A Momentary Stutter Pedal That Doesn't Pop???!!!!!!

[sevenisthenumber]

I want to build a simple box with an input and output that simply kills the signal (like the old Les Paul trick) momentarily when the footswitch is touched. I ahve tried everything out there like Beavis and a few others but they all pop or click when the signal is cut... Is there a way to do this with no pop??

Thanks guys!

[asatbluesboy]

I don't see why a temporary connection to ground would pop, but you could always put a cap between the switch and ground (or hot, it doesn't really matter).

[ashcat_lt]

Any time that you instantaneously change the voltage it will be audible.  How far it changes determines how loud the pop or click will be.

The very first step is to make absofrickenlutely sure that there is no DC offset between the signal and ground.  Even in a passive switch box you could have DC leaking in from poorly behaved active circuits on either side.  Then you will probably not have big pops, but you will still have clicks.  

Why?  Because it is highly unlikely that you will be stomping exactly when the signal is crossing zero.  The instantaneous voltage will be some non-zero value and your foot will cause it to jump to zero real sudden-like.  This will approximate a step function, which includes all frequencies.  This happens with the toggle trick and is a common complaint amongst folks installing kill switches to their guitars.

There is not a good passive solution to this issue.  You need to slow down the change in voltage.  You could use a cap to ground, but that's a low-pass filter.  To the extent that it solves your clicking issue it also sucks your tone.  It is possible to use one of a number active slow witching schemes to "crossfade" to silence and back, but I'm not the one to tell you about those.

[R.G.]

JFET or bipolar series/shunt switch with controlled on/off ramp time, or VCA with controlled on/off ramp time.

As noted, any instant switching, particularly hard-metal contacts that bounce, will cause a transient at the switch point if the voltage is not zero at that instant.

[sevenisthenumber]

JFET or bipolar series/shunt switch with controlled on/off ramp time, or VCA with controlled on/off ramp time.

As noted, any instant switching, particularly hard-metal contacts that bounce, will cause a transient at the switch point if the voltage is not zero at that instant.

Would you mind a simple schematic of this??

[sevenisthenumber]

Anyone?

[ashcat_lt]

With a cap parallel to the resistor to ground (Rg)?

[R.G.]

Would you mind a simple schematic of this??

See "The Technology of Boss and Ibanez Bypassing" at http://geofex.com/Article_Folders/bosstech.pdf for one way to do series switching. A series switch opens the signal line to stop signal passing.

You can also do shunt switching, as in the previous example. Shunt switching "shorts out" the signal by introducing a low resistance to ground. You can do this with JFETs at ground for DC if you use p-channel JFETs, or odd power supplies, but the simplest way to do it is to use two N-channel JFETs like Boss and Ibanez use for selecting dry and effect signal paths, and have the dry path merely be a big capacitor to ground. This two-FET variation has the advantage of being much more effective at turning off.

The slow-down ramping is done with the resistor/capacitor/diode on the gates of the JFETs.

[Toney]

 This worked for me. Use any sized cap that suits you needs.
 For comedy effect try something massive and listen as the JFET swings over in slo-mo.

 


 

[R.G.]

Yep, that's the series JFET switch from the "Tech of Boss and Ibanez Bypassing".

[pinkjimiphoton]

hey, RG,
do you think that trying this may help with the switching problem i've been having with pnp fuzzes and digitech pedals?

[AudioEcstasy]

I've never found stutter pedals I've tried to pop, couldn't you put a resistor on either side of the switch to help with that?

[R.G.]

I've never found stutter pedals I've tried to pop, couldn't you put a resistor on either side of the switch to help with that?

Resistors as pull downs are not a magic amulet.

The real story is that any switch that causes a sudden change in voltage level when turned on or off will cause a pop or click. Period and end.

If you want clickless switching there are two practical ways I read about over the years. One is to only switch as the signal passes through zero volts, the other is to ramp the signal up and down slowly enough that there isn't any audio-band frequency contant in the resulting signal.

Imagine that you have a SPDT switch with the common terminal connected to the DC blocked input of an amplifier system, with a proper pull down to DC on the outside of the cap, so the amplifier sees 0Vdc as an input if nothing else pulls its input around. The two throws are connected, one to a 1.5V AA battery and one to ground. What happens when you throw the switch?

You get a massive popping thump on each switch transition. When you change from ground to 1.5V, that sudden 1.5V step is amplified. BIG pop, with lots of low frequency thump. The thump is controlled by the lowest frequency the input capacitor to the amplifier will pass, as the capacitor loads up to a DC condition of 1.5V. When the switch flips the other way, it connects the +1.5V level on the input capacitor to ground, and the amplifier sees this as a -1.5V transient, and thump/pops the other way.

Now imagine that you have instead of a 1.5V battery, a sine wave which has + and - 1.5V peaks. If you flip the switch back and forth, sometimes you will catch it at +1.5V, sometimes at 0V, and sometimes at -1.5V, as well as a lot of times in between. The input capacitor and amplifier can't tell the difference between the switch turning on a +1.5V (or -1.5V) battery and the peak of a sine wave at the instant the switch makes. It's only later, as the signal voltage starts dragging the input voltage around that the amplifier's input sees this as a signal, not a step in level. Amplifiers have no memory on this time scale, and only see the instantaneous input voltage.

(N.B. Yes,, to all the hifi reviewers, there are things which cause a non-zero transit time and an internal history, but it's not significant for this discussion. Let the beginners get this down before confusing them with transient intermodulation distortion, slew induced distortion, and other tweako concerns.)

So by flipping the switch, you get clicks and thumps depending on where in the wave you hit. If you happen to hit near zero, the changeover is nearly silent. This is one reason a hard-metal-contact bypass switch cannot ever be consistently silent. If you have a combination of DC and signal at the switch input, you get a combination of the effects.  If you have some way of magico-electronically picking the switch time to be only when the signal is passing through the zero point, then you can make silent transitions, and this is what some digital signal switching/modification circuits do.

The other main way to get silence is to force the transients to be at a signal frequency you can't hear. In practice, this means not suddenly switching, but replacing the switch with a pot. The ends of the pot are signal and ground, the wiper is the input to the amp. When you want to "switch" the imaginary pot shaft is rotated. At some speed, the rate of change of voltage at the input is below the frequency that either (a) the input capacitor can pass into the amplifier or (b) your ears can hear. Typically, ramping switching circuits try to ramp up/down over maybe 1mS to 10mS, that being viewed by a human listener as "instant" switching and also slow enough that the transients are inaudible, or nearly so.

So, if your stutter pedals didn't pop, they were (a) not switching a DC level and (b) switching a small enough signal that the transient was inaudible if they were using hard-contact switches.

hey, RG,
do you think that trying this may help with the switching problem i've been having with pnp fuzzes and digitech pedals?

I haven't forgotten that. Still mulling on it. It's a bit mysterious right now.

[sevenisthenumber]

 This worked for me. Use any sized cap that suits you needs.
 For comedy effect try something massive and listen as the JFET swings over in slo-mo.

 

Would you mind sharing all the values you used?

[R.G.]

Start with resistors at 1M, cap at 0.1. Diode can be almost any silicon, but the common 1N4148 or 1N914 works fine. 2SK30A or 2SK117 was common in the Japanese pedals, but 2N5485 or 2N5292 work well for this as well. Your next question will likely be "I have some J201s. Will those work?". Maybe. The low Vgsoff of these devices makes them a bit funny as switches, but try it.

[Toney]

 If you follow the Vero posted with the schem, it is a working switch based straight from R.G.s drawing, with a few small additions such as drainage resistors on the in and out caps, just in case they pop too. So all the resistors are 1m except the two not-drawn Vbias ones (10K - one from power, one to ground + 22uF cap)
 Caps were just general pass-everything-within-reason 100n's and the one that sets the delay Cslow or C1 on the Vero is up to you. Probably best to socket it and experiment with the delay you need to beat the pop.
 Start with anything and increase or decrease as required. Vero is good to go as is, if you do Vero.
 Oh and I used a 2SK30 but many other Jfets will work fine. Diode was 1n4148.

 EDIT:
 There you go - while I was typing R.G. filled in the blanks. Where's my coffee...

[Processaurus]

When hooking a mechanical switch, rather than a logic signal into the jfet circuit, in practice does C_slow smooth over any switch bounces, or would one possibly need to additionally debounce their momentary switch beforehand to get smooth operation?

[R.G.]

Depends on how bouncy the switch is. Some large mechanical switches can bounce multiple times for maybe 50mS. Small dome style tactile switches may bounce only twice, about a millisecond apart. Some switches may or may not bounce.

Ideally, the gate ramp time would integrate any bouncing out. But there is a contradiction in objectives to have ramp for a few milliseconds but to smooth out 40-50mS of bounce.

I guess that means one would possibly need to debounce ahead of the switching. That implies more electronics ahead of it. I like the 10Fxxx series of PICs - aobut $0.40 each in ones, and it can debounce a couple of switches, properly programmed. The flipflop used in Boss and Ibanez debounces the switch for them.  But the need to debounce is in diametric opposition to the desire for a simple stutter switch.

It may be that this is like so many things: good, fast, cheap, pick any two.

[knutolai]

Have you tried this simple passive design?

[Toney]

When hooking a mechanical switch, rather than a logic signal into the jfet circuit, in practice does C_slow smooth over any switch bounces, or would one possibly need to additionally debounce their momentary switch beforehand to get smooth operation?

 That's exactly why I went down this path and it worked great for me.
 I had a circuit that had a "hair trigger" popping with any mechanical switching despite absolutely no DC on the in/outs good ground etc etc. Set the Jfet switch up on a breadboard with a horrible on-purpose leaky output cap from signal and  a 3PDT. With the Jfet switch, you can exactly tailor the cap size to cover the pop or thump in this case. Starting with a very small cap, you still hear the clipped off end of the noise and going up through the sizes you can choose one to exactly cover the pop.
 Set up as drawn on the Vero, the switch is kinda a hybrid with mechanical switching still covering the connection of the input and output signal for true bypass and the Jfet covering the switching of the circuit . It is also possible to wire it up reverse "Clinton Bypass" style and have no mechanical signal switching at all but with the Jfets closed channel (huge mega-ohm resistor) permanently connected to the output.
 During the switching itself, the signal "sees" a closed Jfet channel that opens at a controlled rate, concealing any switching noise from the output. That's to say you hit the switch and it can bounce and shudder as much as it likes, but the signal and output are both looking at a huge resistance via the Jfet channel that slowly ramps up to a connection.