O2 Sensor Bias

[Kurek, Larry]

	<SNIP>

	>I don't know if any of this will help lean out WOT operation,
but the
	>question about the resistor got me scratching my head some.

	>You probably could use a resistor tied to a 1v bias to make the
sensor
	>read a wee bit higher. (I'm pretty sure I've even seen
flaky-looking
	>commercial products designed to do just this.) Picking the
resistance
	>would probably be a pain. Also does an O2 sensor's output
impedance
	>vary much, within the warmed-up temp range? This mod would
probably
	>also mess up the 0.450v bias used to tell if the sensor is
warmed up
	>or not, maybe leading to problems during warm up?

	Nope...the PCM interprets a warmed up sensor as one that is
moving from rich/lean.

	>The clever way requires you to have faith that the O2 sensor
reading
	>is meaningful (not necessarily linear but repeatable) away from
stoich.

	Seems it is. At WOT, I get a fairly linear relationship between
pulse width and O2 readings. Since the PCM isn't going lean/rich at WOT,
it stays relatively flat.

	>Ok this isn't really true of an O2 sensor, but just pretend for
a minute.
	>What you want, roughly, is an O2 sensor who's output crosses
through the
	>450mv range at say 16:1 rather than 14.7:1. 

	Exactly! However, what I am *really* looking for, is something
to create a slight (50-100mv) upward bias to the O2 sensor output. So,
while the PCM isn't looking for a given voltage from the O2, it does
read the times the O2 sensor crosses stoich, and also is looking for an
average reading at a given load. So, if it is looking for an average O2
reading of 450mv and I bias it up 50 mv, an "actual" at the O2 reading
of 450mv would generate a 500mv signal at the PCM. If the PCM is
regulating pulse width to achieve an ideal 450, then it will lower the
pulse width (lean it out) until what it is reading is 450 on average. In
reality, this will be an actual 400mv at the sensor.

	>Suppose your O2 sensor
	>reads 380mv at 16:1 (wild guess). Essentially you generate a
380mv
	>reference, and take the O2 sensor signal, and feed the 2 into a
	>comparator, or even an op-amp set up for maybe 20:1 gain.
Divide and
	>buffer the output as needed for a 0-1v range, and feed that
into the
	>ECM's input. Congrats, your ECM will now see cross counts at
16:1
	>instead of 14.7:1, and adjust the AF accordingly. (If it
doesn't freak
	>out anyway.)

	Should work....any ideas on a schematic for this ? Anyone? :)

	>Problems:
	>This also defeats the sensor-warmup detection via the 450mv
bias.
	>You could add some more circuitry to get around that.

	Nope..see above, no problem here.

	>The sensor voltage corresponding to 16:1 probably varies a good
bit
	>with operating conditions, and the curve is steep to begin
with. If
	>you had a wide-range O2 unit, well, that would be nice. :)

	Sure would be, but I think the PCM looks for an average reading
over time, so any bias would skew the pulse widths where you want them.

	>How lean do you want to run? Detonation could be a real problem
	>depending on the engine.

	A little more than I am now. For instance, at WOT, I'm getting
readings in the 900-920mv range. I'd like to get this down more in the
850mv ballpark. Detonation won't be a problem, since I have an LT1 with
a knock sensor. That's another story...however.

	>I realize this isn't a very useful solution to Larry's original
	>problem... but once the question of tricking the O2 sensor got
	>into my head I just had to think about it until I was
satisfied.
	>Hopefully someone will find it useful or interesting.

	Hey, any ideas out there are useful to throw in the soup.
Granted, this would be a gross fuel curve adjustment, but it's an
idea...

	TTYL!

	Larry