NefMoto

I did a log with what I thought would be my last revision for now, when I spotted this weird/funny looking actual boost pressure `pvdkds_w`. It keeps goin up and down with a difference of ~55mBar.

Does someone know how this can occure?

PID is hunting because your I is maxed out. Not good. You are nowhere NEAR your "last" revision with that persistent lde.

Log all of the PID variables. That log has none of them.

why new threads all the time?

start one thread for your car and your learnings....

Because I have been on some forums a while ago and one moderator likes it all catagorized, and the other does not want any spam.

I choose to make new ones because the subject might helpt the SEO in case other people are looking on google for the same problem.

for the record, I'm ok with this here, or not, either way.

however, to get back on topic, i can't emphasize this enough: persistent lde is a very very very very bad thing.

If you do not know what lde is, definitely, for sure, your tune is nowhere near "done"

I tried to read about PID's, but I am a visual learner, and I find it realy hard to understand what they are and what to do when they are giving problems.

I did a new run today with the PID's logged and I see allot of activity with them, I will attach it. Is there any way you can help me?

Did another carefull read on the wiki, and saw that DIMX was too low. Changed 850 and 1000mBar from 2000 with 10%. Graph looks better but not perfect yet, might have to up it a bit more.

You're still missing lditv ldptv ldrdtv lde

I tried to add those before, also mibas. But ME7 logger couldnt find those. I guess those addresses are different for each ECU or not?

yes, of course.

You dont' really need them atm, we can see pretty clearly what is going wrong - your i limit is far too low, or you can compensate in KFLDRL

Okay, in that case I will read up on KFLDRL. Do you happen to have a log from a 1.8T where P I D are logged? Just as a example? I'd like to compare a good log with the results I am getting

I am having a hard time finding lditv ldptv ldrdtv and lde. Could someone assist me on how to find these? 

bottom line, your i max is way too low.

lditv 381FE0
ldptv 381FE2
ldrdtv 381FE4
lde 380B2F

Not sure if applicable for 002, disassemble to check or go try your luck.

Compared to someone else's log, this looks spot on! Thank you very much!

I have changed the map arround, and thanks to Blazius I now have PID logged. The load, boost and fuel look spot on. I will make a second map for boost capped to 1,2 bar.. I dont think the K03 will have much left at this point(its a old one perfect for testing).

If you agree, the only thing left is ignition.. Are there any tips you can give me when tuning these values as close to 0 as possible?

dont touch anything until you fix boost req and get lde to zero

lde is basically the output of BoostActual - BoostDesired correct? And is this value only used in PID scheduling?

Does this also mean, if I adjust LDRXN to follow actual, that lde will stay or be close to zero?

req boost should always be pretty much under whatever 100WGDC can do

if you let lde be constantly non-zero you'll get trim activity that you dont want.

So that would mean I cant get any more then 1,1 - 1,2 bar out of the turbo without the WGDC increasing and trimming fuel?

wgdc will max out at 100, i-max trim may max out and eventually throw a code.

fuel will be fine.

K03 is a 195-200 HP rated turbo , dont try to push it to do something it cant. http://nefariousmotorsports.com/forum/index.php?topic=20256.0title=

Look at the compressor map if you wish but even from the log you can see 100 wgdc and what it will do up top, first get request there or lower then you can fix other stuff.

I understand that they dont make lots of power, but there are allot of K03's running 180-190 load on 1,3 bar. Thats my goal, only thing I just cant get to work is the dang WGDC..

I have been doing multiple logs a day with small changes. I get fuel, boost and ignition to where I want it to be, its just that last step that is not making any sense to me.

This is basically where I am at.. laptop battery was dying and got fed up, so I stopped for today.

closer.

you're seeing overshoot on ramp, you can bump peak boost if you want, or bring down imx a bit.

prj will kill me for suggesting you fix overshoot/undershoot with imx but it is what it is.

Why would he? Does it not say in the wiki to do the same thing with imx when you see over/under boost?

After fixing the overshoot, could it then be considered good?

2 pages of nothing
There is no overshoot
just move kfldimx axis.. select all axis values higher than 400 or 500 hpa, move them 1 row up so you delete this 400 or 500 and double last with 1100, change last row with 1400 hpa so you came with last 2 rows of stock 1100hpa and new 1400. copy stock map values accordingly new axis. do not change stock values at stock axis. fill last row of 1400 hpa with 95%
run it. if it overshoots - taper values a bit in corresponding area

THATS ALL damn .. no rocket science here

Do not forget to revert back all changes youve made cos it gives you nothing with 0 ignition @mid and those tiny turbo

Letting shit run on I limiter is dumb, especially coupled with adaptation, it will just adapt it higher over time, and the problem will be back.
Overshoot should be controlled with P/D mostly.

Of course if I limit is skyhigh then you also will get problems.
IMO in the OEM PID the I part is in general way too aggressive.

Sort of. It depends how far off it is, in my experience. If it is close, adaptation will do exactly what it is supposed to. If it is too far off, adaptation will never do the right thing.

This is where I landed, and frankly I am happy and done. The turbo is allready due for a change soon. Results are from a very hot day (30C is pretty much very hot for our country).

Thanks to nyet and Blazius for actualy helping me. I know its not perfect.

This seems like a perfectly reasonable result! Good job being persistent.

Like you said, its not perfect, but definitely good enough.