After repeated VE learn sessions,*and YES I am doing it correctly* and even after trying to hand smooth these wacko results that are all along the edge of the graph, im starting to think my sensors/harness/ECM may have a problem with the way they tell the LAPTOP how to do the VE learns. It creates weird random high and low peaks that stand out significantly, and in NO way runs well when run on the street.

This is always the result (see pic below) that I get right after a VE learn. And if I run the chip, I can notice it as the engine "falls" into that "valley" created by the VE learn. It seems to "jump in and out" of that valley about once every 1-2 seconds which causes power surging regardless of manifold pressure.

Could this be my IAC going bad as it tries to adjust when my foot is off the gas, or its dirty causing it not to be able to "catch up" with the changes it is supposed to be reacting too? Meanwhile, the fuel is being added/taken out as the IAC is falling behind it, and therefore causing the lean/rich condition that would make the VE learn go funky?

Why does my VE learn do this?


SEE PIC ---->>>>http://img178.imageshack.us/img178/3...msolvedsm2.jpg




/url]

 

The IAC wont really affect fueling, just causes teh car not to ide properly. The surge is likely due to the VE table. Id tune by hand rather then using teh autotune, and pay close attention to the number of samples. Only a few samples isnt a good dataset and shouldnt be used for corrections.. Could be a lean condition due to AE (if updates are allowed when going lean) or something similar.

 

I use a mixture of hand tuning and autotuning. I usually run a couple of autotunes, then hand-tune based on them.

AE however is a mystery to me. I have yet to find ANY information on how to properly tune AE. Will the VE learn make corrections while AE or DE are in effect? I suppose that could create some problems.

 

I dont know how closed loop handles AE in RBobs code. Maybe he can answer. As for AE, its there to mainly handle the transient fueling. During manifold changes, the fuel can both settle out during high MAP. Thats why theres AE. Its also there to cover for delays in the hardware as well as to generally richen the fueling during acceleration.

Id also avoid making all changes that autotune recommends, especially at low MAPs, high RPMs as youll be entering decel enlean and DFCO, which can screw up things if the parameters arnt set up right. Also make sure you have lots of BLM samples before tuning a VE cell. Probably at least 40 before you tune a VE cell.

 

Check that the 2,000 RPM row in both VE tables have the same values. The only other thing I can think of is that the injector timing is switching between sync & async. This can cause strange sudden lean & rich jumps. If you want send a data log of the drive that created the above posted graph. I'll take a look at it. If using a WB please inlcude the channel and which WB.

Items such as AE, DFCO, DE, PE, BLM active is taken into account for VE Learn. So unless AE or DE is way off, they won't affect the learn process.

RBob.

 

I dont have a datalog, but I will try and get one ASAP and I will send it to you.