NefMoto

Hi, I've been lurking on this forum for quite some time.
It has helped me tremendously with ECU tuning basics and I would like to thank everybody that contributes here.

I have a background in tuning diesels and old school carbureted petrol engines so I thought I could try to tinker with my daily driver a bit :)

Engine is 2.0 TSI 200 hp CCZA and it is completely stock

I have succesfully tackled "stage 1 tune" which resulted in 260hp 420Nm on 100 octane petrol.
Now I thought I could try to tune it on E85 and I have encoutered a broblem which I'm unable to solve by myself.

I dialled in KRKATE so that idle and partial trims are all within 1.5% and car runs great except for for WOT at 4500rpm onwards.
From idle to 4500 rpm everything is fine - requested lambda follows actual as well as rail pressure and engine pulls very strong.

But at 4500 rpm something weird starts to happen.

Rail pressure follows request as it is supposed to,but injector opening time suddenly goes up and actual lambda starts to lean out to an extent that I get a nasty torque intervention - boost cut and sometimes throttle cut.
Does anyone have any idea how is this possible?
How can mixture get leaner when fuel pressure is OK and injector stays open for longer?
I will be grateful for any suggestion :)

Log, original and tuned file attached.

You are out off fuel, You need to work on HPFP and LPFP

You need 30% more injected fuel mass for E85, when you squeeze it even more.

Consider that in your thoughts, just a fast switch to that fuel will not work well mostly. Many hours reading needed, sometimes one ore two new cylinder heads ;)

So is it really a hardware issue? Nobody found something i screwed up in the software?
I read somewhere that stock HPFP and LPFP combo is good for ~350hp on petrol.
So, if i flow enough air for 260hp on petrol and want to use E85 which needs 30% more fuel to achieve same lambda, that should be equivalent of 340hp on petrol (260*1,3) and therefore stock HW should be fine, although pushed to the limit.

I would understand if problems started at higher rpm than 4500, or rail pressure would start to drop when actual lambda deviates from requested. That would indicate that i reached the limit of fuelling system but none of that seems to happen when lambda starts to go way off.

Am I missing something or have I been misinformed about the capability of stock fuel system and I really need to upgrade my pumps?

I did not opened your file, i speak in general.

You forgot about eficency. Lambda and Ignition eficency. You can make rough estimations like that but no layout of systems.


If you squeeze the system to get 360 HP on pump gas, eficency is bad. You will probably need 1,4 times of fuel to get stable combustion without too high peak pressure. Watch lamiar flame front velocity  charts over lambda for alcohol fuels for that.

Maybe someone on sea level on a nice cold day managed to get 360 HP out of the pump, but how sure it is that ot is always possible under all conditions?


So calculate maybe 340 Hp as maximum with 100% pump gas duty cycle. Substract 40 percent and you are already below your current power output.



Remember: i speak in general, have no experience with that engine or specific ecu.

Thank you for your reply, I didn't think of it like that.
Speaking of duty cycle - is it possible to increase pump's duty cycle?
Mine doesn't go over 80% so that might be something to try just to see if it does anything.
I have looked torough the file but didn't find anything that would resemble "pump duty cycle map".

Yes, if you lower the FPR pressure.

This is TSI.

I have not bin able to look on your file, but I'm doing 320-330whp with K04 turbo on stock fueling parts and 98 pump fuel.
So you should be able to run E85 with stock turbo. As I say before you need to look on LPFP and HPFP tuning.

What!.?

Could you be more specific, please?
What maps did you alter in your file that enabled you to safely run 330whp which is surely over 360hp at the crank.

I'm very new to tuning MED ECUs so only thing I was able to do was alter KRKATE and slightly increase rail pressure in part load.
I was unable to increase pressure at full load, although I modified all seven fuel rail maps and everything that even resembled a limiter, but rail pressure stayed at stock value of 150 bar no matter what I did.

I also tried decreasing rail pressure but problem remained the same only thing different was longer injector opening time, but lambda leaned out exactly the same.

Sorry I'm home so don't have my winols her, I'm back at work on monday then I can take a look on your file.
A note is that you can't have long injection times, you should stay under maximum 9ms. Also you should log your low fuel pressure with a external gauge.

So, I've had couple of sleepless nights over this, but I can now proudly say that I solved the issue ;D

Problem indeed was in HPFP and LPFP calibration.
Firstly I increased LPFP duty cycle but that has helped only slightly. (Note: in VCDS - Advanced measuring values there is a "group" with Rail press spec., Rail press actual and fuel pump duty cycle. Spec and actual are values for high pressure fuel but said duty cycle is low pressure pump duty cycle. Maybe this is common knowledge but it sure had confused me at first.)

Then I tried to increase Rail pressure but it never went over stock 150 Bar no matter what was specified in the map. I increased every limiter I could find but none of it had any effect.
Then I found out that "main" temperature dependent fuel pressure limiter was somehow screwed up in the damos so I altered it directly in hex code and now I finally had 170 Bar of rail pressure which fixed my fuel delivery Issue.

Now, (apart from slight dip in rail pressure around 3000 rpm) I have stable fuel delivery throughout the rev range with reasonable injector timing so it seems that it really is possible to support equivalent of 350 hp on stock hardware :)

It appears I celebrated prematurely.
Apparently rail pressure request or some kind of limiter is gear dependent or something like that, because in 4th, 5th and 6th gear everything is fine, but in 1st, 2nd and 3rd gear rail pressure request stays at 150 Bar up to 5000 rpm and only after that rises to the 170, although map requests 170 Bar from 4000 rpm.
Has anyone experienced this?
I looked through the file for some kind selector but didn't find anything, although I dont' know if I can trust my definition file after that mistake with rail pressure limiter.

Sounds painful!

.
In decos it runs of RL not sure on this

I will take a look on you're file tomorrow.

Thank you very much I really have no idea what is going on because rail pressure request seems to be totally random.
Sometimes it is 170 sometimes 153, sometimes 141 and there doesn't seem to be any method to this madness.

I'll upload the latest bin that I'm running in the car at the moment.

Oh, and I managed to put my car on the dyno today and when everything (semi)worked  it put down 284 hp and 492 Nm of torque.
Torque is really nice and I think if fuelling gets sorted in the upper RPM range there is a real possibility for 290+ hp on stock HW :)
I'll post dyno plot later when I get home.

So, here's the dyno chart comparison of stock and E85 tune. It picked up 64 hp and 192 Nm over stock :)
If fuel delivery issues can be solved, I'm hoping to fix the power dip over 5000 rpm with new exhaust and some fine tuning because on E85 stock exhaust gets very restrictive very quickly.

Is there not a rail pressure map? / RL?

Is this MED17?

It is MED17.5.2
There are five rail pressure maps which I all modified together with corresponding limiters but at full throttle ECU seems to pick spec rail pressure at random completely ignoring values specified in the maps.
In part load everything works as it should but as soon as I floor it in first three gears it is 50/50 chance I won't get the proper spec rail pressure but some random lower value instead.

Which maps did you change looking at the FR prist_w is the var

I changed:

KFPRSOLHKS
KFPRSOLHMM
KFPRSOLHOM
KFPRSOLKH
KFPRSOLSCH

and

PRMAX
KLPRMAX
KLPRMAXTF

I even tried to alter % axis in KFPRSOL maps but that didn't change anything.

There's at least 4 functions on the go there, I can't read German well enough to help as I haven't studied it sorry but that var comes later on by the looks and goes though filtering etc.   

You need to look at 58406 (KLLFPRSG) and change the axis, also I would not run LPFP att 100% (ore 92% as limiter is still stock) as this will overheat the controller.
 Also look at 580C2 (FNGPRS) I have some more stuff but try this first.
A good thing is to messure the low fuelpressure with a gauge.
When you are running so high timing you need to extend the knock meassurement windows and also look at optimal timing.
Have you loged your timing? does it follow your maps? as you have same timing in regular as in optimal.




What is the Changes at 254B0 for?

Let me know how it goes.

Take a look at gradient limitation for requested rail pressure.

dream3R: where can one get hold of a MED17 FR? I only have some short crappy version for MED9 and some maps are different in MED17.

S2evo1: I looked at KLLFPRSG several times but decided not to change the axis because all map values are the same.
I'll rescale it to 17 Mpa and see what it does.
Regarding LPFP duty cycle is it not advisable to run it close to 100% at all, or just not all the time as I do at the moment?
FNGPRS looks interesting. Stock value is 1500 rpm ... if I understand its function correctly, should I try to change it to say 500 rpm and see what  effect it has on rail pressure request?

For timing maps I just dumped KFZWOP values into regular KFZW maps until I'm able to tune them properly at the dyno. Actual timing follows requested prefectly and I get no intervention or CFs.

As for changes at 254B0 I have no idea what it is.
I didn't do that but it may have something to do with the fact that my car has an engine swap. It came with 1.8 TSI CDAA which was later swapped for 2.0 TSI CCZA but ECU stayed original, so mabye that's the reason for those changes.

prj: gradient limitation for rail pressure you mean FNGPRS map or are there some other maps that have something to do with this ?

It is "Security Access Label"

Funny how you make changes in a file but you have no idea what some of them are :D

Asi I said at the beginning of this thread I'm new to MED ECUs.
I know how to tune an engine in general but I don't know much about this particular control system and since there is very little info out there about MED9 and exactly nothing about MED17 I'm amazed I even got this far without any major screw up :D

Trial and error is my only option at the moment since I don't know anyone that could help me with this so I really appreciate all the help and guidance members of this forum are giving me :)

Keep digging :)

500rpm is what I use :)

For testing you can have LPFP settings like that but the limiter you have still stock at 91.99%.

I think you done a lot off progress to be new at MED17!

Let us know how it goes.

Good job.   :) 

Thanks guys, I'm trying my best :)

Next thing I'll try will be rescaled axis in KLLFPRSG and FNGPRS reduced to 500 rpm.
Unfortunately it has been snowing non stop past two days so I can't do any proper testing now.

In the mean time I'll try to find the limiter for LPFP duty cycle.

And regarding fuel pressure - do you think it is safe for pump and injectors to run 170 bar rail pressure in 4000+ rpm?

I think your limiter for that is 52D18 , its 92%.

Do you have functionsrahmen from med17?

Johan

Lol so you changed security access yet you have no idea about this ecu?

I did not change security access since I don't have any idea what it is nor what it does.
Those weird changes here and there (security access?) were done by the guy who did the engine swap.

I'm only tinkering with the maps for boost/fuel/spark but I haven't done any progress in the last month. I've been very busy lately so TSI is sitting in the garage and I'm driving my TDI every day.

Throw it out and start from scratch.

At minimum, if you INSIST on using his file (you shouldn't) you should know what EVERY one of his changes does.

If you do not, and you don't intend to learn, and you don't intend to start from scratch either, forget it.

Is it really that bad?
It's ECU for 1.8 engine cross flashed with SW for 2.0 liter engine and since everything works fine I just assumed all those little changes were made to make it work like this. I was given .ols project with completely stock program (apart from said little modifications).
I'm not a professional chiptuner and I don't know any of the electronic wizardry to "make things work". I just like to play with the calibration :)

So could you please tell me what exactly is wrong with my file or link me to some reading material that would allow me to find that out for myself?
Thank you :)

It's not about good or bad. It's about ensuring you understand cause and effect.

If you don't understand the changes, how will you make sense of the results?

I understand changes that I make.
I'm under the impression that nyet's comment was about the security access thing and not the whole file but I can be mistaken and he thinks that whole file is sh*t.

Again, I was given stock file with little tweaks that enabled 2.0 SW run in 1.8 ECU.
Car ran fine so I didn't mess with that and only modified things that affected engine's performance and other behaviour. You can criticise me for those and I will be happy for any constructive criticism but that security acces thing - you could be speaking in chineese and I still wouldn't understand :)

It doesn't matter if it's security access or not, it was changed and you had strictly no idea why or what it was, I think that was nyet's point here. There could have been other changes, code or else, which you would also not have understood, and which could have affected the way the engine runs, how would you have done a decent job tuning the car in such a case?

I understand your reservations.
Guy who did it for me tunes diesels for a living and has been doing it for the last decade, so I expect he did at least Ok job and didn't screw anything up too badly.
Anyway, engine ran fine stock.
Engine ran fine tuned on regular petrol.
Engine runs fine now on E85 apart from obvious rail pressure issues but I'm almost 100% sure it is caused by my lack of experience with tuning this type ECU than enything else.

I realize that this situation is not ideal but I'm not an auto technician and I earn my money pushing papers so if I want to learn how to do this I have to rely on other people doing some things for me at least for now.

Maybe it's just me, but most people who have been doing diesels for the last decade (or two) have no f*** clue WTF they're doing when it comes to tuning. Yes they get power out of their stinky diesels, but most of them don't have the slightest clue why or how.

Well, I'm certainly not in a position to judge that :)

So, to get back on topic - from diagnostic standopint engine runs Ok.
Do you see anything in particular in my file that needs to be adressed and can potentially cause any issuses?

+1.

You can get more power out of a diesel with +10% on a selected square in boost and +15% square in duration map.
Which is what 99% of these do. They have no idea about tuning whatsoever.
Also, Motronic is VERY different from EDC. Just because someone has tuned diesels does not mean they have any clue about tuning Motronic.

Start from scratch or go home.

The other way around i've seen also LOL.
Big turbo, big injector setups aren''t always easy on diesels..

Motronic is way different compared to edc.
I really enjoy learning motronic now.

Bitmap search for all prsoll  addaptations in functionsrahmen.

@bitmap, can you please send me your damos? I've a Golf Mk6 CCZA with 1K0907115AA