Originally Posted by aa3jy
What is the average life of an Oxy sensor before one would expect problems?
On a car, you can expect at least 100,000 miles before it degrades to a point where the engine management system cannot "read" it. Of course, they tell you to replace it much sooner than that, but that is just a blanket statement.
The most common type of O2 sensor (the Zirconium Oxide type) can easily be tested using a standard digital voltmeter.
First, you need to know whether the sensor is a simple type with a single wire, or a "3-wire" type with an integrated heated. If the latter, you need you determine which wire is the one leading to the actual sensor (not the heater). Tap off the signal wire so it is externally available (e.g. stick a pin through the insulation and into the copper wire), then plug the sensor back into the system.
Start the vehicle, and drive it until it is thoroughly warmed up. Connect the (+) lead of the digital Voltmeter (DVM) to the sensor's signal wire "tap" as above, and the (-) lead to a solid engine ground.
Now, watch the meter. What you SHOULD see at idle is the digits jumping around with a low number occasionally reading near zero (less than 0.2V), and a high number above 0.5V (preferably occasionally reaching 0.8 to 0.9V).
This means that it is working acceptably well.
This type of sensor is in effect a miniature fuel cell. It is exposed (on one side) to the atmosphere, and (on the other side) to the exhaust. When there is O2 present on one side and not the other (i.e. when the exhaust contains no residual O2), a new sensor generates about 0.9 Volts. When there is even a small amount of O2 in the exhaust (and therefore the senor "sees" O2 on BOTH side), it generates NO voltage at all.
In operation, the system is constantly dithering between slightly lean (a little residual O2 in the exhaust) and slightly rich (no O2). As soon as the sensor detects a slightly rich condition, it commands the system to make it slightly more lean, which then results in the system detecting a alightly lean condition and commanding the system to make it a little richer.... and so on. The system constantly dithers between very slightly lean and very slightly rich.
Note that the actual VALUE of the voltage generates is irrelivant. All that matters is that the system can detect a voltage has been generated (meaning a rich condition is detected) OR no voltage has been generated (meaning a lean condition has been detected).
As the sensor wears out, it generates less and less output voltage when the exhaust contains no residual O2 the "high" number will get less and less. Eventually, as the thing wears out, the voltage generated as a result of a rich condition is not high enough to be able to be interpreted properly, and at that point the sensor is useless.
By actually testing the sensor like this, you only need to replace it when it ACTUALLY NEEDS REPLACEMENT, which can be as much as 2 to 3 times less often than what the manufacurer states.