replacing a multi-gang pot

[johncronan]

Okay, so here we have a simple slew rate limiter:



Actually, this is from the YuSynth "Dual Slew," simplified for the sake of discussion.

I'd like to build 4 of these and control the rate limit on all 4 from a single pot. I doubt I'll ever find a quad-gang 1M log-taper pot! And it's floating. So I think I need something active to simulate that resistance.

My best idea at the moment is to try building the example circuit for a voltage controlled resistor that you can find in OTA datasheets, such as the LM13700. Is there a better option here? Assuming I really do need a VCR (or CCR), any tips on building one?

Thanks a bunch.

John Cronan

[elmegil]

Yah, a 4 gang pot is going to be impossible to find.

Mike, E.A.S. over on Muffwiggler, ended up getting custom 4 gang pots for his Mind Reader (aka MS-20 external signal processor) PCB, but they're 100K anti-log. You might still try to get hold of him though, I'm sure he has some idea of how hard it was (took a good long while....)

[comrade_zero]

What kind of precision do you need for this circuit? (i.e. how closely matched will you need these resistances?) The simplest (though not very precise) solution might be driving 4 LED/LDR pairs (poor man's vactrol) from a single pot (which means you can also include an external CV in for external voltage control.) You can tighten up the precision by using components from the same batches, and limit your resistance to 1m by placing a 1m resistor in parallel with your LDR, but ultimately this is not a precise solution.
Just a thought...
Hope this helps,
c_z

[bubzy]

+1 to cz's idea

[johncronan]

Hrmm. This is for a MIDI-CV interface that, although it will be open source, I'm interested in selling as a kit. I don't actually have too much experience with vactrols, but I've heard the tempco is pretty extreme, and as you say, they're not terribly precise. I'm open to being convinced, but there are also issues with sourcing... I should have mentioned the context.

[johncronan]

Aha, I just discovered this:

http://www.all-electric.com/schematic/slewlim.htm

[isak]

So funny your looking for this, this was my idea to do few week back
I was looking for 3 gang pot to control 3 glides simulteneouly.
Didn't found the proper value I need, If you find a 4 gang one please let me know.
Btw, your circuit is going to be log, as I understand it's better to use Lin operation for this matter.

Cheers

[stewpye]

You could use a quad anaold switch and PWM the control inputs...

[johncronan]

[Boogdish]

If this is part of a Midi to CV converter, wouldn't the cheapest/easiest solution be to implement the glide before digital to analog conversion?

[kkissinger]

[johncronan]

[comrade_zero]

/slaps self in the forehead/
If a VCA can be used for this application (don't know, haven't tried it) check out the coolaudio clone of the ssm2164 quad VCA. (Mammoth has them for $2.80 a pop, which is about the same price as 2xlm13700.) Driving all four from a single source should be no sweat. It is a current in/out device, with a wide supply range, so it should probably work for your application.

@kkinssinger-- I've done it like that before, arranging the components into a 12-pin IC socket, but depending on the LDR you use the light from a single LED can yield a very limited range. I got tired of rolling my own, so I got a couple of bags of optocouplers from Electronic Goldmine. (Untested lots of 20 will cost you two bucks.)

Anyway, food for thought. Good luck, I hope this helps.
c_z

[johncronan]

Food for thought, indeed. Thanks, guys.

I've heard of implementing a VCA using a voltage-controlled resistor, but not the reverse. I could be way off here, but wouldn't I still need something to sense the difference in voltage between the input and the cap? It seems like I'd be better off with an OTA chip that implements the VCR directly.

Did you see the link I posted yesterday? I'm about to simulate that circuit, or something close to it. It appears the OTA even does away with the need for a separate input buffer.

In any case, it is very useful to know that there is an SSM2164 replacement available. In the past I have bought them off ebay for $6 or 7 a pop.

[prgdeltablues]

Could you use a switched capacitor to replace the pot? (With four driven off the same VC clock, and caps matched as best as?) I've no idea if that would work in practice for your application - the only time I've come across one in electronic music is in the ultrasound filter in Peter Blasser's Rollz5

Peter

[johncronan]

That seems like a totally valid approach, and I'm interested in trying it out.

But as I continue to research this I've been going in a different direction. If you put a current into the input of an op-amp integrator, that changes the cutoff frequency. I found this document, which confirms that, yes, you can use a VCA to implement, not a voltage-controlled resistor, but variable slew rates directly (same thing as a VCF, really):

http://mutable-instruments.net/static/documentation/vcfs_with_vcas.pdf

The MOTM-820 uses this approach, with a neat type of exponential converter that even lets you control rise and fall separately. And the slew response can be linearized with feedback.

Anyway, I'll post a schematic when I get further along.

[johncronan]

Okay, here's my attempt at a circuit that's almost as simple as the original, but is voltage controlled, so that you can control multiple slew rate limiters with one pot:



Thanks for the suggestion on using a VCA -- that was a really good idea. I eventually found this design note from THAT:

http://www.thatcorp.com/datashts/dn130.pdf

But I wanted to use the SSM2164, so I just had to change to a differential input to get the feedback negative. This circuit can do slew rates (per volt) of about half a millisecond to tens of seconds. While staying within the VCA's compliance limit.

I think. I've only done simulations, so I still need to test it for real. But first I'm going to try to add a shape control.

Oh, by the way, this says TL072, but U1 should probably be a low-offset part. Maybe LT1013. Also, the input might need to be buffered to prevent offseting the input bias currents, since that would be another source of offset error in the output voltage.

Btw, is there any way to change the topic title?

John