Bypass pop and current limiting resistors

[DIY Bass]

I have a couple of pedals made using the standard 3PDT stomp switches that pop quite noticeably when turned on.  I am looking at ways to limit the pop.  I am tempted to go full on and do a relay bypass with optical mute, but I am looking at easier/cheaper options first.  I have looked at https://www.mrblackpedals.com/blogs/straight-jive/6629778-what-really-causes-switch-pop a few times.  The pedals I am looking at all have input caps and pull down switches already. 

One of the ones I am looking at is this:  https://docs.pedalpcb.com/project/Arachnid.pdf

Is there any point in trying to add in current limiting resistors at the input/output?  What values would be in the ballpark to work but not reduce volume/tone?  If I switched R22 and R12 to the other sides of their respective capacitors is that viable?  Any other ideas that might help?

Thanks

[antonis]

I'll not again express opinion about Mr. Back's aspects, so I'd propose a IN-OUT grounded wiring for 3PDT switch..



Same with LED anti-popping configuration..



edit: Pics need resizing..

[amz-fx]

Sometimes a switch will make a pop noise that is difficult to silence without extra circuitry. Here is an article about why switches make a noise when actuated:

6 Reasons Guitar Pedals Pop When Switched

regards, Jack

[antonis]

6 Reasons Guitar Pedals Pop When Switched

Continually forget to attach your comprehensive article, Jack..

[GibsonGM]

Rather than relay bypass - how about the Boss method, flip-flop and FET switching?

[R.G.]

Audio switching pops any time the switch changes between two different voltage levels instantly. Hard-metal contacts do this as badly as any instant-switching electronic means, whether JFET, MOSFET or optical. Fancy $$$ analog switching ICs have internal circuits to sense that the two voltages being switched are instantly at the same level and switch RIGHT THEN when they happen to be the same.

Instant switch devices also have feed-through, the effect when a control voltage leaks from a control pin(s) into the audio path. Relays are prone to this unless you're careful about the speed of the on/off signal; so are JFETs and MOSFETs through their internal capacitances. Muting devices can be timed to mute before the induced click on relays.

There are two practical after-the-fact means to eliminate switch voltage transitions. One is to set things up so the from-to voltage levels are very, very close to one another. The classical example is the pull-down resistor on input capacitors, or the middle-of-power-supply bias on MOSFET analog switches.

The other is to slow down the transition between two different voltage levels. With JFETs, you do this by slowing down the gate movement with an RC network so the ramp speed is below human hearing. Something about 50-60mS is slow enough to be inaudible. Things using LDRs are usually inherently slow enough just because the LDR can't change resistance very quickly.

Note that instant acting mute circuits may be almost as bad as instant switch-betweens. If they happen to switch at the instance that the audio is at a peak, that is clearly audible. In fact, electronic organs (at least back when these were not just specialized computers) used this effect to introduce "chiff", faking the bit of air-leakage before a note started on the pipes.

Bottom lines are: (1) don't introduce clicks and pops from a control signal (2) mute the clicks and pops when you just can't avoid introducing them and (3) minimize the voltages switched between and how fast the switch makes and breaks.

There are a lot of articles on how to do switching on geofex that include bits of this, as well as other bits of finesse about switching.

[DIY Bass]

I had completely forgotten about "pedals don't play nicely together" syndrome.  The pedals in question are sandwiched between a Maxon EQ and an ARia flanger.  I can see from the EQ schematic that there is a pull down resistor on the output (but not the input).  Not sure about the flanger.  I will need to pull it off the board and investigate.  I am betting at the moment that it doesn't have one on the input and that is my problem.  I will probably add something to the input of the EQs as well, even though the pedals before them are mostly always on or hardly used.  If that fails I am tempted to muck around with RG's CD4053 bypass schemes - he says they are quiet and the parts are cheap and plentiful enough to breadboard it and play.  Will be cheaper than a relay and an optocoupler and an ATTiny.  Anyway, thanks all.  I have some things to go on with.

[antonis]

IMHO, input pull-down resistor neccesity depends on input cap both value and leakage..
(for a 10μF electro is a must where for a 10nF metal film one might make no difference..)

[amptramp]

There is an easier way to get LED popping reduction.  In the off position, have the switch short across the LED so it stays off.  In the on position, open the switch.  The disadvantage is continuous current drain in both positions.  The advantage is that if you are running, say, a green LED with a 2.05 volt drop from a 9 volt supply, the current only changes from 9/R to 6.95/R which is not near as much as 6.95/R to zero.  it saves an electrolytic cap as shown in the second schematic by antonis.

[amz-fx]

Sometimes it is mechanical noise, and you will need to change the switch to one of the soft-click models:

https://lovemyswitches.com/pro-3pdt-latched-foot-switch-solder-lugs-feather-soft-click/

regards, Jack

[GibsonGM]

^   Those are really nice.  I got some just because, and now prefer them. 

[R.G.]

I am tempted to muck around with RG's CD4053 bypass schemes - he says they are quiet and the parts are cheap and plentiful enough to breadboard it and play. 

Done properly, they are quiet. Some of my CD405x designs are used in commercial pedals, so there is a lot of field experience with this technique.

The important thing is to bias the inputs and outputs to half the DC power supply voltage on the CD4053. This costs you a couple of capacitors and resistors to block that DC level from the rest of the circuit, but it balances the internal MOSFETs in the chip so any control feedthrough clicks are balanced out internally.

This technique doesn't require nearly the attention to careful layout for current flow that a relay circuit does.

[amptramp]

One of the few threads I started was about silent switching:

https://www.diystompboxes.com/smfforum/index.php?topic=120006.msg1122294#msg1122294

Some diagrams disappeared, so I added them i the last post on that thread.  The thread introduces three switching methods: the 4007 linear switch, a diode bridge switch and a 4066-based switch where additional circuitry postpones the switchover until one input matches the other in voltage.  All of these are set up to use a momentary SPST pushbutton switch, just like Boss pedals.

[R.G.]

Is there any point in trying to add in current limiting resistors at the input/output?  What values would be in the ballpark to work but not reduce volume/tone?  If I switched R22 and R12 to the other sides of their respective capacitors is that viable?  Any other ideas that might help?

I realized that no one had addressed the current limiting resistor concept.

Is there any point in adding current limiting resistors? On the output, sure. Particularly for opamp-output circuits, a small resistor, 100-1000 ohms right after the opamp output pin helps ensure that the opamp stays stable for certain situations where the output is shorted. Some opamps are sensitive to a capacitive load, and that's what the output capacitor becomes when the output is shorted. But it's a good idea for reasons other than switch bounce, and in this schematic, it ought to be where it is - between the opamp and the output capacitor, not outside the output cap. You could add another series cap outside the output cap if you want to play with things, but I don't think it's needed, for reasons I outline below.

On the input, its more subtle. I have never used series current limiting resistors on inputs, over a long period of designing for both DIYers and commercial pedals, including on ones with hard-contact switching. I have never noticed input pops that were not taken care of by input pulldown resistors and proper switching, including both waving-in-the-air and shorted-input wiring. I can't say that there is never a situation where switch bounce can't make a pop, but I haven't seen one in some decades of pedal building and design.

The idea of the input wire being an antenna while the switch bounces is novel. There are two ways that pickup could occur: RF and capacitive. RF is unlikely, because the length of the wire puts some frequency limits on how low an RF frequency could couple to the antenna. Wires of a couple of inches long mean that RF would have to be hundreds of MHz or higher to couple effectively to the wire. If you have poor external shielding outside the jacks, yeah, you'll pick up RF, but you'll also get the RF in normal operations, not just on switch bounce.
Magnetic coupling to an open wire is nearly impossible. Sure, special circumstances, very high magnetic fields, and so on might, but I've never seen it in a pedal.
Capacitive loading of charge onto the antenna wire is possible in environments where there is a high electrical charge in the air around the wire.  In a conductive enclosing box which forms a hum-excluding Faraday shield, there is remarkably little internal electrical field to pick up. Capacitors store charge ruled by the equation Q = C*V, where Q is the charge in coulombs, C is the capacitance in farads, and V is in volts. The voltage deposited on an input capacitor is then V = Q/C. The charge available to a wire waving in the air depends on how much charge the air has to transfer to the capacitance and how big the capacitance is.

There are two capacitors here: the capacitance into the wire from the air, and the capacitor at the input of the effect. These are effectively in series. The self-capacitance of a short wire is down in the 10pF or lower range, and even including the stray capacitance of the metal parts of the jack, solder joints, PCB trace and so on to get to the pull down resistor and input cap does not increase that much. The input capacitor of the effect is up in the tens of nano-farads at least and probably more like hundreds of nF to microfarads. The input cap and stray capacitor act like a capacitive divider to any charge picked up by the switch wires during the milliseconds of bounce. Stomp switches do bounce, and it tends to be millisecond-range bounces for maybe 10-20mS. The input wire is only left open for some milliseconds on a bounce, and whatever charge it picks up from the air inside the shielded effect box is divided by the capacitor divider and loaded by the pull-down resistor.

Obviously, about now I would usually be screaming for numbers measuring picked up voltage/charge and pop size in a controlled test setup. But given that I have never seen switching pop caused >only< by switch bounce in pedals, and not removed to inaudibility by the other standard pop-reducing measures, I don't see the need for input current limiting. Not that a need might not exist; I can imagine a situation where the input impedance of a pedal is so high, the input capacitance so low, the input wires so long, the shielding so leaky, the air so full of electrostatic charge that it might be detected by sensitive instrumentation.

But as a practical matter, get the other, well-known stuff right first. Then, if you still have pops that can't be fixed any other way, add a series input resistor. My suspicion is that you will not need it. Just a guess. It hasn't been needed in pedals for the last few decades that I know of.

There is a large number of articles on geofex.com on effects switching...

[PRR]

I figured the "current limiting resistor" for pop was just made-up.

A series resistor in small audio can only be a Low-Pass. So it is gonna cut audio as much as pop. There may be a corner case with a half-dozen dB difference, but in general I cry moose-hockey.

Super-high circuit impedance will get even less benefit from a tolerable series resistance.

The use of 34K + 120pFd at a guitar-amp (or equivalent) input is well proven in high-RF environments. Truck-stops, taxi-stands, police activity, and the old-old days when the live radio show went-out directly under the AM transmitter antenna. AM doesn't work that way today but cell/wifi may still need an R-C. US police have largely moved way up the spectrum, from 2MHz in the 1930s to 600MHz today, and that won't go far in 20KC audio amplifiers. But that is not "pop".

Yes a loose end on a high-enough-gain input is sure to oscillate. (Barkhausen says certain conditions, Murphy overrides that.) An obvious technique is not having loose ends, always terminate in a quiet node. (This can also be a relatively-Zero DC node, cuz that's the usual cause of pop.) Even if it costs more poles.

[DIY Bass]

Well, I added a 1M pull down resistor to the input of the flanger that was after, as there wasn't one there.  Now through a 15W practise amp at full volume there is a barely audible pop that I can certainly live with.  I won't be playing through a PA for a few weeks yet, so i will have to wait and see on that but it sounds very promising.  If all else fails it does occur to me that the 30+ year old input/output caps on the pedals around could probably also be a very cheap change and may also help improve behaviour.  I am definitely going to look at more easy fixes before I delve into more complicated solutions anyway.

[antonis]

If all else fails it does occur to me that the 30+ year old input/output caps on the pedals around could probably also be a very cheap change and may also help improve behaviour.

You can easily check it by significantly lowering pull-down resistor(s) value..
(at the expense of input impedance/output load but we now mind for caps health..)

The lower the pull-down resistor value the most effective the caps grounding..
(smaller RC discharge time constant..)

[GGBB]

I recommend a much higher bottom resistor to avoid a bright flash and possible LED destruction. In a typical pedal - 9V supply and 1.8V LED - the current surge when the LED is turned on will be >20mA through the 330ohm resistor. This is because while the ground path is turned off the cap voltage floats up to the supply voltage 9V so effectively you have 9V through the bottom resistor alone to the LED. This is even more important if you use a higher supply voltage.

[antonis]

I think we've dealt with that in the past, Gord..

4k7 + 330R could be of about equal splitted values but what bothers me is the fact that for a fully charged cap (via pull-up resistor) up to supply voltage, the whole configuration should be useless 'cause there shoudn't be any delay for LED current flow..

I think it should be better for cap GND to be wired on LED cathode..
(Cap should permanetly stay at a voltage level of about 400mV higher than LED anode..)

[GGBB]

4k7 + 330R could be of about equal splitted values but what bothers me is the fact that for a fully charged cap (via pull-up resistor) up to supply voltage, the whole configuration should be useless 'cause there shoudn't be any delay for LED current flow..

I think the value is the reserve charge in the cap which the LED draws from instead of the main supply.

In my builds, I connect the LED power directly to Vin, and the pedal power to Vin through a series Schottky diode. I think the diode protects the main supply caps for the pedal from LED current draw which helps to further "stabilize" the pedal voltage when the LED is turned on.

I think it should be better for cap GND to be wired on LED cathode..
(Cap should permanetly stay at a voltage level of about 400mV higher than LED anode..)

Not sure how long a 10µ cap will hold charge in that configuration. Only "perfect" caps would hold the charge permanently. Real caps especially aluminum electros will eventually discharge. Seconds? Minutes? Hours? ...