NefMoto

Heya!

So, the patient is - heavily modded RS6 C5.
Custom hybrid turbos, 550cc injectors, wagners, 6-speed, etc etc.
It's been fitted with EBC and N75 for now just hangs in the air. So, basically ECU is just running blind in terms of boost control.

The bin is 0002 (366304 IIRC) of the 4D1907558.

The problem I encounter is a random throttle cut after a few seconds WOT run. Attached is the log which shows 1-2-3-(TC) pull, then after some time it "recovers" and allows WOT again. The second TC occurs after ~5 secs in 4th gear full throttle. Excel file has those moments highlighted. CSV is for your convenience.

Now, before you start insulting me and saying that I should stay away from tuning cars - I didn't build, nor tune this car.  ;)
I'm more than willing to "do it properly", to return to N75 control & get rid of the silly EBC, "polish" the tables and so on. There are also numerous other problems, such as the need for 5120 mod, bigger MAFs, probably bigger injectors or bigger FPR - that's still a long way from now.

I'm not exactly a complete noob at tuning, I spent around 2 months a couple of years ago creating a "proper" OLS file for this bin, trying to define as much tables & variables as I can. Also spent some time trying to reverse the bin in IDA, but didn't get as far there as I'd like.

Anyways, what are the possible causes for the TC?
Ruling out knock & temps since logs do not show anything bad there, leaving:
positive/negative boost deviations according to WIKI(which we will see A LOT of with standalone BC).
The second pull (single gear) rules out (or does it?) the underboost condition - there are whole 5 seconds where actual boost is pretty much at requested level.
For overboost, KFDLULS has been FF'd out.

EDIT: Details (http://nefariousmotorsports.com/forum/index.php?topic=8622.msg77253#msg77253)

Any codes?

Nopers. Well, apart from N249 disconnected and some random lean/rich codes related to part throttle fueling. Occasionally also MAFs max out (on cool weather mostly).
Nothing related to boost.

Continuous knock, switches to LDRXNZK which is presumably much lower, you get throttle cut with no codes. Knock subsides, you are back to full on retard mode.

LDRXNZK is only ~10% lower than LDRXN in this bin, not 2x+ times.
rlmax_w  goes from 217+ to 99.75.
And it shouldn't knock anyways, - plenty of fuel, compression ratio has been lowered, angles are not too aggressive.
Not that easy! ;)

Although I'm fairly sure that it's some safety feature intervening indeed.

Oh, EGTs are also not that bad, under 1K deg C.

EDIT:
Checked KFLDHBN - it allows 3.0 pressure ratios at and below 60 deg C IAT.
I guess I will log some more "internal" boost-related params.

Here's the "jpg" version of the most interesting stuff happening in that first log (clickable):
(http://i.imgur.com/cQ80HiKh.jpg) (http://i.imgur.com/cQ80HiK.jpg)

Took another log today with bunch of "b_*whatever*" variables, but the only coincidence is  B_furena going from 0 to 1 when TC occurs. But then again it switches at some irrelevant points as well.

EDIT: nevermind, it's clearly related to overboost somehow still being triggered.
LDORXN limits the load at exactly 99.75.
Will try to temporarily upscale it to 227 and test again.

Please post an unedited ME7L log so it can be parsed by ecuxplot

This is coincidence then?

I agree that there are some severe ignition retards happening, but it's only after throttle has been cut already. Most likely due to part-throttle maps being sub-optimal. WOT it's still fairly safe, up to -6 worst case. And I also agree that it should be dealt with, but it's not the case for the TC itself.


Attached is the original unedited log.

EDIT: Did I understand correctly that B_furena means that any safety mechanism is activated?
EDIT2: Any idea what can trigger overboost if the KFDLULS is FF'ed out?

Forget about running near MAP limit without 5150 or a bunch of MAF hacks.

Log ps_w. You'll see it is a mess.

Thought so. This whole car is a mess. :D
But I gotta give credit to builders - they made it run, boost up to 1.87 bar and haul ass. Now I just gotta make it run properly.
I will start work on 5120 in near future, once I get to know this bin a bit better.

Throw out the EBC, run at 20psi with the N75. Problem solved. You'll be able to run a ton of timing and the car will be fast.

I guess I'll have to start with this, yep. Then implement the 5120 if I'm capable enough and turn it up to 27-28 psi peak.  ;D
Anyways, I'll test the upped LDORXN tomorrow and post back the result.

this, for now


Don't bother running this tune unless you are going to drop boost to below 22psi, or do some MAF hacks (KFKHFM/KFLF games). You are WASTING your time.

If anything, I'm spending this time learning. So it's not all wasted. But I get your point. ;)

Unless you get ps_w under control (and req boost near actual boost), you won't be learning much :)

Unmodified logs would be stellar, it is hard to sort out those. I see ps_w is maxed the entire time though.

So, did a couple of things today to experiment further:
At first I tried to FF the ELDOB (1A081 in 366304 RS6 bin, right? I know that it's listed at the other address in that public xdf floating around, but still). And well, that didn't go as planned. Still getting the TC.
The second "check" was to change LDORXN (FF is only at 191.25 load, unfortunately) - and surely enough, the TC is not that severe anymore, with throttle closed "only" halfway. :)
Log is attached.

I guess I will try zeroing TLDOBAN next and see if that helps. :)
I'm just trying to do one thing at a time (and for now is to completely get rid of throttle cut in current config) before moving on to the next thing. So, yea, effectively I'm wasting time - but that's not the first time I do that! ;D

STOP ITALIAN TUNING.

Seriously. At this rate you are going to break something. There is a good reason all of those safety measures are in place.

Are you running manual transmission? )) just get back AT ECU online(on connector) and fuel controller - your problem will be solved.  ;)

Could you expand upon/clarify what you are saying here?

Thanks!

You are wrong about the initial source of the problem.

EDIT: Actually, it just so happened, that he isn't wrong. The whole problem was caused by this (http://nefariousmotorsports.com/forum/index.php?topic=9220.msg81718). 366304 RS6 software uses the LDRLMX (therefore, LDORXN) mechanism to limit the load if it can't get any feedback from fuel controller. It usually happens under prolonged WOT runs. Basically, there are some maps that control this whole behavior, but for now I just patched this function out in ASM.

TLDOBAN "hack" worked by the way.
Couldn't test it fully yet (with log), but a quick 1-2-3 WOT pull didn't trigger the throttle cut.

EDIT:
It actually didn't work, but I was under the impression that it worked due to LDORXN being FF'ed, so I didn't notice the intervention.



Next on schedule are:
- MAFs maxing out. Trying to decide which route to go here.
- 5120 hack & going back to N75 boost control.

nyet, with all due respect: I do completely agree that in properly built and tuned car, all "stock" safety measures should be utilized and scaled accordingly. But this car is built in such a way that triggering of said measures does more harm than good. If you think that I should stop documenting my efforts here - I will do so.

You have to realize, the 5120 was just finished recently. It marks the first time we've been able to do things "the right way" since inception. As a result, we get enthusiastic when the opportunity to use it is presented. It's nothing personal, just as a model we tend to try to encourage good practices.