NefMoto

OK, so I need to start looking into this.

KRKTE is for how long the injector stays open?
what is the injector latency map? and where?
MLHFM is for MAF scaling?  RS4 MAF values are slightly lower than stock S4..hmmmm

I know Tony has a little info to share once he reads this.

I don't have my laptop with me at the moment, so I will post what I can remember for now.

KRKTE is the constant for converting a reference load in % into fuel injection open time ms. So the units for this value are ms/%. The ecu calculates a reference load % based on the desired AFR. If the ecu wants to run rich, the reference load is higher than the actual load. If the ecu wants to run lean, the reference load is lower than the actual load. If the ecu wants to run at stoich, then the reference load is the same as the actual load. So if the reference load was 100% then the injection time would be KRKTE * 100. If you increase the size of your injectors, or increase your fuel pressure you need to scale this number down. If you decrease the size of your injectors, or decrease your fuel pressure you need to scale this number up.

As for the injector offset (latency) maps, if I recall there are three of them. The values in the latency maps are added to the calculated injection times. There is a latency map based on battery voltage, which allows the ecu to correct for injector opening times based on voltage. Then there is another map that adds injection time offset based on estimated fuel temperature. There is also an injection time scaling map based on intake manifold pressure. I will need to double check all of this when I have my laptop in front of me.

MLHFM is the table for converting MAF voltage in volts into air mass in grams/second. It should be a 1x512 map in the B5 S4. If you increase the size of your MAF housing you will need to scale this map up, and if you decrease the size of your MAF housing you will need to scale this value down.

If the RS4 MAF values you are looking at are lower than the S4 values, my only guess at this point is that you are comparing the table for a Bosch MAF to a Hitachi MAF. The RS4 only came with the Bosch MAF, while the 2001 and newer S4 came with the Hitachi MAF, and the 2000 S4 came with the Bosch MAF.

I will try to pull some more info up on this later.

OK so i guess im on the correct path with KRKTE, oh yea forgot about the RS4's were all Bosch sensors, makes perfect sense.

ill look for the 5x1 maps tonight if i get a chance and post up what i find and where.

Here are the maps I was talking about.

TVUB - injection time offset based on voltage - 5x1
TVTSPEV - injection time offset based on estimated injector temperature - 4x1
FRLFSDP - injection time scaling factor based on predicted vacuum relative to outside pressure  - 11x1

having issues finding these, care to take a looksee?  another set of eyes helps when staring at hex sometimes.  I did find TVUB  im using hbox

Here are the maps with Mbox addresses:

TVUB - injection time offset based on voltage - 5x1 - 0x814EEE
TVTSPEV - injection time offset based on estimated injector temperature - 4x1 - 0x81C2CA
FRLFSDP - injection time scaling factor based on predicted vacuum relative to outside pressure  - 11x1 - 0x81C2A6

Sorry for the wait!

coool, thanks, lemme see if thats similar to my hbox, at least i got addys for us spec ecu now :)

IDK if its useful to anyone yet, but KRKTE is for WOT, need to play with this to get WOT proper.

the other 3 are going to control your part throttle situations.  still working on that.  Now i can see how my buddies old unitronic tune on his 1.8t was such crap after 15 revisions! I am having a semi hard time getting things working how id like it.

also would like to see more about scaling the maf.  its really hard to do unless we have a flow bench pretty much, tony any input? 

scaling the injectors to the approximately same size maf is the way to go, this is the way i see some US tuners do it. 

however i do not see a problem with this as long as its A/F are proper it runs well and no issues. 

KRKTE is used for idle and part throttle as well. But during idle and part throttle you have O2 sensor feedback, so the injection time is constantly adapting. Make SMALL adjustments to KRKTE and keep testing; you can drive yourself crazy if you make large adjustments and keep overshooting. Keep an eye on long term fuel trims; idle trim means you need to adjust latency; part throttle trim means you need to adjust KRKTE. You can use the short term trims, but remember they will constantly fluctuate as the AFR oscillates between rich and lean in idle and part throttle.

As for scaling the MAF, you should try to get the injectors dialed in perfectly before dealing with the MAF. So either use stock injectors and a different MAF, or get your new injectors tuned with a stock MAF, and then switch to a different MAF. If you try to tune the MAF and injectors at the same time you can do it, but you are not tuning them for actual values; you are tuning them to be in the correct relation to each other. You can set your injector scale (KRKTE) and you MAF scale to the theoretically correct values, and go from there. If you end up underscaling the MAF though, you will have abnormally low load values. Some tuners prefer to work this way, since you don't have to correct a lot of the maps dealing with load limits. But if load reads abnormally low, then all maps that protect your car under high load, are essentially disabled until you retune them. Not to mention it sucks not being able to compare MAF and load readings from your logs to other cars anymore.

If you log your car with a stock MAF and record MAF air mass, MAP, and RPM. Then you should be able to switch to a new MAF and correctly rescale if by comparing a before and after log. You would just need to match up the two logs and see what the difference in the MAF reading is, when the MAP and RPM are the same.

Currently I am running Siemens 630cc injectors and a stock MAF while I get all of my low level fueling settings sorted out. In the B5 S4 you should be able to get up to around 400hp before you reach the flow limits of the stock MAF.

Does that help at all?

yea i kinda figured some of this out, its just a 512x1 map to rescale seems like a lot to work with you know?  well i wont need to scale the whole thing, but most of it im assuming.

like injectors first is the way im playing with now, i have 440's im just messing with.  seems OK.  and great with the e85!

however with the MAF, should i just do a percent increase (stock vs larger) in cross section and increase values that way as a baseline?

If you change the size of the MAF, you should just scale the 512x1 map up or down based on the change in cross sectional area. Unless you have a flow bench, that is the best you can do. If you are sure your fueling variables are correct, then you can use the short term fuel trims to determine if certain parts of the 512x1 map are scaled wrong. But this is more work than I have seen any tuner out there do.

So if you are running a larger MAF, scale the whole 512x1 map by (New Area / Old Area). This will tell the ECU that more air is flowing into the engine for a given MAF voltage. There is also an air mass correction map that allow you to scale the air mass reading based on RPM and air mass. You would use this if at certain RPMs and air mass you get abnormal turbulence in the MAF tube and need to correct for it. Some tuners also use this correction map to tweak the idle AFR.

goooood info.  ill start playing with that once i get a maf.

so dont feel like shelling out 300 for a "billet maf housing"  i will just gonna fusor the maf flange to a piece of ABS possibly, or my buddy does carbon fibre work, he would be able to help me he said.

Just to let everyone know what I have been doing. My S4 has K04 turbos, Siemens 630cc injectors, stock MAF. I started with a stock tune that had the fuel injector scale corrected for the larger injectors. I then did numerous 10 to 20 minute test drives giving the car time to learn the response of the new fuel injectors. During the drive I would try to have idle, part load, and WOT, and multiple types of driving types. After each test drive I would read the long term idle fuel trim and the long term partial load fuel trim. Using those trim numbers I would adjust the fuel injector scaling and fuel injector latency map.

The long term partial load fuel trim is a percentage correction of the reference load for fuel injection. So if the long term partial load fuel trim is 5%, that means the car is scaling up the load by 5% to make the car run richer. The fuel injector scale is in (ms injection time / % load). So to adjust for the car needing to add 5% fuel to run at the correct AFR, I would then scale up the fuel injector scale by 5%.

The long term idle fuel trim is also a percentage correction of the reference load for fuel injection. So if it is -1% that means that the car wants 1% less fuel at idle. To adjust this change the injector latency map. The latency map is in ms, not ms/% like injector scale. To convert the percent fuel trim into a ms value you have to multiply by the fuel injector scale. You take the fuel injector scale in ms/% times the long term idle fuel trim in % and that gives you the long term idle fuel trim in ms. Then you use that value and either add it or subtract it from your injector latency map.

Keep repeating the test drives and tweaking of the injector scale and injector latency until the long term trims are sufficiently close to zero.

Hope that helps. :)

When it comes to MAF scaling... would it be wise(er) to keep a stock fueling system intact that you know works well and just slap in a bigger MAF housing that you're planning on using and tweeking the MLHFM until you get back to what you're seeing with the OE MAF?

Once that is sorted out, you would be pretty much set on using larger injectors and would only have to worry about scaling them, correct?

Correct. It is easier to change one part at a time, since you then only have one variable to deal with.

cool, you on here, get your cable yet?  shoot me an email once you do or start tuning the injectors, i can help you with that a little bit ;)  you still working with an h-box or switch to an m-box?

I don't see those tables for an 01 S4....pretty sure they are needed since he upgraded to larger injectors and an E85 tune.  How could the tuning shop send him a file to DL/flash without those tables being existent?  I have a hard time beleiving that you can just slap in some larger injectors without rescaling and adjusting latencies.  Any ideas?

Most tuners just adjust the fuel injection correction map, and never touch the injector latency.

Tony, tell me please, for what bosch number ECU are these addresses valid?

M-Box.....


everyone is using M-Box for the most part, I am using H-Box however.

All the addresses I give are for the M box version 0002.

That is the most popular S4 ecu type.

I am on here... although I don't seem to follow up as much as I should with this..

I finally got my KWP2000 cable a little while ago, so I think I'm gonna dive into this a little this weekend.  At the minimum I want to do a few timing tweeks and del the rear O2 CEL that has been on for wayyyy too long lol...

As for the injectors, I don't have any bigger injectors yet (I don't really have the need for them just yet... still running K03's and no E85 available here --worth while anyway--), but I am thinking of picking up a bigger MAF and playing with some scaling to 'get ready' for when I get bigger turbos I won't be totally in the dark :P

Doing a lot of reading, I'm not sure that this is so true anymore... Finding the equation for the correct calculation of KRKTE under the RKTI section of the FB PDF there is the formula: 50.2624 * displacement * Qstat (flow of the injectors with n-heptane) = KRKTE

50.2624 * 2.7 * 458 (Qstat for Siemens 630cc) = 0.296306725 <--- what the KRKTE should be I believe according to the calculations.... but it seems ppl are using a value closer to .11 with 630's...

What do u think?

^^^ ignore that last calculation... I forgot that it's displacement of ONE cyl, not of the motor lol... so the correct calculation would be:

50.2624 * 0.45 / 458 = 0.049384454

oopsies  ;)

Remember to correct Minimum TE also when you change the injector Size. ;) or elso you will run pig rich at idle no matter what KRKTE is set to. On the 1,8T minimum TE is approx 0,75 mSec. Hope somebody can use this info. :)

minimum TE?

I know that changin the KRKTE alone will not work, the injector latency's will have to be modified as well (if that's what you're referring to :P )

That is the absolut Minimum opening time in ms of the injectors. So if you use big injectors you have to lower this value to run stoich at idle.

What are the map names for min TE and inj latency? Are they one in the same?

edit:

rereading thread....

inj latency = maps TVUB, TVTSPEV, and FRLFSDP
min TE map = ?

when using larger injectors, do all 3 latency maps really need adjusting? i.e. 630cc or 840cc dekas


which of the 3 latency maps are you referring to here?

Thanks,
Rey

And what about Hysteresis for lambda control?

Can anyone tell me, please, the name of these map (min. TE)? It is ok to adjust this value with a new value (0.85) if i use siemens 630cm3 injectors?

This is what I found:

minimales TE bei VA (TEMINVA)
minimales TE (TEMIN)


I believe VA refers to Overrun, but I'm not sure.

Thanks,
You know which is correct MIN. TE value for 630 cm3 injectors?

That depends on fuel pressure. You should not have to change it from the stock value. It is sufficiently low.

The easiest way to do it, that most people use, is to scale the constant by the difference in injector flow rate.

new min injection time = original min injection time * (flow rate original / flow rate new)

That way if you put in larger injectors that flow more, you end up needing a smaller min injection time.

KRKTE always must be calculated when larger injectors are fit!

We do not set KRKTE randomly but there is a formula from Bosch that calculate it! You analyze it before but you have to watch that this formula
do not use the injectors gasolines flow rate at cc/min but uses the gasolines n-heptan flow rate with units g/min!

So watch out!

I usually tune larger injectors only with KRKTE without having problems, injectors till 600cc! Over 600cc you have to adjust injectors lag time and some more that you have found and you explain before!

Well done for your well job on that topic!

Regards
Fotis

Great forum Tony.

Sorry if I've missed it, but is there a constant for the fuel consumption display?  Or will this affect the tune as well.

I have 3 cars running the same program at the moment and tweaked via unisettings; with minor differences between fuelling and MAF setups.  All are running perfectly now, but the fuel consumption reading is way off (~40% lower than in reality).

yes,

go and set KVB properly

;)

Yes, there is a separate constant for fuel consumption. It has no effect on fuel injection, just consumption calculations.

that constant is KVB that I say before
;)

Yes thats another important part to get good idle with big injectors.
Sometimes i have bouncing idle issues with injectors 630cc or higher, even when in same time lambda correction is +/-3%, on VCDS i see that injectors pulse is changed in example from 1.36ms to 1.70ms and idle speed is changes in such time(bounce), so i think if the correction is not 0.34ms but 0.15ms in example than bounce issue will be eliminated or lower.........unfortunately i still cant find the values in ME7 software what are referred to lambda hysteresis on idle.......

One more thing to look into is the warm-up fuel maps.  In my 1.8t these are:
(KFFWL_0_A)   12X12
(KFFWL_1_A)   12X12

My car ran very rich on cold starts prior to playing with these.

I'm kind of cross-posting this as I think this may be a more appropriate place to discuss:

http://www.nefariousmotorsports.com/forum/index.php?topic=170.0title=

(cliffs notes for the above link:)

I took Tony's 85mm + 60lb/hr with idle tuning file, and it ran terrible on my car, idle hunting and small throttle angles would really upset it.

Ignoring the additional load requested at idle, I looked at TVUB, and what Tony did with the latencies made sound logic and sense - the siemens injectors are faster than the stock injectors and the changes Tony made are congruent with the latency difference.

For some reason, my car prefers much higher values in this table than Tony's, but lower than the stock M box values... It also does not exhibit any idle misfires.

My question is, does the ME7 use a dynamic flow model outside of the injector constant KRKTE?  Obviously these injectors open faster - but they also close faster - and outside of these two differences, I'm sure the slope of the fuel delivery is considerably different (at least from what my brain thinks).

Obviously these differences have little to no effect at WOT, but at idle, I think the injector close time is going to play a big role in idle fueling, because if the injector closes faster and the ECU is not aware of this, the pulse-width calculated is going to be too short and the result will be underfueling at idle - something that no amount of latency tweaking is going to solve.

Thoughts?

I started with theoretically correct values, and then slowly changed the injector scale and constant based on my long term fuel trims at partial load and idle.

If the fuel trims are happy, then I am sure everything balances out.

Here is a good reference for fuel injector latencies: http://injector-rehab.com/shop/lag.html

That is exactly the site I used to reference the latencies...  I'm really curious to know what could be different between our cars.

Do you know what the voltage values are for the TVUB axis?

I'm on a stock MAF, with 630cc injectors but not sure what make they are as i didn't fit them.  Latency's are stock m box, i rescaled them to very similar to what Tony did and they are very good.  No misfires detected but there is a slight unevenness to the idle when the AC compressor is off.

Rick

Hammersword reminded me that a better way to correct the idle misfires is with map FKKVS. This map allows you to correct injection time at different RPM and requested injection time.

He recommended reducing the injection time requested at low RPM and low requested injection time.

Tony,

surely your fuel trims would have the same effect?  You will always arrive at the same pw due to closed loop operation. 

However, changing torque reserve i.e. idle ignition angle will have an effect on idle quality.

RIck

In my experience too much idle timing advance can affect the O2s causing fuel trims to go excessively rich.  Too much idle timing will also result in higher NOx - something to consider if you have to pass a sniffer test.

The fuel trims effect fueling at all RPMs and load. The injection time for idle with the 630cc injectors only needs to be offset at low RPM and load.

LTFT's have two independant areas of adjustment don't they?  Idle and part throttle?

Regardless, my trims are within 1% now.  Reducing the idle timing by increasing idle torque reserve slightly does smooth out the idle very nicely. 

Rick

There are separate idle and part throttle long term fuel trims.

Are you doing most of this tuning with TVUB or are you doing these changes to TVTSPEV and FRLFSDP as well?

Hi all.
I tune 1.8 t (auq) seat leon whit franken hybrid turbo, 3inch s3 maf and siemens deka 630cc.
Maf calibration was easy because I copy maf calib from s3 sw.
Then I calculate needed krkte and I got 0.048396 and I use 0.000111 factor.
Then I look tvub and I put 1.14 ms like siemens data says.
Car idling bad and lamda adaptation was -25%.
Now I have 0.043956 krkte and tvub is 0.613ms car works ok and fuel adaptation is -1.2%.
Car is ok but why I have to put erong values?
my tvub factor is  0.002667.
Fuel pressure is stock and all hoses etc are ok and new maf.
Thanks for help.
Regards: janne

Was it the long term idle adaptation or the long term partial load adaptation that was off?

Short term idle adaptation, channel 33 if i remember right.

You will want to read measuring block 32 to see your long term fuel trims.

The short term fuel trims in block 33 can fluctuate a lot, and they don't show you want the ECU has learned.

If the problem is at idle though, and not at partial load, then your problem is with the injector latency values. These can be effected by fuel pressure, voltage, age, etc.

But if your setup is  far away lamda is always -25 or +25 and long time adaptation dont work.

Is siemensa deka latensy 1.14@14v?
Someone says it is fast and someone says it is slow, but siemens says that it is 1.14msek but it does not work on that latency ????
regards janne

If you are getting an adaptation of -25% at idle, and not at partial load, then you need to reduce the latency times.