No, EAE is there to ensure the commanded amount of fuel actually gets into the cylinders.

So your VE for a particular region is 50, but (numbers are all pulled from my ass) '2' is stuck to the walls when acellerating and '1' is sucked from the walls when decellerating.

Once correctly tuned, EAE adds in the extra '2' and removes the '1' to ensure you get the commanded '50' into the cylinder. Which is why it's so important to have your VE table completely perfect before even considering tuning EAE.

AE can be used to complement EAE during sharp throttle transitions as the rate of change in VE is too quick.

ok, great. i have been trying to fix everything with suck from walls. G
The graph is looking a bit like a sawtooth. I think ill make a little more progress. Thanks

no. both are for both. AE is event based. If you surpass your TPSdot threshold the MS will simply increase the pulsewitdh based on the tps rate/rpm table. Likewise when you lift, it will just pull whatever number is in the decel fuel amount.

EAE is constantly running and it's like a ballet dance between the adhere and sucked tables.

Adhere-to-walls values will cause the pulse-widths to increase when opening the throttle. When you tell the MS you have a higher adhere-to-walls value, you are telling it that it needs to inject more fuel because more fuel stuck to the walls and didn't get burnt off.

Sucked-From-Walls values is how much fuel is coming off the "puddle" so it doesn't need to inject as much fuel.

If you widen both tables you might be constantly be in EAE because, like i said, it's constantly running. If poorly tuned you could be telling the MS that the "puddle" of fuel is constantly increasing and never decreasing...or that you are pulling more fuel out of it than you are putting in.


Turn the EAE gauge on when you tune and watch it's behavior. Regular AE is running at the same time as EAE FWIW. Tuning the EAE isnt about stomping on the throttle, you must maniplute the throttle slowly, hold it in place and watch how the EAE gauge reacts. Tuning in 5th gear is a must. When you get it right, driving in 5th you should be able to manipulate the throttle slowly and/or aggressively and see very little change in AFRs.

I turn off TPS based AE altogether to begin with, then get EAE doing what I want for slow-medium speed throttle movements in 5th gear on the highway. Then I tune the RPM curves driving at different RPMs, then I tune the CLT curves based on differences during warmup.

Finally, I tune quick blips by adding a tiny amount of TPS AE until the engine responds quickly. Only enough TPS AE to jumpstart EAE is necessary.

Ken

How mcuh did you guys move your curves fromt the default? Mine is pretty significant at this point. i wonder if I should go back to the default values and start over.

I wasn't too concerned about this because I have a little head work.

mine doesn't look anything like the default.


http://www.boostedmiata.com/gallery2...cked_walls.jpg

http://www.boostedmiata.com/gallery2...1-4/adhere.jpg

Mine is definitely increasing, but not as high as yours. After the tips I received last night I started to re-tweak my curves. They are changing a lot from the default. The default feels good while driving, but my logs tell me it isn't ideal. Would you mind posting a screen cap of your log? I want to see how your AFR is responding to the throttle input. A nice big throttle jab and an easing on of the throttle would be great ;)

Unfortunately I had to take my xterra in for repairs, so I haven't been able to test my current tweaks.

don't have any handy. When I look at logs I still get lean spikes when going from 0% to 100% TPS...but it's no where in the magnitude that I would see them with MS-I and AE only. The way I see it, that will only get so good. You could tell it to dump in 10ms of fuel, but it has to react first, and there will always be a delay.

Otherwise I tuned for feel under regular driving circumstances mostly, and just making sure the AFR pretty much stayed constant when manipulating the throttle. Like for example giving it 30% TPS, holding it there while I build RPMs, then slowing dropping back to 10%. I could probably tune it better, but spent a good week just focusing on it when driving to work and back and said "good enough"

I assume you have ego correction off during this? I think part of my problem is trying to determine what is a delay reaction spike. I may go back to base settings a restart with 5th gear pulls.

Well the MS has to see the throttle change, then react. There's a mechanical delay. I don't believe any car out there in factory form or the best tuned car can stomp on the throttle without lean spikes. When I drive my Altima or my Prelude, it confirms this. Otherwise mines very smooth at really any rate off on/off throttle. And then there's so things you cant tune out like it hovering rich after lifting...it's all mechanical at that point: BOV dumps out air and the MS has to react...even if you cut fuel completely at a lift, the air was already dumped and the MS can only react to the event.

That is true with a regular throttle body, the only thing the EMS can do is respond.

With throttle by wire like the factories are using on all the new cars, when you put your foot down, it knows what you want to do before the throttle is opened and can exactly match the amount of fuel as the transient occurs.

Unfortunately though, I don't want to write code for throttle by wire!

Ken

yeah, you might end up crashing poeple into walls.

How about an autotune for warmup, AE and EAE? You know, nothing big.

he didnt write the autotune code. that's outside the MS.

Yeah, having autotune algorithms active all the time will kill the flash on the chip eventually too.

That's the main reason for not implementing them now.

Also, there's no way for one piece of autotune code to automatically know whether ASE, EAE, VE, etc... are to blame for the AFR not being correct, nor is there any way to know for sure that the wideband is actually accurate at a particular load... As I understand it none of the current wideband controller manufacturers actually account for increased pressure in the exhaust when on boost or at high load, and they don't tell the MS when the sensor gets too hot to be accurate either.

Ken

And the ones that do output some codes (even if not straight to the MS) when the sensor is too cold indicate that widebands generally are too cold in the time the ASE is active in order to send a real signal.

This is why I have a 25 seconds delay for my EGO.

Interesting... Where do you set this delay? Do you set it in decel over run assuming that you are going to be accelerating after you have entered the decel code? This sounds like a good idea.

MS3 has a EGO delay after start(sec) parameter. I've timed how long my LC-1 takes to start taking a reading and it's about 20 seconds. I gave it a little more for good measure...not that it matters as don't use EGO at idle.

I don't think MS2 has back-ported that little one yet.

What do you think about using the delay int eh overrun fuel cut for ego in MS2? This would probably have the same effect most of the time? Any downsides that you can think of?