more O2 sensor BS

[Chris Conlon]

Because I can't resist putting my $0.05 worth of confounding info
into the fray, here it is. Yes most of this is just me BSing so the
rigorous in the audience might feel free to delete already.

>Nominal volts: ~-40mv in free air, returned to this
>after each test by fanning above can with piece of paper.

This is probably just the thermocouple voltage from the cell being
heated. It's in the right range anyway for ~800+C and some rough
average thermocouple outputs.

The first thing is that the cell probably develops a voltage due
to more than one reaction, or could do so. If it does have a
platinum/platinized electrode, why couldn't it act as a fuel cell
and oxidize stuff (HC included) to produce some voltage? This
would give a voltage dependent on redox potential, sort of, and
might explain some of the confusing results.  Note that I'm not
saying "it senses HC", I'm saying it might sense an oxidation potential
gradient acoss the sensor membrane. HC on one side and O2 on the other
could cause this, but other things could too. Has anyone flooded both sides
of a warm sensor with a small HC (methane or acetylene)? I bet you
get ~40mv, i.e. no output.

Also the size of the HC most likely makes a big difference. In the
exhaust stream you're looking at lots of small HC and alcohols, ketones,
etc which might be more likely to participate in a membrane-located
oxidation process.

In my mind a good first test would be to use nitrogen (air minus all oxygen)
on one side or the other of the cell. I think this will tell you if the sensor
responds to an oxygen gradient, without any other confouding HC, CFCs,
etc. The argon test is close to this but not quite there. Also Bruce's
test with pure O2 vs. air (if I understood how he ran the test) suggests
that the sensor might not respond much to just an O2 gradient. (A partial
response, yielding a different voltage or lower current output at the same
voltage, is a possibility, I think.)

Ok my producer is signalling me to wrap, so here goes:

>Thinking about this, it makes sense. Once you reach stoich (450mv),
>how can you go higher? Only if the excess hc takes it there...

First do we at least agree that the O2 sensor does basically *work*?
That it does sense something, and lets you measure rich vs. lean, if
not very precisely? Given that...

My guess is the sensor responds to O2 partial pressure gradients (as
claimed), with a medium to low sensitivity. I don't know if it also
responds (with a higher sensitivity) to O2/HC redox directly (a la
fuel cells), or if the presence of small HC on a hot (possibly platinized)
surface is just very good at creating a bigger O2 gradient by consuming O2.

This would allow response on both sides of stoich; on the lean side there
is excess O2, up to the amount found in air, which lowers the O2 gradient
and the output voltage. (Also in the case of rich misfire, where the HC is
mostly uncracked larger HC stuff, which I'm guessing is too large and slow
burning in this situation to consume as much O2. Remember where I said size
might matter? ;)  On the rich side there is spare HC (small HC residues) to
sop up extra O2, and create more of an O2 gradient than just N2 + H2O + CO2
(which is what you would get at stoich, in theoory) would.

So there you go, fwiw.

   ---Chris C.