rusefi.com

I checked ignition during starting.
The OEM flywheel (with its vr sensor) is accelerated from 0 to 500 rpm. The dwell is programmed to 4ms. The two vertical marks of the scope are set at this time for easy reading.



For me, it's not normal that the dwell is so small at low rpm, it should be as programmed.
Is something set wrong ?

I'm not currently aware of any bugs around dwell while cranking. I've been working on related spark logic recently, and I haven't noticed any issues, though that doesn't mean there aren't any. Does it look right at higher RPM (say 2000)?

Can you share your tune? The easiest way is to press this "Create Debug Package" button in Tunerstudio, and upload the zip file it spits out. A datalog from a test run of it happening would also be helpful.

JPh wrote: ↑

Thu Mar 26, 2020 8:23 pm

For me, it's not normal that the dwell is so small at low rpm, it should be as programmed.
Is something set wrong ?

I agree, I believe that "dwell" TunerStudio gauge should match oscilloscope dwell.

Probably not your case but: rusEfi relies both on clock and trigger tooth information to schedule events. If trigger fronts happen not at the expected angles that would cause loss of precision with ignition.

Just a reminder that "engine sniffer" tab on rusEfi could be another source of data for you.

Can you add decoded digital trigger signal as another channel on the oscilloscope? You have "crank" test pad on microRusEfi if you can solder a test wire to it?

PS: your oscilloscope displays data on the PC? What oscilloscope is that?

Some answers :

I have measured this behavior only in the beginning of spinning. As you see on the video things go progressively better. The phenomenon is very stable, it goes exactly the same way at each attempt. So I don't think it's a parasitic related problem. I didn't go higher in revs because at 500 + things settled down.

My scope is old fashioned (like me). It stores raw samplings so I can share the files. But there are only 4096 samples in the file, so you must make good use of the trigger to have the data you want. It's a Velleman pcs 500.

Here is the debug package.

I agree the engine sniffer would be a great tool but it lacks of some basic explanations of what it is displaying.

Some data are obvious like injector and coil activation times but the timing data on the top of the screen are not clear for me.

I will be glad to post an example of using it on my engine if a have some basic knowledge of how it works.

I think I can be a good example of a basic user using your tools on real worl stuff.

New video with the crank signal picked from the tab on the board (top trace).
The ignition signal is on the bottom. The programmed dwell of 4ms is indicated by the vertical bars. To be sure the bad timing is not a scope problem, I dobbled the mesure with my old analog scope. Both readings were the same corrected by the scale factor.

This I my first reading of the crank signal from the mre, but it seems obvious that the big teeth decoding is not stable and causes the problem. What's the opinion of the experts?

The funny part of that is that my good results at 700+ rpm obtained yesterday were done with that setup!

Your trigger pattern looks a little funky, but it's hard to tell if its an artifact of your scope (since the sample rate looks pretty low) or an artifact of the setup. It looks like the first tooth after the gap comes early, which can happen if the VR sensor isn't spaced correctly off of the wheel. For this reason many OEM setups have the missing tooth far away from any important events - for example Bosch usually puts it 60-90 degrees BTDC cylinder #1, so that everything important is happening far away from the large gap.

If that tooth indeed arrives early, then the ECU will assume that it's in the correct place, so it'll fire the ignition coil at the correct duration after that first tooth.

Can you share a snip of the trigger logger page in tunerstudio? You should see 58 perfectly equal height lines, followed by one line that's 3x the height (2 missing teeth).

edit: oops, didn't watch the video all the way through. Yes, it looks like something is indeed wrong with your trigger, or you're running inverted. Can you change this option to "true", and repeat the test?

Oh, looking at your photo in the other thread, it looks like your wheel has a long tooth, not missing teeth. That produces a slightly different timing for the first tooth (especially at low speed, as you've observed). I think that should be okay as far as maintaining trigger sync is concerned, but may produce inaccurate timing of events scheduled between the first and second teeth after the long one. In essence - install the flywheel so that the long tooth is between two cylinders (not near TDC for any cylinder), and you'll be fine.

mck1117 wrote: ↑

Fri Mar 27, 2020 6:53 pm

In essence - install the flywheel so that the long tooth is between two cylinders (not near TDC for any cylinder), and you'll be fine.

For the test right now, please try running with trigger offset 90 or -90.

Many thanks for your clear explanations!

I tried inverting the vr polarity but it didn't work : no stable sync at any rpm.
I will check were renault put the sync tooth in the cycle.

normally you have on large '1' and then one large '0' (or vice versa), but you have one large '1', then one large '0' and after that another large '1'?

The big teeth may be to long at low rpm and the voltage varies enough to make a bad triggering.

what are those teeth in the background?
why did you put the sensor this way, not a quarter round? i'd suggest looking at the initial signal on the scope.

The teeth behind are those of the starter wheel.

I put the senor this way because it's oem like and I want to switch between oem and mre ecu.

That's the signal at slow speed :

if you google for an oscillogram of a 'normal' missing tooth, you'd see it's somewhat different. I'd suggest twisting the sensor 90º round as an experiment.
if i remember correctly, there was a guidance on how to choose the sensor and how to locate it in relation to the trigger wheel. The metal notch on the sensor should have similar width as the gap between teeth (and the size of the teeth).

New video with ignition set at an angle of 90 degrees as Andrey suggested. This is representative of oem setup on the renault engine. Not had time to check if it is +90 or-90.
Fixed advance to 0 degrees.



As you see, the results are perfect now.
How it behaves : the sync doesn't come before 100 rpm. Precisely at 100 rpm, ignition starts and all is functionnal then.
But better : once youy had a sync at 100 rpm you can go as low at 70rpm without loosing sync. Below 70 rpm, the signal is to week to be used.

At the end of the video I filmed a strange thing with this setup : the gauge indicates -3,6 degrees at all rpm wich is not coherent with the fixed 0 advance programmed. Somebody has a idea what to check ?

@puff

90° flipping of sensor is not possible because of mounting fixtures on engine.

But your idea reminds me there is perhaps a problem of 180° flip of the sensor. I will go and check. Thanks for the help !

Just checked, I'm lucky it's the right way!

The difference in advance from what's programmed is probably from the CLT/IAT advance adjustment tables. Zero those out and you'll get exactly the requested timing.

I am surprised that these corrections are not overrun by the fixed advance setting.

I would think strictly that the fixed timing setting should override everything and set the final output value to the desired fixed value.

If that is not the case then I wonder if that needs to be raised as a github issue.

Not the case.
A GH with relevant screenshots would help

That trigger wheel has both a missing and added tooth.....the scope trace looks correct I think for the wheel design I think, does the software support this style trigger wheel? It does look like the sensor should be rotated 90deg as mentioned though.

To be totally exact it's and added tooth and then a missing tooth. It turns anti clockwise seen from above

OEM setup :

@AndreyB

"A GH with relevant screenshots would help"

GH ???

GitHub issue tracker with relevant screen shot.

OK will try to do it.

I'm not 100% sure there's an issue - all missing tooth wheels can be a little weird near the missing tooth, which is why no manufacturer puts the missing tooth anywhere near anything importnat. If the ECU syncs ok, and ignition timing/dwell look correct away from the missing tooth, everything seems fine to me.