Yet another EAE rpm dependency question

[mrx79]

Hi,

i just want to give EAE a new chance, since i had tuned the EAE in the past quite ok, but had some trouble with the idle and transient from idle rpm to cruising rpm calibration. Therefore i decided to switch it off again and stick with 50/50 MAP&TPS AE for a while.

But i think (especialy for my setup/driving style) normal AE has the huge drawback of just looking once at the gradients and then stick with it. Therefore i want to switch back to EAE

But back to EAE and the intended question:
When it comes to rpm dependent tuning, you read two different oppinions:
1.) Adhere-rpm dependency whould increase with lower rpm's (physical reason, at low rpm, air velocity is lower, so more fuel can stick to the walls - logical!)
At the same time, StuckFrom-rpm should decrease at lower rpm's (same physical reason as above)

2.) Adhere & SuckFrom-rpm should go down as closer you come to the idle rpm to avoid EAE oscillations in idle.

As i had problems with EAE stability in idle (especialy when there are load changes in idle) i went with solution No. 2, but i'am not quite happy with it, as it just cures the symthom and gives a very steep gradient at rpm's between 1000-1500rpm where the transient to normal rpm dependency in cuising is.

I had both rpm dependencies to 50% at rpm's around idle and 100% at cruising rpm.

now the quesion: How have you done it? and how did you solved idle instabilitiy issues?

Best regards

[mrx79]

Noticed this while scanning my log. Can someone tell me why i'am getting no EAE here at this load jump?
The others seem fine, even with a lean peak also, maybe need some finetuning on the EAE or is this already the part which can only be covered with normal TPS based AE?

Here is the Screenshot from the measurement below:


Hope someone can give me a hint.

[16vboost]

Before you can tune the RPM dependent EAE behavior, you have to get the MAP dependent EAE behavior right. Low manifold pressure means fuel evaporates faster (think of water boiling in a vacuum). See the graphs below:



Also take a look at the values. The highest I go is 20% on the Added-to-walls and 9.1% on the Sucked-from-walls. I've found that you can get unpredictable EAE responses with very high values. Once you tune these two graphs to give you the kind of EAE response you need at a constant RPM (say 2500 RPM, 5th gear) then you can get into the RPM adjustments. You'll find that you'll want more EAE at lower RPM and you won't have any problems with EAE oscillation.

Once EAE is tuned the best it can be, you'll want to add TPS-based AE for fast throttle transients.

-Alex

[mrx79]

Before you can tune the RPM dependent EAE behavior, you have to get the MAP dependent EAE behavior right.

Hi Alex, thanks very much for your reply. I know, and i will tune rpm dependency later, this log above was my previous EAE calibration, but i was not totaly happy with it.

I also know my Adhere Values are high, but as you can see from my log's i definitly need more fuel.
But i thought about your approach and it seems logical to me, also reducing Suck-From-Walls values with higher MAPs will also give me a bigger difference between Adhere and Suck, which should make the mixture richer.

I turned it around similar to yours and see tomorrow what i get.

Once EAE is tuned the best it can be, you'll want to add TPS-based AE for fast throttle transients.

So you think my lean peak's can be covered with EAE and without usage of TPS-AE?

I will keep you updated.
Best regards

[16vboost]

I also know my Adhere Values are high, but as you can see from my log's i definitly need more fuel.

High Adhere-to-walls values don't necessarily get you more EAE correction. It's the change in Wall-fuel that gets you the correction. You should plot wall fuel along with your EAE correction to see how it works. It's the change in MAP from a spot where there was low wall fuel to high wall fuel that causes the extra injection.

So you think my lean peak's can be covered with EAE and without usage of TPS-AE?

No, not at all. I always had a pretty sharp lean spike when getting on the throttle with EAE only. A quick shot of TPS-based EAE works great to fix that lean spike. But get as far as you can with EAE only first.

-Alex

[mrx79]

Hi Alex,

of course you are correct, i've tried your approach now and it looks promising.
Today in the morning i had:
- overly lean mixture at letoffs (thats good because i could not get rid of these rich peaks all the time) but it comes together with your approach of more suck-from-walls at low MAP's.
- and overly rich mixtures at TipIns, fast WOT TipIn's still have the lean spike, but i think that can be cured with TPS-AE later easily.

I'am not in the approach of lowering Adhere (like you saied, my values are too big).
And i think it could be that i end up in the region like you have, but i make small steps and improve while driving to and from work.

I will tweak a few days and then post another log for you guy's to look at if you like.

[16vboost]

Sounds good, post another log when you can.


-Alex

[mrx79]

Hi,
as promised, here are my latest log from today with the corresponding MSQ.

From whe i see i have 3 main issues remaining, where i need a little push in the right direction:

1.) Normal Upshifting (first falling load to max. vacuum, then shortly after medium load increase)
First it's getting very rich at tipout and then overly lean at tipin.


2.) Positive medium-fast load change to aprox. 80kPa
First it get's very lean but total EAE is active too long (following rich peak)


3.) Normal letoff/tipout to max. vacuum
Initial rich, but then very lean -> EAE should be stronger (but injection is almost off) but much shorter, i think it's too much too long, which causes a total wallfilm removal even if wallfuel is still falling


4.) I guess during medium normal transients it looks quite good, maybe a little too long EAE and initial EAE could be a hair more

[mrx79]

As i have my next chance for a testdrive tomorrow in the morning, i decided to:
- insert another brakepoint for 15kPa MAP and delete the 150kPa brakepoint for it.
- Increase the suck-from-wall at 15kPa accordingly to my current line shape
- Decrease the adhere-to-wall at 15kPa and accordingly (have now also around 1% here) and reducde the max. end of the line to about 21%
- Increased the Adhere gradient at low loads up to 40-50kPa to give it a bit more initial EAE.

My values and shape is now very similat to your's Alex.

Hope this will bring me in the right direction of shorter EAE interventions, but still i'am not sure about the Upshift problem's, but one after the other.
Also a big problem is driveof with a small TipIn, then LetOff together with falling rpm due to clutch engagement and then TipIn to accelerate.
You can find severyl of these in my measurement above.
I always get two big lean spikes... hope they will get smaller when the EAE interventions become shorter.

Also i will try to finetune the VE at very low kPa's a bit more, as it seems i'am a bit rich there (but injection linearity becomes a problem there already) but this might be a reason why the EAE at very low kPa's is off by some %'s

I will keep you updated.

[16vboost]

Looking good man. My comments below.

1.) Normal Upshifting (first falling load to max. vacuum, then shortly after medium load increase)
First it's getting very rich at tipout and then overly lean at tipin.

This looks fine. AE will take care of that. Notice that you had a TPSdot of 219. Plenty to trigger a good AE response.

2.) Positive medium-fast load change to aprox. 80kPa
First it get's very lean but total EAE is active too long (following rich peak)

Also looks fine. AE will trigger here TPSdot is 437. You may actually have a bit too much EAE. Load goes from 21 to 86 kPa. You may need less ATW at 86 kPa. (maybe)

3.) Normal letoff/tipout to max. vacuum
Initial rich, but then very lean -> EAE should be stronger (but injection is almost off) but much shorter, i think it's too much too long, which causes a total wallfilm removal even if wallfuel is still falling

Yes, maybe EAE is active for too long. In other words Wall Fuel takes too long to come down. Try much higher SFW values at low MAP (19 kPa in this case). But sometimes, you'll have EAE at 0%, injectors off, and you'll still go rich for a bit. This is OK. The best you can do is get rid of the lean spike after. Again, Wall Fuel needs to come down, either by pulling it down faster, or not making it so big to begin with.

4.) I guess during medium normal transients it looks quite good, maybe a little too long EAE and initial EAE could be a hair more

This looks as close to perfect as you can get. Low TPS dot so I wouldn't expect AE to do much here. But it doesn't need it. Don't try tuning this any better or you'll make it worse.

Also a big problem is driveof with a small TipIn, then LetOff together with falling rpm due to clutch engagement and then TipIn to accelerate.
You can find severyl of these in my measurement above.

OK 2 things.
1.) ATW and SFW RPM correction is wrong. At low RPM, more is Added To Walls. You need more EAE at lower RPM. So instead of 80% at low RPM, try 120% (maybe keep SFW RPM correction at 100 for now)
2). TPS-based AE will help here also.

Here are some tips for AE tuning. Set MAPdot vs TPSdot blend to 100% TPS (EAE is already MAP-based).
Set the Accel time to .1, Accel taper time to .1, and end pulse to 0 (to start with a really short AE pulse)(you can tune these later)
Try your Accel TPSdot threshold at 40%/s (based on your TPSdot noise in your logs).
Leave Decel fuel amount at 100% because EAE takes care of this between shifts.
Leave your cold accel settings for now and only tune on a warmed-up engine for now.
Tune your TPS-WOT curve. Start with all 90% or something conservative at first. Read up on this. It is a powerful feature for AE tuning.

Keep it up. Good luck.

-Alex

[mrx79]

Thanks Alex,

it's pretty decent now. I have no new log right now, because i'am fiddeling here and there a littlebit.
Still have some rich-then-lean peaks after TipOut's, but nothing seriously. Try to fight them with a bit higher low load SFW.

Tried activating 100% TPS based AE today with great success. ~1,0ms was a bit much and i had to increase TPSdot threshold.
I will reduce the TPS AE tomorrow to about 0,5-0,6ms and see if it's better, but initial lean spikes are completely gone, car drive's awsome with it
Strangly i got TPS-AE trigger but in the log TPSdot was way below 45%/sec. (my choosen thd.)
I've attached two pictures and a log according to the pic's, have you experienced that? Is the logging maybe missing some data which the code sees?

And currently i fight with stalling right after cranking. Propably rich misfire which cause the stall, because EAE almost doubles the PW right after Crank to run.
I extended the EAE RPM brakepoints down to 150rpm with fine resolution below idle speed and hope to bring the WallFuel up enough to reduce that huge PW spike after crank.
Also increase the Crank Speed a bit to keep it a bit longer in crank mode.

Wish me luck ;-)
Best regards

[16vboost]

You can probably raise the TPSdot threshold or lower the TPS lag factor to avoid the unwanted AE triggers. You may still have too much AE fuel. You could shorten the duration of the AE event or lower the value in the AE table. But it's better to be a little rich than too lean when you need AE. I'm sure the car drives great the way it is.

-Alex

[mrx79]

Got it with a little less AE.
But i find TPS AE triggering regularly from time to time even if the TPS is not moved and the TPSdot is much below the calibrated threshold.
I see only two reasons: the Datalog is missing something or the code does a false triggering.

Also the negative rich-then-lean peaks are the last remaining issue with the EAE.
But the car is now in the garage over winter, so i can only start new things next year.
But i will post my last MSQ and corresponding Log later.

Thanks very much for your hints... they helped a lot.