DIY Vactrol

[amptramp]

You could implement OR gate logic if you have two LED's with one photoresistor.  Turn either LED on and the photoresistor goes to low resistance.  This may require a proper design of LED holder to ensure light gets to the photoresistor equally from both LED's.

This could be used for fullwave rectification of an incoming signal with LED's strung an LED diode drop from ground and V+ for a compressor.  Since an AC-coupled signal with a lot of even-order harmonics tends to have an average voltage that is much higher on one side of the average than the other, there is a need for responding to both signal polarities.  This would allow you to make an optical compressor that actually works reliably regardless of signal polarity from the guitar.

[Matthew Sanford]

It sounds like you may have a way to implement isolated popless, clickless switching that everyone would like for their pedals.  True bypass has its uses but the major manufacturers have avoided it for a reason, mainly expensive and unreliable 3PDT switches.  You can switch from a series resistance of several hundred ohms to something in the megohm range and after a suitable delay, switch the output of the vactrol in its high impedance state to ground with a saturated transistor and get better on:off ratios than most switching systems.

Interesting about the optical compression, I've yet to jump into that, but getting rid of 3pdt would be amazing! I have to figure the slight delay to switch to ground, but need to set up a flip flop and try

[Matthew Sanford]

@amptramp , for switching with a vactrol did you mean like this? Using a flip flop to send the signals, turning the vactrol to high resistance with one side off and the other side going across a C to slow it into the base? And should it have a large resistor there or effectively the vactol becomes that resistor? uh...also, would there need to be anything between the transistor and the circuit, or is it just pulling the input of the circuit to ground more than the now open input jack connection?

[antonis]

[amptramp]

@amptramp , for switching with a vactrol did you mean like this? Using a flip flop to send the signals, turning the vactrol to high resistance with one side off and the other side going across a C to slow it into the base? And should it have a large resistor there or effectively the vactol becomes that resistor? uh...also, would there need to be anything between the transistor and the circuit, or is it just pulling the input of the circuit to ground more than the now open input jack connection?

Yes, this flip-flop is a circuit that has been used by Boss and others because it is cheaper and more reliable than a 3PDT switch.  The change I would make to your circuit is to go with FET drivers for the LED and a FET as the grounding transistor because otherwise, the flip-flop would be overloaded by base current and would not operate.  You can do a slow turn-on for the grounding transistor with the series R-C as you have drawn but you can turn off the grounding transistor rapidly by putting a diode across the resistor with the anode toward the FET gate (which should be an n-channel JFET or MOSFET).

The vactrol is an alternative to the 2SK30 JFET's used by Boss as the series switching element.  You can ground the input for most effects, but if you have a delay / chorus / reverb / flanger as an effect, you would not want to ground the input as there will be a gap between switching the unit on and getting any sound out.

Note that the signal should be biased so the it never goes below ground, so you will have a DC level at the output that has to be removed by capacitive coupling.  Most pedals use Vcc/2 as the bias point and this would work here also.  As with the Boss pedals, you don't have to worry much about output switching as there are always two signals, the unmodified bypass signal and the effect output that meet up at the output buffer so you always have a low-impedance output through the "ON" vactrol to overwhelm the signal from the "OFF" vactrol with an impedance ratio of hundreds of ohms to megohms between the two vactrols or almost 40 db.  It would be comparable to a pair of 2SK30's switching between bypass and processed signal but the selected signal would switch more slowly without any need for added time constants.

[Matthew Sanford]

Ok, good to know. I want to try setting up switching to have a few busses to plug effect in and out into so I can switch their order on the fly too, so this will be good not just for the effect on/off. I have some 5458s so I will try those instead of the 2222s...though I may start with the BJT so I can see.

First though I'm filling up the breadboard for other demonstrations stuff.

[Matthew Sanford]

It sounds like you may have a way to implement isolated popless, clickless switching that everyone would like for their pedals.  True bypass has its uses but the major manufacturers have avoided it for a reason, mainly expensive and unreliable 3PDT switches.  You can switch from a series resistance of several hundred ohms to something in the megohm range and after a suitable delay, switch the output of the vactrol in its high impedance state to ground with a saturated transistor and get better on:off ratios than most switching systems.

Ron, I got some SPDT stomps and plan to route the 9v to each side for the switching. Pretty please will you take a look at this and see if it makes sense? I think got it today re-reading that the BJT takes the input side of the non-lit vactrol to ground, so each side does a vactrol and the BJT on the other side.

(Quick edit: paint picture had NPN arrow wrong way)

[PRR]

see if it makes sense?

Isn't Input shorted to ground by TR1 or TR2 in any switch position?

[Matthew Sanford]

see if it makes sense?

Isn't Input shorted to ground by TR1 or TR2 in any switch position?

When I was out shopping I started thinking about that and the tied input. Maybe 22k between emitter and ground? 1M even? Makes me think that I only need the bjt on when bypassing, so just one, no need to ground the bypass

[Matthew Sanford]

I think this works, the bjt turns on but still has plenty of resistance to ground for bypass to work, vice versa that bypass path is blocked.

[amptramp]

I concur with Paul, your original design shorts the input.  In your latest design, the transistor is never switched off because the base is either pulled up to the 9 VDC supply or pulled up to the +7VDC at the bottom of the LED, which is still enough to turn it on.  If you are using a transistor to ground a signal, it cannot be on the input side of the vactrol because once the switch turns on, the input is shorted so nothing goes out.

[Matthew Sanford]

I concur with Paul, your original design shorts the input.  In your latest design, the transistor is never switched off because the base is either pulled up to the 9 VDC supply or pulled up to the +7VDC at the bottom of the LED, which is still enough to turn it on.  If you are using a transistor to ground a signal, it cannot be on the input side of the vactrol because once the switch turns on, the input is shorted so nothing goes out.

Ok, so move the transistor to the effect side? I'll try it...I was hoping an emitter side large resistor would prevent clean from going to ground...and on the LED I assume separating their grounds with separate resistors might cure that?

[amptramp]

I concur with Paul, your original design shorts the input.  In your latest design, the transistor is never switched off because the base is either pulled up to the 9 VDC supply or pulled up to the +7VDC at the bottom of the LED, which is still enough to turn it on.  If you are using a transistor to ground a signal, it cannot be on the input side of the vactrol because once the switch turns on, the input is shorted so nothing goes out.

Ok, so move the transistor to the effect side? I'll try it...I was hoping an emitter side large resistor would prevent clean from going to ground...and on the LED I assume separating their grounds with separate resistors might cure that?

You need the transistor to clamp the signal so the ~40 db of effect output or bypass input can be cleaned up to maybe 60 db or more.  You can set up the vactrols to shut off the connection from input to output (bypassing the effect), from the effect circuitry to output and effect circuitry input.  You may need a T junction of vactrols so the branches of the "T" can go high impedance and the transistor is turned on but it does not affect what is connected to it.

Effect circuitry input is a special case - if it has a memory element, you may want the input always connected so there is no empty output while the memory is filling with new information.

[Matthew Sanford]

You need the transistor to clamp the signal so the ~40 db of effect output or bypass input can be cleaned up to maybe 60 db or more.  You can set up the vactrols to shut off the connection from input to output (bypassing the effect), from the effect circuitry to output and effect circuitry input.  You may need a T junction of vactrols so the branches of the "T" can go high impedance and the transistor is turned on but it does not affect what is connected to it.

Effect circuitry input is a special case - if it has a memory element, you may want the input always connected so there is no empty output while the memory is filling with new information.

Ok, so maybe switch output? I was thinking to set them to switch input only, so two vactrols from input to output or Fx input, idea being if there are tails they go through. Some, like Christine, could still be noisy even with input grounded I guess...I'll have to try.

For the voltage from SPDT, I'm thinking bjt (between vactrol and effect in) on same as bypass vactrol, status LED on same as effect vactrol.

Perhaps 3 vactrols, one for effect out if needed. I guess I could use an optotransistor for parts that can go faster, I do have a random 50 pack of PC817s that need use

[PRR]

As long as you got the holiday-pack of opto-resistors, just do it all with slow mellow optos. This will give off ratio >60dB, on loss <1dB. Output resistance in Bypass using cheap optos and not-the-britest LEDs may be some over 1K but that's fine in stage-gear.

[Matthew Sanford]

As long as you got the holiday-pack of opto-resistors, just do it all with slow mellow optos. This will give off ratio >60dB, on loss <1dB. Output resistance in Bypass using cheap optos and not-the-britest LEDs may be some over 1K but that's fine in stage-gear.

Actually sat down and blue taped 28 together last night then silicone tape and stretched shrink this morning. At less than $0.30 materials each plus a few hours it feels worth it.

On a breadboard with 2 optos & 1 led a side w/1k clr it read 550,1k on left, 580, 550 on right. I figure using 470 clr would get them lower, and realize my QC is not on point yet. I guess I'll try 4 per switch for noisy things (like the DRM and Christine that I need to get boxed) and 2 w/bjt to switch Fx in to ground for flowy things. Board mount SPDT so I'll make little veros, easy peasy, but this weeks busy!

I have some 3PDT left, but I don't love those melty things!
(Edit: I realize you have 4 for input, not fx out. Would tying any effect's input to ground assure no sound?)

[Matthew Sanford]

As long as you got the holiday-pack of opto-resistors, just do it all with slow mellow optos. This will give off ratio >60dB, on loss <1dB. Output resistance in Bypass using cheap optos and not-the-britest LEDs may be some over 1K but that's fine in stage-gear.

Had to make it 3 vactrols, the second Fx one was creating a 400 ohm to ground on the output since effect out and bypass all go to the same place. After cutting it the switch works great, high resistance to ground on bypass with little resistance to ground effect in, and noiseless switching. A bit bulky, but with diy vactrols cheap and effective. Thinking I should strap 2 LDRs to one LED for bypass, Antonis said flattops shine brighter sideways. Anyway, thanks PRR for the direction!

[Matthew Sanford]

Just some notes after usage.

On Christine the bypass signal is solid with the tens of M to ground, but it interacts with the volume control in that about 3/4 up it starts losing treble to major signal cut with volume full. Strange, as I'd figure volume at 0 would provide a low resistance path to ground...maybe it is. Other route to effect out hits a coupling capacitor. For that one's solution I'll tie the bypass to the output buffer to separate it. I am figuring it has parallel paths to ground from the output jack to ground, so maybe it's reducing the input impedance the dark fx input ldr (to ground through second lit bypass ldr) so there isn't much left there.

How it's wired

[amptramp]

As the level control goes to the top, you have a voltage divider between the output from the bypass vactrol and the output impedance of the effect.  If it is an op amp output, that may be only a few ohms.  Quite often, a circuit output will have a 100 to 1000 ohm resistor in series to minimize damage from output connections backdriving the signal or to isolate a feedback stage from the destabilizing effect of cable capacitance.  Your maximum volume will occur near the midpoint of the level control.  If you have the bypass vactrol on, you need another vactrol to disconnect the output of the effect circuit.

[Matthew Sanford]

As the level control goes to the top, you have a voltage divider between the output from the bypass vactrol and the output impedance of the effect.  If it is an op amp output, that may be only a few ohms.  Quite often, a circuit output will have a 100 to 1000 ohm resistor in series to minimize damage from output connections backdriving the signal or to isolate a feedback stage from the destabilizing effect of cable capacitance.  Your maximum volume will occur near the midpoint of the level control.  If you have the bypass vactrol on, you need another vactrol to disconnect the output of the effect circuit.

I think I understand it now. I had one on the output tied to Fx on but it was sending everything to ground, so really it should tie to the Fx output (on the bypass vactrol line) BEFORE the volume pot, creating a parallel to reduce the volume pot resistance to very low for the voltage divider issue. Makes sense