Bug: VEAL goes haywire on VE2 with MAF

[subwoofer]

There is something very strange going on here, saw the same problem in the 2.2 series of betas.

VEAL adds fuel everywhere when run at an alpha-n correction table for a MAF system. As you can see from the screenshots it just adds fuel, which EGO in turn removes again. After less than 10 minutes of activity VEAL had made the mess you see in the second screenshot. Apart from the upper left corner, I would consider the initial tune (attached) a good one, it is well within single digit % correction territory.

2013-09-10 07_15_14-Greenshot.png

2013-09-10 07_15_31-Greenshot.png

What is going on here?

[subwoofer]

Do you need anything more to try to reproduce the problem?

[LT401Vette]

No... I think I know why. I believe it is looking at AFR2 for VE table2.

[subwoofer]

Mkay... I have never touched that table, so I don't know what it looks like (forgot my tuning PC at work today), but at least it's a plausible explanation.

Is there anything I can do to work around it? Except tune by hand, that is?

[LT401Vette]

Use the attached custom.ini

Rename it to custom.ini and put it in the dir:
[YourTunerStudioProject]/projectCfg/

That should set it straight.

[whittlebeast]

Create a new field in MLV

MAPxRPM and define that field as [MAP] * [RPM] / 100

Then scatter plot MAPxRPM on the X and DutyCycle on the Y and AFR on the Z

If the plot comes up close to a straight line, it is a MAF calibration curve issue and the autotune is doing what I would expect.

Note: This trick is only valid up to about 15% throttle on a well tuned ITB moter. It works textbook an almost anything else.

Andy

[subwoofer]

Andy: Not suspecting the TF here, VEAL was adding fuel that EGO was taking out. VEAL's target was for some reason off.

Phil: custom.ini downloaded, merged with my existing custom.ini and loads without complaints. Will try it in roughly 5 hours time.

[subwoofer]

No avail... Same behaviour as last time, VEAL adds fuel while EGO takes it back out.

Maybe time to make a log while it is working, so you can take a closer look?

[jsmcortina]

Perhaps this is where the option to use the firmware's AFR target value instead of using a PC-side lookup of the table would help?

James

[elaw]

Perhaps this is where the option to use the firmware's AFR target value instead of using a PC-side lookup of the table would help?

James

That would be *very* helpful for those of us who are using AFR table switching/blending! It takes a lot of messing around to use VEAL otherwise.

[subwoofer]

+1...

[whittlebeast]

Did you try the MAPxRPM trick?

Andy

[subwoofer]

No need. When the manual (straight 100s) run with EGO = 100 and VEAL makes it run at EGO = 85 and decreasing, the TF is good...

[jsmcortina]

Did you try the MAPxRPM trick?

Andy

That's unrelated to what's going on here.

James

[LT401Vette]

It really needs to get the right egoCorrection too.

It would be good to have over-rides like I put in MLV.

[subwoofer]

IMO it would not break the concept to have VE target and egoCorrection overrides in the Advanced section of VEAL. Adds steepness to the learning curve once you venture into oddball configurations like I have, but would be functional.

Would it be much work for you to implement anything like this in the coming beta?

[subwoofer]

Not meaning to be a pest, but will you implement this feature or should I start hand tuning the final bit? It is not like I am stuck without it, just that it would make it so much easier to get the final 2% of the tune in place.

Keep up the good work!

[LT401Vette]

I did spend a few hours on adding afr, ego and afr target over-rides the other day, it is going to touch a lot of layers so I became hesitant to add it now when close to a stable promotion. But this must be addressed one way, this or get the ve analyze ini set up properly.

[subwoofer]

I am fine with waiting for the first beta after promotion to stable, just that I know it is coming one way or another. A lot has changes since 2.0.8 so I guess a promotion really is way overdue.

[subwoofer]

Phil, what is the status on this now? It may have turned out to be a major redesign, but I think it is worth doing - I am probably not the only one messing with different load axis for different tables.