Got a weird one. When the car runs, it runs perfectly. However after a heat up, off, then hot soak 20-30 mins it will fail to start due to CKP signal issues then is prone to stalling. It has a new Delphi sensor in it - possibly replaced twice but unsure on the 2nd time. It has been through a number of hands trying to fix this.
Caught a fail to start on the scope, connected at the sensor.
Blue is the crank signal, green is 12v supply at sensor, red is gnd at sensor, blue is sensor output 5v from ECU pulled to gnd, gold is the 5v firing signal to coil 1
There is a LOT of noise and this does seem to be real as I have moved leads around etc. I also captured a non start with NO crank sensor activity at all (ie stuck at 5v)
Next a really interesting capture a minute after a long crank start
First thing I notice is that a lot of the noise disappears at the same time as the crank signal, but the 12v and GND are still there. Also note that coil 1 double fires as the ECU is confused. I have borescoped the reluctor, manually tested the sensor with a bit of paper for an air gap and a screwdriver - then heated it with hot air and did it again -both good clean square waves. I also re-routed the sensor wires out of the harness next to the coils.
I did notice an excessive amount of oil on the sensor, but not sure if that would affect it.
The reluctor is keyed to the shaft and I would re-iterate that the engine is smooth as butter before a hot start.
Ideas ?
Ah, of course, missed that. All automotive components make noise, its the job of the battery to filter it. Acts as a damping capacitor. Check capacity and cable resistances to body and alternator and fuse boxes.
Sounds mad but I look at the noise and I see that all the spikes are at exactly the same time as the crank sensor pulls the 5v to gnd. - Not only on the 12v power and gnd to the sensor but on the coil signal wire too. This makes me wonder about the ground on the ECU. However all the traces are relative to battery ground. I did scope B+/B-ve still noise on there but not as much.
However the noise is not causing the sensor not to work, I thought originally that the engine was seeing noise as extra teeth, but from the above waveforms something much more serious is going on.
I have given in and ordered another sensor to completely rule it out as the next step that I can see is that the reluctor wheel is loose which is a big investigation I am not ready to call yet.
Update
Customer reports today after a week of driving no issues. It was on the way to the scrap heap when I was involved as the last resort. Where would you start without a scope ?!
Not all scopes are the same, so it'd have to be a ppppp PicoScope
Sorry for gate crashing but reminds me of a Montego back in the 1980's when the engine was experiencing a misfire while driving but not at idle or reving in the garage. I used a Crypton device (forgot the name now) attached to my Crypton 335 motorscope of the day, which was used to backprobe injectors and display their waveforms. When I scoped the crank sensor I could see the flywheel pins distorted, hence a recent clutch had been fitted by someone. Would never have found that either without a scope!