Trash
Joined: 06 Aug 2011 Posts: 7
|
Posted: Thu Aug 18, 2011 12:23 pm Post subject: |
|
|
Bob doesn't know either regarding the larger table or changing the row/column values.
He said it interpolates in between cells and off the top and out the side of the table for low MAP/high rpm situations. I haven't been able to fully verify that yet.
Others have explained it this way:
"If this MEFI stuff (that I have not used myself) is anything like the GM and Ford stuff that I have seen, then it will take the last known value and run with it. So if you have PW of 10ms at 5200rpm and 100kPa, then go to 5500, but still at 100kPa, it will keep using the 10ms value until the kPa changes...it which case it will then move to whatever cell intersects the highest RPM value and actual kPa.
Think of the "cell trace" feature of many software packages, where a bouncing ball or highlighted cell moves around and following the current operating location...when you fall off the table somewhere, it just sits at that location in one axis direction, and follows the other axis as long as there is resolution to do so."
So I really don't know what the system does but both of the previous explanations seem plausible.
My current table has rpm points starting at 600 and stepping to 800, 1000, 1200, 1600 etc. in 400 rpm increments up to 5200. Map columns are 40, 60, 80, 90, and 100 kPa. Total table is 14 rows and 5 columns. I under stand from our previous conversations that by increasing the number of rows and columns it would not work with the current memory address locations. I haven't tried it yet but have a table built that I might just for kicks.
What I would like to do is keep the same table size, 14 x 5, and even keep the same MAP values while changing the RPM increments from say 400 to 475 rpm increments.
So the RPM rows would be 600, 800, 1000, 1200, 1675, 2150, 2625, 3100, 3575, 4050, 4525, 5000, 5475, 5950.
This way I could have a BPW value for the ECM to look up during higher rpm events while keeping the total table size the same, thus preventing any memory address issues.
I guess I'm out in uncharted waters on this. I'll try a few of the new tables and see how things go. |
|