Stereo hard panning repeat percussion

[nocentelli]

I made a vox RP-like pedal which i really like using moosapotamus's skippy schematic for the 2N6027 PUT implementation and the hummingbird jfet audio stage. Whilst playing earlier this evening through a pair of amps, i was gripped by a powerful desire to make a circuit that could do the percussion-style "trem" thing but alternating rhythmically between stereo outs.

i have only a very vague idea of how this could be achieved (with an analog circuit - i imagine it would be relatively straightforward in dsp). I have done a stereo panning tremelo before, where the signal is rhythmically crossfaded between the outputs using a single LFO controlling the output of a pair of circuits, and the LFO output being inverted for the right channel. However, for a panning repeat percussion, i feel that the left output would need to be muted when the right passed that reverse sawtooth plosive thump, then immediately reverse, so right is silent whilst left emits the thump. Obviously this would need to be tightly sync'ed so the initial, sudden attack of each signal burst was not missed when the output signal swapped sides.

I have not seen any RP-soundalike circuits using different components that i could modify, and i get the impression the characteristic vrp sound is entirely due to the mechanism of the raw capacitor discharge shape controlling the output. SO my only idea to provide a panning RP would be to have the guitar signal split into two channels, and have the UJT/PUT based LFO signal trigger some kind of flip-flop arrangement, so each cycle of the LFO would swap which audio path was triggered AND apply that cap discharge to the appropriate channel...

...and that is as far as i have got. I have little idea of how the "inactive" channel could be muted, or whether the RP lfo could even be tapped for a sync signal to trigger the flip-flop panning/muting implementation. Any advice, suggestions or party-pooping will be gratefully received.

[Kipper4]

Mind sauce. I'll have a look at a RP scheme later and see what ideas I come up with.

[anotherjim]

The LFO can be fed to a frequency divider/ flip-flop. That used to alternate the outputs every cycle. You could divide by more than 2 with a counter chip or make one of the basic step sequencers with a CD4017 for an actual, programmable, rhythm.

[nocentelli]

.... Or make one of the basic step sequencers with a CD4017 for an actual, programmable, rhythm.

Mind. Blown.

So, take three outputs from the emitter (2N2646) / anode (2N6027) through three separate 330k resistors, two feed the bases of the control BJTs in the left and right channel, the third is the clock trigger for the 4017?.....

Therefore I need to find a voltage controlled panner, and use 8 micro pots as voltage dividers to set the pan position of each step. This is going to be wild. Many, many thanks to anotherjim.

[digi2t]

Check out 4ms Stereo Panneur. It's two Trem Lune circuits that can be linked together, or split and run independently when panning. The latter is really cool when using different time spacing. The stereo effect can be almost disorientating. It's a stereo in and out deal.

I realize that it doesn't include the percussive aspect a la Ludwig Phase II, but it might help you get around the panning/tremolo angle.

[anotherjim]

Yes, but you need to have the divider/sequencer output deciding which of the control BJT's work. A BJT used as a switch (logic - resistor - base), emitter to 0v and collector to base of the control BJT should do it. When the logic is high, the control base is clamped to 0v and that control is off and will ignore its LFO. When logic low, the control BJT is free to follow the LFO waveform.

Most of the 4017 sequencer schemes select CV pots, this one is close to what I was thinking...

That is a 4 row, you only need 2. As (I think) the idea is that only one output swells at a time, you can use 1 centre off (on-off-on) SPDT switche in place of 2 SPST switches per step. So one or the other channel can fire or neither on each individual step.

The clock for the counter might be better coming from R9/Q2 if that's closer to a square wave.

[nocentelli]

Yes, but you need to have the divider/sequencer output deciding which of the control BJT's work. A BJT used as a switch (logic - resistor - base), emitter to 0v and collector to base of the control BJT should do it. When the logic is high, the control base is clamped to 0v and that control is off and will ignore its LFO. When logic low, the control BJT is free to follow the LFO waveform.

Ah. I was thinking along the lines of the RP lfo triggering both control BJT's simultaneously, and the LFO also clocking a stepped panning circuit that controlled the exact blend of outputs at each step.

[nocentelli]

So after a bit more research, I came across the schematic below for a voltage controlled panner. I'm going to breadboard a circuit with the guitar going to a single RP circuit, feed the output of the RP to the G Forrest Cook panner input, then use the LFO from the VRP to also clock a 4017 acting as an 8-step sequencer which will have a voltage divider pot on each output. I'm hoping I can just connect all 8 wipers from the pots to give a stepped 0-9v control voltage for the panner.



I've got to order a few parts, so the actual breadboarding won't happen immediately, and there are a couple of uncertainties (mainly centred around the fact the panner circuit appears to be aimed at the synth crowd):

1) I want to convert the bipolar +/-15v supply to a single sided +9v. Is the circuit still likely to function effectively if I did this?

2) Assuming the answer to 1) is an unqualified "yes", I guess I just need to replace -15v on the schem with connections to ground/battery negative, and replace ground connections on the schem to a 4.5v vref. But how to produce the vref? Will a simple e.g. 100k/100k divider suffice? I also have a TLE 2426 rail splitter, might this be a better choice, or a buffered voltage divider vref?

3) If a single sided supply is a no-no (e.g. because it's reducing total supply voltage by 70%) would a charge pump providing +/-9v be a better bet?

4) The 330r value of the resistors on the transistor bases of the matched pair seem very low (to my inexperienced eyes). I was wondering if I should increase them if I'm using it with lower level guitar signals?

5) Since it will be a stereo pedal, I'm hoping I can just use a 4PDT footswitch for bypass, with the stereo output bypass lugs simply jumpered together: Is there a better way, e.g. a buffer between the input and outputs with separate output caps feeding the stereo outs in bypass?

6) Might the voltage outputs from the sequencer pots need to be connected to the CV input via diodes, or can I just sum them together?

I'll post a proposed schematic asap, but I'm away this weekend so I'm using the travelling time to do some thinking (a rare treat).

[PRR]

> I want to convert the bipolar +/-15v supply to a single sided +9v. Is the circuit still likely to function

No.

Or not "trivially".

Look what is happening. Input is applied near "zero". Transistor emitters are biased by a voltage from about -0.5V to whatever. If you redefine "zero" as "+4.5V" you have to figure what the new bias range is for the transistors' emitters. It will be temperature dependent.

BTW, what that circuit does can be done better by a LM3080. Same gain-cell, more linear bias.

> travelling time to do some thinking

Good plan.

[nocentelli]

So I could use a 13700, maybe like in the mfos panner:

http://musicfromouterspace.com/analogsynth_new/STEREOPANNER/STEREOPANNER.html

I've only used OTAs in filter circuits with a single-sided, 4.5 or 5v vref and positive only voltage to pin 1+16. Some more research needed.