Ok, I'm fed up- anyone have a good electronic switching design?

[R.G.]

One question though: I'd be using the JFETs as relays - kind of like in your 'programmable effects loop' article. This means that the impedance would drop to a few hundred ohms or less if any 1 effect was engaged. Would this cause any noticeable signal degradation, or is it only the amp's imput impedance that we have to take into account?

I must be having a senior moment - I can't translate that. Can  you amplify on it?

[tempus]

I may be looking at this the wrong way, but here's what I'm wondering. The guitar output Z is say 10K, and looks into a load of about 1 meg. I plan to use the JFETs to switch in or out 4 different effects. With everything bypassed, there will be (4x60 ohms) 240 ohms added to the guitar output Z, which is a tiny percentage. However, as soon as I switch one of the pedals into the loop (say the 1st one) the output Z will drop to around 100 ohms (roughly the output Z of the pedal). Now I still have 3 more JFETs to go through, so there's 180 ohms added to the 100 ohms, which is  a huge percentage. Does that make sense, or is it strictly the input Z of the amp that I need to worry about when adding these resistances to the signal?

Thanks again

[R.G.]

I may be looking at this the wrong way, but here's what I'm wondering. The guitar output Z is say 10K,

It's actually something like 8K to 20K resistive in series with a one to four henry inductor, shunted by a variable amount of capacitance. That inductance is what is responsible for much of our misery because it's impedance ramps up steadily across the audio frequency range, forcing us to use big (e.g. 1M) input impedances to not lose treble.

and looks into a load of about 1 meg. I plan to use the JFETs to switch in or out 4 different effects. With everything bypassed, there will be (4x60 ohms) 240 ohms added to the guitar output Z, which is a tiny percentage.

So far so good.

However, as soon as I switch one of the pedals into the loop (say the 1st one) the output Z will drop to around 100 ohms (roughly the output Z of the pedal). Now I still have 3 more JFETs to go through, so there's 180 ohms added to the 100 ohms, which is  a huge percentage. Does that make sense, or is it strictly the input Z of the amp that I need to worry about when adding these resistances to the signal?

And now we're down at the root of the issue. Impedances pretty much only matter when you have two of them. Talking about the impedance/resistance of your switching setup makes no sense unless you talk about the relation of it to your guitar or the amp it feeds.

There are three different transfer efficiencies we look at in electronics. These are voltage transfer, current transfer, and power transfer. You need equal source and load impedances for maximum power transfer. However, in audio we are almost always concerned with maximum voltage transfer. We want the voltage from the source to not be loaded down by the load. This is why I did the parable of the inductor first. We have to make amps with input impedance of ten or more times the impedance of the guitar at its highest point to avoid losing signal. It's that voltage divider thing again - you look at the source impedance, which looks like it's in series with the source, and the load impedance, which looks like it's parallel to the load, and that gives you how much of the signal voltage you lose in the impedances before it gets to the load.

Generally a loss of less than 10% is OK, if not great, which is the origin of the 10:1 rule. You want the load impedance to be greater than ten times the source impedance. The amp has an input impedance of about 1M in most cases because it's designed for sticking a guitar directly into it. So do your effects.

The output impedance of every effect is generally low, below about 10K. Sticking a few hundred ohms in series with a 100K (guitar) or 10K (effect) makes no material difference at all compared to the 1M load of the next effect or the amp.

[tempus]

OK thanks for that R.G. I knew a lot of that already, but I was getting a little muddled with the different fx impedances thrown into the mix. Some of the JFETs I've been looking at seem to exchange a low R_on for a high V_gs. It seems that for our purposes we would prefer the low V_gs over the low R_on, so that the effect is switched completely out of the loop. A J105 for instance has a nice low 3 ohm R_on (which as it turns out isn't that desirable anyway), but suffers from a fairly high 10v V_gs. Its relative, the J107, however, has a respectable 8 ohm R_on and a more usable 4.5v V_gs. So for the purpose of stomp box switching, the J107 (or some other JFET with an even lower V_gs) would be more suitable.

Am I on the right track?

[Processaurus]

http://experimentalistsanonymous.com/diy/Schematics/Phasers/Danelectro%20Pepperoni%20Phaser.gif

Danelectro does it the same way as the GEO article, they buffer the signal once though.  I had a Fab series pedal (metal) apart today and they used a discrete buffer rather than opamp.

EDIT- spoke too soon, my fab metal clicks a little just when it's turning on (but is dead silent when switching off, hmmm), I think because the LED is sucking so much current suddenly.  Could probably fix it by making a decoupled power supply for just the LED and or not making it so dang bright.

[tempus]

Hmmm. I think mine clicks more turning on than off as well. You mentioned the buffer - might that have some effect on whether or not it clicks when switching?

Thanks