Why the Clean Channel goes quiet with the Footswitch connected
Why you get some clean boost with f/s but with the knob pushed in

How the Effect Switching Works; Reverb

Without Footswitch
V2 drives the reverb transformer, which wiggles the reverb springs. The motion of the springs is picked up and amplified by half of V3, which passes this pure reverb signal to the reverb control. That sends the chosen amount of pure reverb via 1.2Mohm resistor into the main signal path at the input to V3b.

With Footswitch
Same as above, except the pure reverb signal is also sent all the way out to the footswitch. "F/s reverb off" switches the pure reverb signal direct to ground (actually the f/s casing) - a zero resistance compared with  the high (1.2M) resistance it would have to overcome in order to get into the main signal path; so you don't hear any reverb. "F/s reverb on" opens the f/s reverb switch, so the ground path isn't there and the pure reverb signal behaves as though the f/s wasn't there at all, ie, it heads for the main signal path through the 1.2M resistor. In this setting the f/s also allows the -6.2V supply, through an LED and a 390ohm resistor, to go down to ground (all inside the f/s), so you get an LED indication that reverb is on.

How the Effect Switching Works; Lead

The overdriven, distorted sound is made by taking a huge, boosted signal from V2 via the input (primary) side of the reverb transformer and sending it to the main signal path at V3b, thus distorting V3b and everything after it. But how does that big signal get there?

For this you have to understand how the optocouplers work. There are 2 identical optocouplers, each one shown on the schematic like this. I'll label the 4 terminals a,b,+ and -.

Inside the optocoupler, a is not connected to + or - at any time. Also b is not connected to + or - at any time.

When there's a positive DC voltage from + to  - , a connects to b by a low resistance. It's not quite a perfect zero resistance (like a good mechanical switch) but it's close. Let's call this "on".

When + and/or - is disconnected, the resistance from a to b goes very high. It's not infinite (as in a good mechanical switch) but it's high enough to be called "off".

In the PRII these optos are connected with - permanently wired to -6.2V. That's 6.2V BELOW ground level. So connecting  + to ground will turn the opto "on" because + will be 6.2V ABOVE  - . (The -6.2V is also used to power the LEDs in the footswitch; it comes from a 6.2V zener which is connected to the bias supply circuit for the power valves/tubes.)

I'm going to call the opto drawn near V3a "OC1" and the other opto  "OC2".

Without Footswitch
The 'pull for lead' knob closes 2 mechanical switches on the back side of the volume knob. The schematic is drawn with these switches 'on', that is, the volume knob is pulled out. One switch is drawn near OC1 and when 'on' it allows the boosted signal to go straight to the input of V3b via a 220K resistor. The other switch is drawn near the red f/s socket. When 'on' it connects OC2's + to ground, via the ring terminal of the red socket, turning OC2 on. This connects the non-signal end of the lead level pot to ground via OC2's a and b, thus allowing the pot to act like a pot  - so it becomes a volume control for the whole amp in conjunction with the master volume.

When 'pull for lead' isn't pulled (ie you want the clean sound), both switches are open. The boosted signal never gets into the signal path. OC2 is off, so the bottom end of the lead level pot goes nowhere - it just 'floats'. Therefore it can't act as a volume control and its setting doesn't matter.

Without the footswitch OC1 is always off and so never affects the signal routing.

With Footswitch
For operation according to the manual, "Pull for lead" should always be pulled when the f/s is connected, so both pull-switches in that control are on. When the f/s red plug is connected, this operates an extra switch in the red socket so that the ring terminal isn't always grounded. This means OC2's + can only be grounded via the f/s 'lead' switch.

With the f/s set for 'clean', OC2's + isn't grounded, so OC2 is off and the lead level control is out of circuit as described above. Meanwhile OC1 is turned on by having its + grounded by the lead switch in the f/s. Therefore the boosted signal goes down to ground via OC1's a and b. So you don't hear the boosted signal and the lead level control has no action.

With the f/s set for 'lead', OC2's + is grounded via the lead switch in the f/s, so the lead level control is in operation. OC1's + is disconnected from ground via the lead switch in the f/s, so OC1 is off and the boost signal heads up to V3b instead of down to ground. Also the lead switch in the f/s grounds another f/s LED, which is fed -6.2V via another 390ohm resistor, so you get an LED showing lead is on.

To summarise the footswitch action (if you're following this on the schematic, remember the pull-switches are 'on', which is the way the schematic is drawn, and the grounding-switch built into the red socket is 'off' because a jack plug is in place)

Why the Clean Channel goes quiet with the Footswitch connected

Problem
The clean sound is much louder and fuller without the footswitch. When the footswitch is connected, the clean sound is quieter and thinner.

Reason (1) my original theory... but see below for reason 2, new Sep 2017
In clean mode without the f/s, the pull-switch is open (so no lead sound gets into the signal path) and OC1 is off, so there's no path to ground for either the clean or lead sound.
In clean mode WITH the f/s, the pull-switch is closed (allowing lead sound to go further into the circuit) and OC1 is on, sending the lead sound straight to ground so it still doesn't get into the signal path. My theory is, if OC1's on-resistance is too high (in a perfect world it would be zero) therefore not all of the 'lead' signal is shorted to ground. Therefore, it arrives at the input to V3b at low level (ie not enough to distort V3b, so the sound you hear remains clean). The clean signal (the signal we want, ie the signal that's come from the tone stack, not the reverb stage) also arrives at this point. Here's the problem; because the reverb driver inverts what is sent to it, the two signals are out of phase, making the end-result weak. (The reverb driver inverts that signal.)

Reason (2) as offered by Ben on the TDPRI forum - more convincing than my own idea!

See post #5 on this page. I have not tried this yet but it makes sense to me - if it works, you only need to add a 1M resistor. 

Why this problem is bigger in some PRIIs than others
Several factors affect this problem;

1. the gain of V2
2. the 'on' resistance of OC1
3. the input impedance of V3b

Why you get some clean boost with f/s but with the knob pushed in

This is the 'Andrew W of Canada' operation mode described on the main footswitch page. Connect the footswitch, select lead on the f/s,  DON'T pull the volume knob, and on some models you get a slightly fatter clean sound. I have been pondering why this should be.

The two modes in question are
(a) knob in, f/s connected, lead selected on f/s
(b) knob in, f/s not connected

(a) is a little louder and has a mid boost compared to (b) . Why?

I get the same effect with V2 removed. V2 is the source of the lead boost; therefore the boost is nothing to do with the extra gain introduced by V2. What's left which could make a difference?

If you compare what's going on between these two modes...

(a) has OC1 off and OC2 on. One f/s LED is on.
(b) has OC1 and OC2 both off. No LEDs are on.

I  don't see what possible effect this has on the -6.2V rail, and even if it did vary a little I don't see what difference that would make.

The only thing that's left is the action of OC2.  In (a) mode, OC2 is on, grounding the bottom end of the lead level pot and the 0.003 capacitor in parallel with it. In (b) mode, the OC2 is off, allowing the bottom end of the lead level pot to float and passing the signal to V4 through part of the lead level pot, and through the 0.003 capacitor which is in parallel with it.

I think the 0.003 cap is the thing making the difference - it's either in the signal  path  (mode b, normal clean no f/s) or it's a path to ground (mode a, gives some mid boost). Is this a phasing thing - boosting volume by grounding some frequencies which would otherwise reduce volume by a phase-cancelling effect? I'd be grateful for any comments on the logic (or otherwise) of this analysis