The symptom is that while cranking the enigne sputters and missfires, it usually does start but sometimes it takes a while (>10-15 attempts).
While crankning the RPM gauge is jumping all over the place.
I have read through the threads I could find that relates to this issue and have tried the different suggestions so far.
I rerouted the power wire to the starter away from the rest of the harness to not allow for emi to affect the sensor signals.
I also checked the ground wire on the transmission to ensure that it is clean on both the transmission as well as on the chassis side.

I bought a new RPM sensor (i call them crankshaft position sensor) which did nothing to improve the situation.

I tried a few experiments that led me to believe that the CAM position sensor is bad. I disconnected the CAM sensor and tried to start the vehicle without it connected. The symptom stayed exactly the same. The ECU did not throw a CEL while trying this, which I thought was weird. The engine ran just like before, even though the sensor was completely disconnected.

Before I went and spent even more money on a new sensor that might not improve the situation, I hooked up my portable oscilloscope to the CAM position sensor to see the signal output and it looks clean. The signal reflects the same notch pattern on the end of the cam.

I am measuring the signal at the connector on the front of the block and was thinking that the next step is to trace the wiring to the ECU and the pin so I can ensure that the signal the ECU sees is as clean.

I guess this is a really long introduction to ask if someone has the ECU pinout for this model?

Grateful for any ideas.

 

You'd need a breakout box for the ECU first.

 

I found the pinout of the ECU.
Regarding the breakout box, it sure makes things easier to hook into but it is not really necessary. I tap into the back of the ECU connector directly to see the signals.
I used to work for the company that designed that break out box in Sweden. It's a company that used to be called Autodiagnos AB. They are now called autometric AB it seems. The break out box is nice since it has some signal filtering to not affect the signal you are measuring.
I still haven't found the root cause for this issue, but I'll update this thread for futre reference.
Cheers

 

Sure enough, the signal looks like crap at the ECU. I'll see if i can route a shielded separate wire from the sensor conector to bypass whatever is affecting it in the wire harness. I'm thinking it might just be the starter EMI that is affecting it. We'll see if this works I gues.

 

Mystery solved!
After doing a lot of signal measuring and starting attempts it became clear what the problem was. The motor housing on the starter had worked itself loose over time because the nose drive support on the starter gear had become completely worn out. The starter gear could easily move 1/4" sideways. Since the motor housing is moving around while starting and the housing is the ground for all that starter current, arcs were always flying between the motor housing and the aluminum housing. Arcs create huge EMI and it also made the starter pull amps intermittently. This wreaked havoc with the electronics. I machined out the nose support on the starter, made a bushing for it and put it together. Now it starts flawlessly!
I guess I consider myself lucky to not spend the $$$ on anothe new sensor that wouldn't have fixed the problem.
I am posting this here so that the info is availble to maybe help somone in the future.
So I guess this adds another thing to check to the list when you see the RPM gauge jumping all over the place when you are trying to start the car.
1) Check engine/transmission ground connection to the chassis.
2) Reroute the heavy gauge starter cable away from the rest of the wire harness.
3) Check the condition of the starter.
/David