Popping Kill Switch. Please help

[Irish Luck]

Hey everyone,
I'm fairly new to building pedals. I've made about a dozen. Channel switchers for amps, a feedback looper, and an effects loop bypass. This Killswitch was one of the first that I made, but it makes a popping sound every time I hit the switch. I'm using the Beavis schematic and the right switch and have read multiple forum threads but figured it's time just to ask again based on my trials and problems.

It seems so simple. And it does work. I just can't get rid of the popping sound. I read that adding a resistor between the two lugs of the switch would fix it, but it didn't. I used a 1m resistor and soldered it from one lug to the other.

What am I doing wrong?

Should I solder the resistor from the incoming and outgoing wires from the switch? What size resistor? I saw one website say use a 10m soldered between the lugs. And I just have a 1m. I just don't want to go spend more money on this if it might not correct the problem.

This is the simplest pedal I've made but I just can't seem to nail down the problem. I'm about to give up. I don't want to use it if it pops when it goes silent. That drives me nuts!

Can anyone please give me some help here? I would really appreciate it!!

Thank you!!

[LightSoundGeometry]

I have come to the conclusion its the mechanical pop being amplified at times in certain circuits. I think you can use a 1meg tie down from out put to ground?

or get a momentary switch which is really smooth "soft touch" , I have one mounted on my scratch plate tied to lug 2 and lug 1 of my volume pot:




http://www.mammothelectronics.com/electonic-switches-s/21.htm?searching=Y&sort=7&cat=21&show=30&page=2

[drummer4gc]

Could try this:

http://www.diystompboxes.com/smfforum/index.php?topic=111547.0

[Keppy]

It's inherent to the effect. You are immediately shorting the signal voltage to ground. That immediate voltage change is what you hear as a pop. The only ways to get rid of it are:
1. Switch only as the signal crosses zero volts. This requires non-trivial circuitry.
2. Slow down the voltage change. This also requires non-trivial circuitry (though the link in the post above is pretty simple), or lowpass filtering which alters your tone.

There's a reason this effect is usually used with lots of distortion. The waveform is pretty square, so the instantaneous voltage change doesn't stand out, and there's usually lowpass filtering as well.

One addendum to all that: If you have any DC leaking into the killswitch circuitry from a pedal, the inherent popping might be made worse, so rule that out before you give up or get complicated.

[Rixen]

I have come to the conclusion its the mechanical pop being amplified at times in certain circuits. I think you can use a 1meg tie down from out put to ground?

This is true, I investigated this on another thread and found that there appears to be a piezo electric effect in the switch body, so even perfectly balancing DC offsets won't remove pops and clicks (only minimise).

Even hitting the switch body with a screwdriver generates a voltage on the (open) switch contacts..

[LightSoundGeometry]

I have come to the conclusion its the mechanical pop being amplified at times in certain circuits. I think you can use a 1meg tie down from out put to ground?

This is true, I investigated this on another thread and found that there appears to be a piezo electric effect in the switch body, so even perfectly balancing DC offsets won't remove pops and clicks (only minimise).

Even hitting the switch body with a screwdriver generates a voltage on the (open) switch contacts..

I can hear the PING when a jack is put in or out lol ..the one switch was a little loose and it you jiggled it the sound rang through amp with that same high PING..reminds of a golf ball coming off a club at with high velocity 

I had a lumpys tone shop pedal that did this , it was the 18v zoso boost.

[thermionix]

I've never installed a killswitch in a guitar, but with a Les Paul you can turn one of the volumes down and use the pickup selector toggle this way.  Same thing, it grounds out the signal.  I don't recall ever hearing a pop when doing this.

[Bishop Vogue]

I went through a similar frustration with lots of popping problems.  The 1M pulldown didn't work for me.  Eventually I built a killswitch using an LED and LDR.  It's the best design I have been able to make for this kind of unit.

[ashcat_lt]

I've never installed a killswitch in a guitar, but with a Les Paul you can turn one of the volumes down and use the pickup selector toggle this way.  Same thing, it grounds out the signal.  I don't recall ever hearing a pop when doing this.

Then you were lucky, or using quite enough gain, or whatever you were playing masked the pop, or you just weren't paying attention because you weren't expecting it and tweaking out about a relatively minor issue.    This was a pretty common question over at the GuitarNutz forum for a little while there.

Keppy is right.  If it ain't one thing it's another, and there's no good way to be absolute sure to avoid all of them with passive circuitry.  It seems like a huge shame to have to add active components for such a seeming simple thing, but...

[Rixen]

If you really want a passive solution and are mechanically minded you could use a force sensing resistor as part of a voltage divider: http://www.interlinkelectronics.com/datasheets/Datasheet_FSR.pdf,

..although it won't go to zero (about 1k)..

I don't think there would be much pop if any with this due to the gradual change in resistance.

[Bishop Vogue]

I forgot to add the link to the sound demo on Youtube.  Here it is: https://www.youtube.com/watch?v=2FRKYf8kwDs

[rutabaga bob]

Nice!  Any spec on the ldr?

Edit: Welcome to the forum, Irish Luck!

[rutabaga bob]

Anybody want to take a guess?   

[R.G.]

It is really hard to put "specs" to LDRs because what they have is variability. They do have certain specs, but the specs are in terms of minimum resistance and maximum resistance, speed in changing resistance, resistance curve (the light-to-resistance relation is not strictly linear), light sensitivity and both light and dark adaptation.

Your problem is that you want the minimum resistance to be zero, the maximum to be infinity, speed to be several tens of milliseconds (for this application), a linear light to resistance curve, and no light or dark adaptation. Obviously, there aren't any of those.

So the specs come down to what you can stand. For a series application like the example circuit shown, you want the "off resistance" to be quite big, even compared to the 1M or more input resistance of an amplifier or pedal following it. 10M dark resistance would be a minimum, before dark adaptation. If you use the LED as a shunt to ground - my favorite for this setup - you want dark resistance of a meg or more and light resistance of as low as possible, but several hundred ohms to several K would be all right. Time, something in the 10mS to 100mS range for dark-to-light speed would be good. If it's instant, you have the same problem as with metal switches. If it's longer than 100mS, you don't have much stutter, it's a fade in/out.

Of course, LDRs are so highly variable that you may get lucky. I'd say get one, try it, then see if it works good enough.

By the way, there is (or was) a commercial device that used an LDR in series with a guitar signal, built into a guitar cord plug. It had a hole open to exterior light, and one did "swels" by blocking the opening with a finger. They claimed owning a patent, but this seems like one of those things that are too obvious to be patented. But what do I know.

If I were doing this, and I had already made it over the hurdle of using a battery or power supply to run it, I'd use a shunt JFET, not an LDR. This can obviously be done quietly, Boss and Ibanez do it in all their pedals, and it can be much lower power than running an LED. The battery might well last its entire shelf life.

[blackieNYC]

Along the lines of the clickless killswitch, I'm thinking about a clickless momentary fx loop insert.  For a very smooth, brief effect, then right back to the first sound again.
If the LDRs in the clickless killswitch thread works best in that application, perhaps I'd better plan on having to use them.
A DPDT needs to be created for TB loop insertion.  With LDRs, that could be done to some extent with 4 LDRs. So a Bypass LED makes two "switches" go one way, and the reverse-state Loop Insert LED would make the two switches go the other way. Darkness and lightness extreme ratings would probably work better I'm sure.

Update, post-RG: Or FETS like in Boss pedals? There's a DPDT right there. That seems a little cleaner.
   Would I need a buffer before and after the bypass FETS and before and after the effect-on FETS, like I'm seeing in the boss schematics?  They seem to kick on and off quite smoothly - is that the cap charging at the switch input to the two-transistor switching circuit?  I'm going to find a description of that circuit. 

[rutabaga bob]

Is this what you have in mind, RG?

[R.G.]

Is this what you have in mind, RG?

Pretty much, except that I'd use a P-channel JFET like the J175 or J176, and make Vb be 0V. Diode d1 would need reversed too.

That reverses the sense of the switch as well.

J201 is not a good choice for switching, as its very, very low Vgsoff makes it have distortion issues with smaller signals. The Boss and Ibanez pedals use JFETs with Vgsoff of about -3 to -4V, probably because they use a bias voltage of half the 9V battery to turn it off.

Your battery voltage only needs to be bigger than Vgsoff plus the biggest signal you have to block. You need the off-bias to be large enough that they signal voltage never pulls the channel out of cutoff.

If the LDRs in the clickless killswitch thread works best in that application, perhaps I'd better plan on having to use them.

Nothing is problem-free. LDRs are hard to find, illegal because of their cadmium content in the EU, and hard to specify. The

Or FETS like in Boss pedals? There's a DPDT right there. That seems a little cleaner.

Well, it's a buffered bypass. See "The Technology of Bypass Switching" at geofex. It's not quite a DPDT.

Would I need a buffer before and after the bypass FETS and before and after the effect-on FETS, like I'm seeing in the boss schematics?

It depends on the unseen elements in circuits - whatever is before the circuit that drives it with an input signal, and whatever is after it, loading it down. If the ends of a circuit terminate in a jack, people can and will insert anything that fits the jack, plus anything that can be forced into it. Using buffers just makes sure you know what you're getting. Some things are fine without them. The FET will simply change from a giga-ohm resistor to a few hundred ohms in any case.

They seem to kick on and off quite smoothly - is that the cap charging at the switch input to the two-transistor switching circuit?

Well, it's the time constant of the R-C on the gate circuit compared to the Vgsoff of the JFET and the V+ driving the gate circuit. But yes, the cap is the slow-down item.

I'm going to find a description of that circuit. 

You probably are. One good one (IMHO) is in "The Technology of Boss and Ibanez Switching" at geofex.com. Here:
http://www.geofex.com/Article_Folders/bosstech.pdf

[rutabaga bob]

For a dummy like me, if the bias is 0, what do the 1 meg resistors (Vb) connect to?  The +9v at the switch is okay?        Thanks!        Larry

[R.G.]

For a dummy like me, if the bias is 0, what do the 1 meg resistors (Vb) connect to?  The +9v at the switch is okay?        Thanks!        Larry

Change "dummy" to "beginner".

The 1M resistors connect to Vb - which in this case is signal ground.  +9V at the switch is OK, because that voltage is what pulls the P-channel gate high enough to turn the P-channel JFET off. The negative side of the 9V source must connect to 0V / signal ground for it to work.

I picked a P-channel JFET because it functions perfectly well with its source and drain at 0V / signal ground and its gate pulled to +V, the opposite of N-channel JFETs, which must have their gates pulled negative to turn them off.  That means that an existing +9V source will work for the gate supply, and a -9V source is not needed like an N-channel JFET would need. That's the next question that will pop up here if I don't include this explanation.

[rutabaga bob]

It won't pop up from me!  Thanks, R.G.!