Boss GE7 bypass problem

[HeaD]

Hi, I have a problem with my boss ge-7:   It works correctly, but sometimes bypass does not work. I thought that the problem was the switch, and I replaced it with another one that i bought from smallbear, but the problem stays. Can it be fault of some damaged component in the electronic-bypass system? I've found this schematics http://img267.imageshack.us/my.php?image=ge7pg28oc.jpg , but I don't know how the boss switching system works, what components should I replace according to you? Thanks!

[R.G.]

Here's how it works:

Q2 and Q3 form a toggle flipflop that flips once with each press of the footswitch - or should. Either Q2 or Q3 but not both is always on, the other is always off. Notice that the collector of the "on" transistor turns the "off" transistor off, and the collector of the "off" transistor turns the "on" transistor on. When the footswitch is pressed, it couples a little spike of "off" to both bases. The already off base literally could not care less, it's already off, but the "on" transistor gets that little blip of "off" and starts turning off a bit. This is amplified through it and starts turning the "off" transitor on a little, which amplifies THAT and turns the "on" transistor more off. And it flips. The next press of the footswitch does the same thing, but now the roles of the two transistors are reversed.

The collectors of Q2 and Q3 are tied through 1M resistors and diodes to the gates of Q4 and Q5. Q4 lets through the effected sound, Q5 lets through the buffered input. JFETs are normally on, and pulling their gate to ground through the 1M turns them off, as their sources and drains are held above ground by the bias voltage. So the collector of Q3 turns off Q5, letting effected sound through Q4. The collector of Q2 turns off Q4 and lets buffered input sound through Q5.

If your unit is not switching reliably, the problem is somewhere around Q2 and Q3, possibly the caps that couple in the "flip" pulse.

[HeaD]

Thanks for all R.G.!