Re: O2 SENSORS

["Scott M." <scottmo@teleport.com>]

Rolf,

I wanted to add some additional information concerning the
operation of the O2 sensor and the operation of the
computer controlled closed loop mixture control system.

> The computer puts out a reference voltage
> of aprox.  .5 volts. This voltage is compared to the
voltage coming out of
> the sensor to determine the stoichiometric (sp?) ratio of
the fuel/air
> mixture. If the ratio is rich, say 13-1 (14.7-1 being
ideal) the voltage
> from the O2 sensor will higher than .5v and the computer
will alter the
> mixture. Vice/versa, 16-1 is lean and the poooter will
richen the mixture.
> 
The typical O2 sensor (0-1V output) unfortunately can only
provide an accurate voltage output right at the 14.7 to  1
air/fuel ratio, outside this 14.7 to 1 A/F ratio the
sensors output voltage is non-linear and is not accurate at
determining the actual air/fuel ratio. The O2 sensor
voltage will transition high to low passing through this
0.45 V output (14.7 to 1 A/F) when the mixture changes from
a "slightly" rich mixture to a slightly lean mixture. The
voltage will also transition in the other direction, low to
high as the mixture goes from slightly lean to slightly
rich. The computer does not try to calculate the actual A/F
ratio by reading the O2 sensor voltage, it only tracks
whether the voltage has transitioned past this 0.45
stoichiometric value.

Rolf said:
> If the sensor is deleted, or is bad, the computer will
likely make up it's
> own value, right around .5 volts, which if the engine is
in a decent state
> of tune, will run almost as well. 

This is true, on the CIS equipped vehicles, if the basic
idle mixture CO% is set correctly and the O2 sensor is
disconnected, the computer will choose a 50% duty cycle for
the frequency valve operation and the car will run very
well with little notice from the driver. This is done on
purpose in the event of a failure of the O2 sensor or
wiring.

Rolf said:
>To check a sensor, unplug the lead wire
> (always black) and hook up to a simple volt meter. The
engine exhaust must
> be at least 600 Degr. F. The voltage should be jumping
between .3 to .7
> volts or there abouts. if it moves very slow or not at
all, the sensor is
> dead. 

With the heated O2 sensors the exhaust temperature does not
matter as much, these heated O2 sensors usually begin
working after 1-2 minutes after the car is started. When
viewed on an oscilloscope the O2 sensor voltage normally
will transition between ~0.1V and ~0.8 V. This cycle of low
to high and high to low voltage transitions (oscillations)
occurs around 1 every two seconds at idle and 3-5 per
second at higher engine speeds (higher exhaust gas flow) .
If the O2 voltage remains high (rich fuel mixture
condition) say at ~0.8V with no oscillation, it does NOT
mean the sensor is dead! The same is true if the O2 sensor
voltage is stuck down low at ~0.1V indicating a lean
mixture. The O2 sensor is only responding to what the
exhaust mixture is and if the computer can no longer adjust
the mixture correctly within the range it was designed,
then the O2 sensor will output a nearly constant voltage
either high or low and is not at fault. 

The early closed loop computer controlled fuel systems
could not correct for a gross problem in the fuel mixture
caused by a vacuum leak or by leaky injectors or incorrect
fuel pressure. The later systems are "adaptive" in that
they can adjust the mid range set point for the fuel
mixture and adjust for "some" problems that occur in the
engine mechanical systems or in the fuel system.

The O2 voltage oscillations occur because of the following
scenario: The computer reads the O2 sensor voltage and sees
that it is high say ~0.8 V, this indicates a slightly rich
mixture so the computer leans out the mixture, then the O2
sensor voltage drops down to ~0.1V due to this leaner
mixture. The computer again reads the O2 voltage and sees
that it is low indicating the mixture is lean and then
adjusts the mixture slightly rich and the cycle continues.
The cycling of the mixture back and forth rich/lean,
lean/rich is done in part to allow the three way catalytic
converter to operate efficiently to oxidize the HC and CO
and reduce the NOx.

One  test to run on the O2 sensor is to connect an
oscilloscope to measure the O2 sensor voltage and then
force the mixture rich briefly with the addition of propane
into the intake system. The O2 sensor voltage should rise
up to at least 0.85 volts. Then force the mixture lean by
creating a huge vacuum leak and measure the O2 voltage
transition time when the voltage drops from high to low.
Typical transition times are around 25-50ms,  the rule of
thumb is that the O2 sensor transition time from 0.6V to
0.3V should be under 100ms. The O2 sensor will have a
different transition time going from rich to lean than from
lean to rich, if I remember correctly the rich to lean
transition is slightly longer.

One problem that affects the O2 sensor operation is
contamination or poisoning by silicone, this shows up as a
fine white powder on the tip of the sensor and will reduce
the voltage output of the sensor when the mixture is rich,
this will cause a loss of fuel economy and increased CO an
HC emissions. You may also see a negative voltage developed
under lean operation when the sensor is poisoned by
silicone. This poisoning causes the sensor to see a lower
proportion of Oxygen. 

The slots in the tip of the O2 sensor can also get
partially clogged with carbon which will reduce the
response time and will show up as a lower number of cycles
or transitions.

There are several Society of Automotive Engineers  (SAE)
articles that have been published since the early 70's that
describe in the utmost detail the operation of these O2
sensors. Each year SAE publishes 2-3 thousand articles
covering the various automotive areas and is required
reading for anyone looking for in-depth technical
information on automotive systems.  Article 860478 covers
the silicone poisoning of O2 sensors, most university
libraries have the articles going many years back on
micro-fiche and they should have an index that lists all
the articles by subject, author and number.

HTH
Scott M.
89 200TQ
CHIPDIPPINDUDE