Understanding ECU's interpretation of narrowband O2 sensor

[slipnfall]

Hi Folks, I'll cut right to the point: I'm trying to understand the values coming out of the CONSULT diagnostic port in regards to closed-loop O2 feedback. I have a datalog, linked below(around 300k, you'll need to unzip it), that has oxygen, A/F base, and rich/lean fields. The rich/lean is easy to grasp, but what are the other two? I understand a little bit about how an O2 sensor works(specefically the nernst type, or two-state narrowband), and have a background in electronics. I am not however, a chemist! Are the other values derived from the timing of the rich/lean, or are they temperature/MAF compensated by some means?

Any basic insite would be appreciated. It's nice to have datalogging, but it's worthless if you don't understand what all the values mean.

Thanks.

BTW the Excel workbook has a second sheet with a basic graph. The beginning was trimmed off, but it was basically a semi-warm startup, onto the HW to cruise at around 45-50mph(where I'm having some hesitation). S13, KA24DE stock BTW.

[NISTECH]

I would look it over but my comp doesnt accept that file extension. I am running 98se

[slipnfall]

I forgot to mention it's a zip(compressed) file. The file is large, that's why I zipped it.

I made the file below self-extracting, so you don't have to have the program(Winzip) to open it. It will ask you where you want to put the file, just pick a place(desktop for example).

http://home.pct.edu/~klijam64/...d.exe (around 0.35Mb)

You can alternately download the file directly, but it's big:http://home.pct.edu/~klijam64/...d.xls

Sorry about that.

[NISTECH]

nah I got a zip wizard its the .xls extension my comp dont work with.

[UK-SRi]

Well those other figures are calculated from the duty cycle of the nernst cell signal, a moving average of sorts.

Without a wideband its all guesswork, even for Nissan.

Mike

[slipnfall]

Oh, I take it you don't have microsoft Office? What type of spreadsheet program do you use? Works?

[UK-SRi]

Well without a wideband signal pulled into the spreadsheet for comparison you are pissing into the wind anyway.The stoich point and slope moves with temperature quite badly.

Mike

[slipnfall]

Ok so the A/F Base in the log is the 'period' of the O2 cycling rich/lean? And the H20 field is the duty-cycle of sort? (time on/period).

[UK-SRi]

I assume it must be, I dont know at this stage if they are modelling the assymetry in the Zirconia sensor or if they assume a pure step response, it would take a lot of effort to work out that one i think.

Mike

[Magnes]

You may be able to use the free Excel Viewer available here:

http://www.microsoft.com/downl...ng=EN

They say it is 'supported' on Win 2000 and later, but that doesn't mean it won't work with Win 98 AFAIK.

(EDIT)- I uploaded a graph of the data here: http://img188.imageshack.us/my...k.jpg
Modified by Magnes at 12:04 PM 11/15/2005

[Magnes]

Oxygen is the O2 sensor output. It outputs between 0 and 1V (here it reads between 0 and 100 on the data) and a healthy O2 sensor should switch back and forth between 2-5 times/second depending on RPM.

A/F base is the air/fuel mixture target relative to stoechiometric, which is based at 100. So an A/F value of 105 is 5% leaner than stoechiometric, and the PCM meters the injectors to acheive that A/F goal.

[NISTECH]

A/F is the Alpha If so A/F alpha is a calculation of primary sensor inputs i.e. maf ,coolant temp ,crank angle sensor,and O2 readings and possibly more. It has a base setting of 100 as mentioned above. The ECm calculates how much fuel it needs to add or take away from base to aquire the required fuel ratio to maintain stoechiometric as mentioned. The reading is not a direct sensor input its sole purpose is for diagnostic reasons. It is there to tell you if the ECM has to adjust the fuel mixture to meet stoech because something may be wrong. If it is able to maintian a semi steady fluctuation between 91 and 109 at 2000 rpm not driving it is considered ok. My personal opinion is if it hangs below 95 or above 105 and just doesnt come back close to 100 you have something going on.

[slipnfall]

Thanks for the link above for the viewing software, good find!

A long time back when I brought it to Nissan, I had them do an O2 sensor test. They basically held it at 2k, and had to wait untill it warmed to 190* IIRC. In fact we had to blanked the radiator to get it that warm. The CONSULT tool wouldn't even allow them to do an O2 test untill it reached that temp. I'm guessing this was to allow the O2 to warm up(since it is a non-heated type).

Anyways, I think I'm getting a better understanding of the values. I just assumed that the readings were a direct reflection of wether the mixture was rich or lean, not factoring in other variables. So if the reading is dependant on so many other things, is this why a wideband is so superior? I take your explaination to mean that, for instance, a 'lean' condition at one MAF voltage/condition may not really be lean in another condition. Make sense?

It's getting late, enough for one night.

Thanks everyone.'slip

[Magnes]

I don't know what 'wideband' or 'narrowband' is.

I also don't have any experience analysing engine datalogs, so we would need to compare yours to a datalog from a known good vehicle for comparison purposes. Also, if you can indicate the exact period during that data run when you experienced the hesitation, that would help.

Other than that, the only thing that caught my eye was the 'noise' recorded at throttle tip-in, where the throttle% line shows a rapid opening and closing of the throttle in a period of less than a second. Maybe a worn TPS or loose/damaged wiring?

[NISTECH]

Nissan uses a traditional zirconia sensor for its narrow band sensor. This sensor must fluctuate from below .175 volts to above .8volts. It will use that whole range during operation. This sensor simply moniters oxygen content in the exhaust and reports it back by a voltage generated within its inner "plates" Although it has been very successful in keeping fuel economy and emissions in check a new and improved sensor has hit the auto industry. Actually I believe GM was the first to use it a few yrs back. Anyhow this is a wide band sensor. This will be about the most simplified explination of it. It is essentially 2 zirconia sensors in one. one of the sensors moniters oxygen content while the other attempts to keep it at .450 volts all the time, not allowing it to fluctuate like the traditional sensor. This is done by a chemical reaction with in the sensors [ the same reaction that happens in a traditional sensor generating a voltage.] The difference is voltage is applied to the other sensor having I guess what you would call a negetive effect on the reaction the diminishes the voltage when oxygen content is high in the exhuast. The ecm is watching the miniscule ampreage checnges it takes to keep the other sensor at .450volts. This thing about the wideband is it cant be monitered with a scan tool,voltmeter or DSO for any ACCURATE diagnostics. When functioning right you will almost never see it leave a steady signal [no fluctuation AT ALL]. There are controdictory test methods on the internet but at this point none have been proven to be accurate methods of testing.

I could go into more depth on how they function but that would be an entire article and for all practical purposes knowing how it works does nothing for diagnostic purposes since it cant be tested accuratly.

[slipnfall]

I found some FAQ's/articles at http://www.wb02.com to be very informative. (Especially the manufacture's datasheets on the specefic devices.) The principal is very much like how a hot-wire MAF works, in regards to the ballancing affect. From the little I have read, electrically speaking, it is a wheatstone bridge of sorts.

[RobertsnewQ]

Dunno if it helps, but Nissan ECUs don't use any kind of curve for the 02 sensor. Basically there is a "rich" voltage level and a "lean" voltage level. There is also a "test" voltage level (.4v) that must be reached before the O2 sensor can be considered ready.

The ECU tracks the number of times the sensor(s) hit the rich and high voltage limits and use that as a way to tell if the engine is rich or lean. The relative number of hits of the rich or lean voltage limit determines the amount of correction from the ECU.

[NISTECH]

Dunno if it helps, but Nissan ECUs don't use any kind of curve for the 02 sensor. Basically there is a "rich" voltage level and a "lean" voltage level. There is also a "test" voltage level (.4v) that must be reached before the O2 sensor can be considered ready.

The ECU tracks the number of times the sensor(s) hit the rich and high voltage limits and use that as a way to tell if the engine is rich or lean. The relative number of hits of the rich or lean voltage limit determines the amount of correction from the ECU.

This is good info for early model nissans but when OBDII came around it no longer used that stradagy. You can veiw O2 sensor curves with consultII on all consult compatable nissans all the way back to 90. So the ECM does use a curve on some pre OBDII nissans as well.

[Q45tech]

Still just a switch function............to help ecu zero in on a combined exhaust bank 14.7 AF except it is fooled with oxygenated fuel to produce an overly rich mixture by whatever % the oxygenates adds..........with 10% ethanol typically runs a real 14.2AF..........why oxygenate fuel consumption is typically 5% worse at cruise than non oxygenated fuel.

Remember the ecu looks up a point in a table or calculates a point from all sensor inputs that might vary 30% at a specific cfm intake flow [temp, barometric]......plus/minus any form of protection [knock sensor/coolant temp] plus ac on signal and specific gear program.

Without an egt and O2 on each individual exhaust port it is an average before OBD2, with OBD2 the ecu can view each port [moment in time for MIL] but doesn't trim each individual injector to optimize........still just an average.