lost sync while cranking

[Aemad]

Hello guys, I am using MS3 3.57 board mounted on KTM 540 EXC engine, from a while the crank shaft position sensor connector broke so it left me with in a dilemma of the polarity but I think I managed to solve it, after then a problem of losing sync happened during cranking( rpm keeps oscillating between 200 and 0 rpm) I figured out that the starter was failing and we solved its problem, then the engine started without problems after a couple of days when I went back for starting the engine I have found that the same problem happens again, I started doubting my sensor wires polarity again so I inverted them with no changes in the behavior, when I returned them back to their original position guess what it worked !!! I run back the engine after stopping and trying to crank it again the problem returned back
attached is the link for msq file and a log
Thanks in advance
link for msq and log
https://drive.google.com/open?id=0B0lnl ... Ew2aDk3WDQ

[Drew442]

Website neither wire on the sensor is grounded with the ECU unplugged. Also if the sensor is movable (i.e. Screwed down with no key) make sure the gap is in spec for a vr sensor.

Sent from my MI 5 using Tapatalk

[Aemad]

I checked the wires with ECU connector unplugged, neither is grounded I also checked the resistance between the 2 VR sensor wires and it is between 80-120 ohm as the OEM specified and checked with a multi-meter that I receive a correct AC signal while cranking, one thing I also tried is to change the ignition input capture from falling edge which was set before to falling edge and I got same problem but I suppose looking at the log that the RPM is partially more synced than before. I attached the new log and 2 pictures of tooth logging with falling edge and rising edge
what do you think guys do you think adjusting my pots again might do any good ?
link: https://drive.google.com/open?id=0B0lnl ... Ew2aDk3WDQ

[Aemad]

Alright after doing this dirty trick of inverting the wires and putting them back the engine started again, that's an attached log for this run, so how do you think of this weird problem !!!