O2 Sensors

Tal Dayan - March 18, 1997

I spent this weekend fixing the  O2 sensor in my RX7 (TII, 87, 155K) and
here are the findings (standard disclaimer applies). 

The problem was that the wire between the O2 sensor and loom got broke at
too places, one near the loom (about 1" long) and the other at its end near
 sensor. A year a go I 'fix' the break near the loom using a spade. Today I
found that this wire is actually shielded (it shows in the wiring diagram,
and you can verify it by carfuly examining the cut wire) and that the spade
I connected just shorten the shield (ground) and the signal. (About 6 month
ago I visited  Sunnyvale Mazda for another issue and the slick guy at the
front desk  came back and about my o2 sensor and who 'fixed' it but he did
not tell me  anything about the short circuit).

LESSON #1: the wire from the loom is shielded ! if you naively reconnected
it, chances are your close loop is not functioning.

Mazdatirx have an O2 for about 150$ and local Kragen has one (Bosch) for
less than 30$ so I decided to go with the Bosch (#11025) and  as afar as I
can tell, it works just fine. I also had to get an O2 socket and found it
at Kragen  with not problem (around 5$).

LESSON #2: Bosch O2 sensor (seems to) work just fine.

(Note: the Bosch sesnor comes with anti seizing grease, if for any reason
you use your own grease, make sure it is 'O2 sensor safe' and that it
specified for  high enough temperatures. Also, the wire connected to the O2
sensor gets very hot so use a wire that can sustained high tempertures (I
used the one from the old sensor)).

I also monitored the signal from the O2, when is connected and disconnected
from the  ECU. I used an oscilloscope (with 12VDC/110AC convertor) and  DMM
(TekDMM150, very handy, about 60$ at local store).  Note that  an
electromechanical voltmeter will not work because its impedance is too low
(usually 20K/V or lower).

With open circuit, the O2 generates DC voltage in the range  
0 - 850mv. (I also got negative reading of about 100mv but this might be
due to difference in ground potential and/or noise, I used a 5' non
shielded wire to the scope on the passenger seat and used ground from the
ignition key).  With throtle  released, the voltage was close to zero
(+-100mv),  and when accelerated, it jumped to about 750mv. I could not
make stabilize it on any value in between. More acceleration raised the
voltage up to 850mv (this time the change was more or less continuous).  It
seems that the sensor has some kind of threshold and a very high gain
around this threshold so more or less it signal is binary (yes, I know that
this may sound strange). (If you want to repeat the experiment, note that
the o2 sensor functions only when it is hot so you  need to warm up the car). 

Then, I repeat the experiment with the O2 sensor connected, and now it had
three states. With throttle released, it had low voltage close to zero.
When accelerating,  the voltage jumped to 750-850mv as before. When
cruising (with or without cruise control) the signal changed alternated the
high and low states in a rate of about 1 Hz. Roughly, it looks like a sine
wave, and I guess that this it shows how the close loop tracks the optimal
level with small errors above and below it. The error in the A/F is
probably small but the sensor has very large gain around the threshold and
that why the output goes between the two extremes.  The DMM makes a sample
about  once every second and the values changed all over the place, I guess
it depends what was the value at the time of the sample. However, the DMM
has also an analog scale that is updated five or ten times per second and
on that one, I could see the  one hertz fluctuation. Probably, this is
similar to the LED A/F meter sold by K&N.

LESSON #3, even when the close loop is on, the voltage on the O2 sensor is
either high or low (with a DMM with long time constant, this might look
like a stable 400mv but it isn't realy).

LESSON #4, if you want a cheap A/F meter that can do many other things, get
 TEK DMM 150 (its main disadvantage when used as an A/F meter  is  its
automatic shut off).

This summarizes my experience with the O2 sensor. The close loop works and
I am happy. I cound't notice however significant differences with the O2
sensor  connected or disconnected. I hope though that it will increase the
MPG.

If you have any comment/addition please let me know, I might compile
everything and submit it to the Turbo II site. Spelling and syntax  style
comments are also welcomed.

Tal 

Back to the Articles Page
Back to the Main Page