Tuning part throttle, 2D lines crossing
Posted: Tue Jul 31, 2018 6:27 pm
I'm new to tuning and have been reading and experimenting with tuning part-throttle (closed loop is disabled) but I don't understand why the suggestions that KPro is making would result in fuel curves that cross (something that I understand is never preferable).
For the first few revisions, the adjustments I made were all less than +/- 3 points per cell and I ignored any cells completely that had less than ~300 samples for my drive back from work (~30 minutes drive-time). From there, I increased my adjustments as I saw repeated suggestions / patterns and took cells into account that had less samples. I tried to adjust adjacent cells on the same curve to keep it smooth while still maintaining the general character from the basemap throughout all revisions, while giving more weight to cells with a very large number of samples.
As I'm new to this, I figured I would see if anyone from this forum, which I have already learned so much from and seems full of very knowledgeable people, could provide any guidance for me going forward or maybe even pinpoint something incorrect in my approach. Any feedback is welcome, thank you in advance.
Please see the attached latest rev .kal and my latest datalog: 07311600.kdl
For the first few revisions, the adjustments I made were all less than +/- 3 points per cell and I ignored any cells completely that had less than ~300 samples for my drive back from work (~30 minutes drive-time). From there, I increased my adjustments as I saw repeated suggestions / patterns and took cells into account that had less samples. I tried to adjust adjacent cells on the same curve to keep it smooth while still maintaining the general character from the basemap throughout all revisions, while giving more weight to cells with a very large number of samples.
As I'm new to this, I figured I would see if anyone from this forum, which I have already learned so much from and seems full of very knowledgeable people, could provide any guidance for me going forward or maybe even pinpoint something incorrect in my approach. Any feedback is welcome, thank you in advance.
Please see the attached latest rev .kal and my latest datalog: 07311600.kdl