after fitting my dawes afr i found the oxygen sensor wasn't working correctly so with help from a friend we used his diognostic equipment to make sure everything was how it should be with the new sensor on(universal titanium 3 wire type).
after fitting it he attached the guage and found the new oxygen sensor was sending a 12v signal out(instead of a up to 1v).
took it back and got a new one,same problem!!!
the type of sensor that i am using is the type the supplier said is the right type for the car(95 wrx)
sensor type-intermotor(1 volt) 3 wire,red(12v live)black(ground)white(signal)titanium universal.
anyone else had problems with their oxygen sensors or know anything about the system used on the impreza.

HELP its starting to get to me now!!!

 

Sounds like you've mixed a heater and signal wire somewhere - I guess the heater's just a coil that heats up when you've got 12v on one side and ground on the other.

If you were to mistakenly connect the signal wire to the lambda's heater ground, you'd get this problem. And I'm not sure the ECU will like having 12volts fired at it.....

Richard

 

got the wires in the right place as they are the same colour as the ones on the standard oxygen sensor(if the universal one uses a different configuration how can i tell???) no instructions!!!

 

jonny,

When i read the instruction for the dawes AFR meter it said you must take the ground from the negative terminal battery only....

Im sure i read that somewhere..!! could be wrong

Dan

 

You should be have a fairly low resistance across the 2 heater wires, but nothing between the heater 12v supply and lambda signal....but not sure about between lambda signal and ground....is the signal wire shielded/grounded? The OE my00 one has a signal ground wire at the ECU.

Good Luck
Richard

 

dan,the afr is working ok,its the oxygen sensor that the afr takes its signal from i am having a problem with.
yours must be working ok as you are getting a signal to your afr.

 

havnt got one yet i was just reading how to fit it..!!

Have you ever had if running with your standard o2 sensor.... did you replace the sensor for the current one... if so what symptoms did you have to replace it..??

You could try running it on someone elses scoob do you know anyone.... this way you know for sure if it is the sensor... if its the same then get on the phone to grade A subaru and get a genuine one...!

Where was it from and was it expensive..??

Dan

 

the universal one is £64 trade

the subaru one is £300 + VAT

not sure how much a 2nd hand one is but i will give grade A a ring tomorrow.

 

the problem i was having with the original one was it was cutting out over 4000rpm(flatline on the test equipment)causing the car to run very lean!!!
would have never found the fault if i hadn't fitted my afr.

 

thought the o2 sensor is not used at WOT....??? and the car uses the maf and other sensors..?

Dan

 

WOT??? whats that stand for.
I was hoping someone that knew how the system worked would give me an explanation as to how it works.

 

anyone?????

 

WOT stands for 'Wide Open Throttle'

How the ECU uses the Lambda signal depends on what set-up you have i.e. What model car is it, year, what type of ECU i.e. Standard - Link - GEMS - Motec etc?

We've had Lambda (or O2 as you call them) sensors pack in, but never only above 4000rpm. The sensor ha no idea what rpm you are doing, so this sounds odd.

Post some more details and i'll do my best to help.

 

The lambda operates all the time - but the ecu only uses the signal while in closed loop fuelling operation to ensure the mixture is best to be cleaned by the catalyst that you may (or may not ) have fitted.

I don't know the exact figures, but once you're above a certain rev or load limit the ecu switches to open loop mode and does fuelling from the maps stored in the ecu.

So a lambda that stops operating at 4k isn't a problem for your ecu operation (except that it probably isn't working correctly at lower revs/loads either!), but is no good if you want to use the signal to get an idea of fuelling.

As mentioned above, sounds like you've mixed up the wires if you're seeing a constant 12v on the signal wire....

Also note: lambda output starts to under-read after extended WOT, it's only really accurate for a single power run.

Richard

 

thanks,thats what i needed to know(owe you one)

 

is it ok to run the car without a lambda if you have a full de-cat system?
fitted the new one today,the car is now hesitating on very light throttle!!!
disconected the feed to the ecu and the hesitation goes away???

 

Lol! What MY car do you have? If it's 99/00, maybe you've just inadvertently found the cause of the light throttle hesitation Does it occur most between 2.5 and 3k, more noticeable when cold? It afflicts all of these MY's to a greater or lesser degree....

Unfortunately, running without a Lambda will (I guess) give poor fuel consumption when in closed loop fuelling at light throttle/loads - it normally uses the lambda to swing the mixture over the stochiometric point (you'll see your monitor leds cycling up and down at idle, if it's connected too). Without lambda it will use map based fuelling (probably pig-rich) in open loop mode and log an error condition in the ecu. You wouldn't pass an emissions check either.....

How bad is the hesitation?

Richard

 

bad but i don't think the sensors i have been using are suitable for my car(my95 wrx)even though intermotor say they are!
(you were right,the signal on the new sensor was the black not the white)direct replacement!!!

on the the diagnostic gauge its reading at half a volt at tickover!!!(should be much lower)

does the my98 use a 3 wire system,if so i can get hold of a sensor from one.

 

Ho hum...

OK first point to note, you bought the wrong sensor. There are two types of basic sensor, zirconia and titania.. well that's what they normally call them anyway even though it's not really that accurate. Back to the point, you need a zirconia type sensor, one with three wires, not the titania which is resistance rather than voltage based.

When you get the right sensor you should have three or four wires. The colours will be two whites for the heater, a black for the signal and perhaps a grey for the ground reference (which is not used normally, but I run mine through to the ECU well for the benefit of the LambdaLink).

Normal sensors like this are narrow band units. They have a sharp swing of output voltage from 0 to about 0.8-0.9V as you go from lean to rich. They cannot tell you how rich you are running, only that you are running rich. Devices like the LambdaLink purport to being AFR meters, they are not, and you cannot rely on them for accurate determination of air fuel ratio. What you can do is, given sufficient experience, use the readings from the LambdaLink to give you an idea of what the fuelling is like. I repeat, they are not accurate in absolute terms and you have to make mental compensations for certain operating conditions.

Given that they are not accurate at determining AFR, it would be stupid to rely on them under boost when getting enough fuel into the cylinders is critical. Hence they are not used on boost. The one good, consistent characteristic of these sensors is that they have a large voltage swing as the AFR passes 14.7:1. This happens to be just what a 3 way catalytic converter needs in order to work. Closed loop lambda control "rocks" the mixture around this point...

When the sensor is reading lean, the fuelling is slowly increased until it swings rich, then it is slowly reduced until it reads lean... and so on, so that the average position ends up being 14.7:1, which means the catalytic converter works as it should. Typically these cycles will happen between one and ten times a second. It is not possible to hold the fuelling at 14.7:1 with these sensors because they response is too sharp at that point, plus the absolute voltage corresponding to 14.7:1 is not constant.

When an engine is run under boost it is much more important to get enough fuel into it than to worry about silly things like emissions, so the engine management ignores the lambda reading and just throws a lot of fuel in. This prevents the catalytic converter from working, but more importantly it also prevents you from getting a 2000 pound engine rebuild fee every time that you boot it

For stock engine management systems, I would recommend the Bosch LSM-11 sensor, part number 0 258 104 002, which should cost about UKP 100 plus VAT. This is also the same sensor that MoTeC incorrectly sell as a wide band sensor for UKP 150 plus VAT. It IS NOT a wide band sensor. A proper wide band sensor can only be used with an engine management system capable of using it, such as MoTeC M800, Pectel T6 etc... notably the MoTeC M48Pro cannot directly support wide band sensors, even though MoTeC claim it can. Wide band sensors are made by Bosch and NTK, they have at least five wires, and CANNOT be used with the stock ECU.

The Bosch LSM-11 is a high quality "wide range" narrow band lambda sensor which is lead tolerant (so octane boosters based on MMT like Millers shouldn't give it any grief either, whereas these boosters can damage less robust sensors). Wiring is as described above.

Hope this helps,

Pat.

 

pat-no wonder the sensor didn't work

thanks very much this information.

forever in your debt

jonny

 

Pat - what, exactly, was wrong with my description, lol?!

Is the LSM-11 more 'temperature tolerant' too?

Thanks
Richard

 

There appears to be some confusion here. According to Pat the lambda sensors on our cars is zirconia type and not titanium. However, in my research thre three wire (red, white, black) sensor is a titanium type (if Subaru are using the universal colour coding). Zirconium types use the colours white, grey and and black. According to Fuel Parts I (we) have titanium sensors. I'm really scratching my head now. For anyone interested I have uploaded a pdf file containing quite extensive information on lambdas at http://www.subarunutter.com/WFP36.pdf

The universal sensors are zirconium jobs and are much cheaper than the titanium. So the the ultimate question is "what really in on our cars??"

Pat, what year car you got?

 

Looks like mines a zirconium type on a MY97 grey import - its got white, grey & black leads.
Still the original one but is showing signs of failing (idle zipping up & down when thrashed; also the idle CO just scrapes in for the MOT)

Nick

 

Thanks - where's the info on how to do smileys?

 

when driving the car with the old sensor on,it would give a normal reading (pulse) then when you put your foot down (open throttle) the sensor would read normally (constant)then cut out completly,it would give no signal at all until you eased off for a few minutes then it would come back on,sometimes it wouldn't come back on at all???

is this correct???

the guy that tested it for me said it is a self generating unit and should give a signal at all times.

the car is a my95 wrx with standard ecu and full de-cat.






[Edited by jonny gav - 7/11/2002 12:04:39 AM]