NefMoto

So.. im working with an 032HN ecu, ME7.5.
I setup my fueling based on this method discussed here in the forums, and i was very happy with it.. atm i was running about 3CF of pulled timing and everything worked 110%.. fueling was SPOT ON without delays.

NOW, i modified the tune for E85, and everything seemd happy, but, when i was done with the timing, and the CFs dropped to 0 (zero, as imo them should be using this fuel), the KFLAMKR/KFLAMKRL maps deactivated completely and my fueling solution was totally scrapped.. why? ???

i was under the impression setting the first column to 0 would make it allways active and use this map as a baseline, but from what i can find in my recent logs this is NOT the case.. or i am completely missing something here.

i think prj mentioned something earlier here:

and yeah, that statement seems 100% correct.

so.. is there a way around this?
..or will i need to setup my fueling baseline using another method.. perhaps it possible to use the KFLF map as a baseline? can it be used for this?

any input would be appreciated =)

prj was saying that the axis didn't even need to be modified because the first row would get used when retard is from 0-2.25* which is the opposite from what you are saying.

What variables are you logging?

well.. im thinking the map are not even looked up unless knock occurs.. hence the first column could be 0-3 it wouldnt matter.. oooor im missing something here.

i was logging AFR using vag-com at the time.. "requested" and "actual", load%, CFs and rpms, and also using a standalone WBo2.. same results.
it is setup to ~12afr at a wot-pull, but once the CFs drop to 0, i loose the fueling and end up at 13.2 somewhere all the way to redline, which basically suggests it runs on LAMFA enrichment alone. (ive deactivated the BTS for troubleshooting purposes).

thing is, adding 2 degrees of timing and causing 2 CF on purpose cures it and then it follows the baseline around 12ish as it should.

I also experienced this exact symptom when I attempted to run KR fueling, without any CF's my lamfawkr = 1. I went back to BTS fueling as I was running Maestro and didnt have access to anything to further investigate why this happend.

There is however 2 KFLAMKRL maps in my binary in difference from the 2.7tt binarys.

518M ME7.5

What do you have for CWLAMFAW and how is KFLAMKR setup?

Why not use ME7L and log variables from this function?

i would love that, but atm my instrument cluster blocks the k-line so i cannot use the ME7L, i will run a separate k-line directly to the ECU later on and do some more in-depth logging.. just trying to get it running smoothly atm  :)

just as Bische wrote above, seems it ignores the maps unless CFs are >0.. ugh.. major setback if this is the case..

CWLAMFAW sais "0", did i miss something here..??

KFLAMKR is setup as in attached picture.

Not my experience on 7.1, can you post your bin?  I have run low boost race fuel and e85 and everything worked off the 0 row like normal.  Possibly a wideband thing?  What is requested AFR doing?

requested AFR is requesting only 0.93 something.. while it should stay close to 0.8.. and as soon as a CF is present it follows the KR path again.. annoying.
actual follows requested, very accurately.. i just wish requested would follow the "-0.0" row in KR.


attached bin is the base bin im using, before i started messing with it.
KR-related tables in post above.. and yeah, its a wideband ecu.

is my setup of the tables wrong? the KFLAMKRL is the multiplication factor for KFLAMKR, right?
(im thinking i perhaps swapped the maps in my head.. and that KFLAMKRL should be the AFR, and the KFLAMKR the factor.. not that this should matter, but perhaps it does..?)

KFLAMKR is the factor for KFLAMKRL.

KFLAMKR (load, rpm)
KFLAMKRL (~KR, Load)

shouldnt KFLAMKRL be KW, Load with KR in the cells,
and KFLAMKR be load, rpm, with KR-factor in the cells?

so, wishing to have 12.0 afr, i could set up KFLAMKRL in column 0.0 to 14.7.. and then KFLAMKR to be 0.8163.. am i understanding things correctly?

this "load", can it be over 100%? else i could see where things fail.. as i start my maps at 95%.. the original map is never over 100% in these ecus.. axis should be located at 0x18288 in the bin, 8 bit, with X * 0.75 as conversion.. is this even correct?

KFLAMKRL (load, ignition angle offset for knock and temp) - lambda
KFLAMKR   (rpm, load) - weighting factor for KFLAMKRL


Not sure what factor your are using, but KFLAMKRL is naturally lambda with a factor of 0.007813. If you prefer AFR you can view it that way with the proper factor (*14.7).

I just put target lambda/AFR in KFLAMKRL and use KFLAMKR to enable/disable the function basically (0 = off, 1 = on). I already use LAMFAW for slight enrichment prior to LAMKR and the interpolation between axis values provides a nice transition.


Yes, it is engine load (rl). Never over 100? That can't be right.

I'm not sure the location in your file, but both maps share the axis, it is 8 bit, and 8 bit load factor is 0.75.

hmm.. this.. can KR only be used to enable/disable? 1/0.. or can it be used as in my example above to scale the KRL output, 0.95 for example..?

the axis seems right.. rl.. and all conversion factors are right.
the original axis reads: 20, 40, 60, 70, 80, 90, which is why i was curious..

thanks for all your input on this, its greatly appreciated  :)

KFLAMKR is used to scale KFLAMKRL, it basically provides a percentage.


That doesn't look right compared to the M box axis, but it could be. M box increments by 20 and maxes out at 140 iirc.

that was my understanding also.



regarding the "rl" axis, the definition i based this upon could be wrong, ive noticed a few minor errors earlier, but it seems close to 100% accurate.

map KFLAMKRL: located at 19051, 8 bit, 6x6, conversion: x * 0,007813
Y-axis: 1908A, 8 bit, units: KW, conversion: x * (-0.75)
X-axis: 18288, 8 bit, units: % rl, conversion: x * 0.75

maby there IS a difference in how this map is looked up between the 7.1 and the wideband 7.5?
perhaps it could be "reverted"/altered in the code.. im very novice to code patching and c167 asm, but i would love to use this fueling method as it has no delays and the accuracy is spot on.

032HN and 551F appear to have the same routines that call these tables.

then why wont it enrich further in the 0.0 column.. while when the least bit of CF is present, it enriches just as it is supposed to.. is there something limiting the enrichment from KR when there is no knock present?


matchew: any pointers as to how i find this routine? i have dissassembled the 032hn bin..

If your definition and changes are correct, then the only thing it could be is that the conditions that enables the entire LAMFAW function are not being met.

This would be much easier to diagnose if you could actually get some real logs.

It is very easy to find LAMFAW by just x-ref lamfa_w.

found it..


what do i need to log? im going to give tonys logger a go as seems to be able to connect atm.

lamfawkr ? where do i find this one..? =)

hmm.. the axis "rl" for KFLAMKR and KFLAMKRL should be at the same adress, right?

i had a look at nyets map-pack for the mbox, and tonys tune using KFLAMKR/KRL-enrichment, and in those definitions the KFLAMKR map has a different reference to rl then the KFLAMKRL..

from what i can tell, reading the funktionsrahmen, these two share the same axis, "SRL06GKUB"..

Same axis 0x18292 in M box.

Look, set KFLAMKR to 1.0 across the board and KFLAMKRL to your target lambda.
Then test and log.

One map at a time.

In the screenshot you posted, you have KFLAMKR and KFLAMKRL backwards.

KFLAMKRL should contain desired lambda and KFLAMKR should be the factor.

Also, make sure that you have the "Major Order" set properly in your XDF definition.

phila_dot: yeah, i think the "major order"/"populate by row/column" was set backwards in this case, i have rearranged it now.. and i will try to log it tomorrow.

actually, i set the KR/KRL maps up backwards on purpose ;D
..but the values should still read lambda OR factor around 0.8-1.0, so it should work just fine this way.
making the KFLAMKR your 0 CF "base map" based on rl and rpm.. and then just multiplying it by factor when CFs rise.. instantaneous enrichment in predefined % no matter where you are in the map..

..if i get this to work with the 0.0 row, this would be THE way to set up fueling in these ecus.. no delays.. having knock enrichment intact, the BTS working as intended and then some lamfa for pre-emptive WOT-enrichment only.

Ahh...ok. I see what you're doing.

You may want to follow prj's advice if the major order doesn't solve your problem.

so.. after some digging into this i find my approach of setting the maps up backwards didnt really work.. i took prjs advice and now i got it working 100% after some flipping of major order etc.

the correct way to do this is as per FR and setup KFLAMKRL to the target lambda, and to use KFLAMKR to add/retract fuel in needed regions.

What does 'flip major order' mean? Im not using tunerpro, is it something done to the definition?

Yes, if you define the x and y axis backwards then the major order needs to be changed from ROW to COLUMN because otherwise the z data will not be laid out properly. I don't remember if there is an equivalent in WinOLS.

it has to do with how the data populates the cells.. if it fills in from left to right untill the grid is full, OR if it fills in from top to bottom.. =)

Thanks :)

Problem solved :P

I have also some problems with KFLAMKR/L based fueling. I have set KFLAMKRL to what i want but KLAMKR wont work as it suppose to. I ques there are something wrong with my map pack. Is KFLAMKR map in 190BC and rpm axis at 1A3F0 in this file?