That's it..someone has to decide at which point the O2 sensor is to be considered bad..and balance that between Federal requirements and not having it be too "sensitive" and perhaps forcing them to replace a lot of those sensors under warranty or cause their cars to develop a reputation for being unreliable because the check engine light comes on.
Benefits of Changing O2 Sensors>?
- Thread starter ULVER
- Start date
- Status
O2 sensor switch counts are a function of engine rpm. Faster rpm produces more lean/rich changes and more O2 sensor switches. I don't think switch counts mean much in terms of performance. The rise/fall time when a rich/lean mixture is encountered is supposed to be 300ms or less. I've never seen one this slow in lookig at the 6 cars I try to keep maintained. I monitor O2 sensor w/ an OBD2 scanner/monitor package (software + laptop).
I ran one O2 sensor 80K miles in a Plymouth Voyager, removed due to "reminder light" and stored. I later wired the same sensor into a 90 Accord and ran another 100K miles. The Accord ran fine the entire time w/ commuting mileage a constant 25 mpg.
They do go bad, but how to determine a weak, subpar O2 sensor is not clear. I suspect a ton of money is spent replacing perfectly fine O2 sensors. It's a nice piece of business for gaarages as they mark up the expensive part by 100%.
I ran one O2 sensor 80K miles in a Plymouth Voyager, removed due to "reminder light" and stored. I later wired the same sensor into a 90 Accord and ran another 100K miles. The Accord ran fine the entire time w/ commuting mileage a constant 25 mpg.
They do go bad, but how to determine a weak, subpar O2 sensor is not clear. I suspect a ton of money is spent replacing perfectly fine O2 sensors. It's a nice piece of business for gaarages as they mark up the expensive part by 100%.
I always thought it'd be an interesting experiment to take a sensor out of say a 1996 Ford Contour that declared it "bad" and stick it in some other make/model of vehicle to see what happens...
Last edited:
^ Had an unheated o2 in a 95 saturn that worked "okay" and the car went into closed loop off idle... but went into open loop at idle occasionally. (ok by GM) No codes or issues but I pulled it just the same and held on to it since I had a nice low mileage one from the junkyard.
Brought home a 98 saturn with the same part/connector but more advanced on-board diagnostics. Declared its own o2 was lazy then declared the replacement (from the 95) also lazy. New bosch universal cleared this up and made it run better to boot!
So to extrapolate, your contour sensor will probably work ok in an older car... but why bother?
Incidentally, the thread pitch and electrical characteristics of that unheated o2 match a zillion cars from 1980-2002 or so. So the computers got better... the sensor stayed the same.
Brought home a 98 saturn with the same part/connector but more advanced on-board diagnostics. Declared its own o2 was lazy then declared the replacement (from the 95) also lazy. New bosch universal cleared this up and made it run better to boot!
So to extrapolate, your contour sensor will probably work ok in an older car... but why bother?
Incidentally, the thread pitch and electrical characteristics of that unheated o2 match a zillion cars from 1980-2002 or so. So the computers got better... the sensor stayed the same.
Supposedly there is a difference between Bosch and NTK sensors in terms of something called the "Spinnel layer". See this: http://www.asashop.org/autoinc/may98/techtips.htm
In the case of the Contour, it had NTK sensors, I replaced them with Bosch sensors (it in fact was only complaining about one but I figured if one had gone bad the other would soon follow). The Bosch sensors are working fine. So I'm not sure if these differences really make much difference in the real world.
In the case of the Contour, it had NTK sensors, I replaced them with Bosch sensors (it in fact was only complaining about one but I figured if one had gone bad the other would soon follow). The Bosch sensors are working fine. So I'm not sure if these differences really make much difference in the real world.
I'm impressed Brian. Remind me to be careful tangling with you in the future
As TexasHonda said lots of these things get changed without clear reason. I look for at least 8 cross counts per 10 seconds at 2500 rpm, a full swing both rich and lean, a midpoint around 450 mv and a max of 150 ms response time when driving it both ways using a vacuum leak or propane. I also check to see if it's opposite the injection waveform. That way I know the sensor is really in command of the system and something else hasn't taken over. I always scope the raw signal when doing this. Using a propane torch to test a sensor off the car isn't foolproof but fwiw a good majority I've seen pass that test also pass in the car.
Unheated sensors will often fall out of closed loop at idle when the exhaust stream temp drops off. Plus they take longer to get into closed loop. I've retrofitted a few older cars to heated sensors. I connect the heater to the fuel pump signal. Not as nice as having the PCM control the heater as many modern ones do but it works fine.
The Bosch universals usually work ok. After all (iirc) Bosch invented the lambda sensor. There are times you've got to be careful about the number and configuration of slits though. I suspect cars of the future will all be going to A/F lean burn sensors with built in Nox measuring. I've heard many firms are either working on or have already developed such a sensor. If you think current sensors are expensive you don't want to hear what those are going to cost
As TexasHonda said lots of these things get changed without clear reason. I look for at least 8 cross counts per 10 seconds at 2500 rpm, a full swing both rich and lean, a midpoint around 450 mv and a max of 150 ms response time when driving it both ways using a vacuum leak or propane. I also check to see if it's opposite the injection waveform. That way I know the sensor is really in command of the system and something else hasn't taken over. I always scope the raw signal when doing this. Using a propane torch to test a sensor off the car isn't foolproof but fwiw a good majority I've seen pass that test also pass in the car.
Unheated sensors will often fall out of closed loop at idle when the exhaust stream temp drops off. Plus they take longer to get into closed loop. I've retrofitted a few older cars to heated sensors. I connect the heater to the fuel pump signal. Not as nice as having the PCM control the heater as many modern ones do but it works fine.
The Bosch universals usually work ok. After all (iirc) Bosch invented the lambda sensor. There are times you've got to be careful about the number and configuration of slits though. I suspect cars of the future will all be going to A/F lean burn sensors with built in Nox measuring. I've heard many firms are either working on or have already developed such a sensor. If you think current sensors are expensive you don't want to hear what those are going to cost
Good thread! Lots of info. Maybe too much, ulver?
May I say, Quickbeam, if I lived in TN you'd be first in line to look at my car if I had emission trouble. Lots of intelligence in your posts.
May I say, Quickbeam, if I lived in TN you'd be first in line to look at my car if I had emission trouble. Lots of intelligence in your posts.
I say change it when you have a MIL or CEL that indicates 02 sensor failure...or if you track gas mileage and it takes a big #@$%! for a while (
They're too expensive to change in set intervals like spark plugs.
They're too expensive to change in set intervals like spark plugs.
But isn't it possible for the sensor to behave perfectly normally, but just be wrong? IE, it says the mixture is at stoich when it's actually a touch rich?
Had a MAF do that a month or so ago. Every sensor reading was in the middle of the range for several different RPMS just the MAF was WRONG.
#@$%! butterflies!!!
Cleaned and all is good.
#@$%! butterflies!!!
Cleaned and all is good.
- Status
