set-up trouble with 4093 nand to 4013 flip flop ("Wicked Switches" related)

[any]

Hi All,
I'm working on a massive stompbox delay/sampler combo that I've been wanting to build for quite a while
and I've finally got my head around what I want. (a digitech 20/20 multiplay and a ISD1420 based loop-recorder in one box).
I'm using a nand gate to trigger a flip flop, which in turn controls a 4066 to do all the switching and routing.

I've used the set-up as per Andrew's "wicked switches" article, more specific, this particular one:

I had some problems with the set-up where the 4093 was not clocking the 4013 properly.
(I tied set and reset to ground on the 4013 and VS=9v+)
They both seemed to be working (4093 gave pulse and the 4013 switched if clocked directly with a momentary to ground)
but hooking them up to each other produced no flipping or flopping,,, 

I've done a little probing and It would actually work after placing a 12k resistor between nand output and the clock input
while touching the 12k resistor lead?... after a while it seemed to work without me touching the lead as well (just the 12k resisitor
between nand/out and clk).

Then I proceeded with the second switch set-up and noticed that the first one stopped working the moment i put the 220k resistor
shown in the schematic on the nand input of the second switch I was building.

Can I solve this problem by changing resistor values within the set-up or is the set-up shown unstable by nature?
Eventually I planned on using 2 4093's driving 3 4013's driving 2 or 3 4066's (I know, it's probably slightly mad...  )

Help is very much apprieciated!!

[any]

Huh, Now both switches work as per above schematic?! (a bit more fiddling)
Does that mean the breadboard is the cause of this erratic behaviour?
(I remember reading about breadboard capacitance somewhere..)

It would be problematic if something works on the breadboard but stops working
when soldered on pcb... any experience on that?

[George Giblet]

With CMOS gate you should connect all inputs on the unused gates to ground.  If you don't you can get all sorts of weird behaviour (for example excessive current draw).

Because one gate of the 4093 is driving an LED the relatively high current could have been causing glitches on the other gate.

If you want to know the answers you have to keep playing around until you can find cause and effect.  For example adding supply decoupling.  Temporarily pulling the LED out.

Yet another possibility is the switch debouncing is marginal.  The 20ms time constant is usually OK.   *For experimental purposes* I would use 2 to 5 times the cap for example.   I have doubts this is the case because of the 12k + body touch trick you did - that actually points to some sort of glitch on ground or power.

To find problems you have to make guesses then do mods to prove or disprove your hypothesis.

[any]

Think that might be it right there with the non-proper grounding of unused inputs.
I did manage to stumble upon a proximity control in the process.
(like what earthtone's been posting lately, but then with a 4093...)
It gradually lights up a led by proximity quite nicely, I'd say sensetivity starts at approx 10 cm from the antenna.

I'll investigate that later on though....

Thanks for the tips!