Testing your O2 sensor
Andrew Evans
There's been a lot of talk on this newsgroup lately about oxygen
sensors that might be bad, but I haven't seen anyone suggest how you
might test it, other than to replace the sensor and see if that seems
to make any difference.
An O2 sensor generates a small voltage which is proportional to the
amount of oxygen in the exhaust. Typically this voltage ranges from
0-1 volt. The leaner the fuel mixture, the more oxygen in the
exhaust, the higher the voltage from the sensor. On Audis and VWs
(don't know about other cars, but they should be similar), the voltage
for a properly working O2 sensor with the fuel mixture set correctly
is about 0.5 volts.
So, you can easily test your O2 sensor's operation by getting a
voltmeter that is capable of displaying the 0-1 volt range with decent
resolution. Connect the negative probe of the meter to some suitable
ground, and connect the positive lead to the O2 sensor output; leave
the sensor connected to the computer. You should get a reading of
about 0.5V, which should fluctuate slightly as the computer varies the
mixture constantly and monitors the results.
If your sensor is bad, most likely it will be putting out little or no
voltage. In most cases, this will cause the computer to adjust the
mixture to an excessively lean setting.
Hope this helps those of you who want to actually check your sensors
before you blow $40 or whatever on a new one.
Paul L. Fortman, Jr.
In a previous article, and...@airs.com (Andrew Evans) says:
>
>So, you can easily test your O2 sensor's operation by getting a
>voltmeter that is capable of displaying the 0-1 volt range with decent
>Hope this helps those of you who want to actually check your sensors
>before you blow $40 or whatever on a new one.
How much does a good voltmeter cost? How often would I use it--what
else can I test with it to justify the cost. Is there a combination
voltmeter and tachometer? This combination could be useful. I think
I'll visit some auto parts stores. What is a good brand, etc.
Thanks.
Paul
--
fort...@dayton.bitnet
fort...@udavxb.oca.udayton.edu
Rick Kirchhof
>
>There's been a lot of talk on this newsgroup lately about oxygen
>sensors that might be bad, but I haven't seen anyone suggest how you
>might test it, other than to replace the sensor and see if that seems
>to make any difference.
>
>An O2 sensor generates a small voltage which is proportional to the
>amount of oxygen in the exhaust. Typically this voltage ranges from
>0-1 volt. The leaner the fuel mixture, the more oxygen in the
>exhaust, the higher the voltage from the sensor. On Audis and VWs
>(don't know about other cars, but they should be similar), the voltage
>for a properly working O2 sensor with the fuel mixture set correctly
>is about 0.5 volts.
>
In addition, the sensor voltage should vary quickly and with gusto.
Voltage around 0.45 is the mid point (neither rich nor lean) and
indicates a cold (under ~600F) or defective sensor. It should vary from
under 0.2 to over 0.8 volts as the engine runs. A better test is to
watch it on an oscilloscope, but running the engine above 2000 RPM for 2
minutes then watching a voltmeter attached to the sensor output wire will
do. Note that you want to sample the signal wire on a assembled system.
Do not disconnect to test.
>So, you can easily test your O2 sensor's operation by getting a
>voltmeter that is capable of displaying the 0-1 volt range with decent
>resolution. Connect the negative probe of the meter to some suitable
>ground, and connect the positive lead to the O2 sensor output; leave
>the sensor connected to the computer. You should get a reading of
>about 0.5V, which should fluctuate slightly as the computer varies the
>mixture constantly and monitors the results.
>
>If your sensor is bad, most likely it will be putting out little or no
>voltage. In most cases, this will cause the computer to adjust the
>mixture to an excessively lean setting.
>
>Hope this helps those of you who want to actually check your sensors
>before you blow $40 or whatever on a new one.
You can also create a known rich and known lean condition and watch for
expected changes. Remember to preheat the sensor with some higher RPM's
first. If there is suffient interest, I will follow with a procedure to
test sensors on the work bench.
--
Rick Kirchhof Austin, Texas |
Domain: ri...@posms.cactus.org | Someday...
Bang path: ...!cs.utexas.edu!dogface!posms!rick |
===========================================================================
Frank Mallory
PL> >voltmeter that is capable of displaying the 0-1 volt range with decent
PL> >...
PL> >Hope this helps those of you who want to actually check your sensors
PL> >before you blow $40 or whatever on a new one.
PL>
PL> How much does a good voltmeter cost? How often would I use it--what
PL> else can I test with it to justify the cost. Is there a combination
PL> voltmeter and tachometer? This combination could be useful. I think
PL> I'll visit some auto parts stores. What is a good brand, etc.
Some person here claiming to have special knowledge stated that you can burn
out your O2 sensor if you try to test it with an ordinary voltmeter; that you
should use a VTVM or equivalent (presumably any meeting the old standard of
20,000 Ohms per volt).
* Origin: Silver Bullet - Silver Spring, Md. - 301-622-2247 (1:109/417)
ven_polychronakos
inaccurate reading. A Digital VoltMeter (DVM) with 10 MOhm input impedence
is best. They are fairly cheap these days. J.C. Whitney's sells a
combination DVM, dwell-tach for about $50.
A faulty oxygen sensor, by the way, should set an error code. For GM
cars, for example, error code 13 will set and the "Service Engine Soon"
light will come on if the engine is at operating temp (>150F), at
least two minutes have passed since engine turn-on and the Oxygen sensor
shows a steady reading of about 0.35 to 0.45 Volts. The the Engine Control
Module (ECM) will enter the "open loop" mode which simply ignores the
Oxygen sensor readings and calulates fuel delivery based only on
the engine temperature, air intake manifold pressure, and throttle posotion.
In my experience the car continues to run well although I'm not sure
about fuel economy or emmissions compliance.
A guy offered to describe a bench test for oxygen sensors. I think he
should go ahead, it seems it's a fairly common failure for cars with
more than 60k or so.
Happy New Year to all, Ven
Paul L. Fortman, Jr.
In a previous article, Ven Polychronakos () says:
>In article <69454...@blkcat.FidoNet> Frank....@f417.n109.z1.FidoNet.Org
>(Frank Mallory) writes:
> A faulty oxygen sensor, by the way, should set an error code. For GM
>cars, for example, error code 13 will set and the "Service Engine Soon"
>light will come on if the engine is at operating temp (>150F), at
There is a light on my '81 Corolla with the word "sensor" on it and
that light is on and stays on (drives me crazy driving at night).
I looked in the owners manual and it says that the light is supposed
to come on at 60K miles, which it did, as a reminder to take the car
in for a 60K check-up. I haven't called a dealer yet to ask about
the price of this check-up. Since its an '81 with only 61K on it I
do want to keep it in good running shape, but I really don't want to
take it to a dealer. Is the dealer the only place that has a computer
that will plug into my car's computer? Should I go ahead and replace
the sensor whether it needs it or not (it does run VERY lean).
>least two minutes have passed since engine turn-on and the Oxygen sensor
>shows a steady reading of about 0.35 to 0.45 Volts. The the Engine Control
>Module (ECM) will enter the "open loop" mode which simply ignores the
>Oxygen sensor readings and calulates fuel delivery based only on
>the engine temperature, air intake manifold pressure, and throttle posotion.
>In my experience the car continues to run well although I'm not sure
>about fuel economy or emmissions compliance.
I don't think my '81 Corolla is that sophisticated. It actually has
a carburator on it. The above description sounds like a fuel-injected
engine. Although, I am not a mechanic or very knowledgeable about cars.
>A guy offered to describe a bench test for oxygen sensors. I think he
>should go ahead, it seems it's a fairly common failure for cars with
>more than 60k or so.
Maybe I should just replace the sensor...
-Paul
--
fort...@dayton.bitnet
fort...@udavxb.oca.udayton.edu
Jonathan R. Lusky
>out your O2 sensor if you try to test it with an ordinary voltmeter; that you
>should use a VTVM or equivalent (presumably any meeting the old standard of
>20,000 Ohms per volt).
>
big fold up display) connected to a cheapy (cheif auto parts) GM 1-wire oxygen
sensor for about 2 hours of city/highway driving and about twelve hours in the
shop idling and on the chassis dyno at loads of 25 - 75 HP. 4,000 city miles l
ater the O2 sensor readings still agreed with the readings from a 4-gas analyzer
. The fact the engine was running excessively lean above loads of 75 hp
and was running on natural gas (much higher EGT than gasoline) shouldn't have
improved the life of the O2 sensor any.
Jon Lusky
lu...@gnu.ai.mit.edu
>
mel...@skyler.mavd.honeywell.com
> A faulty oxygen sensor, by the way, should set an error code. For GM
> cars, for example, error code 13 will set and the "Service Engine Soon"
> light will come on if the engine is at operating temp (>150F), at
> least two minutes have passed since engine turn-on and the Oxygen sensor
> shows a steady reading of about 0.35 to 0.45 Volts. The the Engine Control
because, as you stated it looks for a steady reading. They really should look
at the cross counts not the O2 output. For those of you who don't know what
cross counts are, it is a parameter used by the ECM and denotes the number of
times the sensor output has crossed the .45 volt value. The last two sensors I
had fail, never set a trouble code. They both had steady readings of .33 Volts.
The cross counts never changed. I go through about 3 sensors per summer when
racing my GN using leaded racing fuels. If using a scan tool, a good way to
tell if a sensor is bad is they are sluggish, and the car takes forever to go
closed loop.
>
>
Craig Leres
> A faulty oxygen sensor, by the way, should set an error code. For GM
> cars, for example, error code 13 will set and the "Service Engine Soon"
> light will come on if the engine is at operating temp (>150F), at
> least two minutes have passed since engine turn-on and the Oxygen sensor
> shows a steady reading of about 0.35 to 0.45 Volts.
Not all GM error codes cause the "Service Engine Soon" light to come
on. My Beretta was doing neat idle tricks a few months ago. The last
problem they found was a bad O2 sensor. They said they found it by
looking at the stored error codes. (This made me wonder why they didn't
look at the stored error codes during any of my previous four visits.)
I don't have the manual in front of me but it was one that has to do
with the engine running lean. I checked the manual at the time and sure
enough, it doesn't set the "Service Engine Soon" light. It also gets
purged after so many starts.
If you're gonna work on these cars, you gotta git a scan tool.
Craig
Ven Polychronakos
mel...@skyler.mavd.honeywell.com writes:
...... They really should look
>at the cross counts not the O2 output. For those of you who don't know what
>cross counts are, it is a parameter used by the ECM and denotes the number of
>times the sensor output has crossed the .45 volt value. The last two sensors I
>had fail, never set a trouble code. They both had steady readings of .33
Volts.
>The cross counts never changed. I go through about 3 sensors per summer when
>racing my GN using leaded racing fuels. If using a scan tool, a good way to
>tell if a sensor is bad is they are sluggish, and the car takes forever to go
>closed loop.
>>
>>
>
mention of it. In the O2 sensor section they say that "scan tools"
can provide, in addition to the sensor's instantaneous value, an
integrated one and "block learn" expressed in counts. Is the latter
the same one as the "cross counts" you are refering to? Thanks, Ven.
mel...@skyler.mavd.honeywell.com
> I was not aware about the cross counts. My GM shop manual makes no
> mention of it. In the O2 sensor section they say that "scan tools"
> can provide, in addition to the sensor's instantaneous value, an
> integrated one and "block learn" expressed in counts. Is the latter
> the same one as the "cross counts" you are refering to? Thanks, Ven.
and the integrator. These are both correction factors to the Air to fuel
mixture. The integrator reflects short term corrects and the BLM is for long
term corrections. Cross counts are different, but influence the integrator and
BLM. The higher the int and BLM, the more fuel the computer is adding to the
system. This is usually in response to what it perceives as a lean condition.
The computer can remember the BLM after you shut off the car, this is why it is
important to clear the computers memory after replacing things such as the O2
sensor.
On a scan tool, a vacuum leak may cause the BLM to be maxed out (255) while the
integrator may be around normal (128). An excess fuel condition (stuck fuel
injector) may cause a low BLM and normal or high integrator.
With a scan tool it is possible to monitor both the O2 sensor output and the
integrator. As the O2 output goes up, the computer will correct by leaning out
the mixture (lower Integrator). Since the computer can never achieve the ideal
.45 volts for very long, this max, min O2 sensor voltage is constantly going
on. Each time the output crosses the .45 volt point, a cross count is tallied
in the cross count parameter.
For an indepth discussion of this and the other sensors on most GM cars a good
source is a book titled something like "How to Repair and Modify Chevrolet Fuel
Injection Systems". The book covers both PFI and TBI systems. It is pretty well
written using language that everyone can understand.
>
>
Codesmiths
> How much does a good voltmeter cost? How often would I use it--what
Go to Tandy / radio Shack & spend a very few dollars on a multi-meter
(aka VOM (Volt / Ohm Meter)). This measures voltages over a wide
range, resistance & small currents. The current ranges are pretty
useless for auto electrical work, they're too small. The resistance
ranges are mainly useful for simple short circuit detection & fuse
checking (and very useful they are too). Voltage ranges will do
anything you need on a car, even complex things like alternator
testing.
These things are really cheap (I'm in the UK, so wouldn't like to
quote a local price). Even a very cheap one is quite good quality,
paying more gets you a stronger case & higher current capacity. Buy
it from a hobby electronics shop, rather than a garage equipment
shop, and it will be half the price.
To measure the currents that are commonly found in cars, you'll need
something like an external current shunt. Better meters can use
these, either bought with the meter, or bought separately. It's not
feasible to use shunts designed for one meter with a different meter.
Digital meters cost far more & are no better for most work - stick
with moving needles. A continuity buzzer is a useful gadget, very
handy if you ever need to trace your way through a wiring loom. These
are either built into a better quality meter, or can be bought (or
made !) separately.
The Gunson (UK maker, equivalents are probably available in the US)
range of test gear includes a combined tacho / dwell / volt meter. I
don't like these things; they're expensive, the tacho is inaccurate,
dwell is pretty useless these days & the voltmeter part isn't a patch
on the cheap meters sold for electronics work. As a set of garage
tools, I'd recommend a cheap VOM (because it's cheap), then a xenon
timing light with a built in tacho. It costs only slightly more than
a combined garage multi-test-meter, but the meter part is better &
you have a timing light too.
Whatever sort of meter you get, make sure you have 2 sets of test
leads to go with it; pointed probes & crocodile clips.
Andy Dingley din...@cix.compulink.co.uk +44 91 230 1695
Jurek Rakoczynski
>
Text deleted
>
> So, you can easily test your O2 sensor's operation by getting a
> voltmeter that is capable of displaying the 0-1 volt range with decent
> resolution. Connect the negative probe of the meter to some suitable
> ground, and connect the positive lead to the O2 sensor output; leave
> the sensor connected to the computer. You should get a reading of
> about 0.5V, which should fluctuate slightly as the computer varies the
> mixture constantly and monitors the results.
DON'T DO THIS!!!!!!!!!
First off, check a service manual for the proper way and equipment to
test the sensor. Some sensors require the lead disconnected and a
voltmeter with at least 1 megaohm resistance. If you don't have the
correct resistance, you could damage the sensor, or get erroneous
readings.
While were at it, follow the directions for installing a new sensor.
Some sensors (that have a single terminal?) come with electrically
conductive anti-seize compound on the threads. If you use the wrong
anti-sieze, you may have a grounding problem.
>
> If your sensor is bad, most likely it will be putting out little or no
> voltage. In most cases, this will cause the computer to adjust the
> mixture to an excessively lean setting.
>
> Hope this helps those of you who want to actually check your sensors
> before you blow $40 or whatever on a new one.
--
Jurek Rakoczynski, AG Communication Systems, POB 52179, Phoenix, AZ. 85072
Voice: +1 602 581 4867 Fax: +1 602 582 7111
UUCP: {ncar!noao!asuvax | uunet!zardoz!hrc}!gtephx!rakoczynskij
Inet: gtephx!rakocz...@asuvax.eas.asu.edu
Gerald E. Riechert
>
> There is a light on my '81 Corolla with the word "sensor" on it and
> that light is on and stays on (drives me crazy driving at night).
> I looked in the owners manual and it says that the light is supposed
> to come on at 60K miles, which it did, as a reminder to take the car
> in for a 60K check-up. I haven't called a dealer yet to ask about
> the price of this check-up. Since its an '81 with only 61K on it I
> do want to keep it in good running shape, but I really don't want to
> take it to a dealer. Is the dealer the only place that has a computer
> that will plug into my car's computer? Should I go ahead and replace
> the sensor whether it needs it or not (it does run VERY lean).
>
> a carburator on it. The above description sounds like a fuel-injected
> engine. Although, I am not a mechanic or very knowledgeable about cars.
>
>
> -Paul
> --
I have an 81 Celica with the same Sensor light (however, it comes on
every 30K miles). The service manual contains a very simple procedure
for checking the sensor. If I remember right, just connect a voltmeter
to the specified underhood connector, run the engine at some RPM value
for 1 or two minutes and verify the voltmeter needle is varying at
some rate. Then you flip a switch hidden behind the driver's kick
panel to turn the light off for another 30K miles. So far, I've
checked the sensor twice and it has been OK twice.
-----
Jerry Riechert
AG Communication Systems, Phoenix, AZ
UUCP: {ncar!noao!asuvax | uunet!zardoz!hrc}!gtephx!riechertg
internet: gtephx!riec...@asuvax.eas.asu.edu
voice: (602) 582-7332
Dan Eldred
>
>In a previous article, and...@airs.com (Andrew Evans) says:
>>
>>So, you can easily test your O2 sensor's operation by getting a
>>voltmeter that is capable of displaying the 0-1 volt range with decent
>>...
>
>else can I test with it to justify the cost. Is there a combination
>voltmeter and tachometer? This combination could be useful. I think
>I'll visit some auto parts stores. What is a good brand, etc.
>
Make sure it's a high impedence voltmeter (i.e. 10 meg-ohm input impedence)
since the impedence of the O2 sensor is very high. The $10 analog meters
floating around won't do. Most digital meters will do; I've seen them
selling for around $30 at a variety of places.
--
-----------------------------------------------------------------------
| THE WAR ON DRUGS IS AN ASSAULT ON OUR CIVIL LIBERTIES. STOP IT! |
-----------------------------------------------------------------------
| Dan Eldred | Jet Propulsion Laboratory, 198-326 |
| eld...@rrunner.jpl.nasa.gov | 4800 Oak Grove Dr.,Pasadena, CA 91109 |
| Phone: (818) 354-6519 | Fax: (818) 393-4440 |
-----------------------------------------------------------------------
John De Armond
>> So, you can easily test your O2 sensor's operation by getting a
>> voltmeter that is capable of displaying the 0-1 volt range with decent
>DON'T DO THIS!!!!!!!!!
>First off, check a service manual for the proper way and equipment to
>test the sensor. Some sensors require the lead disconnected and a
>voltmeter with at least 1 megaohm resistance. If you don't have the
>correct resistance, you could damage the sensor, or get erroneous
>readings.
Where does this stuff come from? NO! I repeat NO! An oxygen sensor
is NOT damaged by the load presented to it. If you had an inkling of
how the sensor works you'd know better. A very low impedance voltmeter
may cause an erroneous reading but it will NOT damage the sensor.
Indeed one mode of operating a sensor is the "oxygen pumping" mode
whereby about .5 volts is applied to the sensor. This causes the sensor
to pump oxygen across the membrane. The current drawn represents the
work product which is proportional to the ratio of oxygen on either side
of the membrane. This of course, works only for zirconia sensors.
Go to your SAE library and read some of the papers on the subject to learn
more.
>While were at it, follow the directions for installing a new sensor.
>Some sensors (that have a single terminal?) come with electrically
>conductive anti-seize compound on the threads. If you use the wrong
>anti-sieze, you may have a grounding problem.
All sensors, whether 1 or 3 terminal, use the ground as the signal return.
The other 2 wires are the heater that is isolated from the sensor.
As long as one tightens the sensor securely, the conductivity of the
anti-seize will not be an issue.
My advice is if you suspect a bad sensor, do a quick check (voltmeter or
scan code or whatever) and if there is any doubt, install a new sensor.
If that does not fix the problem, put the old one in and save the new
sensor - you'll need it sooner or later.
John
--
John De Armond, WD4OQC | "Purveyors of speed to the Trade" (tm)
Rapid Deployment System, Inc. | Home of the Nidgets (tm)
Marietta, Ga | "It's not a bald spot, its a solar panel for
j...@dixie.com | a sex machine."
Rick Kirchhof
sensors, perhaps the following information will help.
Comment:
These procedures are only for self powered conventional sensors.
Some very new cars are using a different style sensor that is
powered. *Many* Oxygen sensors are replaced that are good to
excellent. *Many* people don't know how to test them. They
routinely last 50,000 or more miles, and if the engine is in good
shape, can last the life of the car.
What does the O2 sensor do?
It is the primary measurement device for the fuel control computer
in your car to know if the engine is too rich or too lean. The
O2 sensor is active anytime it is hot enough, but the computer
only uses this information in the closed loop mode. Closed loop
is the operating mode where all engine control sensors including
the Oxygen sensor are used to get best fuel economy, lowest
emissions, and good power.
Should the O2 sensor be replaced when the sensor light comes on in
your car?
Probably not, but you should test it to make sure it is alive and
well. This assumes that the light you see is simply an emissions
service reminder light and not a failure light. A reminder light
is triggered by a mileage event (20-40,000 miles usually) or
something like 2000 key start cycles. EGR dash lights usually fall
into the reminder category. Consult your owners manual, auto repair
manual, dealer, or repair shop for help on what your light means.
How do I know if my O2 sensor may be bad?
If your car has lost several miles per gallon of fuel economy and
the usual tune up steps do not improve it. This *is not* a
pointer to O2 failure, it just brings up the possibility. Vacuum
leaks and ignition problems are common fuel economy destroyers.
As mentioned by others, the on board computer may also set one of
several failure "codes". If the computer has issued a code
pertaining to the O2 sensor, the sensor and it's wiring should
be tested. Usually when the sensor is bad, the engine will show
some loss of power, and will not seem to respond quickly.
What will damage my O2 sensor?
Home or professional auto repairs that have used silicone gasket
sealer that is not specifically labeled "Oxygen sensor safe",
"Sensor safe", or something similar, if used in an area that
is connected to the crankcase. This includes valve covers, oil
pan, or nearly any other gasket or seal that controls engine oil.
Leaded fuel will ruin the O2 sensor in a short time. If a car is
running rich over a long period, the sensor may become plugged up
or even destroyed. Just shorting out the sensor output wire will
not usually hurt the sensor. This simply grounds the output
voltage to zero. Once the wiring is repaired, the circuit
operates normally. Undercoating, antifreeze or oil on the
*outside* surface of the sensor can kill it. See how does an
Oxygen sensor work.
Will testing the O2 sensor hurt it?
Almost always, the answer is no. You must be careful to not
*apply* voltage to the sensor, but measuring it's output voltage
is not harmful. As noted by other posters, a cheap voltmeter
will not be accurate, but will cause no damage. This is *not*
true if you try to measure the resistance of the sensor.
Resistance measurements send voltage into a circuit and check the
amount returning.
How does an O2 sensor work?
An Oxygen sensor is a chemical generator. It is constantly making
a comparison between the Oxygen inside the exhaust manifold and air
outside the engine. If this comparison shows little or no
Oxygen in the exhaust manifold, a voltage is generated. The
output of the sensor is usually between 0 and 1.1 volts. All
spark combustion engines need the proper air fuel ratio to
operate correctly. For gasoline this is 14.7 parts of air to one
part of fuel. When the engine has more fuel than needed, all
available Oxygen is consumed in the cylinder and gasses leaving
through the exhaust contain almost no Oxygen. This sends out a
voltage greater than 0.45 volts. If the engine is running lean,
all fuel is burned, and the extra Oxygen leaves the cylinder and
flows into the exhaust. In this case, the sensor voltage goes
lower than 0.45 volts. Usually the output range seen seen is
0.2 to 0.7 volts.
The sensor does not begin to generate it's full output until it
reaches about 600 degrees F. Prior to this time the sensor is
not conductive. It is as if the circuit between the sensor and
computer is not complete. The mid point is about 0.45 volts.
This is neither rich nor lean. A fully warm O2 sensor *will not
spend any time at 0.45 volts*. In many cars, the computer sends
out a bias voltage of 0.45 through the O2 sensor wire. If the
sensor is not warm, or if the circuit is not complete, the computer
picks up a steady 0.45 volts. Since the computer knows this is
an "illegal" value, it judges the sensor to not be ready. It
remains in open loop operation, and uses all sensors except the
O2 to determine fuel delivery. Any time an engine is operated
in open loop, it runs somewhat rich and makes more exhaust
emissions. This translates into lost power, poor fuel economy
and air pollution.
The O2 sensor is constantly in a state of transition between high
and low voltage. Manfucturers call this crossing of the 0.45
volt mark O2 cross counts. The higher the number of O2 cross
counts, the better the sensor and other parts of the computer
control system are working. It is important to remember that the
O2 sensor is comparing the amount of Oxygen inside and outside
the engine. If the outside of the sensor should become blocked,
or coated with oil, sound insulation, undercoating or antifreeze,
(among other things), this comparison is not possible.
How can I test my O2 sensor?
They can be tested both in the car and out. If you have a high
impedence volt meter, the procedure is fairly simple. It will
help you to have some background on the way the sensor does
it's job. Read how does an O2 sensor work first.
Testing O2 sensors that are installed
The engine must first be fully warm. If you have a defective
thermostat, this test may not be possible due to a minimum
temperature required for closed loop operation. Attach the
positive lead of a high impedence DC voltmeter to the Oxygen
sensor output wire. This wire should remain attached to the
computer. You will have to back probe the connection or use
a jumper wire to get access. The negative lead should be
attached to a good clean ground on the engine block or
accessory bracket. Cheap voltmeters will not give accurate
results because they load down the circuit and absorb the
voltage that they are attempting to measure. A acceptable
value is 1,000,000 ohms/volt or more on the DC voltage.
Most (if not all) digital voltmeters meet this need. Few
(if any) non-powered analog (needle style) voltmeters do.
Check the specs for your meter to find out. Set your meter
to look for 1 volt DC. Many late model cars use a heated
O2 sensor. These have either two or three wires instead of
one. Heated sensors will have 12 volts on one lead, ground
on the other, and the sensor signal on the third. If you have
two or three wires, use a 15 or higher volt scale on the meter
until you know which is the sensor output wire.
When you turn the key on, do not start the engine. You should
see a change in voltage on the meter in most late model cars. If
not, check your connections. Next, check your leads to make sure
you won't wrap up any wires in the belts, etc. then start the
engine. You should run the engine above 2000 rpm for two
minutes to warm the O2 sensor and try to get into closed loop.
You are looking for voltage to go above and below 0.45 volts.
If you see less than 0.2 and more than 0.7 volts and the value
changes rapidly, you are through, your sensor is good. If not,
is it steady high (> 0.45) near 0.45 or steady low (< 0.45).
If the voltage is near the middle, you may not be hot yet. Run
the engine above 2000 rpm again. If the reading is steady low,
add richness by partially closing the choke or adding some propane
through the air intake. Be very careful if you work with any
extra gasoline, you can easily be burned or have an explosion.
If the voltage now rises above 0.6 to 0.7, and you can change it
at will by changing the extra fuel, the O2 sensor is usually good.
If the voltage is steady high, create a vacuum leak. Try pulling
the PCV valve out of it's hose and letting air enter. If this
drives the voltage to 0.2 to 0.3 or less and you can control it
at will by opening and closing the vacuum leak, the sensor is
usually good.
If you are not able to make a change either way, stop the engine,
unhook the sensor wire from the computer harness, and reattach
your voltmeter to the sensor output wire. Repeat the rich and
lean steps. If you can't get the sensor voltage to change, and
you have a good sensor and ground connection, try heating it once
more. Repeat the rich and lean steps. If still no voltage or
fixed voltage, bad sensor.
If you are not getting a voltage and the car has been running
rich lately, the sensor may be carbon fouled. It is sometimes
possible to clean a sensor in the car. Do this by unplugging
the sensor harness, warming up the engine, and creating a lean
condition at about 2000 rpm for 1 or 2 minutes. Create a big
enough vacuum leak so that the engine begins to slow down.
The extra heat will clean it off if possible. If not, it
was dead anyway, no loss. In either case, fix the cause of the
rich mixture and retest. If you don't, the new sensor will
fail.
Testing O2 sensors on the workbench.
Use a high impedence DC voltmeter as above. Clamp the sensor in
a vice, or use a plier or vice-grip to hold it. Clamp your
negative voltmeter lead to the case, and the positive to the
output wire. Use a propane torch set to high and the inner blue
flame tip to heat the fluted or perforated area of the sensor.
You should see a DC voltage of at least 0.6 within 20 seconds.
If not, most likely cause is open circuit internally or lead
fouling. If OK so far, remove from flame. You should see a
drop to under 0.1 volt within 4 seconds. If not likely silicone
fouled. If still OK, heat for two full minutes and watch for
drops in voltage. Sometimes, the internal connections will open
up under heat. This is the same a loose wire and is a failure.
If the sensor is OK at this point, and will switch from high to
low quickly as you move the flame, the sensor is good. Bear in
mind that good or bad is relative, with port fuel injection
needing faster information than carbureted systems.
--
Rick Kirchhof Austin, Texas |
Domain: ri...@posms.cactus.org | Someday...
Bang path: ...!cs.utexas.edu!peyote!posms!rick |
===========================================================================
SDOG...@cmsa.gmr.com
should never be used to test an O2 sensor. This sensor functions as a battery,
producing a voltage proportional to Oxygen content in the exhaust. Obviously,
when you short a battery we all know what happens. We apply the same idea to t
he O2. Also, you never put an ohmmeter across a live circuit. I realize this
still leaves questions, yet this is GM's teachings.
Anti sieze is a MUST ON ALL O2 SENSORS. A point over looked is the servici
ng aspect. Antisieze allows future servicing without damaging the exhaust mani
fold. All AC O2 sensors come coated w/ antisieze. When reinstalling one, you
have to add the antisieze yourself. One more point. Let the engine warm-up
for a couple of minutes before removing the O2.
Unfortunately, I did not view the original note on this subject. I hope I'
m not repeating someone else.
Jurek Rakoczynski
In article <6pagd=j...@dixie.com>, j...@dixie.com (John De Armond) writes:
> rakocz...@gtephx.UUCP (Jurek Rakoczynski) writes:
Whole bunch of text deleted throught this posting.
> >correct resistance, you could damage the sensor, or get erroneous
> >readings.
>
> Where does this stuff come from?
> NO! I repeat NO! An oxygen sensor
> is NOT damaged by the load presented to it. If you had an inkling of
> how the sensor works you'd know better. A very low impedance voltmeter
> may cause an erroneous reading but it will NOT damage the sensor.
Well I did my homework (at least cliffnotes :-) ) and I'll add what I
learned to what j...@dixie.com (John De Armond), ri...@posms.cactus.org
(Rick Kirchhof), and mel...@skyler.mavd.honeywell.com have already
posted. I won't repeat what they posted except to add continuity.
I had a telephone conversation yesterday with a senior eng. at Allied
Signal-Autolite Division who works on Oxygen sensors. He sent me
'some' design drawings and charts. I'll paraphrase and summarise what
I received and discussed.
First of all he confirmed that he does not see any way that a
low-impedance voltmeter would cause damage to the sensor. I agree,
but see my comments at the end.
There are two basic types of O2 sensors:
1. Zirconia sensors generate a typical voltage of (figures are
interpreted from a chart):
Theoretical output
voltage @
Air/Fuel --------------------
ratio 800 deg C 350 deg C
13.5 .81 1.03
14.0 .78 1.01
14.5 .74 .94
14.65 .68-.23 .86-.18
15.0 .12 .08
15.5 .08 .04
Plot this in order to get a better perspective. Now you can see what
the .45V crossover point is about. The chart shows a vertical line at
the 14.65 A/F ratio (the y axis, output V starts at 0 and increases to
1.1, the x axis, A/F ratio, starts at 13.5 and increases to 15.5).
1a. Zirconia sensor-unheated
These are typically 1-2 terminal sensors. Some GM sensors require the
2nd terminal for a better ground because a 15-20 millivolt drop can
cause inaccurate readings. The drawings I have refer to applying an
anti-sieze compound to the threads, but that's all. Nothing special
about the compound.
1b. Zirconia sensor-heated
2. Titania sensors generate a typical resistance of (figures are
interpreted from a chart) (this appears to apply only to 'VRO' type
sensors-see 2a):
Typical resistance (ohms) @
Air/Fuel ---------------------------
ratio 850 deg C 650 deg C
13.5 2 2.2
14.0 2.1 2.4
14.5 2.4 2.6
14.65 2.7-4.0 2.9-4.9
15.0 4.3 5.3
15.5 4.5 5.5
Plot this in order to get a better perspective. The chart shows a
vertical line at the 14.65 A/F ratio (the y axis, Ohms, starts at 1
and increases to 6, the x axis, A/F ratio, starts at 13.5 and
increases to 15.5).
Titania is a recently development started use in Jeep's(?) ~2 years
ago.
I have two schematics showing differant methods providing a reference
voltage to the ECU:
2a. 'VRO' Titania-heated, 3 terminal
Provides a reference Voltage parallel to the sensor output external to
the sensor.
You can measure across the titania element without the heater in
series. This seems to be the type of sensor that the above table
refers to.
2b. 'WO' Titania-heated, 3 terminal
Provides a reference Voltage parallel to the sensor output internal to
the sensor.
A 2050 ohm resistor is in series with the titania element. There is
also a heater and 150 ohm dropping resistor in parallel with the
2050ohm res/titania circuit. This one, I'm not sure how you'd test
this other than using the Zirconia chart.
Heater voltage is 11-19 V.
Sensor output Voltage 0-1 V.
Hope this adds some useful info. Now if only someone could measure
the resistance of a Zirconia O2 sensor, we could figure how much a
low-impedance voltmeter would REALLY be off!
> >correct resistance, you could damage the sensor, or get erroneous
> >readings.
>
> Where does this stuff come from?
Damaging the sensor comes from:
1. 1983 Chevy light truck 10-30 series shop manual-HELM.
2. Chiltons Chevy/GMC Full size vans 87-90 repair manual.
3. Haynes Chevy Camaro 82-87 owners workshop manual.
OK! So the last two to aren't REAL book but they do say it; 'The
current drain of any conventional voltmeter would be enough to
permanently damage the [O2] sensor.' I can only guess that this
erroneous info comes from testing the ECU and other components with a
Digital Voltmenter, and someone just got carried away wit the specs.
> >Some sensors (that have a single terminal?) come with electrically
> >conductive anti-seize compound on the threads. If you use the wrong
> >anti-sieze, you may have a grounding problem.
> All sensors, whether 1 or 3 terminal, use the ground as the signal return.
> The other 2 wires are the heater that is isolated from the sensor.
> As long as one tightens the sensor securely, the conductivity of the
> anti-seize will not be an issue.
Info comes from the same reference source.
At the voltage/ohmn levels were talking about, I'm not sure I would
take a chance. I've been using anti-sieze and teflon tape/paste on
many sensors (switches, variable resistance sensors for analog gauges,
etc., but no O2 sensors) over the last 20+ years with no promblem
traced to poor ground. But when your looking for millivolts
differences in a HELL environment, does the same argument apply? For
the problems it might create, I'd go the extra step. Some people
apply anti-corrosion paste to battery terminals. I don't, but then I
also clean them every once in a while.
I like to see good techy type stuff like this O2 sensor thread. I
certainly have learned something.
John De Armond
>I had a telephone conversation yesterday with a senior eng. at Allied
>Signal-Autolite Division who works on Oxygen sensors. He sent me
>'some' design drawings and charts. I'll paraphrase and summarise what
>I received and discussed.
Would you happen to have a phone number? Thanks.
>Hope this adds some useful info. Now if only someone could measure
>the resistance of a Zirconia O2 sensor, we could figure how much a
>low-impedance voltmeter would REALLY be off!
There are several excellent theoretical articles on sensor parameters
in the SAE's "Sensors & Actuators" over the last few year. I know
specifically that source impedance is addressed. Unfortunately I
don't have my files handy.
>At the voltage/ohmn levels were talking about, I'm not sure I would
>take a chance. I've been using anti-sieze and teflon tape/paste on
>many sensors (switches, variable resistance sensors for analog gauges,
>etc., but no O2 sensors) over the last 20+ years with no promblem
>traced to poor ground. But when your looking for millivolts
>differences in a HELL environment, does the same argument apply?
I'd be about as equally concerned about thermoelectric potentials generated
by the metallic fillers commonly found in anti-seize compounds. Many
compounds contain either nickel or aluminum. Either metal in contact
with iron will generate significant potential at exhaust temperatures.
I don't consider either conductivity or thermoelectric potentials to be
of any concern considering the clamping force involved in a tightly
torqued sensor assembly.