FX loop In and Out impedances

[GainFreak]

Hi,

I have some theory blanks that I would need to fill so here's the situation. This is the Soldano Decatone FX loop:



In order to get a line level the Send output level is decreased by the R37/R38 combo. From what I know the output impedance of the stage is R37 in parallel with R38 which is ~2k. My first question is: how much is the Return input impedance? Is it

a/ 470k

or

b/ 150k in parallel with 470k?

or

c/ something else?

Now if I want to set up a Send level pot like on the next pic and assuming the R40 (150k) resistor is not there the 2k2 pot will connect in parallel with R42 (470k) effectively making the stage impedance ~2k? Is that right? Then what would be the impedance if R40 (150k) is put in place?



If my assumption about the 2k input impedance is right what should be done to keep the Send level pot and have the Return stage impedance intact? Should I add a big capacitor in front of the R40 (150k) or ?

I would appreciate if anyone could explain those impedance issues for me.

[Mick Bailey]

The return impedance is 150k in series with 470k. The impedance of the triode grid is so high it can be ignored. The grid is connected to the junction of the two resistors, which form a voltage divider to reduce the voltage the grid 'sees' from the return signal.

The simplest way to implement a send level pot is to replace R38 with a pot and connect C23 to the wiper. This way the integrity of the loop is maintained. With the loop disconnected the pot can be turned right up and the circuit behaves exactly the same as drawn in the first diagram.

[Chuck H]

What Mick said. And to add (sort of) the output impedance of the loop would be roughly the value of the dropper, or 2.2k. It's actually a bit lower because it's being run from an unbypassed cathode so there is generous local NFB. I think R37 (33k) spoils that. So without analyzing too much I'd guess 2k-ish. But it doesn't need to be that low. True that lower is typically better WRT effects loops, but I haven't noticed any problems running loops as high as 22k. Though going that high is a bad idea if you plan to make really long cable runs in the loop. But I mention this because you may (or may not) like to have more drive option for your adjustable loop. The fixed 2.2k dropper value would have been chosen to keep the loop from misbehaving with maximum signal. In other words, it's NOT chosen to give optimal output at lower preamp levels. By increasing the output potential of the loop send you can better optimize for noise and dry/wet mix at more variable preamp settings. I'd simply replace R38 with a 5k pot. Like this:

[GainFreak]

Thanks for your replies.
If R38 is replaced with a pot it will have some DC on it. Wouldn't that result in a scratchy pot or something?

Still in theory if the voltage divider after the CF is as per my second picture and connects directly to the 150k resistor what be the resulting impedance and how it would affect sound if at all?

[J M Fahey]

Yes, it will probably get scratchy.
Since you have a lot of signal (you are attenuating it, after all ) and it comes from a low impedance source, you pay no penalty for adding an extra resistor+pot string and DC insulating it with a coupling cap, like was shown on an earlier example.
And agree that raising pot value to, say, 10K , will still work very well and simplify things a little (as in you have far more choices in 10K than in 2 K).

[g1]

Sorry to nitpick, but a word about your drawings. mF is a legitimate designator for milli-farads. For micro-farads, uF is used.
A milli-farad is 1000 uF.
Most people will assume you mean uF. But with the increase in modern capacitance capability, mF is becoming more widely used terminology. So massive 47,000uf caps can be described as 47mf.
I've been caught up in the error when looking at specs. for capacitance measurement on multimeters.

[Chuck H]

If R38 is replaced with a pot it will have some DC on it. Wouldn't that result in a scratchy pot or something?

Agree. But IMHE when wired this way there is no path for DC through the CT of the pot, so it doesn't scratch. Not only that but "I" don't usually make the adjustment here WHILE playing a tune. Maybe I'm too cavalier. If it bothers you just replace R38 with a 10k resistor and put a 10k pot after the capacitor.

[GainFreak]

Sorry to nitpick, but a word about your drawings.

Note taken but I didn't draw these schematics.

Not only that but "I" don't usually make the adjustment here WHILE playing a tune.

You're right and I'm sure all suggestions so far will work but the main reason I asked those questions is because I'm adding a noise gate ENGL style to a new preamp build. You know that shhhhhhhhhhhh coming from those high gain preamps.
The NG "driver" part takes signal from the tone stack and the "gate" part is a J175 (could be an LDR as well) in parallel with the smaller resistor of the voltage divider in my second picture. The attenuation you get this way is 33k/~80Ohms which cuts most of that noise without influencing the sound and I don't think that JFET can be inserted in parallel with R38. That's why I asked what happens with the Return stage impedance when no effects are present and that 2k2 resistor connects in parallel a/ with the 470k alone or b/ when 150k is there as well.

[J M Fahey]

when wired this way there is no path for DC through the CT of the pot, so it doesn't scratch.

No need for such DC path to scratch.

Think about it: when moving the pot slider, you are going from one point along a resistive track to a different one.

If there is DC applied, said points are at a different voltage.

If slider contact is *perfect*, you will have a "slowly" varying DC level, which means there is an "AC component" which *will* go through that coupling capacitor.

Of course, no big deal because that "slow" movement (your hand rotating the know) implies a very low frequency (sub Hz), which will be hugely attenuated, besides being inaudible.

But if track surface is not perfect (or you have the slightest dust deposit on it), you will *jump* from one "fixed voltage along that track" to a different one.

That jump is scratch.

[Chuck H]

Yup. The old Marshall style presence circuit sounds scratchy. The cap is wired from the pot CT to ground to bypass HF from the FB loop. I had this discussion on another thread years ago and insisted that such a circuit will always scratch and was poo pooed. I build my presence control with the cap between the FB node lug and the CT of the pot and they don't scratch. I don't know why. Then I wondered if perhaps all those old Marshall caps are leaky, but reissues scratch too!?! Then I thought it might have something to do with the cap being held at a steady potential on one end (like the Marshall circuit with one end of the cap grounded). I dunno really, but I've made a half dozen amps with presence controls and none of them scratch.

And I still stand by my observation that unless you're adjusting the effects loop send in the middle of a performance it shouldn't matter. There certainly isn't enough voltage or current there to be dangerous to the pot or anyone turning it.