Question about Ibanez discrete flip-flop switching

[jrc4558]

Hi all! It's not a DIY build but a pedal I dreamed of having for a while, after fixing one a number of years ago for a friend. The Ibanez Mostortion, MT-10. Its a fine pedal with very soft and dynamic overdrive characteristics. I got it NOS at a store that apparently bought it new back in the day, saw that it has switching problems, and left it for "dead".
Pedal works excellent in every respect but switching. When the switch is actuated, the switching circuit changes state in ~30% of the time. In about 20% nothing happens, and the rest - intermittent on-off, as if its a fast tremolo between the states. The switch itself is in bad shape, but the problem persisted when I tried simply shorting the contacts.
Where would one start looking for trouble in such case?
Here's a schematic for reference: http://www.schematicheaven.com/effects/ibanez_mt10_mostortion_dist.pdf

Thank you all!

[R.G.]

(a) C21, possibly other polarized caps going bad. Replace every polarized cap.
If that doesn't fix it:
(b) measure all of the component values in the flipflop, correct any that are more than 10% off
if that doesn't fix it:
(c) replace transistors and diodes in the flipflop section.

[zachomega]

Also the switches Ibanez used weren't the best.  However, any momentary switch will work.  Just wire something up to work as a temporary fix.  It'll at least let you know if the switch is bad. 

-Zach

[Mark Hammer]

I do not know if it is the same on the Ibanez pedals, but I know my old MXR Commande Series Flanger has a little plastic nubbie thing on the actuator that is pushed onto the switch.  It has seen its best days behind it, and switching can be spotty at times.

[jrc4558]

Yep Mark, mine has the nubbie. Its the flange that's broke off, therefore th switch doesn't hold in the opening. Glue gould fix that, but seems to me like the problem is in the flip-flop. There are no electrolytics in the switching circuit. So it's either the transistors that don't hold the state, or I don't yet know what.
Do oyu gentlement know anywhere a good reading on the discrete flip-flop subject?

[plankspank]

Do oyu gentlement know anywhere a good reading on the discrete flip-flop subject?

Read R.G.'s article "technology of the tube screamer" on GEOFEX. Excellent info on the switching flip flop. I seem to recall I fixed a friends T.S. which had similar problems: it had a bad ground connection at the ckt bd. / enclosure point, and also a verticle mounted switching diode lead  was shorting out on another component...

[jrc4558]

read R.G.'s article. Clear, interesting, informative. Now to find something on why the flip flop works the way it does.

[R.G.]

The flipflop is two inverters in series, with feedback from output to input. With the feedback always positive, it locks up in one position or another. The way it's flopped is to feed a turn-off pulse into both bases from the momentary switch. The already-off transistor can't turn any more off, but the already-on-one can. As it starts to turn off in response, its collector output feeds an "on" signal to the transistor that was off, which reinforces the turning off, so the transistors flip between being on and off. The circuit is a standard one, covered in all beginning logic circuits courses - if that's taught any more.

As to why it doesn't flop properly. A bad switch is always possible. I assumed you'd already taken care of the obvious, and I should not have. The switch is the problem most of the time. If it's not the switch, like I said:

(a) C21, possibly other polarized caps going bad. Replace every polarized cap. C21 could make for a marginal trigger pulse.
If that doesn't fix it:
(b) measure all of the component values in the flipflop, correct any that are more than 10% off; components could have drifted, lowering the gain on one or both inverters, until the remaining gain does not sustain latching.
if that doesn't fix it:
(c) replace transistors and diodes in the flipflop section, as they could have drifted or been damaged and not have enough gain left to latch up.

One I should have noted is that the external loads on the collectors of the two transistors could be too heavy for some reason.

This kind of flip flop relies on the forward gain of the two inverters being greater than one, even with output loading and circuit variations taken into account.