My bad guys, I guess when we talk about these things we should ALWAYS specify which platform we're using. David and I are using MS3, so MS2 users should be careful and double check.

Here's why I was saying what I said:

https://www.miataturbo.net/attachmen...ine=1431896910

Thanks deezums! Got it working relatively nicely again, still a bit rich on lower kpa transitions but I can drive, log and make adjustments again.

My bad too, I knew this was about ms3, just didn't realize how different it was from ms2. Thanks for the clarification.

OP, 18psi, etc. Thanks for all of this great information.

I have a NB2 VVT engine running on a Rev built MS3. With AE only (100% TPS or 80% TPS/20% MAP) it drives pretty well, doesn't buck, etc, but I haven't been able to balance out the settings for quick/large throttle changes with smaller transitions, especially with multiple throttle transistions back to back. For example, the first quick stab on the throttle (maybe 50% TPS) looks good, but if I goose it 2-3 times in a row without giving the engine time to stabilize, the AFR's drop way too low. It sounds like EAE would help with these smaller, but more frequent transitions/throttle changes.

Question: What values (TPSdot? MAPdot?) do you use for EA, when EAE is tuned as described? In other words, when should EA take over the bulk of the compensation for a throttle change? It would be great if someone who has a solid EAE tune could also post screenshots of their EA settings, so that we can see the entire picture of EA + EAE tuning.

Thank you!

The values for AE are completely useless across different cars. I can post when I'm near my laptop, but I guarrantee you they won't help you, since I've tried this on my very own NB's and none of them used the same actual values.

The curves on the other hand, do usually stay about the same.

AE only affects the 1st split second or so. Just open up the little AE wizard and look at the graps real time as you stab the throttle. The initial reaction is caused by AE, the rest is your fuel map and EAE. Also EAE is more for the smoother/slower transitions, not as much for stabbing the throttle. Think more driving steady state and increasing the throttle gradually and watching the afr follow the transition, and letting off and watching how it responds. Stabbing fast and letting off fast is what AE is for, that split second reaction.

I can try to explain it to you til we're blue in the face but just watching the graphs/wizard showing you how it happens real time will be better than 1 million words.

PS: a completely tuned and perfect VE map is absolutely cruicial before tuning AE and EAE.

I would want to leave AE on as aggressive as possible, with a TPS trip point high enough to never accidentally trip, but low enough that you don't need much EAE at all.

I currently have mine set somewhere around 200%, the way I understand that is I double the throttle position (or more) in one cpu cycle for AE to add fuel.

If you are stabbing the throttle repeatedly, enough to trip your AE entry point, and getting richer and richer I believe you need to tune decel fuel as well as leaning out the AE a bit. You are well past eae at that point.

200%? I guess that's a MS2 thing
my ms3 is I believe 9 for mapdot and 5 for tpsdot

I think all my dots are percentages in MS2, at least MAP and TPS?

I also have some noise on my TPS line, and I use 10% filtering under general settings to smooth it over a little bit. Were it a cleaner signal I could push it harder, still want to make a new engine harness, one where everything doesn't run in a giant bundle with all the fuel and spark signal wires...

But that's just one reason why these things can be so different on two identical cars!

18psi, I do "get it" and understand that the absolute values don't transfer from car to car. Too many variables in the base tune settings and engine hardware. However, the shapes of the curves, general patterns/ratios between settings, and values in terms of order of magnitude should be similar. It isn't perfect, but is helpful.

I actually found that trying to tweak a Rev supplied basemap to match my setup was very frustrating... Worked much better starting from a brand new tune, then applying the tuning techniques from this forum and other sources.

With my MS3, I am currently using values in the 30-50 range for the low end thresholds, but could go as low as 5-10, since my inputs are clean, with little noise. However, I am going to try using EAE to handle the low end, since it should be better at accounting for less drastic changes.

Thanks again for the great discussion.

I am not running any AE. I cannot see any improvement (at least not on a warmed up car) with the AE added to the EAE.

Here are two logs where I tried to duplicate a throttle punch in 4th or 5th gear, while traveling at about 3100RPM and 45kPa.

Without EA:
https://www.miataturbo.net/attachmen...1&d=1432846092

With EA:
https://www.miataturbo.net/attachmen...1&d=1432846092

In both cases, the EAE adds about 75% to the pulse. In the noAE case, the AFR drops (almost straight) to the target AFR and stays there. In the yesAE case, an additional 6% fuel is added for a very brief time. The AFR drops (straight) to a point richer than the target AFR and then rises back up to the target.

In additon, if you look at the rise of the EA vs the EAE, they go up together. Within the time slices of the the log at least, there is no indication that the EA is responding sooner or more quickly than the EAE.

Perhaps there is timing delay in the plots due to smoothing settings in the MS, but it appears to me that the TPS, MAP, EAE, Inj PW are all moving together over a 50mS or so interval.

To me, this means the EAE is handling the punch very well. Would 3% AE be more perfect?

I may be missing something in my evaluation, but that is how I have come to the conclusion that, on the MS3, EAE is sufficient by itself.

I love that you always post up log screen caps, I really need to stop forgetting to do that haha

It looks like you have the same really slight lean spike on throttle tip in just like I do.

You're just ok with it and I keep pointing it out :D

On another note, regarding you not using AE: I was really curious about this and decreased my AE by half yesterday , and didn't see much difference, so I think you're right, little if any of it is actually needed if you tune EAE fat enough

Right now I'm just nit-picking things because I'm bored (lol), but I'm trying to balance out "throttle lift transients"...fast vs slow. Pretty sure it's all about SFW correction, but during steady state (on the freeway mostly) in top gear I'll give it less throttle and it will "feel" more abrupt then when I'm letting off fast and abrupt during a quick run through the gears taking off from a light. Then it will ever so slightly "rev hang" and dip pretty rich sometimes between shifts. Other times it will hold the AFR so steady it's almost creepy :giggle:
I'm almost positive the rev hang is a result of the afr rich dip.

I think I need to flatten out my SFW EAE a little bit and increase my negative AE.

Thoughts?

On tip-in, if you are talking about the AFR Error curve, then I contend it is not real, but rather a time lag between MAP and the exhaust gas getting to the O2 sensor.

That is due to the fact that my AFR target is richer at higher MAP. So upon increase in MAP, the O2 sensor is reading the 15.2 AFR that the engine burned at 40kPa, but the MS is, at the same instance, reading a MAP of 100kPa, knows that the AFR should be 12.8, and so shows an error of 2.4. If you look at the actual AFR, it never rises, only falls.

The experiment I did to verify this hypothesis was set all AFR targets from 13 to 100 kPa to 14.7. Then the spike in AFR error did not occur during fast throttle stabs up to 100kPa. I think I posted a log of that in the first couple of posts.

On the variability on lift, I think it is those crap ID1000's you run. :giggle:
Which is to say, I haven't a clue.

dropped AE cut way down and no longer the variation :party:

Actually the thing I learned with that, is exactly what I already knew:
"perfect" on the gauge =/= "perfect" feeling on the butt dyno

If I get it to cut fuel perfectly as soon as I lift, the car is jerky and feels really stupid. If I get it to slightly dip rich after throttle lift, it feels great, OEM-like.


On another note:


I'm having a problem with EAE (I think) for the first 5 minutes after cold start if I drive the car immediately after starting. Basically it have severely lean tip-in.

As soon as car is over like 90-100CLT it's perfect.

I've tried lowering SFW CLT correction at cold, seems to help, but then I get serious throttle-lift richness until it warms up. I've tried raising SFW CLT correction, and it felt like crap today.

So is the only solution to just reduce SFW CLT correction when cold like crazy and just ignore that it's rich on throttle lift? Or am I missing something? Raising ATW CLT correction seems to not do anything at all to help this. And car runs perfect as soon as it warms up even a little bit.

18. I have experienced what you are talking about as well. As far as immediate driveoff, which I do, I can correct to good drivability by running ASE for a bit longer. But, I do still get some leaning of AFR when taking off until 160* or so CLT. It goes to about 16:1 AFR (from 14.7 target). There is no drivability issue, only sensor shows lean.

I did bump my VE cells just off idle to try to tame it, but that didn't make a big difference, and I don't like to compromise full warm tables for low temp perfection.

It does not seem to be an EAE issue, but I'm not sure what the issue is.

I think I will try adding some AtW RPM correction in the 1000-2000 range. Thing is, once warmed up, the phenomenon goes away.

Update tonight, after starting this afternoon when CLT will be about 90*.

Ok, feel free to call me crazy and tell me why I'm wrong, but:

I think we may be mistaken about how sfw works.

Today I inverted it, and the car not only feels just as good on throttle lift, but actually looks much better on the gauge too (not nearly as much of a dip on lift).

Here's what mine are at the moment:
https://www.miataturbo.net/attachmen...ine=1433291079

Now, I've yet to test this tomorrow morning when it's cold but with the hot car right now it feels great. Even better I think than previous curve.
Here you go, finally:

https://www.miataturbo.net/attachmen...ine=1433291079

I'm still working on these btw. They feel real good now, and I think I got em really close to perfect.

I recently bumped up my "dead zones" so that it triggers even less, and am using low values and dialing in the rest similar to David (who doesn't even use AE).

Hope that helps.

Now I'm trying to figure out if my SFW CLT enrichment might be off. Might set it to all 100's for now, since I'm not sure if more = it will pull more fuel on lift, or more = it will pull less fuel.

Sometimes I overthink this and have a brain fart lol

On the SfW, some people slope it up, some down. I can attest that I tend to go lean, not rich, on throttle lift, but NO CARE because no load on engine and no drivability issues. And, I'm moving to a part of the AFR target area that is set lean (as well as the VE).

So, to me, it is still doing what it should. I think the code is pretty simple. Wall Fuel X SfW is going to tell MS that less injection is needed, so open time should fall. And, it does make sense to me that at low kPa, the puddle should indeed go to zero.

Let me know how the new SfW slope works for you.

Yes absolutely. I guess the difference is how your main VE map is set up, because mine does exactly what yours does, except with this "inverted" setting. The car is very smooth on throttle lift, and doesn't dip insanely rich. Starts pulling fuel slowly, and then overrun takes over.

If it stays this way on cold starts too I think I'm set :party:

Next will be rpm correction, I am kinda thinking this stuff is not needed at all past about 3k rpm. But gotta test and see

Ok so to follow up, nothing I have tried so far made the weird lean tip in on cold start go away.

Until

I re-did the injector timing table. Now I can get it to peg stupid rich on cold start and stay rich during warm up throttle tip in. This of course also threw off my tune to the point that I would need to re-tune the car if I want to use this current table. But I won't derail this thread with that, new thread incoming soon.

My results were the same, to the point that I had concluded that the issue was not related to EAE or VE. Looking forward to your report on the effects of adjusting injector timing. Especially considering the comments in Ryephile's thread that inj timing changes affected mid-cruise torque by 10%.