Losing sync only during cranking...
 

So the 05 MSM I've been tuning off and on has been running great using the regular v3.0 crank input and ms3x cam input circuits. Tuned them by hand and the car doesn't lose sync at all while its running. The car won the pro-solo and national tour at blytheville (bsp)... And its been doing great on e85. However it does not want to start hardly, I've messed with and messed with the cranking pulse.

So this morning, I noticed that it was regularly, reliably losing sync during cranking, fire fire lose sync, start over. I had just been blaming the e85, as it will start eventually and I'm used to seeing the occasional sync issue during cranking in other cars... But now I'm pretty convinced that I've got a pot or two set a little off and that if it'd just stay synced it'd start better. I didn't have time this morning to really look at the composite logs closely, but I'll do that later. BTW, the warmer the car is the less it does it so it still seems like cranking pulse needs tuning... But my car is also on e85 and I was able to get its cranking pulse tuned to where it starts fairly reliably down to just below 40 degrees. The msm almost won't start below 80 ambient... we're talking 30 seconds of cranking, super not cool.

A light bulb went off last night, I'm not 100% sure how these mazda hall sensors are built... but I'm wondering if maybe the voltage sags a little during cranking, and maybe that's throwing off the opto circuit. Going to try to dial the set point down a little tonight and experiment.

edit: ran composite log and the problem appears to be on the crank signal, the cam signal always shows the right pattern.

And it is fixed. Thanks Ian!

Turns out if you have a similar problem, adjust your crank sensor spacing.. .it might just be off enough to cause this problem.

Did you have to move it closer or just align it?

put the car in 3rd gear and pushed it till a tooth was right in the middle of a sensor, then we found something a piece of metal that was about 20 thou thick (guesstimated based on a spark plug gapper, it was actually one of those radio remover tools) It was slightly bigger, .25 or so... Then tightened the 10mm bolt down, f'r fired right up.

so yes, moved closer, there isn't much 'alignment' other than distance, it pivots on the one bolt.

This is my exact problem, but moving the sensor didn't seem to help. Anyone know how often these fail? Is there a way to change the sensor input in the ms3x to be more sensitive?

The stock ecu doesn't throw a code, so it's an ms3x problem.

The standard gap is 30 to 50 thou... I don't know why 25 thou doesn't work for you.

I will go through this, and report back. I need to take a better log, but I think that my lost sync count does not increment. I just see the lost sync flashing every couple pulses.

Megasquirt-3 Sync loss and remedies

While I do this tonight, anyone else have any suggestions? Outside of startup the car runs great and does not lose sync. I did have an issue with not grounding the MS3x, followed instructions too blindly, but that is fixed now.

Update with datalog
 

I re-checked my crank position sensor, it was within range. I also took a datalog I will post. The result of the burst logging was 20+ sync losses with no power loss (SecL was never 0). The Sync Loss Reason was 32, which is 0 cams seen for 99/00 Miata.

This problem has been happening since I first installed the MS based on my comparison of my first start datalog, doesn't happen with the stock ECU. Any thoughts?

I replaced the cam sensor and I didn't see any change. I am going to open up the MS3 and see if I am sharing any grounds that would give a conflict. Anyone have any input?

I am thinking that the stock ecu would have thrown a code and had the same starting issue if either sensor, wheel, or wiring were bad on the car. I would just replace the crank senor too, but the dealer doesn't take returns on electronics, whatever that means.

I have to install my MS3 knock anyways, hopefully this is something simple.

I went through the ms. No grounding/wiring issues. I will try a firmware update and new crank angle sensor.

edit: I added my MSQ, ignore the timing and fuel. I am working on it for my FI install, so don't try to use it as is!

Same MS3 from the 05msm from OP went into my '04msm and has exact same sync loss issues. Crank sensor is in spec, and was replaced. Once it runs it is not losing sync.

BTW, 1000s and e85 the stock ecu cranks right up. Yes its pig rich but it fires immediately.

Ideas?

Mine ended up being something stupid. I had loaded someone else's tune and didn't notice my rising edge/ falling edge setting had changed. Everything else was identical. Won't make that mistake again.

So... If your dwell voltage correction curve is off, it would appear that the coils won't fire which can apparently cause back EMF to throw off the crank signal? At any rate, they turned the dwell at 10V up a bit and now it cranks no problem.... Bizarre, but something to look for, those settings should have been unchanged from the diyautotune 99-00 base maps. May be causing other people problems too, idk.