Totally Stuck (reset @ idle ?) MS2/Microsquirt

[ErnieJones]

Hi Folks;

I'm having a really hard time troubleshooting this issue - at idle the engine intermittently seems to miss the slightest amount and then it's as if the ECU resets (acts like I just started it up). ASE comes on, it goes rich and then settles down again. It does this intermittently and I've been all around the block trying to figure out what the hell is going on and I'm getting no where.

Single cylinder engine, full hot idle. I've re-done all 48 of the connections to the ECU - they are all good. I've replaced the coil, wire and plug (as I had new lying around) - same issue. Used a different BIP373 - same issue. I replaced the CPS/Stator assembly - same thing.

Intermittently, during hot idle (typically) it will appear to miss and then it's as if the ECU was turned off and then on. Attached are log file of the issue (once) and tune file. It happens right after record 1788.

Any ideas would be appreciated. I can do another log tomorrow probably showing it doing to several times if this one is not enough to get a clue on what's going on?

[slow_hemi6]

It is showing a sync loss on the log. Something is strange about your Ignition options/Wheel decoder settings. You have dual wheel with missing tooth selected. That suggests you must have a wheel on your crank that is missing teeth and another wheel on your cam you have set to trigger from a rising edge. But you also have cam input set to map which indicates no cam wheel or input.

[ErnieJones]

Hi Luke;

Thank you for your reply - I'm really at my wits end so everything is helpful. Those settings are required (as I recall) for Sequential Injection to occur when the engine has no cam sensor and uses the MAP sensor for determining engine position. I just looked at it again and the ToolTip says this;

'MAP Sensor' must ONLY be selected if you are using the MAP sensor for phase detection on a 1cyl engine or V-twin. Used with 'Dual wheel with missing tooth' only. Set the MAP sensor phase threshold on the MAP Sample settings page.

So I think it's like a 'work-around' to enable sequential injection (simulates having a cam sensor). I guess nothing was apparent in the log files? I was kind of hoping there would be some kind of red light to someone in the know (not me, it seems )

[LAV1000]

Are you using resitor type sparkplugs ?
Worked for me.

[jsmcortina]

MAP phase sensing is a feature for single cylinders to enable sequential without a cam sensor.

Use the "Sync error logger" to capture the problem in the crank signal.

James

[jsmcortina]

Your MAP sampling needs work, that MAP reading is bouncing around in your logs due to your settings. As per the manual I would suggest sampling over two events. You probably need a larger window too.

James

[ErnieJones]

Are you using resitor type sparkplugs ?
Worked for me.

Yes, there is a new resistor plug in the engine (Iridium NGK) along with a new NGK Racing Wire and a new OE coil. (same components used successfully before - just brand new).

Use the "Sync error logger" to capture the problem in the crank signal.

James

I am 99.99% certain there are no CKP issues. The OE sensor was replaced and continuity was checked right from the sensor to the top of the 50 pin connector on the uSM.

Your MAP sampling needs work, that MAP reading is bouncing around in your logs due to your settings. As per the manual I would suggest sampling over two events. You probably need a larger window too.

James

This!!

Adjusting the No. Sample Events to 2 made an enormous difference in the quality of the idle. It still stalled but...I'm much less concerned with that at the moment because the idle cleaned up like you would not believe. Huge difference. I then looked at how to set up the sampling as per the docs you mentioned but...I don't seem to have a CrankAngle field. Truth be told, I'm getting pretty close to being over my head. I had previously set the MAP Sample Angle and it should be very close if not correct.

But I did take a MAP log (attached) and maybe it's enough information to make a suggestion? Also, latest MSQ attached.

[jsmcortina]

Have you viewed that log?

Look at:
MAPwindow
MAP
Tooth no

You can see that the MAP sample is too early and quite a small window.
Perhaps 50 deg later would be better and make the window a lot wider to catch the minimum.

With your current settings you are seeing maybe 65kPa reported? But with actual minimum is 47kPa because you sampled at the wrong angle.

While you claim that you don't have a crank signal problem, you obviously do! You are getting lost sync in the datalog you posted before - that means there's a signal problem.

James

[ErnieJones]

Have you viewed that log?

Look at:
MAPwindow
MAP
Tooth no

You can see that the MAP sample is too early and quite a small window.
Perhaps 50 deg later would be better and make the window a lot wider to catch the minimum.

With your current settings you are seeing maybe 65kPa reported? But with actual minimum is 47kPa because you sampled at the wrong angle.

I tried changing the MAP Sample Angle from 300 to 360 but this seemed to move in the wrong direction. So I tried going the other way, to 240 with a window of 60 and got this (see attached). This seems to approximate what you mentioned in the docs;

[ErnieJones]

But there is a problem...at the 240 degree sample angle the engine does not go full sync (sequential)

If I'm correct with the chart above, 240 degrees is a very good MAP signal. It idles very well there and sounds good. But....comes at the cost of Sequential FI ?

Only from 290 => 350 degrees will it go sequential FI

[jsmcortina]

Have you adjusted the MAP phase threshold down?

James

[ErnieJones]

Have you adjusted the MAP phase threshold down?
James

Yes. I tried that in several steps of varying granularity and the engine ran noticeably worse each time with zero effect on sync. So I went back to 300 degrees, with double sampling, the wider window, and then out for a ride (auto VE to see). It really bugged me that I could not nail down the proper MAP pattern with sequential. It seemed like a case of either 300 degrees and live with the signal being off or 240 and live with perfect signal - but no sequential. Yet I know the OE had it......(Keihin design for Honda - very good) How come it works well for them?

Out of frustration, I did a road-side shift from ITB Mode back to %Baro/SD and guess what? The test ride was very smooth, the engine became very forgiving on fuelling & spark, it locked up sequential right away during cranking.Vibrations reduced. When I got home, I did another MAP log at the same 300 degree settings (this time on %baro) and what did I find? The lines had shifted to perfectly meet the DIP in the MAP at the bottom!

I don't understand the science behind it but shifting to SD changed where the engine idled and somehow, this seems to have also moved the sampling area to the bottom of the MAP peaks in the log. It made the adjustment that I could not get. I suspect you would understand why but I don't.

So this is really good. Not only that, but prior to this at idle it would use up to four vertical fields (neighbours) for the idle range or for the timing table. So I would see the 'scribbler' move between those four blocks. Now, the 'scribbler' is a fast moving dot in the middle of ONE of those fields. It doesn't touch it's neighbour. So the accuracy of the idle has changed like you would not believe.

I think James, that we just (finally) broke the back on this thing. And... I learned how to read and adjust the MAP sensor in the process. Tuning now should be very easy as it's so forgiving. Up to this point it had to have incredible precision of settings to make it run right.

Also, you were spot-on about the sync loss. The stalling issue seems to have gone away as these settings came into play and it's smoothed out so well. But, on high speed highway with protracted high-rpm (12,000 or so) every once in a while the engine jerks. I'll do a log but I think it will show sync loss. Many, many thanks for this as I think I finally managed to replicate the invisible settings Keihin used. Those trickky bastards

Thanks for teaching me how to take MAP log and read it!!

[jsmcortina]

Have you adjusted the MAP phase threshold down?
James

Yes. I tried that in several steps of varying granularity and the engine ran noticeably worse each time with zero effect on sync. So I went back to 300 degrees, with double sampling

I was saying the threshold in kPa - not the angle.

As you've moved the sampling window, the kPa reading on the opposite cycle might be too close to the 80kPa you have set. Try 60kPa for example.

James

[ErnieJones]

Have you adjusted the MAP phase threshold down?
James

Yes. I tried that in several steps of varying granularity and the engine ran noticeably worse each time with zero effect on sync. So I went back to 300 degrees, with double sampling

I was saying the threshold in kPa - not the angle.

As you've moved the sampling window, the kPa reading on the opposite cycle might be too close to the 80kPa you have set. Try 60kPa for example.

James

I figured that, and tried various numbers on the threshold as I thought this might be what the code uses to 'trigger' sequential. 60 was tried as well as a bunch of other values (40 etc) to the low side. Nothing worked.

But I think it's cased now using 300 and SD. When I went to SD and did another MAP log the pulses lined up perfectly. I can re-visit it if you like, it's easy enough to change back.