Testing the O2 sensor off the car.

jetjock

creepy-ass cracka
Jul 11, 2005
9,439
0
0
Redacted per Title 18 USC Section 798
I keep getting asked about this so I'll post it up. O2 sensors are one of the most needlessly changed parts on cars. While there are several methods of testing it on the car I keep getting ask about testing it off.

1) Remove the sensor from the engine.

2) Connect a meter and set to the range nearest 1 volt DC. Connect one probe to the sensor wire and the other probe to the sensor's housing. Clamping the housing lightly in a vise and connecting to the vise makes the process easier.

3) Heat the sensor tip with a propane torch. Get it good and hot (near red) but don't melt anything.

4) Watch the meter. When hot the sensor should put out at least 800 mv. When the flame is removed from the tip the voltage should quickly drop to less than 200 mv. Play the flame back and forth across the sensor and look for rapid switching. While modern OBDII systems require switching faster than 150 ms it's not that important on this engine. Just look for a rapid change with temperature.

5) Lastly, keep the sensor tip hot for at least two minutes. Again, be careful not to melt anything. The voltage should remain steady at 800 mv minimum. You're doing this because sensors can fail in such a way as to not act up until they've been hot for a few minutes. It's caused by the expansion of cracks in the ceramic parts inside.

Any sensor that passes these tests is good. Let the sensor cool then ohm the other two wires to be sure the heater is OK or hook them across the battery for one minute and carefully feel for the sensor body to become hot. You can do this before the torch thing if you like.

6) Reinstall into the engine using a new gasket. Any exhaust leaks at or upstream of the sensor will cause it to be shifted lean.

If you want to know more about how zirconium lambda sensors work there's enough out there to choke a goat. Read some and don't stop until it makes sense.

7) Stop hassling me for these instructions ;)
 

87mgte

87 Turbo Targa
Sep 9, 2007
536
0
0
34
Sequim
by "needlessly changed" do you mean they rarely need to be changed or do you mean they're usually one of the overlooked parts that should actually be changed regularly?
 

jetjock

creepy-ass cracka
Jul 11, 2005
9,439
0
0
Redacted per Title 18 USC Section 798
Needlessly changed means there's usually nothing wrong with the ones that get replaced. It happens for two reasons:

1) People blame them for all sorts of problems when they can't possibly be at fault. Time and again I see this. It stems from not understanding how EFI works in general and this car's EFI system in particular.

2) They get changed because people don't know how to test them either on or off the car. Those who do replace them usually are the same technically inept folks who tend to throw other parts at the car when it has a problem.

I'll spare you the technical details but the bottom line is a bad lambda sensor on this car will *never* be the cause of a driveability problem. Ever. Emissions and fuel economy issues are another story. If you have those it may be the sensor. If the car is "acting up" in any other way you're barking up the wrong tree.

That said they do have a finite life and one usually shortened by abuse. For example the use of non-sensor safe silicone sealants on the engine will kill them. So will the silicates in conventional coolants if a BHG occurs. The point is they shouldn't be changed without being intelligently diagnosed first. Course, that applies to just about any part on the car.
 

flubyux2

Madd Tyte JDM yo ®
Apr 2, 2005
1,019
0
0
43
st. pete, fl
www.myspace.com
im less than hopefull, but: will these instructions apply to wide band sensors too? or are the variances too far different?


From JJ: No, sorry. A WB is a different animal and can only be checked in it's entirety by using calibration gas with varying concentrations of O2.