[Diy_efi] WB O2 system

[Garfield Willis]

On Tue, 11 Jun 2002 09:26:39 +0800 (WST), Bernd Felsche
<bernie at innovative.iinet.net.au> wrote:

>Nit-pick: The sensors' output voltage is a measure of the ratio of
>partial pressures of oxygen. AFR can then be approximated if you
>know the fuel type and the relative pressures of exhaust and
>reference gases.  The sensors don't measure AFR. If you're unsure of
>the fuel type (usually the case at the consumer end), then it's more
>sensible to use the normalised ratio of AFR to stoichiometric,
>denoted in the technical literature with the Greek letter "lambda".

No, unfortunately this is not correct. Nor is it a nit. The sensor's
output is a measure of Oxy partial pressures ONLY *on the lean side*. It
isn't even measuring Oxy on the rich side (any excess Oxy on the rich
side is fairly constant). Since the sensor is sensitive to both CO, H2,
& unburnt CxHy on the RICH side, there is NO equivalence for arbitrary
fuels by just using a 'lambda' based on sensitivity to Ox alone. You
must also know the sensor's relative sensitivity to CO, H, & CxHy, which
varies from individual sensor to sensor, as well as type of sensor.

Therefore, you can't arbitrarily use such a sensor 'if you're unsure of
the fuel type'. The 'lambda' indicated will be just as far off as the
AFR indicated.

If you get a real commercial 'lambda' meter and sensor calibrated for
arbitrary fuels, from a reputable source, you'll also get a cal sheet
that gives the relative sensitivities to each of these exhaust
components, that you then punch into your instrument. You also have to
punch in the relative components of the fuel. Without these TWO steps,
your 'lambda' won't be of much use to you on the rich side. This is true
of the Horiba, Bosch, ECM, & EGORtech products at the very least. Of
necessity, it must also be true of any use of these sensors when
measuring mixture with fuels other than petrol.

In summary, you CANNOT determine a normalized lambda good for arbitrary
fuels, with these sensors, unless you know both the relative components
of the various fuel byproducts, AND the relative sensitivity of your
particular sensor to those byproducts. Everybody I'm aware of that
supplies meters for use with alternate fueling, ALSO includes both this
calibration data, AND a way of introducing it into the instrument, as
well as the fuel particulars. This takes special calibration gases and
apparatus.

Whether you call it an AFR or Lambda meter, without these features, you
cannot begin to measure either, IF using any other fuel than petrol.
With petrol alone, these effects are already built into the published
AFR (or lambda) vrs. sensor output curve. But that curve assumes a
particular fuel. Just because you're given such a curve in 'lambda vrs.
sensor output' form, doesn't mean it applies to anything BUT petrol. The
curve on the rich side will be different for each different fuel.
Especially when dealing with fuels containing their own oxygenates, like
the alcohols or nitro, the differences can be *very* pronounced.

Hope that clears things up a bit, about measuring AFRs -or- Lambda's
with fuels other than straight petrol.

Gar Willis
Principal Engineer
EGOR Techno
3491 Edison Way
Menlo Park, CA 94025
650-216-9874
garwillis at msn.com (e-mail & PayPal transfers)
www.egortech.com (best viewed with IE)


_______________________________________________
Diy_efi mailing list
Diy_efi at diy-efi.org
http://www.diy-efi.org/mailman/listinfo/diy_efi