Narrowband Tuning

[Deeker1]

I am very new to tuning and have been reading ALOT. I understand the reason to use a wb but what is the difference with the O2 sensor?
Wouldn't it be more accurate to use a mv signal?
With dual stock exhaust which O2 sensor would I use since it has one for each bank, or is this something that requires two?
Not sure if I am suppose to try and put a wb O2 sensor incross-over pipe?
I know these may sound stupid but trying to answer some thoughts as I know each bank can possibly fire differently. Thanks!

[Mep_q8]

i've looked at the NBs 1000s of time and per vehicle they're all different.. 12.5-13 on one vehicle might read 920mv where as on another vehicle it'll read 890mv.

NBs are good for tuning idle and part cruise tables via the LTFTs (personally) when the vehicle is in closed loop. Otherwise stick with a WB for WOT

[HawkZ28]

Agreed.

I myself disable LTFT and tune solely on STFT as LTFT tend to collect garbage. If you tune via LTFT, make a custom PID for LTFT + STFT to log off of....it will go faster and be more accurate.

[DSteck]

I use LTFTs, because they way they are developed is different from just averaging STFTs... Just drive for a long time to let the trims completely develop. It is a longer drive cycle, but I find myself making less changes to get it right.

As far as which bank, I prefer the driver side bank as it seems to run leaner out of the two.

[DSteck]

Thanks alot guys, this helps. If I understand correctly, by using the chart to convert mv to afr during closed loop the vehicle can be tuned. I know that the ecu goes into open loop at wot but if s reading directly off of the O2 sensor say on the driver's side bank, couldn't the same mv conversion be used at wot sense I am reading my O2 sensor directly?
My G8 uses a resonator/muffler as it's cross-over exhaust circuit for lack of a better term/explanation. Would it be betremove this and install a true x-pipe with the wideband O2 sensor in the center thereby picking up a mix of both banks or is the drivers bank just the better way to go sincseems that you say it seems to run leaner.
Sorry for the long questions but sure do appreciate the help!

Let me make this perfectly clear:



Don't use narrowbands for wide open throttle tuning.

Edit: Damnit, the guy deleted his post.

[HawkZ28]

^x2 on WOT tuning, and LOL!

[BBA]

I approach it this way: If you disable PE and FORCE closed loop, with LTFT disabled, DFCO disabled and fuel trim learn forced on, you will be tuning for stoich (lambda=1). I do this for all load ranges (I usually turn timing down about 4 degress to keep it safe).

Once lamdba is dialed in in closed loop, re enable PE and such and final lambda under WOT is simply a matter of setting PE multipliers. The way to get the best WOT AFR overall is to do about 5 runs while logging calculated torque. Each run done with incremental changes to PE, then saved to excel format. Once the runs are logged and saved to spreadsheets, compare the spreadsheet data and find the PE values that produced the highest calculated torque at each RPM and use the results to produce your final PE values. You will likely get better results this way than using an actual dyno.

BTW: Stop talking about AFR, AFR is a bullshit measure since each region will have different blends of fuel (ethanol) that require different AFRs for both stoich and max power. Lambda is the way to go. Lambda=1 = Stoich no matter what the actual AFR is.

[BBA]

Let me make this perfectly clear:



Don't use narrowbands for wide open throttle tuning.

Edit: Damnit, the guy deleted his post.

Let me make this perfectly clear: You have to qualify your statement for it to be accurate, because the way you present it, it's wrong. If you are saying do not use 900mv as a simulation of wideband reading...then I agree, but using 450 MV for stoich works every time with narrowband.

You can use stoich closed loop narrowbands for WOT to dial in VE and MAF perfectly. Done it myself several times and it works great. (narrowbands at stoich are more accurate than any wideband sensor with over a few hundred runs on it will be).

[DSteck]

Yea, good luck on a 9:1 DCR engine doing that. Why cheap out and do it a ghetto way?

I hope Banish chimes in with this crap.

[BBA]

Why cheap out? How many cars do you street tune that have a wideband O2 installed?

Thanks for the comment though, much appreciated.

[umrjmac]

He has his own wideband that he installs on a car when tuning.

[BBA]

He has his own wideband that he installs on a car when tuning.

That's great but how does he know when he makes max power? If he tunes for AFR he will never know.

[DSteck]

That's great but how does he know when he makes max power? If he tunes for AFR he will never know.

I use a dyno when it's all said and done. The calculated torque in HP Tuners isn't reliable. That said, I've yet to have a car that gains more than five horsepower after going on the dyno for tweaking.

[NICKY BOY]

subscribed!

keep it goin

[Mep_q8]

The calculated torque in HP Tuners isn't reliable.

you can say that again.. Read your calculated torque on an N/A WOT run, then read it on a 150shot run

[DSteck]

you can say that again.. Read your calculated torque on an N/A WOT run, then read it on a 150shot run

If I could ever find a g-meter with a 0-5v output, I'd use that and just graph acceleration, but alas...

[blkscooby]

If I could ever find a g-meter with a 0-5v output, I'd use that and just graph acceleration, but alas...

Has anyone ever used DataLogLab- http://www.flyinmiata.com/dataloglab/

If you can dump out a .CSV data file of time, RPM, and throttle position, you can define any car based on weight, gearing, tires, drag, etc and plot HP and TQ. I've been using it for years to do road dyno plots when tuning Subarus.

[BBA]

Has anyone ever used DataLogLab- http://www.flyinmiata.com/dataloglab/

If you can dump out a .CSV data file of time, RPM, and throttle position, you can define any car based on weight, gearing, tires, drag, etc and plot HP and TQ. I've been using it for years to do road dyno plots when tuning Subarus.

That's what the PCM does. You can do a custum table to add a dyno calculation into it as well. Again, depends on a lot of variables being right, so it's only a reference to use for back to back runs of minor changes made, to be able to directly compare to each other.

[S2H]

Just to clarify for anybody who reads this...

your Narrow band o2's are only accurate at Stoic...
anything outside the small center section of stoic represented by (typically)450mv is a shot in the dark.

the narowband sensor's design makes it very accurate at this one point, but allows for exhaust temperatures to skew any other value..

so one minute your car at 13:1 could be 850mv... with a change in the heat of the exhaust temps.. the next minute it could read 900mv...or 1100mv..

each brand of o2 sensor reacts to heat differently and many different versions have different qualities of heater elements for this reason as well...


there is no way to use a narrowband o2 for WOT...impossible by sensor design flaws

[BBA]

Just to clarify for anybody who reads this...

your Narrow band o2's are only accurate at Stoic...
anything outside the small center section of stoic represented by (typically)450mv is a shot in the dark.

the narowband sensor's design makes it very accurate at this one point, but allows for exhaust temperatures to skew any other value..

so one minute your car at 13:1 could be 850mv... with a change in the heat of the exhaust temps.. the next minute it could read 900mv...or 1100mv..

each brand of o2 sensor reacts to heat differently and many different versions have different qualities of heater elements for this reason as well...


there is no way to use a narrowband o2 for WOT...impossible by sensor design flaws

Good clarification, although incomplete.

I'll clarify the rest: If you want to tune with narrowband, you MUST tune ALL ranges with narrowband for STOICH, even WOT. You MUST do this in forced closed loop mode with PE disabled (using scanner controls).

Once ALL ranges are tweaked for zero trims (INCLUDING WOT) you can then set lambda by PE value and verify by test runs, either logging calculated torque or quickest 0-60 times, or both for double checking.