13700 VCF (schmitzbitz MS-20) voltage question

[nocentelli]

Has anyone tried other VCFs with guitar?

https://schmitzbits.de/ms20.html

Sounds great, works with LM13700 just fine.

I fancy whipping one of these up from the parts I already have. The first issue i forsee is the +15/-15 volts. I'd like to use a bipolar supply bigger than +/-9v and I know +16/-8ish volts can be squeezed from the ICL7660S: Could I use the klon power supply arrangement and use a voltage divider of two equal resistors to give a +4ish volt vref in place of ground? Would this circuit still more or less work? Would buffering such a vref be wise?

[PRR]

Run it on +/-9V. It will work. It may not voltage-control scale properly against a +/-15V synthesizer, but you obviously don't have one of those handy.

[Josh?]

Run it on +/-9V. It will work. It may not voltage-control scale properly against a +/-15V synthesizer, but you obviously don't have one of those handy.

Outside voltage-control issues aside, is it possible to run this as is on +/-4.5v from a single 9v supply? (Sorry to hijack the thread a bit btw, I just happen to be planning on messing with this circuit this weekend )

[anotherjim]

9v only with a divider should work. Schemes I remember seeing use an active low impedance 4.5v reference source from an opamp or a rail-splitter chip, probably mindful of those low-value 220R's into the OTA inputs.

If you only want to hear what the filter core sounds like, I don't think you need the CV input current source circuit up top. Only drive the 10k feeds to the OTA control pins from a pot across the 9v supply just to hear the filter sweep.

Add a high impedance clean buffer feeding the input to test it with a real guitar!

[ElectricDruid]

Outside voltage-control issues aside, is it possible to run this as is on +/-4.5v from a single 9v supply? (Sorry to hijack the thread a bit btw, I just happen to be planning on messing with this circuit this weekend )

Yes, it's possible. LM13700 will run happily down to a handful of volts, iirc. It certainly runs fine on a 9V supply with 4.5Vref.

A low impedance Vref source isn't a bad idea, but it's not the only way. R.A.Penfold used a simple divider (4K7s and 100uF) for his Autowah circuit, and pushed the 220Rs (bottom of the input divider) value up to 1K.

<edit>Jim's comment about the current source not being required for basic testing is a good point. There are various other ways to do that. Just remember about the Iabc input being two diode drops above ground, as PRR pointed out recently in the other thread about the 13700.

[nocentelli]

Thanks for the advice so far, and I have another question...

The buffers that follow each OTA section: A number of 13700 VCF schematics I have seen use the Darlington buffers within the 13700 itself, but some schematics actively eschew the internal buffers and use separate opamp buffers, leaving the Darlingtons unused. Can anyone suggest a good reason not to use them and save a dual opamp?

[anotherjim]

I have an old Maplin publication  - "Audio IC Circuits" which touts a versatile pcb for the LM13600/3700 for making various OTA based circuits and it specs +4v to 30v single supply or +/-2V to +/-15v dual supply. IIRC, most of those circuits shown use the Darlingtons without restriction and provides pads for a 741 based supply splitter.
I suppose that, given you have enough voltage for the parts forward voltage drops and some to spare to actually vary the current flow through those with, you can make it work.

The Darlingtons need two diode voltage drops for the base-emitter junctions, so the reference voltage needs to be higher than the negative supply by at least 1.2V (maybe safer to say 1.5v?), so that +/- 2v I quoted above does make sense.
I suspect many of the schemes with op-amp buffers were originally done with CA3080's and the op-amp buffer was kept to keep it as much like the original as possible. We can get CA3080 quite easily now, but only a few years ago it was deemed obsolete and NOS or pulled examples were best kept for repairs and boutique clones.
If I wanted to use the filter outputs directly, an op-amp might be a better line driver in some respects.

[ElectricDruid]

Can anyone suggest a good reason not to use them and save a dual opamp?

If you're building a 4-pole synth filter, 4 sets of double diode drops adds up fast. So it's better to go with op-amp buffers and avoid the need to AC-couple between stages.

Another reason is that the distortion performance is improved using something other than a simple resistor to ground as an I-to-V convertor. The best way is to feed the output current to an op-amp I-to-V stage. You see this quite a lot with VCA designs - again, better performance without the buffers.

Honestly, I never saw the point of them. They should have put a couple of halfway decent op-amp circuits on the chip instead. Alfa Rpar have recently produced a quad-OTA clone chip based on the 13700, and guess what? - No buffers. They left them out.

The buffers tend to get used in stompbox circuits where minimal parts is more of a priority. In synth stuff where audio quality is more of an issue (add 6/8 voices together and see how much noise you get!) they tend to get used a lot less.