EAE tuning method? Custom gauge? Scatter plot?

[spittybug]

I have successfully gone to 3 squirts and got back to a decent looking and performing VE table. I took the opportunity to change the advance transitions and the target AFRs first, but nothing radical.

I tried EAE with the lag...no way.... kills me instantly with the threshold at 100. I suppose I could try upping that, but why wouldn't it simply kill me then?

Anyway, EAE on and looking at the log it does a very nice job of keeping AFR very close to target, except when I come off the throttle and the MAP plummets. Just like before, so I'm ready to tackle this now. Looking at log, I notice this:

• My WALLFUEL1 variable bottoms out with a slowly declining throttle and MAP and then I go very lean
• While my MAP is between 30 - 40, the WALLFUEL1 keeps falling (though not bottoming out) and I go a bit lean
• When MAP drops quickly, WALLFUEL1 peaks down, but recovers, and I get the BIG lean spike

So, using the above posted advice, it looks to me like I should be adjusting the lower MAP end of the sucked-from-walls curve downward. The < 40kPa area looks like it needs to be very flat, same as the 40 kPa values. Should the same logic apply to the added curve? Should that be very flat at the < 40 kPa zone? Am I on the right track by trying this? Thanks.

[spittybug]

I played and played and played today.... while I like how the EAE helps keep the AFR tighter around the target, I just haven't been able to get rid of my lean spikes upon decel. Every time I change the curves I start introducing oscillation as well. I can see why the recommendation is to do all the testing on the bench.

I killed EAE and went back to AE based 100% on TPSdot. First however I loader gslender's latest code variant to take advantage of the TPSdot smoothing I read about. I'm quite impressed and I've been able to get a TPS curve starting at about 50 that gives me good performance without unexpected herky jerky PW spikes. Yeah, the AFR doesn't hug the target as nicely as with EAE, but I've eliminated the decel lean problem entirely.

I went for a pretty long tuning drive with VE analyze and it never fails to surprise me that the table it generates is not "linear" or smooth (in 3D). It has some zones where it will have little peaks, others with valleys. Nothing too extreme, but just odd. I have it on "normal" resistance and on those areas where I question it, I looked to see if there were a good number of sample points and there were many hundred. I suppose this is a function of the wall puddle phenomena that the EAE tries to cure?

I'm looking forward to learning more and playing with the idle improvements of this code. After all warmed up, I idle very nicely, but during warm up I still get some hunting (oscillating) half the time.

A question for those in the know: I've heard that too much advance at idle makes your exhaust smell and burns eyes. This, even with an AFR in the 13.5 - 14 range. Is that true? I deliberately mimicked my vacuum advance curve from my stock setup including the elimination of vacuum advance at idle (stock has a solenoid to block vacuum). Immediately upon throttle the stock dizzy is advanced from static 13* to full vacuum advance plus static of 33*. I tried to imitate that but with that interface so close to the idle area it contributes to the hunting. I then realized that the stock setup may be for ease in cranking (not an issue for MS) and that once idling, much more advance shouldn't hurt. Doing this (only went to 20*) seems to have increased my exhaust smell however. Any suggestions on proper idle timing would be appreciated.