[zataviou]

No I ran 2 ecu in parallel just to test, the oem ecu provided ignition to be sure of sync because with speeduino it was not clear (i didn't know if the sync loss came from ecu or if the wrong settings (fuel/ignition) induced sync loss)

I was lost and oem ecu provided an igniton control without a doubt

20201227_180452.jpg (94.95 KiB) Viewed 3970 times

I think the problem come from the "signal" I show in the picture above, if it consider like a "high state " it loose the sync

[PSIG]

Prove your hypothesis. Take a tooth log and count the teeth or 'scope the conditioner output. If in-fact the little bump is causing a tooth trigger, then you could act on that, such as increasing the input capacitor value to smooth it out. There will be a balance between that and lowering the signal too much during cranking or causing rpm limitation. Find that balance. There are other solutions.

[zataviou]

I measured the input and output of the conditioner and could not prove my hypothesis, but it is difficult to stop the measurement when the sync loss appears.

I didn’t quite understand how the "trigger edge" setting worked and I was in the wrong direction... now I put on "RISING EDGE" (and readjusted the base advance)

here a measure at idle

20201230_151050.jpg (1.8 MiB) Viewed 3906 times

in attached files you have a toothlogger and the last log.

[PSIG]

I'm not looking to create confusion, but trying to confirm correct data and answers.

Note in your last 2 posted 'scope images, the wave forms are reversed. You have either rewired your VR backwards -or- connected your 'scope probe and ground backwards. Either could be reversed in either image. So I would ask that you verify your sensor connections for correct polarity (test for positive wire, and to VR1+), and your 'scope probe for correct polarity, before we begin suggesting changes or corrections. Are you clear on how to test a VR sensor for correct polarity?

[zataviou]

Yes

I am sorry on the first picture I didn't know how to measure a VR sensor (the probes was inverted)
Second picture the probes are in the right position and the scope is set correctly

[PSIG]

Ah, OK. Then your VR sensor is wired backwards. The trailing edge of the tooth leading into the gap should be fully negative-going, then rising voltage across the gap, as in the first 'scope image. That means the VR sensor voltage swings high as a ferrous object (tooth) approaches, and swings low as it departs. There are a number of reasons the polarity should be this way for almost all setups.

So, reverse the VR wiring into Speeduino, and then select the correct Trigger Edge for Speeduino to see the trailing (falling) VR voltage. This setting would be Rising for inverting conditioners such as the MAX992x, LM1813, etc, and Falling for non-inverting conditioners such as the DSC, LM358, etc. That should make the signal and output coordinated, trigger a proper gap width, and prevent timing drift. Note that using reverse VR polarity and reversed edge would seem to be OK, but does not solve some of the issues.

[zataviou]

I reversed the wires and it's worse... I get more sync loss but I was able to take a measure that proves my hypothesis, the small "bump" is considered like a tooth. How can I filter the signal?

It seems that the bump is caused by an electrical noise because I tested the wheel and the sensor out of the bike (on a lathe) and I didn't get bumps on missing tooth

20210106_191929.jpg (170.85 KiB) Viewed 3778 times

20210106_191841.jpg (142.76 KiB) Viewed 3778 times

[PSIG]

You might be able to filter a spike like that, but that's a big one (wide) and would be a challenge to suppress it without altering the remaining pulse edge timing. Preferable to me, would be to locate the source. Is there an event when the spike occurs, such as ignition firing, injector pulse or flyback, alternator pulse, etc? BTW - excellent that you tested the sensor and wheel separately.

[zataviou]

OK
I removed the alternator (which is inside the trigger wheel) and I get a curve without the small bump.
But I get sync loss again: cry:, I didn't see anything wrong with my Oscilloscope.
I made a composite logger at low rpm and I could record 2 sync loss but I don't understand why. The crank signal seems normal and I get an error.

Sans titre.png (29.49 KiB) Viewed 3737 times

What are the conditions for the software to lose the sync ?

[PSIG]

Well, it could be any of a number of reasons, but @noisymime would be better to answer specifically about conditions for sync loss in the firmware. Do you have a data log with sync losses on the current wiring and setup?