CV / LFO panner

[nocentelli]

I'm planning a CV driven stereo panner pedal using the baby8 sequencer and an LFO derived from the tremulus lune. I'm hoping to have a switch that will allow either the baby8 CV out (clocked by the LFO) or the LFO output direct to the panner CV input.

I'm fine with the control side of things, but I'm looking for suggestions for the audio part of the circuit. Most CV controlled panner schematics are from the DIY synth world and so tend to be +/-15v, whereas I would prefer to use a single sided +9v supply and a 1/2 supply vref. Additionally,  most guitar-oriented panner/trem schematics I have come across utilise a pair of LDRs driven out of phase, which I feel will overcomplicate the LFO/sequencer switching and require matching the LDRs' response etc: Ideally i would like to use the control voltage directly (e.g. OTA).

Any suggestions or pointers would be gratefully received.

[ElectricDruid]

Personally, Id' start with one of those synthy ones and then modify it for stomp box use.

For example, I like this one of Thomas Henry's:

http://electro-music.com/forum/topic-54633.html

The trick with the transistor pair providing opposing currents is very neat (I first saw it in the SSM2045 datasheet). You could keep that, but drop all the other Linear CV an Envelope stuff and just connect R11/22K to 9V.

On the second page, I'd swap the inverting op-amp input buffers for non-inverting buffers to get the impedance up to something more guitar-orientated, and then I'd swap the two OTA inputs over (so input into  -ve OTA input, not +ve like he has it). That maintains the overall non-inverting nature of the whole thing.
And then of course, he's got it set up for two-in-one-out, whereas you wanted one-in-two-out, so you'd have to use one input buffer and two of the I-to-V stages, one for each OTA.

Finally, there's the supply, but we know that LM13700 runs fine at 9V, s no problem. Everything that's marked as -15V goes to ground, everything that's marked as +15V goes to +9V, and all the ground connections go to our "virtual ground" 4.5V bias.

There, after you've done all that, Thomas Henry's circuit isn't Thomas Henry's circuit any more! It's yours, 'cos you just rebuilt the whole thing!

HTH,
Tom

[nocentelli]

Personally, Id' start with one of those synthy ones and then modify it for stomp box use.

....

Tom

Many thanks for your valued input! I've had a go at drawing out something that I hope reflects your advice, and it has prompted a couple of questions: The first is a simple LFO, with the output shape adjusted by pot and the output buffered and sent to a 4017 counting 8 step with each output to a mini pot (a la seekwah/trem) and the wipers connected to the transistor pair to produce inverted CVs. The switch takes either the 4017 pot wiper output or the buffered LFO signal direct (unsure about a depth pot here).




Now for the questions: the 13700 has a pair of (darlington?) buffers that i usually see utilised in similar circuits, taking the output from pin 5 to pin 7 (output from and the same with output pin 12 to 10 and (buffer out from 9). Would these simply be unnecessary in this context?

Secondly, I realise I have no way of "tuning" the OTA's response to the CV: Of course, I'd like to assume that i can just hook it up as drawn and the sequencer pots will set the stereo position of each step from far left at full CCW to far right at full CW, but I assume it is not this simple, and there needs to be an adjustable element i have not included.

[ElectricDruid]

I've had a go at drawing out something that I hope reflects your advice

That looks good!

The first is a simple LFO, with the output shape adjusted by pot and the output buffered and sent to a 4017 counting 8 step with each output to a mini pot (a la seekwah/trem) and the wipers connected to the transistor pair to produce inverted CVs. The switch takes either the 4017 pot wiper output or the buffered LFO signal direct (unsure about a depth pot here).

Yeah, as you say, a depth pot is going to be necessary somewhere. I'd probably leave in the mixer op-amp for the "Crossfade" input and feed that input from your LFO/Seq switch. That gives you your depth pot (R27 on Thomas' schematic) and it also gives you a "initial frequency" adjustment. That'll go between 9V and ground on your version of course, and you might find you need to adjust R32/270K for either more or less range.

Now for the questions: the 13700 has a pair of (darlington?) buffers that i usually see utilised in similar circuits, taking the output from pin 5 to pin 7 (output from and the same with output pin 12 to 10 and (buffer out from 9). Would these simply be unnecessary in this context?

Yes. The typical way you see the chip used is to use a resistor to ground on the output to convert the output current to a voltage, and then the darlington buffer to buffer that voltage. While this works (mostly) and only requires a single chip, it isn't that good, and you get much better results using a proper op-amp based I-to-V stage like Thomas has done in this schematic (that's that inverting op-amp bit with no input resistor). So no darlingtons required.

Secondly, I realise I have no way of "tuning" the OTA's response to the CV: Of course, I'd like to assume that i can just hook it up as drawn and the sequencer pots will set the stereo position of each step from far left at full CCW to far right at full CW, but I assume it is not this simple, and there needs to be an adjustable element i have not included.

I think putting that Crossfade mixer back in solves this for you. That "Initial Freq" control is the one you need.

One last thing - your query over whether the Bias needs buffering. I dunno! Try it and see I guess! Soz!

HTH,
Tom

[nocentelli]

[ElectricDruid]

That looks like an excellent start for a Rev.1. Now it's time to give it a whirl and find out where it needs tweaking! And we have a pretty good idea where those places might be. I'd watch out for:

1) Distortion through the VCAs - the 100K/220R divider needs to limit the input to the VCA to whatever the OTA can handle without unpleasant levels of distortion (note we're not using the diodes here, so we get a soft "tanh" clip). You choose how much distortion you can cope with. The equivalent levels are in the data sheet. Louder is better if you can stand the distortion, since you're trading that against S/N ratio.

2) Maximum OTA volume - that 22K on the transistor pair. Is it to much/not enough?

3) OTA range adjust - the 270K going into the mixer. Too much? not enough?

4) LFO range - Oh!
I notice you've got no input resistor for the LFO/Seq side of the mixer. You should have a 47K or so there. Again, adjust once you've tried it. If that's in place, you could drop the 1K from the LFO before the pot.

I used to always do this on breadboard, but these days I mostly use stripboard for prototypes like this. I find that more reliable for me and it means I spend more time debugging the circuit and not debugging the breadboard. YMMV.

Tom

[nocentelli]

Thanks again for the advice, I have to rearrange my breadboard + bypass board to accommodate the dual output switching and then I'll try it out.

[ElectricDruid]

One other minor thing - accepted practice for the 13700 if you don't use the buffers is to ground the darlington's base. You can leave the output open.

T.

[nocentelli]

So I spent yesterday afternoon slowly putting most of this together on the breadboard, starting with the LFO -> 4017 -> CV opamp -> pnp BJTs

I checked each block as I went along, LFO works fine (along the way I discovered a missing 220k in the schem from non-inverting  input to output), the 4017 works fine, outputting between 0-8.2v from the summed wipers from the pots on each sequenced output, depending on pot setting, and now I get to the "CV conditioning" block, my understanding starts to fall apart....

With the "initial frequency" (bias pot) set to deliver 4.5 volts, a 150k resistor holds the CV output from the sequencer at 8v no matter where each step pot is set (I have them set alternating between max/min, i.e. fully panned left/right on alternating steps). When I increased the bias (R32) resistor up to 330k, the CV voltage from the sequencer pots swung between 7.2 and 8v. My question is, is there an ideal voltage swing range going into the opamp, or I an I better off dropping the multimeter, breadboarding the audio part and setting it by ear for maximum pan?

[ElectricDruid]

Ok, perhaps we need to make sure we're clear what the range going into the LM13700 should be before we start whacking sequencer signals into it. You say you've tried different values for R32, but you don't say what input resistor value you've got coming from the sequencer/LFO side (R13 equivalent). *That's* the resistor you need to tweak to alter the amount of modulation from the sequencer. If that resistor isn't present, odd things will likely occur, since you're feeding current directly into the op-amp. It'd be happy enough, but you won't get the result you expect.
R32 is good for altering the range of the "bias pot" adjustment, but shouldn't/won't alter the range of the sequencer.

If I were you, I'd do a few experiments with the sequencer disconnected and with just the bias Pot and R32 in the circuit. Like that, you know you've got 0-9V going in, and you can tweak R32 to give you a good pan going out. The relative value of R32 to R6 gives you the gain of that mixer, but if you've got 0-8V from the sequencer instead of 0-9V, you could just scale down R32 by 8/9ths to get the value for the other input resistor.

HTH,
Tom

[nocentelli]

R13 is 47k

[ElectricDruid]

According to a LTspice simulation running with 9V power, the voltage at the base of Q1/at the output of the CV op-amp needs to vary between about -130mV and +130mV, referenced around the virtual ground (so 4.37V to 4.63V).

The CV op-amp is an inverting mixer, so the gain on the bias pot is 4K7/270K  = 0.0174, so +/-4.5V in gives +/-78mV out, probably not enough (but then we did drop the supply by a factor of three, so that pot that did have 30V across it now only has 9V). So I'd reduce the 270K to more like 180K or 150K.
Similarly, the 4K7/47K gives a gain of x0.1 for the sequencer signal, which is +/-400mV for 8V input, probably too much. Increase the 47K by at least double, 100K, 120K or 150K.
Finally, the 22K is not letting much current through to the OTAs. 180uA each according to the sim. You could reduce it to 4K7 and still only get 0.8mA per OTA (which should be enough).

HTH,
Tom