I've had a scan through posts on recent pages, not sure if the following points could still be a factor or have already been ruled out:
Old TPS wasn't at the end of it's travel with closed throttle, new TPS was (at first) at the end of it's travel. Could the new TPS have been physically holding the throttle open a bit, accounting for the fast idle? Could adjusting TPS to prevent that still have made the TPS show less than typical (and prior) 0.7v in the idle position, giving unsmooth transition between idle and off idle? Or moving the TPS still a bit more cause higher than 0.7v at idle, still causing unsmooth transition but now also causing stalling on return to idle because the Ecu doesn't see the throttle as being closed?
Would've thought that if TPS was set so that with closed throttle voltage was close to 0.7, drive-ability wouldn't be too bad to start with and transition would only improve?
Having said that, one P38 conversion I did sticks in mind. The guy had received his bosses P38 instead of his annual work bonus and made much fuss about 'will the engine still idle nicely on LPG because I've heard this can be a problem'. Come the day he was dropping his P38 off for conversion he was 90 mins late, even though 85 mins before he arrived he'd told me on the phone he was only 5 mins away. On arrival he went to great lengths to point out how it was currently idling nicely at correct rpm etc... Oh and could I just re-tap the TPS securing threads, secure the TPS properly and sort out TPS wiring that been hacked about...I'm sure the bloke had had loads of prior grief with it and intended to blame me for any further idle/drive-ability problems. I had a fair bit of grief sorting it out but did manage it.