NefMoto

Hi guys,
I've bean reading this forum for quite some time and it's has bean very helpful. However I couldn't find the answer for the best way to tune for loads above 10ms ? In all the definition files for the me3.8.3 the maximum value for the load column is 10ms (KFLF and KFZW ), but the max load value is 12.750(FF - hex), so how does the ECU calculate fuel and timing for values above 10ms, does it extrapolate or use the last value (in this case the value for 10ms load ) ? What is the best method to tune for loads above 10.00ms ? I was thinking about rescaling the load axis so the last column is max load (12.750) and the one before it is 10.00, this way the ECU should interpolate between the two values.

Based on a plenty of guys said you don´t want to reach 12,75ms  ;)

In this point, 10ms will be enough...  ECU will use same value above 10ms

The only interesting point I can see, is if you want to make some kind of protection when actual load is bigger than you expect... in this case you use the last colum just to maker a huge enrichment and advance decay to try to protect engine...

Yes, I've read that if you hit loads 12.75 or more you have to downscale KHFM which leads to changing/scalling of the ignition/fuel maps. I've seen some tunes where the tuner just capped of the load protections and tuned the car for the boost it runs (load above 12.75ms) i.e. the fuel and ignition values at 10ms load were set for the boost/load the car run, but in my opinion that's not good since you bypass the protections . Another drawback is that with this method that ignition and fueling will not be optimal for loads that are really around 10ms  since they will be setup for high boost/loads.

I'm interested in rescaling the load axis since I have the following problem. I'm running an EBC which has two boost settings 1.0 and 1.2 bar (currently, I'd like to go 1.5 bar but after I rescale KHFM) . At one bar I'm hitting loads of ~10ms - 10.5ms, when I tune the fuel/ignition for 1 bar it is not optimal for 1.2 bars since at 1.2 I'm hitting loads ~11.5-12ms. My idea is to have the last load position say 12ms and the one before it 10ms so I can tune the ignition and fuel to be optimal for both boost settings (I assume load values between the two will interpolate?). Does this make sense or am I shooting in the wrong direction ?

P.S. My current setup is a t3/t4 turbo, bigger MAF (calibrated), 380cc/min injectors and FMIC.

Well... you´ll be losing resolution in lower load areas, that you are most of the time

Overspeed, thanks .   I guess when modifying/rescaling an axis you always loose a bit of resolution . So in this case would it be better to just down scale KHFM so the target boost/loads fall into the 10ms region ?

Looked at down scaling KHFM and it looks like a lot of maps must be changed to get the timing and fueling right .
KHFM,FGAT0,KFLF,KFZW.0-2,KFZWS,KFZWES,TLRAN.0-1,KFTLWS.0-2,TLST.0-1 and by the looks of it a few others. Looking at all this maps I get why some tuners just set the protection maps to FF and tune the ignition and fuel (rows at 10ms) to the max boost/load  ::) .

Don´t forget to mention that a lot of maps interact with other (IAT advance correction for example).

My opnion is the AFR (or advance) you want to use for 10 to 11ms will be not that diferent at all... the old "better be safe that sorry"

How are you going to control boost if the load is at 12.75?
If those "tuners" used MBC's then they are a joke.

Basically there are two things you do not ever want to happen on this ECU:
1. Hit 12.75 load.
2. Max out the MAF.

My mistake, I said tuners should have said tuned files. Your right some use MBC others EBC . I also currently use a EBC since I like the options it provides - 2 boost levels switched at the press of a button, easy and fast adjustment of the boost (if I want to modify them), over boost function, overboost warning and protection etc.

You can limit boost with your right foot.

Then scaling or modifing KHFM and changing all it related maps in those old Motronic ECUs could generate a new large and polemic thread.  ;)  ;)  ;)

Bigger MAF + MLHFM =>  Right g/s readings, but load will increase in proportion (and shall be more than 12,75ms limit)
Bigger MAF + MLHFM + KHFM => Rigth g/s readings and load will be lowered again to "normal" levels (10ms axis)
The logic can bem simplified to
  Ld = K x MAF/KHFM
If you raise MAF value (in g/s by using rigth MLHFM) then Ld value will raise it too... and shall touch 12,75ms/umdr
to keep Ld in "sane" values you raise KHFM in same proportion...

I have tried to do many tests to get a good tune but the results are not satisfactory.

I have 72mm maf and 350cc injectors bosch red.

Can anyone give me advice?

Thanks in advance.

It's pointless only scaling KHFM, you then have to remap the entire ECU for it to work halfway decently, which takes absolutely AGES, even with an emulator on the dyno.
Much easier to scale KHFM far enough that you don't hit 12.75 load and then scale FGAT0 to compensate for the rest.

EBC's are for people who can't tune, and when the ECU is tuned correctly you modulate the amount of load you want with your right foot.
Anything an EBC does the stock ECU always does better.
Only point in them is when you have a NA car that you have turbocharged and are using the stock ECU, but most work like cr*p. Prime example apexi, which is utterly useless and creates huge amount of artificial turbo lag…

As for getting a good tune - need proper data logger with boost, rpm, wideband, VCDS, knowledge how the ECU works, and of course a realtime emulator, else it's impossible to get it tuned right.

But wouldn't only scaling KHFM and FGAT0 throw the igntion timing off ? In this case you still would have to retard the ignition tables in order not to have a lot of knock retard, question is could you get away with it if you only scale KFZW tables (log timing and modify these tables accordingly) .

Yes of course your ignition timing tables will be useless.
That's why you have knock logging, headphones and dyno.

It is NOT easy to tune these ECU's to high power.
I have a few cars running 310-350hp on these ECU's since years incl. full boost control. But it took me two weeks to calibrate the first car, and I have a lot of experience with tuning Motronic.

This ECU is really the wrong one to start with. Do an ME7 conversion and make use of the knowledge on this forum, it will be MUCH easier for you than trying to do anything with M3.2x/M3.8x. Impossible without realtime emulator and proper external data logger…

The way I did it was emulator instead of 29F200 chip. Wideband data logger with inputs logging AFR, Boost, RPM, MAF voltage.
Logging via VCDS for either Boost control or Knock control.
I converted the cars to a HFM5 R32 sensor from the old HFM2, incl. rewiring for the 5V reference…
Then many many hours tuning fuelling, timing, boost control and also COLD START, it was completely broken after messing with KHFM and FGAT0…

It's a proper nightmare tuning this, especially when doing it professionally for a customer and there can not be any glitches in the software.

I think that old Motronic Achilles heel, is it poor tunability, basically slow & poor data logging. Maybe I am nuts, but I think that it can be treated like TEMPLATE allowing it work with 10ms+ loads. After that it can be adjusted for proper tuning.

You shall try tuning an M2.10.4 to see what is an slow and poor unit lol !

I tune M2.1, M2.3 etc.
For those I wrote custom firmware with own logging protocol.

I have 15hz data logging with 30 variables.

Never bothered to learn the 80C196 chip's assembly though, because there is only one tuneable engine with this ECU and it's 1.8T.
Not worth it investing the time.

Must of missed this post. Why would load increase in the first case if the g/s readings match after scaling ? I've scaled the MAF and the g/s readings are very close when I logged them before/after, load seems to much also (without changing KHFM).

Here is a graph of g/s before/after, I'll see if I can dig up a graph of the load before/after .

let´s resume it:

The goal is to lower load to avoid reaching 12,75ms:

1- If you change MAF for a big one:
   The g/s readings will be underscaled => load will be smaller proportionally in area increase

2- If you want to keep g/s correct:
    Change MLHFM + g/s will be Ok again, but load will increase to same load before MAF changed

3- If you want to lower load and keep g/s readings rigth:
     Change MLHFM and KHFM  to compensate it.


(By the way, get an Golf done by another guy last week that reach 12,75ms/umdr in 3.500~3.800RPM but worked 100% Ok)

Exactly, so to recap - if say you are hitting the MAF limit but not the load limit you can just modify MLHFM (after installing a bigger MAF) but if you are also hitting load limit then KHFM has to be changed also in order to bring load down below the 12.75 limit. Important thing to know is that after modifying KHFM one must modify the spark/fuel tables .

I've also seen 2-3 cars hitting the load limit but running fine,  on those cars KFLUL.0/KFLUL2.0/KFMLDMX ect. where maxed out, which in my opinion is not a good way to tune them.

Friends sorry to take this thread to life again, but I was investigating some info and have a doubt, anyone knows which are the protection maps that was commented here?

From memory -> KFMLDMX ( airflow kg/h) and LDSMXN (load)

Really thanks, I have saw it but do you know what changes in ECU or Engine behavior are produced when we reach these values ??? ??? ???
Heavy AFR enrichment? timing retard?

From memory too

KFDLUL will cut N75% to 5%
KFMLDMX will enrich (alpha-n)
LDSMXN just lower you N75% (Don´t think it´s a limiter at all)

Thanks a lot as usual Junior...... I will look at it.

As I understand it LDSMXN is maximum permitted load, so yes it's a limiter for the max load before the boost is cut i.e. N75 duty is lowered.

for MY undestanding of this:

KFDLUL is a limiter=> above this limit ECU will cut N75 to 5%
KFMLDMX is a limiter => above this ECU will understand MAF is defective and use alpha-n

For this reason I don´t call LDSMXN a limiter, it is more a threshold.

LDSMXN is target filling limit.

A bit off topic, but is there anyway to have boost/load by gear, i.e. 1-st & 2-nd gear lets say 0.8bar 3-rd,4-th & 5-th 1.2bar (for example)

Hi friends sorry to ask this, but anyone have test rescaling the load axis in any map? I mean extending up to 12ms LOAD for better high load resolution.

Yes, I've been running with 12ms as the last load cell for over a year.

Dropout do you have notice some rare AFR behaviour during boost in your car? I mean some AFR sudden steps in engine operation during boost (high loads).

No, I haven't had such issues. Currently my wideband is off the car but since I've put in H beam conrods I'll be retuning it soon and I'll provide logs.


P.S. my goal is to run 1.8-2.0 bars (hopefully hit around 400 bhp) on the factory boost control which means a lot of rescaling of maps.

Nice to hear that Dropout. Just in case you or anyone will like to comment about it, to illustrate my issues there are --> http://nefariousmotorsports.com/forum/index.php?topic=9903.msg87391#msg87391. I have notice that always occur just after boost ramp finish.