PDA

View Full Version : 02 sensor and Open Loop behavior



BHCevo
11-14-2006, 07:09 PM
Does the front O2 sensor affect fuel delivery at all during open loop (WOT) ?

The reason I ask is because I can't seem to achieve a stable AFR no matter what values I put into the fuel map tables (using ecuflash,evoscan and Zeitronix wideband). I'm trying to run rich (10.0) but half my WOT runs are quite a bit leaner (10.5 to 11.0)
the other half are 10.0. I also had this problem when trying to run quite a bit leaner. The variations are always towards more lean.

Logging the injector duty cycle (idc) shows that the lean behavior is associated with idc's that are lower than they should be, so it is definitely running lean because the stock ecu is telling the injectors to deliver less fuel. I'm hitting the right load cells (as seen in logs).
Stock airbox and maf btw. Why is my stock ecu underdelivering fuel half the time?

I've been running through possible reasons and so thats why I want to know about the front 02 sensor. I'm also using the helix aftermarket 02 sensor housing.

trinydex
11-14-2006, 07:14 PM
you're calculating the load in the datalogs?

the load equations aren't quite correct so which adjacent cells are in the "afr range" that you're actually seeing?

BHCevo
11-14-2006, 07:36 PM
Yeah....I'm aware the load calcs aren't quite correct. I've richened up a huge swath of the fuel map in order to accomodate for possible adj cell excursions. Problem still exists.

Everything on my fuel map, from 180 to 240 load, from 2.5k to redline is set about 8.8 (8.3? its all the same whatever it is) or so. My typical load is somewhere between 200 and 230 depending on boost.

Possibilities I've considered:

1: faulty injector(s) --- Unlikely. Running lean because the IDC is dropping.
2: bad MAF --- Unlikely. bad MAF would show up as anomolously low load. Load during lean events is just like load during rich events.
3: wrong load cell? ----Unlikely. Richened up huge swath of the fuel map see above.
4: bad wideband? --- Unlikely. When it goes lean I get knock and pulled timing and my butt dyno shows no power. This is a real event.
5: bad fuel pump? --- Unlikely. Running lean because the IDC Is dropping.
6: bad 02 sensor? --- Unlikely if open loop involves no feedback. Otherwise a possibility.

BHCevo
11-14-2006, 07:58 PM
The fuel underdelivery problem is irritating as hell. Half the time the car doesn't pull well.

If I pulse (pump) the throttle during WOT I can force it to richen up a little. You can feel the power coming back when I do this. Very very irritating...I know the power is there.

EVOate909
11-14-2006, 10:14 PM
the ECU doesnt get signal for the o2 sensors wen its in open loop

BHCevo
11-15-2006, 12:35 AM
Yeah, thats the strict definition of open loop. But depending on how they implemented the open loop fuel delivery, it might involve some feedback from the 02 sensor. I want to know from someone who has some direct experience...I know the theory as well as anyone.

Regarding what you said about the ECU not getting a signal from 02 sensor in open loop...are you just quoting the theory of open loop to me? Or do you have some direct experience with exactly what the EVO open loop scheme involves?

trinydex
11-15-2006, 06:14 AM
how did you deduce that there might be feedback from the o2 sensor? your wideband should be reading right... but that said why would there be feedback from the o2 sensor? there are cars running around with like afrs in the 9s.

is it possible that you're running off the load chart? have you rescaled your load tables?

EVOate909
11-15-2006, 03:51 PM
if u know the fact of open loop not getting signal from the o2 sensors...y would you think that theres any feed back from the o2 sensor

during open loop the ecu runs rich until the o2 sensor heats up and gives proper signal

C-Spec
11-15-2006, 03:52 PM
you're calculating the load in the datalogs?

the load equations aren't quite correct so which adjacent cells are in the "afr range" that you're actually seeing?


+1000 to what trinydex is saying. which mean no to your questions. Also to figure out what Load/rpm column to adjust. Malibujack on Aktivematrix.com have made a software to firgure all those stuff out and it's free. O0

Alfred@TTech
11-15-2006, 05:18 PM
If you pm me a list of your mods maybe I can be of some assistance. Fueling issues can be cumbersome especially when your dealing with a boosted car. Let me know and we will se what we can do.

Alfred

trinydex
11-15-2006, 05:25 PM
after the problem is solved please post up the solution

BHCevo
11-15-2006, 05:34 PM
triny:

I'm not running off my load chart. My max load column is 260 and my calculated load has never gotten past 240.
So I don't thinks so. Hmm.

EVOate909:

Err...I think you meant to say "during closed loop the ecu runs rich until the 02 sensor heats up and gives proper signal.

thanks for the feedback.

In any case, the narrowband 02 sensor can't give useful data at typical AFR values during WOT. So it seems unlikely that the data from the 02 sensor would or could be used in any meaningful way during open loop operation.

Oh...and the ecu is always physically and electrically connected to the front 02 sensor. So I'm sure that the ecu is seeing 02 sensor voltage at all times...even though it might not be doing anything with it during open loop.

Closed loop is implemented to optimize fuel economy and minimize emissions. Open Loop is implemented for optimizing power...in part because negative feedback is too slow during WOT. There is no intrinsic reason feedback is not useful when seeking optimum power fuel delivery.

If I was implementing an open loop fueling scheme for a car, and I had accurate, if delayed, AFR feedback you bet your sweet ass I would do my best to use these AFR values to help massage the open loop fuel delivery. The feedback would be a minor portion of the fuel delivery algorithm, and thus the character of the fuel delivery would still be mostly "open loop".

earlyapex
11-15-2006, 06:41 PM
Listing your mods would be helpful. Do you have an aftermarket intake?

Most of the times when I see unstable AFR from pull to pull it's either a boost leak or something wrong with fuel pressure, be it the pump or the FPR or hamsters in the tank.

BHCevo
11-15-2006, 07:41 PM
Mods: 03 evo 8

Walbro 255 fuel pump
Buddyclub camshafts 264/272 intake/exhaust
Cam gears at 0/0
TBE
Helix 02 housing
MBC between 1.2 and 1.4 bar (17 to 20 psi).
Stock intake. Stock airbox with HKS foam air filter.

If i had a boost leak my AFR would fluctuate richer, not leaner. I only get leaner. Yes?

Wait...what if my "lean" condition was the good one and the richer condition was the boost leak? Hmm. I dunno. I've varied the values in the fuel map quite a bit and i'd bet 10.0 afr is what 8.5 on my fuel map actually means. Could be wrong.

Good suggestions earlyapex ty.

Fuel pump...possibly...

Just tested the FPR at low load conditions seems to be ok.

Ill post up my fuel maps when I get home.

aniki
11-15-2006, 10:50 PM
the ECU does not use O2 sensor as an input durring open loo,p an O2 sensor switching time is too slow to have any effect on fuel mapping at WOT....hypothetically if you had any type of sensor failure causing it to go lean in open loop i would say MAF sensor or TPS

BHCevo
11-16-2006, 01:37 AM
Thanks aniki. What you say makes sense....I hadn't thought of the TPS sensor. I know an MAF blip will affect the fuel metering and AFR accordingly...not sure how the ecu uses TPS info....all the info that TPS provides you get that and more from MAF no?

aniki
11-16-2006, 02:01 AM
are you able to log ECU input signals with your lap top durring the runs? like record voltage on maf sensor TPS coolant temp sensor ect just compare everything from a good run to a lean run see what is varrying

trinydex
11-16-2006, 03:47 AM
btw i'm wondering why the fuel curve is so rich to begin with. why are you shooting for the 10s? isn't 10.9 like the highest you ever wanna be rich wise? most people tune for 11.1 or so right?

you are mashing the pedal right? tps sensor should be saturated. if it's indeed not then you might have a faulty one altho i'm not sure why it'd only fault at high rpm....

Blak94GSX
11-16-2006, 02:33 PM
Post up some logs and the ROM file.

It seems like purely a tuning issue and not a physical problem with the car since the AFR doesn't change that much. Typically if there is a real problem the AFR will go REAL lean and the car will either misfire or basically stop running.

A boost leak will cause it to go richer since the air is metered at the MAF and then lost along the way to the motor. An intake leak between the MAF and turbo would make it go leaner, but not in the way you are describing and usually at low load only.

There is also going to be some differences in the way the car runs pull to pull. If you are trying to make the AFR consistent under all conditions you are just chasing your tail. The key is to tune it so it runs good on average, instead of perfect once.

Ultimately I think the problem lies mostly in the main fuel table. The fuel curve isn't linear from load column to load column, and the load always varies a bit from pull to pull and gear to gear.

The ignition timing is also a factor and will move the AFR around a bit if it is too far off, but this is rarely a problem unless you drastically changed the linearity of the timing table.

BHCevo
11-17-2006, 05:21 AM
I've tried tuning for 11.0 and low timing...real street driving with these tunes was too erratic in the power delivery. I get much better drivability and repeatable power delivery by tuning for richer and higher timing. Drivability plus repeatability >> single dyno result.

Here is the download link for the ROM
http://www.yourfilehost.com/media.php?cat=other&file=peter102006_rev3

Here is a log showing AFR going from lean to rich in the space of a few consecutive WOT events.
4th gear, stock boost, freeway driving.

http://www.socalevo.net/gallery/displayimage.php?pos=-11294

Here is a log of a bunch of consecutive 4th gear WOT events...yeah for the 110 at night.
AFRs were exactly right (10.0).

http://www.socalevo.net/gallery/displayimage.php?pos=-11291

Here is a log showing AFR going from rich to lean in the space of a few consecutive WOT events.
5th gear, stock boost, freeway driving.

http://www.socalevo.net/gallery/displayimage.php?pos=-11293

Here is a link to jpeg of the fuel map

http://www.socalevo.net/gallery/displayimage.php?pos=-11295

Blak94GSX
11-17-2006, 03:04 PM
Post up the logs and the rom file. The actual files are easier to work with.

BHCevo
11-17-2006, 04:49 PM
Ok. The ROM file link is available above. I provided the jpeg fuel map just for quick access.

I'll post up the (long) logs that the graphs above came from. They will require some serious parsing to find the WOT events out of 30+ min of mostly freeway driving.

BHCevo
01-27-2007, 02:32 PM
Finally...problem corrected. See here for solution : http://www.socalevo.net/forum/index.php?topic=37135.msg6938458#msg6938458

Turns out it was a tuning issue...an obscure table in ecuflash allowed for extra fuel to be added at 4531 rpms...this table has nothing to do with openloop fuel tables.

It looks like some people are happy with tuning with this as is. In my case the afrs would be ok in 3rd gear but too lean in 5th gear (5th gear is below 4531 rpms, 3rd gear is above 4531 rpms).

This made a good tune for 3rd gear a way too lean tune for 5th gear. Too lean meaning big detonation and loss of power.