NefMoto

Hopefully someone can give me some pointers with this. Maybe it belongs in the n00b zone, but i'm trying to get my head round everything.

I've got a hybrid turbo with a 12psi actuator, on a B5 1.8T, running a TT225 ME7.5 ECU.

I'm supposed to be getting it professionally tuned, but the tuner seems somewhat elusive at the moment. He provided a base map to get it running with the injectors setup and some other tweaks. The axis for KFVPDKSD has been scaled to account for the actuator, but it looks odd to me. In the original tune the axis only goes up to 1.2, which doesnt make any sense? The scaled version goes up to 1.5, but again this seems odd. I'd expect much higher values? The car does seem to manage to control the boost with the throttle upto around 10psi (any more and i get limp mode), but i'm unsure whats going on here. Is it maybe like the RS4 files where 1=wastegate pressure rather than 1=atmospheric pressure? Does simply scaling the axis work in that case? Theres much chat on the forum about hacking files to restore the "normal" version of this function, rather than info about tuning the RS4-style version.

I've been tweaking bits myself to try and get it running a bit better in the interim period, The wife drives the car a lot and it needs to be usable. I've mainly been trying to dial in the injectors a bit better (it was throwing system too lean codes), fixing cruise control, sorting fault codes for missing CAN etc. Getting to grips with the whole tuning thing as i go. The load and boost stuff in the map is all stock.

If i apply more than about 75-80% throttle it drops into a limp mode, limits boost to 5psi and doesnt clear after an ignition cycle remaining like that until i go in and reset the fault codes, leaving the P1555 code behind.

I raised LDRXN slightly, as its default setting was very close to the actuator pressure, lower in some places, and i thaught that might have been the cause of the overboost code but that didnt change anything.

I'm not quite sure whats going on, as engine load in the logs looks sensible, but measured boost pressure (pre throttle) is showing pressure way over requested, 2.2-2.3bar at the point it trips out. Boost guage in the car only shows around 12psi when this happens, The throttle appears to only be 25% open when this occurs. At a guess its related to WGDC, maybe i should just unplug the N75 in the mean time? I have a vague idea i need to fix the boost PID, but its all rather alien, and i'm sort of in at the deep end.

I was out logging nvquot earlier to fix cruise control and happened to catch this happening so i figured i'd post the log and see what you think. The overboost code trips at around 300seconds.

Cheers

Need a WOT run to confirm (BEFORE it goes limp) but likely your LDRXN isn't even close to right.

Cant even get WOT, any more than about 80% and its instantly into limp.

Its requesting ~55% WGDC in that log around 300 seconds, despite being way over requested boost? I'm thinking its a boost PID issue? WGDC must have to be a lot lower on the 12psi actuator than whats in the map?

LDRXN tapers up to 170 by 3k, sits flat until 6k then tapers off. Should be easily achievable for the turbo fitted. The logs show requested and actual load was around 140, so well below LDRXN? I cant see why LDRXN could be the issue?

Because your requested boost is WAY under your actual boost for most of the log.

Could also be IRL related, but w/o a wot log it is hard to tell.

just clear the code, go into 2nd, and floor it from 2k to redline.

I'm not sure i follow.

LDRXN is a limit on requested load. I'm not at the limit, its requesting 140ish, and achieving 140ish according to those logs, yet boost is far too high and the throttle plate is only open 25%. How can it be an LDRXN issue?

55% WGDC seems far too much when it only wants 1800mbar. It'll make 1800mbar with 0 WGDC, you can see it doing exactly that earlier in the log... Yet its requesting 55%, which causes the overboost?

I can try flooring it in 2nd, but i dont want to break it :P

If that isn't expected, then your WG line is busted or your N75 is malfunctioning.

Start there. I assumed that you were surprised at the overboost code because of something tune related, not hardware related.

In any case, you said in limp you were seeing 5psi, so something is odd if 0wgdc makes 1800 mbar as well.

It is impossible to diagnose anything w/o actually being WOT.

You wont break anything if the car goes insta limp anyway.

Its got a 12psi actuator, 1800mbar with 0 WGDC is expected and normal? A pull on the old ECU with no N75 connected would usually see about 13psi thru the midrange.

The issue is the ECU is wanting 1800mbar and is cranking out 55% WGDC, which clearly results in a lot more than 1800mbar?

edit:
When in limp mode, the ECU is using the throttle to limit boost to 5psi.

Something is very very wrong with that file.

ps_w is tiny compared to pvdkds_w because the throttle is almost shut completely.
 
Hard to make head or tail of the log with your throttle part closed though.

Just go WOT. If you're nervous, cut IMX

Also, you'll want to add some of the PID variables to the log. Not sure why you're seeing so much integrator wind up if req is so much higher than actual.

Problem is in your KFLDRL and IMX not been adjusted for harder WGs
You need to do PID calibration from sctrach or pull n75 before you get it tuned

Also, if your tuner isn't being helpful, you really should be doing your own file from scratch.


Something is really crazy with the file, the WGDC is not following ldr error properly.

Is this a 5120 file or something?

Sort of. Why is he getting 55% WGDC when req boost is BELOW actual?

He needs to log wg before lin and the the rest of the PID variables.

Perhaps something is wrong in DRL, though, like you said. that would explain the 55% dc.

Well went and tried a WOT log. And as expected nearly blew the bloody thing up.  :-[

Boost guage pegged itself at 25psi and massive missfire at about 3100rpm. That sort of shit is what puts holes in the side of 1.8T's...

I've been fairly sensible upto this point, because i cant afford a knackered engine. Mashing full throttle on an engine that clearly has wonky boost control is obviously a shit idea!

He needs to do a proper log. He is pedaling up and down. I am almost 100% sure that his tuner did not even touch PID maps at this Point.

As i said - in this log you can clearly see that PID is not adjusted for your tighter WGs and it overshoots badly. Do not attempt this again.

ps_w and pvdkds show well below 25psi.

So whatever you think is happening, either isn't happening or...

Don't mess around with 3rd party tunes you have no clue about.

If the tuner won't help, ditch it and start from scratch. Even if you CAN find the problem, you're wasting your time.

Also, I still don't see wg before lin, or any PID variables in your log.

Bottom line, your tuner screwed you, he gave you a fixed WG file.

The PID maps are stock, i said that in my original post.

I pedaled it becuase it spiked the boost guage round to 25psi. Went for it again and then thaught nah this is stupid.

Files are attached.

Original file from the tuner, then v2 is the version i modified to fix FKKVS and some other things.

ORI is here:

http://nefariousmotorsports.com/forum/index.php?topic=2156.0

The original file should have only had the emissions deletes, disabling VVT and EGT, and KRKTE for the injectors. Its not been wildly hacked about with.

So why is it requesting >10 DC even when actual boost is above requested?

Exactly as i said. Original PID maps in your case are realy wrong. Needs to be reworked from scratch.
I understand why you was pedaling it. See it in your log

Botton line dont push it or pull N75 until you get it tuned.

ONE MORE TIME.

Why is it requesting >10% DC when actual boost is over requested?

I dont know, thats what i'm here to find out :)

I presume its pulling that from IMX and the PID controller cant react fast enough because its so far out?

Could it be related to the modifying of the axis of KFVPDKSD?

As i mentioned in the OP, the modified axis on KFVPDKSD looks strange. It may be that this file is using the RS4 style CWPLGU = 0

Perhaps thats causing the weirdness?

No, that is in the throttle plate path, not the LDR path, and for a tighter wastegate, that axis change should be fine (and is recommended)

What does DRL look like? Is it stock?

I dont have a CB kp or xdf so i can't check it myself.

This is the XDF i have so far. I've been slowly building it up from the OLS on here for the 0002 software revision.

Just added DRL and IMX to it for you.

I'm asking YOU to look at DRL, dont' make me download your file and xdf. At least do some of the work; you'll have to tune DRL at some point anyway.

aragorn, un hide your email or pm it to me, im trying to send you something

As i've said a few times, its stock.

My file:
https://www.dropbox.com/s/wd5spc6kma5wbzg/TunerPro%20-%20Audi%20TT%20%28Kev%20Aragorn%298N0906018CB%20v2%20-%20Checksummed.bin%20_%20COMPARE%20BIN%20-_2016-06-01_08-00-43.png?dl=0

ORI OLS:
https://www.dropbox.com/s/rljmm09t7xwru4d/WinOLS%20Demo_2016-06-01_08-01-41.png?dl=0

Log PID vars, esp pre lin wgdc

ok i can do that.

Can you perhaps explain how the ECU selects an WGDC? I've read the Wiki but cant really figure it out.

I assumed it used IMX, but looking at that table it appears to reference WGDC against ambient pressure? I expected there would be a table of WGDC against boost and rpm? The wiki seems to suggest that IMX should follow the WGDC at WOT? yet in the stock file, its up in the 70-80% range across the board? Does the little 5psi actuator need that much DC to produce 12psi?


Is there some way i can cap maximum WGDC to stop it spiking stupidly while i'm trying to figure out why its running so much WGDC? Theres no way i'm going full throttle again with it in its current state. I like my engine in one piece. If i'm right above, i presume i can limit the top two rows of IMX to say 30%?

Cheers
Kev

ok that IMX axis isnt atmospheric is it? Its boost...

Means it only goes up to 1 bar as stock though?

To cap the boost you need to set KFLDRL to the max you want by RPM AND THROTTLE or you cap set an absolute limit with TVLDMX which is the max n75 DC. It wont go above that.

Does DRL have a throttle input? I thaught input was WGDC from the pid?

So if i cap it with TVLDMX, will the logging a WOT pull still show what the PID controller is trying to do?

I've found the big thread on here which discusses using KFLDRAPP to set fixed duty cycles and logging. I think thats probably where i need to start?

http://nefariousmotorsports.com/forum/index.php?topic=517.msg4028#msg4028

my understanding is that imx axis is mbar above atmospheric pressure, could be wrong...

You'll have to read up on PIDs and read the FR. The input to the PID is lde (requested pressure - actual pressure) and there is a ton of gain scheduling also done. The output is WGDC pre-lin, which goes into DRL. If I were to add an explanation of PIDs to the wiki, it would be 100 pages long.


No.


In the absence of all other inputs, a PID will follow the accrued I value, which, in ME7, is limited by IMX


In some conditions, yes.


Yes, you can do that but I think you are wasting your time. You need to start with a stock file.

Cap DRL is the safest, it is the last step before WGDC.


I think you need to start by figuring out how PIDs work before doing ANY tuning.

Nyet ,

Have you tried taming the PIDS on the 1.8t files, unlike the 2.7 files its not that easy and from experience the easiest way to control the actuator when its over stock is open loop via KFLDRL.

It works perfectly well and far easier than trying to set up the PID maps to control the n75.

My suggestion and it is only that, is to raise LDRXN out the way, then disable the maf diagnosis maps and increase the boost deviation, set IRL and IOP enough to cope with your new loads and simply map the boost using KFLDRL ie n75 DC..

Honestly, never had a problem tuning the 1.8t pid.

99% of the work is in IMX and DRL. You don't even really need to tweak Q2 if IMX and DRL are set up right. You can even use "almost" open loop by clamping DRL around the area you want PID control to be limited to..

I agree with this but I think by adjusting the DRL you are really open loop still as the pids aren't really learning or adjusting as they would in a true closed loop system, you aren't setting the ECU up to run the stronger actuator just stopping it from uncontrollably going above, the Pid isn't really controlling in the strict sense, I can see it working though.. My personal choice is just go open loop and set the n75 dc where you want it... and with a little effort you can get a pretty good boost profile..

Except if the temp changes, or the altitude changes, or ... etc. etc. the boost will drift, and unless you numb all the deviation codes, you'll eventually throw a code.

Even worse, fueling might start going all over the place too, unless you're really anal about getting fueling right at a variety of WOT boost levels.

The latter isn't as big a deal on wideband 1.8ts, though so we can agree to disagree.

To each his own. I have no problem with people who decide on open loop boost control, as long as they're aware of the downsides.

It's a base file purely to allow running on the larger injectors.  Boost control is standard and it will almost certainly go straight into limp with the stiffer actuator.  It will run full closed loop when it's finished :)

Rick

Cheers Rick :) I did say quite a few times that the file was stock, some reason they're jumping to conclusions that its been buggered about with.

Thanks for all the help guys, just a little too much too soon i think. It needs time spending learning whats going on and i just dont have the time at the moment. Managed to get it booked in next week, so the wife will stop bending my ear about "messing with her car" :P

I think i'd be better starting with a "stage1" type tune and working my way up, hopefully i'll get my track car converted to ME7.5 in a few months and sort out a tune on its standard turbo, and work my way up from there.

Cheers
Kev

Because lde<0 should not result in DC>10%

You claim you want to learn (and diagnose the problem), but you've ignored me every time i've said this, including my requests to log the PID variables :P

Rick, if this is your base file, what am I missing?

I didnt get a chance to log the PID vars, Work and babies get in the way of tinkering with cars!

Spent this evening replacing some pipes that had been chafing and sorting out the cruise control stuff, which incidentally works perfectly, leaving me quite chuffed that i managed to disassemble the file, find the ram variable and log the required info. Clearly some learning is happening. Just not with PID controllers.

You've said a few times that lde<0 shouldnt result in a DC >10, yet the logs clearly show exactly that happening. I'm not an expert, i'm a noob learning how to tune ME7.5. If you dont know why its requesting significant WGDC when lde<0 then how do you expect me to know!?

Specifically, addressing your statement:

And, generally, one good reason not to use open loop boost control..

I dont really understand your response.

Are you suggesting, yet again, that its got non-standard boost maps and it running open loop?

I've posted the IMX and DRL maps as screenshots earlier in this thread, they are exactly identical to the 018CB OLS file posted on here years ago. So unless Audi mapped it with open loop boost...

No idea. Without logs, we'll never know.

Hey Aragorn, if you'd like to write your own mapping for this hybrid, I would recommend my current file! It is a fully defined A6 150hp base (1024kb), but I can give you my file that is adjusted for 225hp timing and KFMIOP/IRL. Might be a better base to start from since you have a longitudinal engine. Just need to find out if it cross flashes with your current ecu.

We're in the Uk Temp changes and elevation aren't really going to play that big a deal in all honesty!

I've attached a file I recently did for an HTA30 that's running 500bhp, completely open loop and solely controlled by KFLDRL. Works really well and can just as easily be setup or smaller turbo's and injectors.

Conclusion drawn from this thread:
Not one person in this thread understands ME7 boost control.

1. LDDIMNN is the reason the DC can not go much below what is defined in KFLDIMX. This map is used on many 1.8T.
2. When you make a "base file" you usually set CWMDAPP to 16 and LDRAPP to 0 if you have half a clue.

As for open loop boost control - this is for people I mentioned at the beginning of this post.
Yeah, those who can't tune ME7.

Wasn't aware of it. Don't tune 1.8t


A base file should not request 60% DC if lde<0, period.

Very good points!!

Sent from my E6853 using Tapatalk

lde does not matter in this case, look at LDRPID and learn.

I understand, but I see no reason to set a lower bound for I in any PID.