NetNews Usenet Archive 1992 #23/NN_1992_23.iso/spool/rec/autos/tech/13887

Text File | 1992-10-12 | 5.6 KB | 129 lines

Newsgroups: rec.autos.tech
Path: sparky!uunet!stanford.edu!agate!usenet.ins.cwru.edu!magnus.acs.ohio-state.edu!chohan
From: [e-mail] (Harpal Chohan)
Subject: Re: oxygen sensor
Message-ID: <[e-mail]>
Sender: [e-mail]
Nntp-Posting-Host: photon.magnus.acs.ohio-state.edu
Organization: The Ohio State University
References: <[e-mail]> <[e-mail]> <[e-mail]>
Date: Mon, 12 Oct 1992 17:01:36 GMT
Lines: 116

In article <[e-mail]> [e-mail] (Chris Elmquist) writes:
>
>So, the more O2, the lower the voltage out from the sensor ? Or visa-versa ?

Vice-versa. The leaner the mixture, lesser the residual O2 in the exhaust,
lesser the output voltage. Main thing to be noted is that the "V-out" vs. the
"excess-air factor (lambda)" curve is not linear. At the optimum, the
voltage jumps from about .25 to .75. This makes detection of deviation
from the optimum rather easy. .25 or less indicates a lean mixture, whereas
.75 and above shows a rich mixture.

>Mine (my engine) seems to misbehave more before it is fully warmed but
>after it has stepped the idle down from the higher speed cold idle.  It
>is during this warmup phase that it runs the worst.  What kind of O2
>sensor output would be expected by the computer during this time...? is
>the computer even looking at the O2 sensor then ?

It probably is, since the cold idle phase is over. Maybe your sensor is
old, and is not responding as fast with warming temperature as it should.

There's an excellent article on oxygen sensors in the June 1992 issue of
"Europen Car" on page 6. It's kinda' long for me to type the whole thing in,
but here's what it has to say. (...comments...) are mine...


How quickly the oxygen sensor responds with a voltage depends upon the
temperature of the probe. Below about 300C, there is little or no voltage
generated. That means on start up the oxygen sensor loop is out of the
circuit. (I believe that means the ECU doesn't rely on its input, if I recall
what my Bentley manual says correctly). Right about 300C, the response
time is around one second. As the temperature of the probe rises to around
600C, response time drops to less than a millisecond.

    [...]

The ECU looks at the response time of the sensor and sort of averages things
out to try to keep the mixture in the narow band the converter (catalytic)
prefers. (I believe it's talking about when the response time is slow,
between 300 and 600C, because when it is hot, it goes into closed loop
operation).

  [...]

Closed loop operation is when the oxygen sensor has control over fuel
mixture. The ECU reads the difference between output voltage (from the
optimum) and adjusts the mixture. Typically, closed loop is any time
that the throttle is not wide open or completely closed. (Indicated to
the ECU from the throttle switch).  [...]. During closed loop operation
with a hot sensor, the oxygen sensor and the computer are juggling fuel
mixtures thosands of time a second, keeping the mixture spot on so that
the catalytic converter can function properly.

 [...]

(On how they go bad with age).

With time, deposits on the tip of the probe slow things down. Afetr a
number of miles, the sensor responds so slowly that the idle can deteriorate
and the fuel mileage can drop. [...]

Oxygen semsors can be damaged by age [...]
contaminated by leaded fuel [...]
fuel additives [...]

With a brand new oxygen sensor, the sensor reacts quickly, mixture
changes are quick, and the idle varies less. With an old or damaged
sensor, the idle can be very inconsistent and "lumpy", or it will
ground out internally and the mixture will go way too rich and the
car will not idle properly.

(On testing)

If you unplug the sensor and the idle improves, then chances are that
the sensor is bad.

To test the circuitry part of the ECU's lambda system.
Disconnect the wire coming from the oxygen sensor. Hold the
wire going to the brain (ECU) in between your fingers taking
care not to touch any other metal parts of the car. Touch the
positive post of the battery with the other hand. If the ECU circuitry
is good, the small voltage flowing through your hand to the ECU will
cause the mixture to lean down. (I dunno about touching the battery
myself :-). I'd rather feed it about .8 volts from another source
to do that same thing, though). The idle will get rough, and eventually the
engine will stall, showing the lambda control (in the ECU) is working.
Conversly, if you touch the wire from the computer to the ground
(as Richard suggested a couple days ago), the mixture should richen up,
causing the idle to speed up. If your car passes these tests, the
ECU part is probably working correctly.

End excerpts.
---

The article doesn't say how to directly test the sensor itself,
but if I remeber what my Bentley manual says (it's at home),
when the car is hot, disconnect the oxygen sensor input to the ECU,
and hook on a digital voltmeter to the sensor output. The voltage
output should be fluctuating between .2 and .8 volts. (You probably
can't really tell how fast it is fluctuating, since the sampling
rate of yor voltmeter might be no match for the rate of the
sensor's output [put the meter in analog mode if it has one],
but you can still tell if it is so bad that it doesn't change much
at any rate at all).

Then, lean out the mixture by pulling the dipstick or oil cap
and see if the sensor's output drops immediately. Conversly,
ground the ECU end of the sensor input wire and see if the
voltage rises immediately.

These tests should more or less give you an idea if your sensor
is on the ball or not.

Cheers...

-h