PDA

View Full Version : Narrowband 02 readings



Kieran
31-08-2007, 11:14 AM
Peeps...

Can someone just confirm for me, that on the VR-4's engine, A voltage of 1.00v from the stock narrowband sensor equals stupid rich (in narrowband terms, I know they have a very limited AFR range) and that 0.00v is Leaner than a starving thing's lean bits?

Ta.

MikeKey
31-08-2007, 01:45 PM
Unless I'm mistaken, narrowband sensors don't work in this way. They swing constantly between 0 and 1 volts, and the ECU just injects more or less fuel to keep the "center point" of this voltage at a certain value.

If you measure the voltage across your sensor with the engine running, you should see the reading swing back and forth between 0.1v and 0.9v, or thereabouts.

Mike

Kieran
31-08-2007, 03:46 PM
Unless I'm mistaken, narrowband sensors don't work in this way. They swing constantly between 0 and 1 volts, and the ECU just injects more or less fuel to keep the "center point" of this voltage at a certain value.

If you measure the voltage across your sensor with the engine running, you should see the reading swing back and forth between 0.1v and 0.9v, or thereabouts.

Mike

Not quite... Yes, they do normally oscilate around the 0.45-0.5v mark at idle or when in closed loop (0.5v being Stoich), for the reason you say - the ECU is always varying the amount of fuel injected (however slightly) to keep the AFR at around 14.5/14.7 AFR (again, in closed loop) - The voltage that the 02 sensor feeds to the ECU is dependent on the amount of oxygen present in the exhaust stream. It certainly shouldn't swing right across the entire volt range however, even though this is only a very narrow AFR range - Unless there's a problem.

But, what I need to know is whether 1 volt corresponds to 'Rich' or 'Lean'... From what I can glean, 1 volt should equal 'Rich' and 0.0v is 'Lean' - But can anyone confirm this, please?

Gowf
31-08-2007, 10:37 PM
Yes Kieran you are indeed correct. This is why when you disconnect the stock narrowband it will run rich, as it sees it as its most lean condition.
But yeah, it should cycle as its trying to achieve its afr of 14.7 which as you have correctly pointed out, is only in closed loop. As soon as you put your foot down it doesnt care about cylcing to get the desired afr, it just fuels as per the table.
This is why its pointless trying to map for cruise if the stock sensor is connected (unless of course your fuel computer has the ability to map in closed loop) as the ecu will be fighting your imput with the fuel trims.

Went on a bit there, what do you need to know for?

Kieran
02-09-2007, 07:53 PM
Thanks Gowf.

I needed to know because I think I've goosed my 02 sensor like I did my cat converter (MMT Poisoning, article coming up on that). I'll post the charts later but what I'm seeing is the narrowband reporting scary lean running at WOT as the revs climb - whereas the rolling road results showed that my car was running so damn rich the dyno wasn't able to measure it beyond 4500ish rpm - it's long been on my 'suspect' list, but I've not had any data logging untill recently to prove my theory.

Nick Mann
02-09-2007, 08:19 PM
But the O2 sensor is not monitored at WOT. The car uses pre-defined maps at that point? The O2 sensor is only used whilst cruising.

Kieran
02-09-2007, 08:44 PM
But the O2 sensor is not monitored at WOT. The car uses pre-defined maps at that point? The O2 sensor is only used whilst cruising.

Does it not reference the 02 sensor at all then?:inquisiti

If that's the case, 1'm back to square frikkin one!/Grrr Anyone got some rags and a zippo?! /Grrr /Grrr /Grrr

Nick Mann
02-09-2007, 08:52 PM
I am 90% sure that I am right. I am sure someone who has more knowledge will be along soon to confirm or deny.

Nutter_John
02-09-2007, 08:54 PM
Yep Nick is correct once the TPS goes past (30% I think ) and the rate of change is high it goes into open loop mode and ignores the O2 sensor

Kieran
02-09-2007, 08:55 PM
I am 90% sure that I am right. I am sure someone who has more knowledge will be along soon to confirm or deny.

Damn, I think you are right....In fact, I already knew that now I think about it.... /duh

http://www.stealth316.com/2-fuelinjection.htm

Generally nost 'theory' on that site seems to apply to us.... So.... why am I seeing the 02 sensor spit a lean condition then?! /help

A thousand assorted and inovative curses!!:furious3: :furious3:

bradc
02-09-2007, 09:21 PM
when does it throw out that lean condition? what sort of driving?

Also it may just be dirty or need to be replaced.

Kieran
03-09-2007, 12:23 PM
I do think it's knackered too, Brad.

Anyway, have a look at these two graphs. They show the relationship between my narrowband voltage and RPM and Throttle position. I would bung them on one graph but excel doesn't support multiple Y axes..../pan

FYI, I have a graph of knock levels too. The knock sensor is picking up engine noise (you can see this through the voltage), but the actual 'Knock sum' never goes above two, so I'm pretty sure it's not detonating itself to death.

Oh - to make these graphs easier to read, the 02 readings are a 5 point moving average trendline - the raw data draws a messy graph as the voltage is swinging in closed loop running - but even so, the average closed loop readings are high, don't you think?

Anyway - Lambda vs RPM - Blue is RPM, pink is Lambda. As you can see, as the revs climb (I am at WOT), the sensor flips rich (0.9v), then suddenly goes lean!

uploaded/326/1188818185.jpg

And here's the same graph, but this time plotted against Throttle position - As you can see, it's still reading after going WOT (not in closed loop) and then suddenly goes haywire?

uploaded/326/1188818348.jpg



Or - Is this normal behaviour, anyone?:inquisiti

bradc
04-09-2007, 07:20 AM
To be honest, those are a bit hard to interpret. Me thinks you need to go wideband :)

Kieran
04-09-2007, 12:29 PM
To be honest, those are a bit hard to interpret. Me thinks you need to go wideband :)

Sure, but I'd like to pin this one down (if there *IS* a problem) before I just cover over it, if you know what I mean.

Anyone else?! /help

Eurospec
04-09-2007, 01:55 PM
Sure, but I'd like to pin this one down (if there *IS* a problem) before I just cover over it, if you know what I mean.

Anyone else?! /help
Kieran,

I would be 95% sure your narrow band is goosed.

Thats typical behaviour that they show lean at WOT and only partially cycle at cruise when they are fubar'd.

The ECU wont use the O2 sensor once it jumps for open loop, so i wouldnt worry too much.

Aftermarket narrow band will cost a lot less than mitsu, but will only last a year, whereas the mitsu ones go on much much much longer. If i were you, get a nicew plx wide band with narrow band simulation built in.

Cheers,

Ben.

Kieran
04-09-2007, 02:52 PM
Cheers Ben, I thought as much... Especially after seeing what my Cat converter looked like...:speechles

I've got a 4 pin 'Universal' zirconia sensor here as a stop-gap, though I might just bung a PLX in there anyway.:scholar:

Nick Mann
04-09-2007, 09:20 PM
Kieran - you have an email with my first ever datalog! And I would say you and Ben are right. At WOT I always seem to be reading a rich condition. I know my car is stoopid rich at WOT from the last rolling road day, so the narrowband at least seems to agree!!

Kieran
04-09-2007, 09:29 PM
Kieran - you have an email with my first ever datalog! And I would say you and Ben are right. At WOT I always seem to be reading a rich condition. I know my car is stoopid rich at WOT from the last rolling road day, so the narrowband at least seems to agree!!


:iloveyou: :iloveyou: :iloveyou: :iloveyou: :iloveyou:

I'll repay you in kind when I see you at Japshow!! ;) :deal2: /Banana :love: :joker: :smitten: :inquisiti :inquisiti

Eurospec
04-09-2007, 09:47 PM
I know you wouldnt do it guys, but just so that anyone who searches and finds in the future.......

Dont try and tune with a narrowband O2. The characteristic curve is way too steep for any kind of accuracy, plus you will find different (meaning non mitsu) sensors will have different resolution at the extremes.

Think of a narrowband like a switch, like the stock ecu does. all its doing is going 'ooh, ooh, i'm richer than 14.7, ooh, ooh, now i'm leaner than 14.7' etc

Cheers,

Ben.

Nick Mann
04-09-2007, 09:51 PM
LOL! Thanks Ben! I never thought of an O2 sensor going 'ooh, ooh' before!! :P

Anyway, I can't tune it until you give me a MAP2 for ten bob!

Kieran
04-09-2007, 09:51 PM
Good point Ben.... I can just see the posts now...

"Hey guys, I tuned my car so that I see a steady 0.5v on the stock 02 sensor at all times, but the first time I booted it, i left my ringlands behind - is this normal?!"

Kieran
04-09-2007, 10:00 PM
Nick - thanks muchly for the file.... just plotted the same charts as I posted earlier and my 02 readings, even in closed loop are WAY off compared to yours. Explains a lot!:scholar: