NefMoto

Hello,

After learning some basics on a WB 1.8t and simpler S4 setups I tried to tackle on a different car.

2000 B5 S4

8D0907551M Base file

440cc Green Giants
4bar FPR
AWE MAF housing with the Hitachi Element
Stock Airbox
K04 Hybrids from the early 2000s
LWFW
Some exhaust work.

I have spent a considerable amount of time getting the calibration of my WBO2 to be spot on.

Using R, a statistical software, I have made an easier script to help with the Non-linearity of WB02 sensors/controllers, quickly produce plots, and some backfittng algorithms of some tables (KFLF and FKKVS).

I have gotten to the point where my fueling is spot on with my error in requested - actual lambda being negligible in most if not all scenarios.

Some maps I have modified

Air/ Fueling:

MLHFM = Scaled across the range to get idle rims as close to 0 as possible
KRKTE = Same but used generic calculations to get close.
TVUB = to values found on the green giants
LAMFA, LAMBTS, LAM(Ignition) = have changed accordingly to reasonable values.

Load:

KFKHFM = Upper load regions changed to 1 since the stock map leaves these with a .90 input and wanted it to be a little flatter to keep the scaling in KFLF a little less.
KFMRES/K = To help with idle swings and LWFW especially with AC on.
KFMIRL =  in lower pedal positions to request less load. This was done in order to numb the accelerator pedal a bit to increase driveability since too much load was being requested at 25% pedal. Also, more load was requested in the last 10% of the pedal position.
LDRXN = ramps up to 210 at 3k and tapers down to 190 at 7k
LDRXNZK = similar ramp but only to 160 (I don't understand why other tunes I have seen make this the same as ldrxn

Other maps:
KFLDHBN =  RS4 K04 map copied in here.
Pressure ratio Axis = changed based on the ratio of my cracking/stock k03 cracking pressure.
KFZW/2 = Removed some timing where spool up was occurring and pulling 8 degrees of timing to be safe.
Some maps for emissions/readiness

The reason for this self-study and tuning is that the car was originally chip-tuned using a major tuner, but the support does not exist for this type of interface and as well as the stupid rich scenarios on WOT. I realized that maybe the tune wasn't even reasonable for the setup and I am here trying my best with what I have. I know the MAF housing and injectors are not the favorites of the community, but they are what I have and I feel if I can learn to tune ME7 with these, I can do better with more reasonable hardware.

So my issue I've run into now is that the Requested Load/boost seems to be a little bit higher than the actual load/boost on WOT or close to WOT.

Most partial throttle stuff is spot on, but when I go WOT/Higher-pedal positions I don't see them match as well.

Below is one of my latest revisions on fueling and some related Plots

(https://i.postimg.cc/pmc2QTvP/1.png) (https://postimg.cc/pmc2QTvP)

(https://i.postimg.cc/R33vgBCK/2.png) (https://postimg.cc/R33vgBCK)

(https://i.postimg.cc/PLBTcRNc/3.png) (https://postimg.cc/PLBTcRNc)

(https://i.postimg.cc/TLgfXJDP/4.png) (https://postimg.cc/TLgfXJDP)


Thanks in advance and I hope we can all learn from mistakes, or even if it is okay, and if I am overreacting then thanks too.

- DLB

* you are breaking forum rules by not posting logs. Please read the stickies and obey forum rules http://nefariousmotorsports.com/forum/index.php?topic=12361.0title=
* you should be using ECUxPlot for plotting
* compare ps_w with actual boost (which I could have done for you if you posted logs)
* you can always scale down MAF and compensate with fueling to bring load in line to prevent torque intervention

That housing will cause you no end of issues. Your insistence that it is "ok" to use is not supported by the data you are logging.

you're wasting your time with it; all the effort tweaking is going to have to be redone when you get proper injectors and a proper MAF housing.

the notion that "this is ok for now" is insane given the effort you've already put into a setup that just isn't going ever do what you want it to do.

Get your hardware right first. Then do your tuning. Anything else is just asking for pain.

I realized that my current config file for ME7logger does not contain ps_w. So I will rectify that and post an appropriate log when available.

Pain is only temporary. I know that sounds cliche, but the way I see it is if I learn the intricacies of ME7 control using this difficult setup at 60-75% of its true reasonable power potential, then an appropriate setup will be much easier since I would have learned more after struggling and failing.

It has now finally clicked what scaling back the MAF means. The idea is that MLHFM is transformed from stock using a smaller ratio and then use KFKHFM to correct individual areas of MAF readings to get load lines in check. The issue that can arise is that too large of corrections in KFKHFM (greater than (+/- .05 is what I was understanding is a safe limit of the corrections) will hit the ps_w limit. So at this point, I need to adequately log ps_w and go from there.

So the idea is to get the MLHFM scaled (within reason of the diameter) so that reasonable corrections exist in KFKHFM and the ps_w limit is not exceeded excessively. Meanwhile, KFLF and FKKVS can be used to clean up the fueling issues that can arise.

Where KFKHFM can change the MAF readings and ultimately load, my intuition is that KFLF and FKKVS only change the fuel injector pulse by the appropriate scaler and do not affect load calculations. Is this correct?

I had another question on MLHFM but will ask it once I properly get ps_w logged.

-DLB

> KFKHFM to correct individual areas of MAF readings to get load lines in check

Typically, you use KFKHFM to correct individual areas of MAF to get fuel trims in check, since you cannot accurately calibrate KFKHFM directly unless you have access to a flow bench. Fuel trims are a (completely indirect) way to find if your MAF is calibrated correctly, but it assumes your *fueling* is calibrated correctly. Obviously, this is a chicken/egg problem.

In the real world of aftermarket tuning, that just gets entirely ignored, and what ends up informing your MAF calibration strategy (after getting your trims under control) is simply this:

1) avoid the ps_w cap (which is better addressed via the 5120 hack)
2) artificially scale down your MAF to keep load low enough to not have to deal with too much torque intervention issues (which is better addressed by first finding if there are other things causing torque intervention unrelated to MAF scaling and load reading higher than it should)

Both require *again* another round of fueling tuning to get your trims on target (or, more directly, make sure requested fuel during open loop matches any external wideband O2 metering you are doing).

Hope this make sense. As a final caveat: there is always more than one way to skin a cat, and just about every strategy has it's haters.

Also, ditch that AWE MAF as soon as you can. Really. Other than your (laudable) goals of learning the systems, it will definitely cause more confusion than simply finding a housing that is much more linear and deterministic across a wider range of flow.

> KFLF and FKKVS only change the fuel injector pulse by the appropriate scaler and do not affect load calculations

Correct.

Okay, after some PID issues, resolved with the info from prj's thread on it, I have gotten to a place where I have run out of fueling duty cycle (and close enough to my personal limit on the stock MAP).

I have attached a log since my ECUxPlot presets have not really been developed as much as my R prompt.
(https://i.postimg.cc/YhKZy3RJ/Log-2-Fuel.png) (https://postimg.cc/YhKZy3RJ)

(https://i.postimg.cc/68BSBpDc/Log-2-Timing.png) (https://postimg.cc/68BSBpDc)

Some things I see:

I will need to do another refinement of the partial fuel trims.
Some timing can be played with to make it safer/more efficient.
It is bugging me that after the first idle period requested lambda never really truly sits at 1 when it should and can be. Rather it is something like 0.99979. I will check and look at the three fueling pathways and see if any of those are to blame.
I could taper LDRXN a bit more to avoid the possibility of requesting more fuel than exists.