Jason
Hero Member
Karma: +38/-0
 Offline
Posts: 500
Breaks everything!
|
 |
« Reply #30 on: May 18, 2011, 11:40:12 AM »
|
|
|
The oscillation is there for a reason - long story short, 3-way catalytic converters use two processes - reduction and oxidation. The catalyst converts NOx when the mixture is rich, and when lean it converts HC and CO.
The ECU intentionally oscillates the mixture at idle and cruise so one catalyst can convert all 3 gases.
|
|
|
|
|
Logged
|
|
|
|
Jason
Hero Member
Karma: +38/-0
Offline
Posts: 500
Breaks everything!
|
|
« Reply #31 on: May 18, 2011, 11:41:37 AM »
|
|
|
Also, why are you running ECUx in a VM?
|
|
|
|
|
Logged
|
|
|
|
NOTORIOUS VR
Administrator
Hero Member
Karma: +58/-7
Offline
Posts: 1056
|
|
« Reply #32 on: May 18, 2011, 02:05:16 PM »
|
|
|
Also, why are you running ECUx in a VM?
Easier as all my tuning stuff is in a VM |
|
|
|
|
Logged
|
|
|
|
|
julex
|
|
« Reply #33 on: May 18, 2011, 02:26:05 PM »
|
|
|
Also, why are you running ECUx in a VM?
Easier as all my tuning stuff is in a VM Strange that it would disable itself at 4k repeatedly, I sometimes get no logs but it is not a back-to-back event. Are you sure your cable is ok? Let me tell what happened to me once, I bought a several foot extension for odb2 port with 90deg plug so that I don't kick it getting in/out of the car. It worked fine except every now and then I would lose connectivity when flashing/logging so I deemed it "Crap" and wasn't using it. But one day I decided to see what's up with it and continuity/short test on all leads showed perfect cable. So what was wrong with it? With a flashlight in hand I ducked under the dash and plugged the cable... only to realize that the 90 deg bend was just too close to plug itself and a corner of the plug was hitting kick panel and preventing the cable from going 100% in. Sure enough, the 12v lead was not contacting right from time to time. So I just shaved a bit off that corner and since then the cable is 100% reliable. I am just wondering if this might be a factor in your situation. |
|
|
|
|
Logged
|
|
|
|
NOTORIOUS VR
Administrator
Hero Member
Karma: +58/-7
Offline
Posts: 1056
|
|
« Reply #34 on: May 18, 2011, 05:15:02 PM »
|
|
|
^^^ Seems like restarting the VM (instead of always hybernating it) does the trick and it logs again.
On other interesting news, the jerking that I am feeling is something in the NefMoto Stg 3 base tune.... So before when I was saying the AFR's are solid and the car is still jerking means that it's obviously not fuel related (I'm pretty sure it's not timing related either).
I switched to another M-box tune, plugged in some KRKTE and TVUB values and the car drives butter smooth again. So, that makes me feel much better about the ECU and myself lol..
Didn't have enough time to get the WOT fueling under control for the other file (super lean under boost (14.7 lol), but idle and cruise looks pretty good after 2-3 revisions of the TVUB values.
|
|
|
|
|
Logged
|
|
|
|
Jason
Hero Member
Karma: +38/-0
Offline
Posts: 500
Breaks everything!
|
|
« Reply #35 on: May 18, 2011, 06:47:42 PM »
|
|
|
My hybrid K04/K16 car did not like that tune until I restored the stock latency values and stock idle torque reserve. I also nudged KRKTE to .065 with the Siemens 60's @4bar. YMMV... something to try.
|
|
|
|
|
Logged
|
|
|
|
|
julex
|
|
« Reply #36 on: May 18, 2011, 07:12:09 PM »
|
|
|
If you think Tony's stage 3 tune is weird you should have seen that other tune I initially had from *cough* you know who (see very loud thread in reverse engineering section).
For some reason, the tuner threw factory M-box MAF correction table totally off by adding as much as over 20% maf correction in spots. Then they messed the injector correction table just as bad (stock is think 1.00 or 1.02 is all cells) and then to make tings right I think, they adjusted (again to the extreme) the table for main fueling in part load.
There might have been a plan behind the madness since once you combine the three tables for a given load it yields right results but still, what a mess.
Anyway, I tried Tony's Boost PID tables responsible for max/over the MAP values and PID and I overboosted bad to the amount of 27psi so I think he might have very inefficient setup, but in general I started appreciating tuner's job to get custom tunes dialed in.
|
|
|
|
|
Logged
|
|
|
|
NOTORIOUS VR
Administrator
Hero Member
Karma: +58/-7
Offline
Posts: 1056
|
|
« Reply #37 on: May 20, 2011, 02:26:44 PM »
|
|
|
Well did two more revisions on my new M-box file... 1st one was too much KRKTE, wayy too rich and also the boost limit was too low so I was getting throttle cut.
So I revised it again, lowered the KRKTE and voila the car drives amazing! AFR's look pretty good under full boost, but coming into boost it might be too lean in a few spots for my liking so I will start to massage the MAF correction table next.
But for now the car drive like it's stock. Idle back to ~800 rpm, no surging/oscillation at all either...
All in all, I'm very happy right now. I have no logs so after the long weekend I will try to get some to see how the car is reacting and where to make changes in the file.
Once I get that settled in I might try out the LC/Flat shift code, order up some larger meth nozzles, install my 044 and start to tweak the timing advance and boost curve to my liking.
Thanks to all for the great suggestions btw... needed to help me get my head on straight and back on track. I was really getting frustrated with the way the car was driving.
|
|
|
|
|
Logged
|
|
|
|
carlossus
Sr. Member
Karma: +38/-0
Offline
Posts: 394
Leon Curpa Stg1+
|
 |
« Reply #38 on: May 20, 2011, 03:51:50 PM »
|
|
|
One thing I will say, it's really nice that you take the time to post back and update the thread. Even if it seems quiet, I'm following with interest.
|
|
|
|
|
Logged
|
|
|
|
|
judeisnotobscure
|
|
« Reply #39 on: May 20, 2011, 06:48:49 PM »
|
|
|
Thanks for sharing your work in this thread.
I've been checking daily.
Looks like you're making great progress.
Do you plan on trying an e85 tune?
I hope to get my hands on a 605 car next month... I'll let you know how i make out.
|
|
|
|
« Last Edit: May 22, 2011, 05:31:55 PM by judeisnotobscure »
|
Logged
|
I have a b5 s4
but i just want to dance.
|
|
|
Matt Danger
Full Member
Karma: +17/-1
Offline
Posts: 116
|
|
« Reply #40 on: May 20, 2011, 08:05:07 PM »
|
|
|
My hybrid K04/K16 car did not like that tune until I restored the stock latency values and stock idle torque reserve. I also nudged KRKTE to .065 with the Siemens 60's @4bar. YMMV... something to try.
I found similar positive results (smoother on/off throttle) by reseting these maps to stock M-box values. |
|
|
|
« Last Edit: May 20, 2011, 08:11:49 PM by Matt Danger »
|
Logged
|
|
|
|
NOTORIOUS VR
Administrator
Hero Member
Karma: +58/-7
Offline
Posts: 1056
|
|
« Reply #41 on: May 24, 2011, 10:52:52 AM »
|
|
|
Alright so I've got WOT fueling pretty much exactly where I want it now... and I'm super happy about that.
I've got req. load to be inline with actual load as well...
But Idle AFR and Part throttle/cruise AFR's are really starting to piss me off.
Can someone explain to me how this is supposed to work?
From what I've gathered is this (simplified).
Idle and part load lambda is always supposed to = 1. Fine.
TVUB has the greatest impact @ Idle and some part load. The least impact on WOT fueling.
KRKTE is the overall offset for fueling but is the exact opposite effects from TVUB.
KFKHFM is for correcting the MAF scaling during low loads... possibly idle? What is the Axis for the MAF g/s? Is it directly what I see in VCDS? Or is there some conversion needed to use the MAF and STFT's values from VCDS logs?
I've been told to try stock TVUB values... forget it, I just tried and the car run's super lean just at idle and hardly wants to rev (WB shows 18-19:1, hardly possible but still bottom line it's lean).
I've now got stock TVUB values with a multiplication factor of 2.1 added... Thave gave me values around where I used to be with other tunes but slightly more all over... This is giving me a very rich idle as described above, but when around part load from say 12in hg. vac to 0 and into boost values that I would be comfortable with and the car seems to be smooth.
Thing is, extreme light loads is very touch now because it's very rich. Even throttle transitions suck now and make the car jerk.
I'm wondering if I should now start correcting with the FKKVS map for low pulse widths, to try and smooth things out for me seeing that almost everything else is where I want it to be.
Thought's, suggestions?
Thanks for the encouragement as well guys.
|
|
|
|
|
Logged
|
|
|
|
|
nyet
|
|
« Reply #42 on: May 24, 2011, 11:13:01 AM »
|
|
|
What is the Axis for the MAF g/s? Is it directly what I see in VCDS?
Yes. Thing is, extreme light loads is very touch now because it's very rich.
You mean your STFS are very negative, or they are *maxed* negative, so you are *actually* rich? I'm wondering if I should now start correcting with the FKKVS map for low pulse widths, to try and smooth things out for me seeing that almost everything else is where I want it to be.
Maybe. But all this really tells me is that I still think something is wrong with your MAF set up... in which case KFKHFM is there to compensate for badly non-linear MAF setups that the owner is too lazy to fix properly  Either that or you have a boost leak. |
|
|
|
|
Logged
|
|
|
|
NOTORIOUS VR
Administrator
Hero Member
Karma: +58/-7
Offline
Posts: 1056
|
|
« Reply #43 on: May 24, 2011, 11:47:03 AM »
|
|
|
Yes. Good to know, thanks. You mean your STFS are very negative, or they are *maxed* negative, so you are *actually* rich?
Correct. Maybe. But all this really tells me is that I still think something is wrong with your MAF set up... in which case KFKHFM is there to compensate for badly non-linear MAF setups that the owner is too lazy to fix properly Either that or you have a boost leak. Boost leak? I guess it's possible. I will pressure test tonight. But I hear no leak, I make good power for the boost level I'm running, MAF curve looks good in the logs, etc, etc. non-linear MAF setup? Tony's file is supposed to be properly scaled, he claims it as such anyway and the curve does look good under power. But you're right there must be something there. As the other M-box file I tried (with underscaled MAF values) was running butter smooth. So maybe you are right and the MAF values in Tony's file aren't very good in the lower end. Any suggestions on getting it scaled properly? Is it really going to be a matter of driving and logging STFT's and MAF values, correcting them and rinse/repeat? If so that is going to suck. |
|
|
|
|
Logged
|
|
|
|
NOTORIOUS VR
Administrator
Hero Member
Karma: +58/-7
Offline
Posts: 1056
|
|
« Reply #44 on: May 24, 2011, 12:05:45 PM »
|
|
|
I just compared the MLHFM between the two files. They're both linear in my eyes.. just the scaling is different. But between that file (nefmoto) and the one that isn't giving me idle/part throttle issues the only difference (in the MAF table) seems to be the scaling. The one that is smooth can't be because of the underscaling, that just doesn't make sense.  |
|
|
|
« Last Edit: May 24, 2011, 12:07:26 PM by NOTORIOUS VR »
|
Logged
|
|
|
|
|
