Voltage controlled resistor circuit

[pichenettes]

No this doesn’t make more sense

You really have to get out of this “voltage controlled resistor” mindset!

I’ll take an example. If you take the schematics of a digital Mutable Instruments module, you’ll see many pots used as voltage divider and connected to ADC inputs of the MCU. How would you turn these into (GROSS!) CV inputs: just by routing a jack to the MCU pin, because the function of the circuit built with the pot is just to generate a voltage between 0 and 3.3V, so the same thing could be achieved by directly sending a voltage in that range.

My point is that pots can be used in all kinds of circuits: as mere dividers connected to rails for generating DC voltages within a certain range, for adjusting the gain of an op-amp through which a signal pass, for crossfading two signals… You have to consider the circuit globally, and know its function to understand how to make it voltage controlled.

So one needs to see the whole circuit and understand how it operates. In some cases, you wouldn’t even need op-amps and transistors. In the worst case, you’d need several OTAs or VCA chips.

What does this circuit do? Even the part with the 4067 doesn’t make much sense to me (why is only one address pin connected? where do the outputs go?

[jlmitch5]

Okay, I understand…will come back with the full schematic in a bit. TBH my analog circuit knowledge is very limited, basically nonexistent before I started this project.

So far I’ve been making progress by treating it sort of like swappable blocks. For example, instead of just buttons I wanted gate ins, so I basically hooked that into where the buttons were in the circuit…so I found a gate in circuit (marbles actually) and just put it in there and it’s working great!

Was trying to approach the cv in a similar way and I think I need to actually understand what’s going on.

[jlmitch5]

Okay, finally got through this!

Here’s the schematics EDIT: updated with component number annotations

---

What does this circuit do? Even the part with the 4067 doesn’t make much sense to me (why is only one address pin connected? where do the outputs go?

So basically this is a 3 button, 7 note binary sound generating keyboard. A0, A1 and A2 are the buttons, the output channels are connected to pots which control the pitch of the particular note. (the PitchPlus and PitchMinus work additvely to determine the pitch…raising resistance on either for the particular channel will make the note lower. Note 0 is the “off” note. When notes 1-7 complete, the decay portion will be at whatever pitch note 0 is set to.

Honestly…still pretty fuzzy on how it is generating sound, it seems like pin 1 outputs a square wave on the scope which changes frequency when the buttons are played (if the pitch knobs are set to different notes…also seems like that B2M master pitch pot can control this square wave frequency. My guess is this is the main oscilator signal, and somehow A3 and Z on the 4067/three PNPs/second op amp are shaping the wave into something more triangle-y and adding the envelope/decay, and switching between the note pressed with the button and the 0 note. I think it clicked what you meant about understanding the whole circuit…I’ve poked around and I have not found a voltage that goes up and down based on turning the master pitch knob (nor have I found one for the decay circuit…the other thing I’d like to add cv to).

[pichenettes]

Ok, I get how the thing works.

Depending on which button is pressed, one of the PitchPlusN voltages charges the “hairy cap” selected by SW? (number your parts!), so an increasing ramp is observed at pin 1 of the 4067. This voltage is buffered by one of the TL082 (U?B, number your parts!), and further enveloped and sent to the audio output. This voltage is also sent to a comparator made with U?A, so that, when it reaches a certain level, the 4067 selects one of the PItchMinusN voltages instead and the capacitor is instead discharged - during this phase, you should see a decreasing ramp at pin 1 of the 4067. So square output on A3, and triangle output on Z (or buffered at pin 7 of the TL082).

The pot you want to voltage control, RV? (number your parts!) changes the threshold at which the comparator flips. It has an impact on the triangle wave’s frequency (since a lower threshold will mean that the signal changes slope earlier in the cycle), but also on the triangle wave’s amplitude (a lower threshold means a lower amplitude). You could probably confirm this experimentally, but is this really what you want?

Some things you could try:

1. Altering the voltage at the 22k R? (number your parts!) resistor connected to pin 3 of the TL082. This is another way of changing the threshold directly from an external voltage. You’ll need to experiment with different values for R? and for RV? (number your parts!). Caveat: if you go this route, changing the pitch will change the amplitude, and the range of pitch control might not be great. It’s okay for a slight vibrato+tremolo effect maybe, not for more.
2. Buffering your CV source and connecting it in place of the +9V of all the PitchPlusN circuits. Caveat: the waveform will be asymmetrical, since it’ll change slope only during its ascending phase. It’ll have a constant slope during the descending phase. And obviously you’ll have a messed up relationship between control voltage and resulting pitch!
3. The “right” way of doing it is to alter the current charging the cap, at pin 1 of the 4067. You can do that with a 2164 cell (the 2164 can be used in single-supply operation, use Bias? for its GND pin, GND for Vee, and +9V for Vcc) ; or with an OTA cell. Both have current output and this is what you need. The 2164 cell offers you expo control, the LM13700 needs some more work for expo control. Since the 2164 output can only go to a virtual ground, you’ll need to slightly change the topology of the circuit (replace cap+buffer by an op-amp integrator made with U?B.

For decay control, I’m probably very wrong, but you could try replacing the variable resistor by a fixed one, and change the voltage it’s connected to (not 9V, but a buffered CV). This seems to change the rate at which the capacitor goes back to full charge.

PS: number your parts!

[jlmitch5]

Thanks so much for all this info, and apologies for not having the components numbered…I just updated my post so they would have those.

I understand most everything you mention, and you are right about the amplitude being changed in coordination with the pitch…it does so with the variable pot too.

The “right” way of doing it is to alter the current charging the cap, at pin 1 of the 4067. You can do that with a 2164 cell (the 2164 can be used in single-supply operation, use Bias? for its GND pin, GND for Vee, and +9V for Vcc) ; or with an OTA cell. Both have current output and this is what you need. The 2164 cell offers you expo control, the LM13700 needs some more work for expo control. Since the 2164 output can only go to a virtual ground, you’ll need to slightly change the topology of the circuit (replace cap+buffer by an op-amp integrator made with U?B.

If I go the 2164 route, I think pin 3 here is the cv input:

I am guessing I could use a pot to control the amplitude of the CV voltage getting into pin 3.

The only part I’m not exactly understanding is:

replace cap+buffer by an op-amp integrator made with U3B (NOTE: updated with component number)

What is the cap and buffer and how do I replace it with the U3B circuit?

[pichenettes]

Yes pin 3 of the 2164 is the control voltage input. 0V and Iout mirrors Iin (the VCA is “wide open”), then for each increment of about 0.16V of this voltage, only half the current goes through. So you would effectively get a -0.16V / octave response (an inverting op-amp circuit can give you the proper 1V/octave response, you can copy Ripples’ schematics).

@jlmitch5

What is the cap and buffer and how do I replace it with the U3B circuit?

The cap is C5, C6 or C7 (as selected by the switch), and the buffer is U3B.

You need to replace this buffer by an inverting integrator built with U3B, so the topology changes a lot around U3B.

• C5 (or C6, C7) is no longer connected to ground or Z, but goes between pin 6 and 7 (feedback path).
• Pin 5 is only connected to BIAS.
• Pin 6 is connected to Z.
• R25 is still connected to pin 7.
• Pin 2 of the schmitt trigger (U3A) is only connected to pin 7 now.

And because this op-amp is now inverting, you need to swap all the Minus and Plus connection on the 4067 side.

A side effect of this change is that you’ll get a perfect triangle waveform. Good or bad?

Maybe at this point you can try drawing the schematics with all these changes!

But with all these changes in place, you can now insert a 2164: Vin to Z, Vout to pin 6 of the integrator.

At this point, I would guess it would make more sense to just throw that 4067 nonsense and replace it with a simple triangle VCO. Then a design a circuit that mixes an external CV with the NoteOn0 voltages, each with its own trimmer, and send it to that VCO. Adding voltage control to a circuit that can afford to be simple only because it doesn’t need voltage control is a much more complicated route than designing something voltage-controlled in the first place.

[jlmitch5]

I think I got everything you mentioned (part numbers may have changed slightly because I had to reannotate)

@pichenettes

• C5 (or C6, C7) is no longer connected to ground or Z, but goes between pin 6 and 7 (feedback path).
• Pin 5 is only connected to BIAS.
• R25 is still connected to pin 7.
• Pin 2 of the schmitt trigger (U3A) is only connected to pin 7 now.

@pichenettes

(the 2164 can be used in single-supply operation, use Bias? for its GND pin, GND for Vee, and +9V for Vcc)

---

I read:

But with all these changes in place, you can now insert a 2164: Vin to Z, Vout to pin 6 of the integrator.

as breaking

• Pin 6 is connected to Z.

basically allowing a variable amount of current between 6 and z.
Was that assumption correct?

---

I’m working on PitchCV with @brianleu now. I think what I want is this circuit from ripples (F_CV = PitchCV in my schematic above), but I’ve learned it’s slightly more complicated because I’m just using 9v and don’t have access to + and - voltages to do this. Hopefully we will have that part figured out tomorrow and I can share.

[pichenettes]

You almost got it right (though in a way that is a bit hard to parse visually!). The R26/R32 divider should be connected to pin 7 of U3B – you’re showing to the comparator a fraction of the voltage at the integrator’s output, not some current!

@jlmitch5

Hopefully we will have that part figured out tomorrow and I can share.

Ah yes, you’ll have some work to do here! I don’t remember how the V2164 is wired internally – if the control voltage has to be Vbias for an attenuation of 0dB to be obtained (I think so!), or 0V.

[jlmitch5]

EDIT 5:

Small update, I was missing connection to A3 on the 4067

[jlmitch5]

And so far as the inverter/attenuator, here’s what we (well really what @brianleu) came up with:

simulator

In terms of supply pins for the 2164…we’re not sure if it should be this V/3 bias (necessary to get the math right) or the V/2 bias the rest of the gerassic organ circuit uses, but that’s easy enough to experimentally figure out.

---

The CV for the new pitch input responds to 0…+9v, which we think is fine? In terms of euro land, it seems more common to accept -5v…5v, but that could be achieved with a negative offset if needed.

[pichenettes]

Four observations.

1. I assume the 40 Hz sine wave is your CV source. Then you don’t need the buffer. It’s actually doing more harm than good since it has to be a fancy rail-to-rail part to work as expected and output voltages as low as 0V. Obviously the circuit simulator simulates ideal op-amps! A benefit of removing that op-amp is that you can adjust the circuit for a wider CV range that will extend below 0V and above 9V.

2. Related to the above: it’s very rare for a CV input circuit to be something other than a 100k resistor going to the inverting pin of an op-amp.

3. I checked the V2164 datasheet, and the voltage on the VC pin must vary between 4.5V and about 6V for an input range of 9 octaves. 4.5V for the 2164 to let all the current flow through, about 6V for the 2164 to attenuate the current by 2^-9.

4. I would recommend buffering VBias before sending it to strategic places - at the very least the 2164 pins. It’s going to be used in many places in the circuit. You’ll lose some of the original tone character, though, since in the original circuit, many subcircuits are loosely coupled and interact with each other imperfectly through that wobbly voltage.

[brianleu]

@pichenettes

Then you don’t need the buffer

The only reason it may be needed here is that this is intended to use banana jacks, which makes it a similar case to this where switched jacks (connected to GND when no jack inserted) can’t be used.

[jlmitch5]

@pichenettes

I would recommend buffering VBias before sending it to strategic places - at the very least the 2164 pins. It’s going to be used in many places in the circuit. You’ll lose some of the original tone character, though, since in the original circuit, many subcircuits are loosely coupled and interact with each other imperfectly through that wobbly voltage.

Bias does very slightly go up when one of the buttons is depressed and back down when it is released (it goes from like 4.1 to 4.25 as measured on my old scope). I can investigate buffering the whole circuit, as well as just buffering the 2164 pins.

I do really like the tone of the organ in the original circuit! fingers-crossed making the circuit better doesn’t lose the vibes…I’m really excited to able to mess with pitch changes in time with modulations doing the gate in stuff, I think it could potentially make some pretty strange and interesting sequences.

---

So far as the buffer goes, I was just thinking and we could add a physical switch that grounds the input in addition to the jack. Not sure about the tradeoff there.

[jlmitch5]

Also apologies it wasn’t clear about the banana jack thing. the plan is actually to do all the jacks with this combo thonkiconn/cinch banana part that electric noodle box has shared with schematics for their modules https://electricnoodlebox.com/diy-pages/

with the idea that people could have a choice if they want to build this with 3.5mm TS jacks or bananas (like the rest of the Ciat-Lonbarde stuff)

for that to happen i need to figure out how to get their eagle part to work with kicad and i was trying to take it one step at a time (getting the schematic more or less figured out before i start importing custom parts and doing the PCB layout)

[jlmitch5]

Here is what I’m thinking w the physical switch idea

simulation

[pichenettes]

Ah thanks for pointing that, I’m way too much Eurorack-minded

[jlmitch5]

I’ve made these circuits on the breadboard and am getting no sound or any sort of voltage change out of J3. Things I’ve attempted to troubleshoot:

• I’ve went over the circuit traces several times
• I attempted to buffer all of v bias because it kept going up in voltage as things were connected to it.
• I’ve tried excising the 2164 part of the circuit, connecting pin 6 directly to z

I’m going to try to redo this whole part, but I want to check to double check things are right around U3 and Q6?

[pichenettes]

There’s something wrong around R26/R32.

It should be a divider that takes the voltage at U3B’s output, and the divided voltage is applied to the base of the transistor.

[pichenettes]

[jlmitch5]

After a period of burnout and taking things a part/putting them back together again several times, I’ve finally got it working. Need to dial in the response a bit, but it’s working pretty good. Thanks for all the help!! Side note, really need to figure out a more organized method of breadboarding than the mess of wires I’ve got going on, makes it very difficult to debug.

I currently have bias buffered to all destinations and the sound is still pretty gerassic organ-y. One interesting thing I noticed is that fast sequences can cause some build up in the amplitude which results in “clipping”. Raising C3 to 100uF helps mitigate this problem, but the response is too slow for normal playing…will probably put that on a switch. It seems like C5 doesn’t seem to really do anything any more. maybe that is due to the bias voltage being buffered.

As I get into PCB designing mode, is there any good resources out there for figuring out the best SMD equivalents to the through-hole stuff I’ve been using on my breadboard (particularly op-amp ICs, transistors and diodes…I feel like resistors and caps should be more straightforward.).