Yep Paul, it's the wiring to the EDIS module itself - even when running at the fixed 10 deg. with just the module itself ('limp' mode) its rather sensitive to interference on the lines from the pickup and to the coilpack(s).
I don't know whether Ford route the wiring on the Escort/Mondeo/Fiesta etc so that it's as far away from possible sources of interference, but it seems that when you move it, or use non-original wiring, it becomes quite susceptible to noise.
We tried three setups with our megajolt kit:
Megajolt, EDIS, bike-carbs (slightly tweaked 7000rpm 3d map): base timing was jumping around at idle (8-24 deg.) and no.2/no.4 spark was always poor
EDIS (no PIP/SAW/TPS, so in 'limp mode' with static 10 deg. of timing) , bike-carbs: timing jumping around (not as much as with MJ), poor spark on 2/4
Sierra dizzy, spark amplifier, bike carbs: Runs great!
The TPS was calibrated spot on, new plugs and leads, 3 different coil packs, correct values for static testing of the wiring... but when running you could see the spikes on the runtime display from the noise on the lines.
I've since found that it's quite a common problem when using the EDIS system with either MSnS or MJ (same principle) - the wiring, earths and voltage levels are absolutely critical or it's absolute bobbins. I'll dig out the test values that you should get when it's wired up correctly....
Although the side affect of this is that we can state that R1 slide carbs idle much, much better than 40DCOE's, even with a simple dizzy... with virtually no fettling needed at all :)