NefMoto

Hello Guys,

I need a bit info about this map. Map id KFLDS.1 and KFLDS.0. Both of same values.

i think is it boost pressure regulation constant.therefore, this is releted with injection. it should be tune via engine load. am i right ?

How can i tune it correctly ?

as well as,my max throttle body position angel is 83 degree, but maximum vertical value looks 16.041. is it strange ? or thats normal...

And lastest rows has a same value.. why ?

thank you

and idea   ::)

What is your raw axis data and axis factor?

Hi,

Please find in attachment.

Thank you for your comments.

really I hate to Motronic 3  :)

You need to click X-axis data source from "Eprom" to "Eprom, subtract".
Does make much more sense that way :)

Perfect Terok!!
Can i use for other maps ? i applied my TVUB map.
it now seems more sense !!

Thanks for your tip.

I have a some new info about that.

lastest rows has a same value..Because, each motronic 3.8x 1.8T engine has a different max throttle angle values. some one 77degree,some one 86.. it looks always this range. Therefore lastest rows should be has a same value.

Another information, This maps generally related with fueling. But i had seen, it affect to ignition. Because motronic 3 iginition maps working for engine load. Engine load related with Turbo Pressure and MAf reading. when i reduced lower Throttle body angele values my cruise retard timings gone out.

I spend a lot of time for my car and motronic 3.8x tuning.

If you have a any experience about that, please share it.

Thank you !

Greetings,

KFLDS has nothing to do with Timing and fueling to me.
It is only related to boost control.
I understand it as requested "boost/load" (since ME3.8 doesn't have boost pressure sensor).

Exactly... these are the desired load maps.

Hi guys
I know this values for requested boost so engine load..
If it desired with engine load- it could be affect ignition. Because motronic 38x using engine load values for ignition.

I don't have a motronic 3 function sheet, therefore I never %100 sure.

Just to clarify:
The timing maps and fueling maps are sharing a load axis that is directly linked to the actual load, not the load requested by this map (KFLDS).

Thank you for your clarifcation. But when requested load higher the actual load I had seen 0,7-1.5 degree timing retard. I solved this issue with load request map. I had reduce requested load and problem solved. This is strange..

Not strange at all...

When you increase load it is natural that the advance is less than before (the original maps use loads smallers than the one in the last colum)... besides, when the pressure goes up the temperature rises and the IAT corrector lower the advance

What's the IAT correction map name ? is this map should be tuned for high advance ignition ?

Thank you

=> Zündwinkelkorrekturkennfeld: KFZWNA.0

Yup, found that.

this temp value is IAT or Engine Temp ? 

We can increase selected area values. I think engine doesn't have a timing retard.

BUT,

i think if it IAT values, when we changed values, it may lead to detonation.engine cannot reduce ignition sufficiently.. Isn't this a risk?

Yes, it is intake air temperature.

Tuning any map is a risk  :D

hehe yes right.. but not always  :P

I'm just working on known maps.  :D

x-axis mirror map

Friends sorry to take thid thread from death but then this Kennfeld LDR-Sollwert KFLDS is directly related to the N75 solenoid as boot control or Injection time?

it´s desired load [ms/Umdr]

Load in earlier Motronic is related to injector time (not actual, but theorical to reach lambda =1) wich is in the end torque


N75% will change to achieve this desired load value, if you put too high values in low RPM the PDI control will throw 100%DC and you´ll have a overshot in some RPM later...

Out of curiosity.. I had a look, the M3.8 .OLS I found laying around has map TV - Kennfeld für LDR // TV - Map for LDR.  TV should be the abbreviation for Tastverhältnisvorsteuerung, translated "Duty cycle pilot control".  If M3.8 works anything like M4.4, the duty cycles in the TV LDR map define the starting duty cycle values before the PID logic is applied.  Basically, if you don't get the TV map duty cycles somewhat in line with what's actually required to meet the target loads in the LDR-Sollwert map, there will be a large system deviation between actual and target loads.  This tends to result in under/overshoot, boost spikes, or just plain unexpected behavior.  It's difficult to get these maps right without proper data logging.

You can also neuter the ability of the PID logic to do it's destruction by greatly reducing the values in maps I-Anteil Ladedruckregler and P-Anteil Ladedruckregler.  Then it's up to you to tune the TV duty cycle map to get you in the vicinity of target load.

I think this map is post-PID WGDC linearization map, it's not pre-PID. It actually translates PID controller result into physical n75 duty cycle which needs to be applied to WG valve to open/close WG to reach target boost. Configuration of this table usually varies depending of WG spring squeeze force. Although in real life most of tuners use the table to reach target boost which is not quite proper way way to do it.

I have read many assumptions about M38x, in some forums says that is similar to M33x and here similar to M4.4, then I will like to know if some one have documentation or funktionsrahmen about it???????

Interesting.  I haven't been able to get a proper disassembly of M3.8 yet, but in M4.4 the PID logic directly outputs the duty cycle to the hardware.  I'm not sure why they would need a table _after_ PID, but only the logic would say.

Eliotroyano, there's quite sparse documentation on these older fuel systems.  Best documentation is a disassembly..

Speaking of that, it seems M3.8 may be running an 80C196 (intel MCS-96) or variant there-of?  Is that correct?

I haven't time to play with M38x firmware and IDA. You are right M383 is based in the Intel MCS96 Processor family. Here I left you all the identified ICs in a 06A906018CG M383 Ecu. The other fact is that those processors was used in many applications in the past and I saw some documentation about it usage and dissasembling it software.

Nice! Thanks :)

Very intersting, the SCL4402v4 is also on M4.4.  This attaches to the external data bus (P0/P2) of the 80C517 (used instead of the 87C196 in M3.8), appears to be a sort of control ASIC -- Appears to handle a lot of electrical control functions, especially anything PWM related, such as controlling the duty cycle of the turbo valve - In M4.4 the program simply moves a data byte to the SCL across the external data bus to vary the duty cycle.

I shall have to see if I can get a M3.8 image to disassemble in IDA, I'm not familiar with MCS-96 assembler but it doesn't look too difficult to comprehend after a cursory glace. 

Interesting to get a look at some other older ECUs :)

I have been playing with M383 and M592 ECUs since some years, well the time that work allow me. Also here I left you the OEM flash content a that ECU of the images if you need it. I also have much more info at home.

Sweet, thanks for that.  I honestly don't even have a M3.8x car, but an M4.4 car, mainly curious about the similarities, learn and share a little bit in the process :) MCS-96 ASM looks interesting too. 

Cheers!

Do you have some pics of the M4.4 Ecu? Can M4.3 & M4.4 ECUs be read & write through OBD2? I have read something about that.

Some pics @ http://volvospeed.com/vs_forum/topic/159506-tuners-rejoice-free-tuning-for-m44/page-3 -- I can take some additional ones if those are not sufficient to get part numbers off. 

- Both M4.3 and M4.4 must be in boot loader mode to read/write the flash, for both, applying voltage to pin B8 before powering up the ECU accomplishes this
- Both M4.3 and M4.4 are programmed via 9600 baud K-line, however M4.3 uses pin B5 and M4.4 uses pin B36 (the later shared with OBD-II).  On M4.3 cars, pin B5 connects to pin 3 of the OBD-II port.  On M4.4 cars, B36 connects to OBD-II pin 7, the normal pin for OBD-II K-line data, and provides both the programming interface as well as OBD-II and proprietary diagnostics protocols (VST, McMess)
- M4.4 has a split address flash arrangement (128kbye flash chip on a 16-bit address bus, a GPIO pin is used to toggle the MSB) and can't be read via serial as there's no way to toggle the GPIO pin while in bootloader mode
- M4.3 uses a 64kbyte flash chip and can be read completely across the serial interface.
- To program in-car via the OBD-II port, one has to somehow get voltage to pin B8.  For M4.3, the OBD-II cable needs to be wired to use pin 3 instead of pin 7 for data.  For M4.4, voltage must be at least 14v, for whatever reason, so most people connect a battery charger when flashing to raise the voltage on M4.4 cars.

Personally I just program them on the bench.  It's less trouble, IMHO.  M4.3/M4.4 cars have the ECU in a basket under the hood and it's very easy to remove/reinstall, unlike some cars which have the ECUs in difficult to access locations, such as tucked behind the dash, etc.  I have two ECUs and simply alternate, pulling one to replace with the other and bringing the other one back to the bench for the next round of tweaks. 

Use this factor on this map - 10,24
and precision - 2

It will show you values in absolute pressure in mbars

It works fine,as requested and actual, ONLY IF the MAF is well calibrated in bigger applications.
No doubt that it works excellent in Stage 1 files

;)

Hi sorry to take this thread from dead, but I was reading and studying and found this info.
Hammersword do you mean that we can use a factor of 10.24 to convert any ms/Umdr Load axis to absolute pressure in "mbars"? X, Y or Z axis??? Also does it can be used for any map, I mean ignition, fueling or boost or just boost maps?

Load is directly related to torque, not pressure...

That´s the same "load/4,266+300 = pressure" in ME7 systems, in normal conditions it´s mathematical next enough for a guess, but is not the real thing.

It will be afected by IAT, ambient pressure, compressor eficiency, ignition timing, and so over...

Thanks a lot for elaborate that. I understand your comment. Then for older Motronic M38x & M59x, that Load (ms/Umdr) is directly related to engine torque, but is it the same measure that we can log using VAGCOM as Load (ms)?????
Finally many people try to find a relationship between Load (ms) and Load (%) or Load (pressure) as a comparation for tuning purposes. That is reason of my previous question about convert or just transform Load (ms/Umdr) in other unit like pressure for the typical RPM/MAP Relative Pressure/AFR table. That give me to my other question, what is the difference as torque, boost or engine load itself found as X or Y axis or as a result in a Z axis table?

Yes,  "ms" is actually "ms/umdr" But you have to be carefull as you have another things thar are measured ins "ms"

The rigth to say is, dont´try to simplify what bosch did, start thinking in load values... they are by definition "injection time for lambda =1", if you enrich (and you will in WOT) the correlation will decay... the same thing if you decrease advance.

Could you care to tell a little more about how decreasing or increasing advance is going to affect load being injection time to reach lambda 1?

Don't overcomplicate things. In this ECU: LOAD = MAFLIN(MAF_VOLTAGE)/RPM*K
Done. None of your ignition advance or torque yadda yadda has anything to do with it.

As for pressure - this ECU does not know what pressure is. It pulses the N75 valve to reach req. filling, it uses feedback control from the MAF instead of using a pressure sensor. Really damn simple system, works great as long as your MAF is big enough, not so great if you exceed it's size.

Anything decent power you have to decalibrate load anyway, as it just goes to FF. This ECU is best treated with the "garbage in, garbage out" method. Never fails to work, and not any different from tuning any other old motronic.

Excuse me, english is not my first language, let me explain again in other way...

Advance will have not to do with load requested and it´s correlation with MAF readings, but when you decrease advance of its ideal point (what was made by bosch in OEM calibration) you have less torque, that´s what I said, the correlation load<=>torque will no remain constant as long as you change advance and AFR.

Try lowering advance and you´ll get less torque evem with same load... try enrich more than 0,75 lambda and you have decaying troque too... thas what I was saying.

Yes, I know all this, but this has nothing to do with this ECU.
The torque model, effective ignition angle and all that came in ME7.

M3.8 and older have no notion of torque, there is only cylinder filling (load). What you do with that cylinder filling is up to you.

Motormoment beim optimaler zündwinkel und hochdruckphase does exist however

Not used for anything really important.

read again the topic in order to get what I was saying:

First I said load => torque ( to make him know that you can´t mess with factors in WinOLS and "think" in boost)
second I said load = K x torque in one specific condition (lambda =1, and obviously advance in MBT)

I just advice him that He can´t keep the correlation if he change too much AFR and/or advance.
To make it more clear:
one example:   Passat 4B0907557B soft version 359157
in KFLDS.0    TPS = 94% RPM = 2000 you have load 7,45ms  if you change it to 9ms you´ll raise torque ABOUT 21%

If you know the torque in this condition (WOT at 2.000RPM) you can get a round value for convert load in torque, BUT will be afected for lambda and advance (because you will be in another point in fuel and advance tables).

that was just my point.


Why say it ?  I Live in Brasil, and here we have 25% ethanol in fuel, it make the AFR raise a bit and make the engine suport more advance without detonation... an 100% stock A3 150AGU will reach here even with all the heat about 160c~165v in dinamometers




 

allmost forget !! Indeed ECU did´t translate load to torque, and the cause is just what you said (there is no torque model as ME7)...ECU can´t know what is the detrimental loss with lower/higher advance, AFR... or even with elevation or IAT (there is correction load tables, correct requested load to keep the same torque output).

We are agreeing in everything...

IN my other post I said:

"in KFLDS.0    TPS = 94% RPM = 2000 you have load 7,45ms  if you change it to 9ms you´ll raise torque ABOUT 21%"

torque will raise about 21% IF K03 would give pressure for this ( I would have picked a higher RPM to keep it clear, but the idea is the same)

No, this is not how an ICE works in the real world. Otherwise making huge horsepower would be extremely easy.
Correlating load and torque is a fairly pointless excercise.

Less talk and more tuning ;)

a little offtopic but it regards 3.8x ,

does in anyway the ECU limits injector timing @ 16.32 ? I am getting mad with these logs.

You are probably hitting max load, not injection timing.

PRJ and OverSpeed I really appreciate your help, experience, knowledge and tricks with us. As you know the main problem with this old Motronic is that we don't have to much documentation. In other worlds with have a lot of variables (WinOLS Damos/MapPack constants, maps, etc...) but we don't know exactly the formulas (Motronic Funktionsrahmen docs) to know which variables are computed to give a result. Well at least we all have some guess or estimation about it.

From your comments (thanks for that) we know that these ECU computes LOAD = MAFLIN(MAF_VOLTAGE)/RPM*K as injection time for lambda =1 or in other words cylinder filling, knowing that these old ECUs doesn't have an explicit torque model in it.

Then how we know what MAP/Contanst/etc...do what in terms of LOAD (ms/Umdr) as a result (Z) or as a variable (X or Y) using the axis nomenclature. That ways how do you know a MAP or Constant manage fuelling (injector pulse as ms), load as (ms/Umdr) or N75 as solenoid pulse (ms).

I appreciate your tips with that.

Regards,
Elio

I fully disassembled M2.x, labelled every subroutine, function, knock control, boost control, ignition, fuel, maf etc.
M3.x is very similar to it, so this is how I know all this.

You need to dial in your MAF curve and then scale the MAF and injectors at the same time using FGAT0 and KHFM until you stop hitting 100% load with bigger turbo.
KFLF is your main fuel correction map on this ECU.
KFZW is timing map.
You also must calibrate the pilot duty cycle maps for the turbo.

After you're done, you are going to need to do all the cold start factors and cold start maps, as your FGAT0 and load is fubar now.

And after that it will work perfectly fine.
I have a customer with a car with GT28RS and 1.7 bar flat to redline with a R32 MAF and fully controlled boost by the ECU without any issues.
It's been running fine for a long time now.

Assembly knowledge helps. German knowledge helps and knowing how these ECU's work helps a lot too.
But really, compared to ME7, these things are trivial. But of course, IMO you need an emulator to tune this properly.
Without an emulator (flashing blah blah) is extremely time consuming, as you need to make a huge amount of changes... The Moates Roadrunner works great.

PRJ that is an excellent comment about the forgoten but still capable old Motronic systems. From your comments seems to me that you need to start from zero without any info about old Motronic.  ::) ::) ::)

But from your knowledge with M2.xx do you find or develop any diagrams or algorithm (process flow chart) of the factors/functions/constants involve in those calculations? will be awesome to see some info about it, if you have it.

1 -  LOAD = MAF_VOLTAGE/RPM*KHFM
2 -  MAF_Voltage x K1 = Power (that´s the same way Bosch know torque in ME7)
3 -  Power = Torque x RPM x K2

from 1 and 2:  LOAD = Power/( K1 x KHFM x RPM)

from 3:  LOAD = (Torque x RPM x K2) / (K1 x KHFM x RPM)

LOAD = Torque x (K2/(K1 xKHFM))

If You call (K2/(K1 x KHFM)) = K

LOAD = TORQUE x K

Just Math pal...

Power/RPM = Torque

Torque = K x Load

just math, but forget... that´s only my blablablá again...

That`s what I am talking about (its always better with loggs huh ? )

It is obviously tuned ,specified load @ 11,6 from 2.400 rpm ,actual load is struggling and lots of Knock Retardation.

But what about Injectors ? They are stuck at 16.32 .

Part of me wonders if you have ever seen an internal combustion engine inside.

With your logic an engine would always make more torque with more load. But this is not the case in the real world.

As far as I know we were talking about DESIRED LOAD, not if the engine, turbo, etc etc can achieve this value (ACTUAL LOAD)...

I can ask a monkey to jump over the moon... he can´t do it (DESIRED JUMP >ACTUAL JUMP)... but If I ask... and he will jump as high as he can... if I ask the monkey to jump 3...4 or 5 inch he can do it... the job is done (DESIRED JUMP = ACTUAL JUMP).

Simple like that...

But, I don´t know anything about engine or tuning... or even how a management system work, don´t listen to me... ::)

I am talking about actual load.
Do you really think that more actual load is always more torque?

Have you ever tuned a high powered car on dyno? Or any application with more than 300hp/liter?

The entire topic is about SOLLWERT (Desired Load) and what Load can be translated...

I Hate "my dick is bigger than yours" talk... but

Just cars With more than 400cv, all E100, All street legal (emissions), All daily cars...justa cars made in last 12 months

(http://i828.photobucket.com/albums/zz208/overspeed_2010/S3Adrianodyno485whp_zps88ab121b.jpg) (http://s828.photobucket.com/user/overspeed_2010/media/S3Adrianodyno485whp_zps88ab121b.jpg.html)


(http://i828.photobucket.com/albums/zz208/overspeed_2010/MareaBrunoHX40_zps5e01861f.jpg) (http://s828.photobucket.com/user/overspeed_2010/media/MareaBrunoHX40_zps5e01861f.jpg.html)

(http://i828.photobucket.com/albums/zz208/overspeed_2010/S3Adrianodyno485whp_zps88ab121b.jpg) (http://s828.photobucket.com/user/overspeed_2010/media/S3Adrianodyno485whp_zps88ab121b.jpg.html)


(http://i828.photobucket.com/albums/zz208/overspeed_2010/GOLF20-GT3076R13e155Bar.jpg) (http://s828.photobucket.com/user/overspeed_2010/media/GOLF20-GT3076R13e155Bar.jpg.html)


(http://i828.photobucket.com/albums/zz208/overspeed_2010/Golf150K27Brammer_zps7b7058d6.jpg) (http://s828.photobucket.com/user/overspeed_2010/media/Golf150K27Brammer_zps7b7058d6.jpg.html)

(http://i828.photobucket.com/albums/zz208/overspeed_2010/VR6105Bar426cv.jpg) (http://s828.photobucket.com/user/overspeed_2010/media/VR6105Bar426cv.jpg.html)

(http://i828.photobucket.com/albums/zz208/overspeed_2010/golfgt2811Bar.jpg) (http://s828.photobucket.com/user/overspeed_2010/media/golfgt2811Bar.jpg.html)

(http://i828.photobucket.com/albums/zz208/overspeed_2010/golfgt2811Bar.jpg) (http://s828.photobucket.com/user/overspeed_2010/media/golfgt2811Bar.jpg.html)

let´s make a deal... You said what you believe it´s the true, I said what I believe it´s true... when some Bosch engineer state one is wrong or rigth we start talking these topic again.

I´m not here to prove anything, just learn and teach one or two things if I can...

Friends take a look of the LOAD concept for M-Motronic document from Bosch (page 28) --> http://www.mediafire.com/view/94c1hy99xa6c76v/Bosch_M-Motronic_Engine_Management_89403362.pdf (http://www.mediafire.com/view/94c1hy99xa6c76v/Bosch_M-Motronic_Engine_Management_89403362.pdf).
Maybe this info could be useful.

Page 52 shall make things clear:

"A linear relationship between torque
curve and throttle-valve angle is also obtained,
with improved sensitivity at the
accelerator pedal"

Well ,because last several posts are not about M3.8  ;D I will bump this up 8)

max load value = 256/20 = 12,8ms

Never search specific M3.8.3, but normally lambda systens can enrich 27,5%  (25% in normal conditions plus 2,5%)

12,8 x 1,275 = 16,32ms

Obvioulsy it´s the same "monkey Jump" thing...   the moon can´t be higher than 16,32ms.

Your formulas with desired load are incorrect. Your translation of load to torque is incorrect also.
This is not how ICE's work. More load does not always equal more torque on a gasoline engine, just as more injected quantity does not always equal more torque on a diesel engine.
There are mechanical components in an engine, and your "load is higher 15% so torque is higher 15%" formulas are utterly ridiculous, as they have absolutely nothing to do with a real world scenario. In reality, you can in many cases increase load by 20% and the torque will increase 1-2%.
This is also the reason why KFMIRL in ME7 is a map and not a single line. As well as the reason that doubling cylinder charge never equals double the torque on any internal combustion engine. Same reason why nm->iq maps on diesel are nowhere close to linear.

Also, speaking about torque on M3.8 has absolutely no relevance to anything about the engine. There is no torque used on this management system for spark and fuel. Yes, it can output it in measuring blocks by going through a conversion map, but that's it. It does not request torque, it reacts to the cylinder charge which you create by pulling the throttle cable, just as every system before it.

There is no translation from load to torque without taking into account a plethora of other factors.

There is not a such cap on enrichment on this ECU. Also last I checked the maximum 8 bit value was 255, and not 256.

However, if you know something about KWP1281...

IPW = 0.001*a*b
a is fixed to 64.

0.001*64*255 = 16.32

I recommend to stop caring about this number and tune afr via wideband.

ok thank you both for your replies , I have never seen IPW "capped" before ,so I had to ask

Hello. I have a question. What is N75 map called? As the ECU controls the valve N75 in me 3.8.3?