Quote:
Originally Posted by toohighpsi
Nice work - looks like you have a good handle on it now!
Time for more E and more boost!
 |
Thanks! I could use a refresher course on the Lambda P/I/D values. It was very lightly touched in your video, but I'd like a deeper insight into what those values do.
I'll take my educated guesses here for fun and public humiliation:
If I recall correctly, the P (Proportional) value has to do with the amount of correction applied and can be tweaked based on the % of PI. You had said that when on 30% port, the theoretical ideal value is 2.0, but that setting to 2.2-2.5 allows more (or maybe more precise) correction. If % of PI were higher, 40-50% etc... then that value should be lower, no doubt because a lesser amount of correction needs to be factored in when PI is a greater % of total fueling. Not being on a dyno makes it difficult to make, test, and analyze a bunch of changes to see the real-world impact of changing these values.
The "I" being the integral, I assume as described by Integral Calculus having to do with finding the area under curves, but it isn't my strong suit, and I'm not clear how it is being used in the calculations, or what the real-world impact of changing that would be. Based on the formula in TunerPro:
ITerm = ITerm + (ki * error)
I could surmise that this factors into a carryover effect to ipw calculations from moment to moment by the Reflex to the injectors at run-time. I have observed that the final inj pulse width by the Reflex is not always what it says in the "after lamba correction" value... In other words, there appears to be a smoothing effect based on prior lambda correction values for ipw...
No clue how the derivative term would be used.