Wicked switches

[Gringo]

Since i have a couple of projects ready to box and no mecanical switches, i decided to finally bite the bullet and go un-true bypass :evil:

So I set my hopes in Andrew's 4066's wicked switches

I used a momentary switch, and controlled the 4066 with a 4049UB. Worked right off the bat, and the board i drew was less than half the one i did for RG's 4053 based switch, BUT... there's distortion in both paths (bypassed and processed).

In RG's switching article, i quote:

the analog inputs and outputs need to be held somewhere near the middle of their power supply. The 4053 does have a "Vee" supply pin that lets it control signals near 0V, but that is often a fair amount of trouble to use. the simplest thing to do is to AC couple the in/out pins and bias them to the middle of the 4053's power supply.
...
Yes, this is a pain, eats up board space, and is complicated. It works, though. When biased this way, the CD4053 has little or no distortion and no perceptible switching pops - exactly what we want.

I'd like to know if the distortion i'm experimenting is something to be expected and i NEED to bias and decouple the 4066's I/O, or i just screwed up my build in some way.

Schematic used:

http://gr1ng0.tripod.com/Downloads/mom_switch.zip

edit: go with the usual "copy link-paste in new window" to get the file.

[gez]

There shouldn't be any distortion.  You need to bias each end of each switch at around half supply.  Your schematic doesn't show the switching in relation to the effect, i.e. how each switch is actually wired to the effect, so it doesn't help much.

[Gringo]

There shouldn't be any distortion.  You need to bias each end of each switch at around half supply.

You say i should not have any distortion only IF i bias the ends of each switch, am i right?

Your schematic doesn't show the switching in relation to the effect, i.e. how each switch is actually wired to the effect, so it doesn't help much.

The distortion happens also in bypass mode, so i would think it has no relation to the fx being bypassed.  :?

Anyhow, i wired it the way Andrew shows in his "wicked switches" article, guitar to "In" tag, amp to "Out" tag, the input of the fx to its matching tag, same thing for the output.

Is there anymore data i should be providing?

[gez]

Your schematic doesn't show the switching in relation to the effect, i.e. how each switch is actually wired to the effect, so it doesn't help much.

The distortion happens also in bypass mode, so i would think it has no relation to the fx being bypassed.

Slight misunderstanding.  Just saying 'FX in' or whatever doesn't tell us anything.  It doesn't show how you've wired up the switches to the actual circuit (as in this output cap goes to this end of the switch, there's a resistor here, blah blah blah).

Have a read of RG's article again.  The ends of each switch need to be biased to half supply (I'm assuming you haven't done this).  This is usually done with a resistor at each end to Vbias, and the switches are AC coupled.  Take a look at some schematics of commercial FX and you should get the idea.

Yes, biasing the switches should cure the distortion.

[gez]

PS  You do have an output buffer at the end of all this don't you?

[The Tone God]

Anyhow, i wired it the way Andrew shows in his "wicked switches" article, guitar to "In" tag, amp to "Out" tag, the input of the fx to its matching tag, same thing for the output.

The in and out tags can be applied to a very wide varity of uses other then audio switching like logic switching. To have that flexabilty I left off the biasing circuitry letting the user decided if it is need or not.

There may or may not be the need for the biasing suggested. It depends on your circuit. Most enginners would insist on it to be safe. I usually do those things on the effect boards I layout so as to keep the switching simple. Alot of the time one can sneak by without it.

Is there anymore data i should be providing?

Yes. What voltage are you running the switch at ? How are you attaching it to the power supply ? What effect are you hooking this up to ? What are the ICs full labeled names (i.e. HCF4049, CD4066, etc.) ? Are you using an appropritate LED current limiting resistor ? Anything else you can think off ?

Andrew

[donald stringer]

Hey gringo  If you dont already have this up and running,Iwould advise you to download the data sheet on the 4066 and breadboard to figure it this thing out.Familiarize yourself with this cmos. There are a number of ways to wire it up. I went through a similiar headache myself and after I understood the path the signal takes through the chip thats when the light came on. It turned out to be some wires where crossed up etc. I was pleasantly surprised to find out just how few components it takes to get this thing to work. And a breadboard is a step in the right direction especialy when your working with something new. I have built this particullar cmos into several effects with great success, even using my own limited knowledg of things related to electronics. ...Ther should be no dist. at all. Mine operated with no noise  and with minamal clicks. Let us know what  you discover.

[bioroids]

Hi

you must decouple In and Out with capacitors (I assume you connect the guitar at In and the amp at Out) (1 uf should be fine). Then you provide bias at half the power supply (with a voltage divider with cap to gnd and 1Meg resistors). There should be no distortion this way.

Luck

Hola Gringo
se entendio lo que escribi en ingles? Si no, avisa que trato de explicar mejor en castellano.

Suerte

Miguel

PD: si conseguis que ande conta que tal es el nivel de ruido (me refiero a si hace pop o click cuando switcheas)

[bioroids]

I forgot, also you have to bias and decouple FX In and FX Out. If the effect input and output already have a capacitor, then you can ommit them here.

Luck

Miguel

[gez]

you must decouple In and Out with capacitors (I assume you connect the guitar at In and the amp at Out) (1 uf should be fine). Then you provide bias at half the power supply (with a voltage divider with cap to gnd and 1Meg resistors). There should be no distortion this way.

If there are no significant DC offsets either side of the switch (i.e. the input buffer, FX out and output buffer are all biased at half supply), and you use FETs and turn them on slowly, it's possible to do away with all the caps and resistors.  

I've done A/B testing and there's no increase in switching noise.  Although FETs distort more and can offer less headroom (depends how you do things), I've found switching to be quieter with them.

[Gringo]

First, thanks for all the replies!

What voltage are you running the switch at ? How are you attaching it to the power supply ? What effect are you hooking this up to ? What are the ICs full labeled names (i.e. HCF4049, CD4066, etc.) ? Are you using an appropritate LED current limiting resistor ?

- I'm using a regulated PS at 9.1v
- The switch is hooked to the ps at the same point as the fx.
- This is the fx i'm using: http://pages.infinit.net/cbriere/userfiles/hipwah.PDF
I'm not using SW2, of course. It's input wired directly to the schem's tag "FxIn", same thing with the output to "FxOut". No coupling caps, no biasing the inputs.
- The ICs are TI's cd4066BE and cd4049UBE
- The current limiting resistor value for the led is 4k7 (as shown in schematic)

PS You do have an output buffer at the end of all this don't you?

No. I tried with an input buffer, same results. I'll give it a try at the output.

Che Miguel, gracias por la data, anduve por tu pagina tambien, recopilando info. Quedate tranquilo que se te entiende perfecto. Me tire a armar este justamente para zafar de todo el parterio que usa el de rg, y ahora resulta que tiene que ir si o si. Eso me pasa por rata, si el pibe los puso es porque van  

I'll follow your directions, and report with whatever results i get.

[David]

Gringo:

For what it's worth, I breadboarded a "Wicked Switch" a couple of years ago.  As I recall, I didn't have any noise problems with mine.  Power was supplied by a 9V battery.

[puretube]

well, the wah`s output is referenced to ground - the input got a floating cap...  :roll: