite get your meaning for your first two questions, so this is my best guess and please correct me if I've misinterpreted you.)

1) I am wondering weather to use the non-inverting input or the inverting input on the second (-1) op-amp feeding the lctl on the first CCA.

Do you mean which input of the OTA? It won't matter from a "changing the volume" point of view, so the choice would depend on whether the rest of the circuit inverts the signal or not. If it does, you might want the OTA to invert it again so it's in phase overall.

Would the unused input simply be sent to 0v?

Yes. If you look at the CCA schematic that's the first thing in the document, the unused input is tied to ground in the same way as the used (+ve) input. So it's already done.

Also regarding biasing, my plan is to use a simple old TL072, any suggestions for the values of those 2 unmarked resistors?

This time I'm guessing you're talking about the two unmarked resistors in the invertor "-1" block. It's a simple unity-gain inverting amp, so the only thing that's crucial is that they're the same. Any value between 10K and 100K would be fine. Outside that range people might start to wonder why you'd gone so low or so high (and there *are* situations where you might want to).

2) In the pseudo-random LFO article, it is mentioned that "while the waveform does occasionally get as large as the full supply voltage and as small as 0V, that it spends most of its time near the middle of the power supply." This would mean I would have to adjust the dc level down to reach the required 0-9v swing, correct?

No, not if you run the pseudorandom LFO off 9V too (CMOS runs at 9V perfectly). Then it's output is guaranteed to be in the right range. It's just a warning that although it *might* go as high as 9V and as low as 0V, mostly it won't, because it'll randomly wander off somewhere else before it gets there!

3) In the CCA schematic, there is an un-identified capacitor before the output. Is there a reason for this? Will a suitable cap be be chosen depending on the application?

It's the typical "DC blocking" cap you see on the output of most effects. It forms a highpass filter with that 2M2 resistor in parallel with the input impedance of whatever you feed this thing into, so it needs to be big enough to allow bass through. If we assumed a worse-case value of 100K for the parallel combination, 100nF would get the cutoff down to 15Hz.

4) Im not sure exactly why I am having trouble with this, but for the 4.5v sources, would I need to create a simple voltage divider to supply those points with the needed 4.5v?

Yes, but *only once*. You don't need individual dividers for each time you need 4.5V. You make a single divider to create the 4.5V "Vref" supply (with a big electrolytic cap down to ground to make sure it looks like ground for AC signals) and then you connect everything that says 4.5V/Vref to it.

I hope this helps you get a bit further. The 13700 is great fun once you get into it!

[john_h]

This was very helpful, thank you for the response to my confusing questions. Please accept my apologies for that!

I breadboarded just the single, mono CCA tremolo and it actually works (or at least somewhat!)

Right now, a square wave LFO is all but useless...just loud clacking with no discernible change in amplitude, but smoothing out that waveform produces some interesting (and dare I say pleasing) effects. Im still undecided as to weather the pseudorandom lfo is of a musical value in this application though...Ill try the stereo trem-pan with this to see if it sounds any less like a bad guitar cable cutting in and out at random...otherwise feeding 2 different effects into each input could make for some fun...

There is a pretty noticeable volume drop (far from unity gain) and quite a bit of noise. Of course, the noise could be chalked up to the breadboard itself, so I am not too concerned about it for the time being. I did change the values of a couple of resistors, just because these were all I had on hand (240R -> 150R, 7.5K -> 8.2K, 4.3K -> 4.7K) and I also had only an LM13600 available to me at that moment. It seems that the 13600 is the same as the 13700 with the exception of the buffer section. Could this make a huge impact on the signal loss/noise?

Also, with the voltage divider, I figured 2x100K resistors in series from +V to ground, tapped in the middle for 4.5Vref would work, but I seem to only be getting about 3.1 volts...Im really not sure what Im missing here! Ill double check that...

Anyways, this little project has been fascinating...looking forward to getting more of a grasp on OTAs.

[ElectricDruid]

The volume drop won't be helped by swapping the 240R for 150R. There's a voltage divider on the input to reduce the input voltage to something the OTA can cope with. The divider consists of the 10K and that 240R (plus a portion of the 1K trim). So what you get (ignoring the trim for a moment) is 240 / (240+10K) which is about 0.023 of your signal! 150R reduce that to 150 / (150+10K) which is 0.015, so a chunk less. What the right value is depends on the signal level you're feeding in and the distortion level you consider acceptable. That's the input end dealt with. The other resistor that affects the volume is the 22K to Vref on the output (before the buffer). That converts the output current back into a voltage. Since V=IR, for a given current, a bigger resistor implies a bigger voltage. So you could increase that a bit too.

Finally, the "Vref" 4.5V voltage. 100K resistors might be ok, but you should definitely have a big electrolytic cap in parallel with the lower one of the pair (so from the 4.5V down to ground). Try 47uF or so. If that still doesn't work, reduce the 100K resistors. See if 47K gets it closer. If not, go to 22K or 10K. The "wrongness" of your 4.5V depends on how much current is being drawn from that midpoint, and less voltage will be dropped across smaller resistor values.

HTH!

PS: I wrote up a walk-through of doing a VCA design with the 13700. It's intended for a different situation, but the basic scheme is the same, and you'll see where some of the values come from.
https://electricdruid.net/design-a-eurorack-vintage-vca-with-the-lm13700/

[john_h]

Once again, your response has been immeasurably helpful.

I had actually stumbled on and read your walk through of the 13700 VCA design in my attempts to understand some of the concepts surrounding these OTAs...really great stuff

Anyways, Ive really fallen in love with it all, so Ill continue mucking around with your suggestions in mind.

Thanks!