Op amping the O2 sensor

[Tom Cloud]

At 11:29 AM 10/24/97 EDT, Bruce Bowling wrote:

>
>What I have heard is that this "switching" frequency is something
>below 10 Hz - an op-amp could easily do this.

that may be the rate at which the ECU changes the mixture -- I know
that on SEFI systems the ECU is able to monitor each cylinder's
O2 content -- so the EGO is pretty fast (read that somewhere, can't
remember it now).

>I guess the op-amp circuit would be configured as to add an adjustable
>DC offset to the O2 sensor output?

this is what I did with the LM-324 circuit that I whipped up and
submitted -- which doesn't work  8^(  ... the LM-324 outputs won't
pull to ground -- no matter what the literature sayeth -- unless
they're "helped" with a "pull-down" resistor or are driving no load.

>Since we are talking O2 sensors, I have a fundamental question about
>the operation of an O2 sensor around stoichiometric. Some people have
>stated that the O2 sensor puts out a roughly linear voltage as one
>mucks with the A/F ratio. Others have stated that the O2 sensor
>acts like a hard "switch" at stoichiometric, producing voltage
>at one side of stoch and not producing on the other.
>
>Which is it? OEM boxes appear to treat the signal as a switch, but I see
>all these gizmos that display the voltage as a variable quantity?
>Does the thing possess some sort of "duality"?

the output is analog (i.e. it's not a quantum leap kinda thingie  ;-)
but it's function is far from linear, with it changing joltage
rapidly around stoich.  Output near zero for lean mixtures and
near 1 jolt for rich, with steep slope from about .2 volts to
.8 volts at stoich -- forget A:F range, but it's very narrow.  So,
A:F mixtures either side of stoich are not well represented by the
standard EGO.  The SAE literature I've seen describes the sensor
as fairly accurate -- if you know the temperature of operation.
Elsewise, it's accuracy is only at it's "switching" point at stoich.
(Also, it's "switching" speed -- or response time -- is directly
related to it's temp and it's condition (i.e. good or bad).)
So, it'd be possible to put a look-up table in ROM to correlate
EGO readings to a fairly wide range of A:F ratios -- but only
if you know the temp of the sensor.  I've heard that there are
some tricks where one uses a constant current through the sensor
and somehow determines its Zout -- which is a function of its temp --
and thereby deriving its operating temp.


Tom Cloud

 Clothes make the man ....  Naked people have little or no influence on society.