So I was reading about EFIE devices, and...
 

I have a question about them, but I'll get to it at the end.

O2 sensors are used by the engine computer to sense the amount of oxygen needed to completely combust the remaining hydrocarbons in the exhaust stream. Rich exhaust causes oxygen to flow from outside into the exhaust stream, and lean exhaust causes this flow to slow or even stop.

http://www.plxdevices.com/products/w...utputGraph.jpg

Standard O2 sensors can detect AFRs between 14:1 and 15:1 (much simplified explanation for the sake of brevity), but they don't do a very accurate job of it because they'll go to an equilibrium oxygen flow, which will cause the O2 sensor signal to become slightly inaccurate. Such inaccuracy may be tolerated with post-cat O2 sensors, where their primary job is to monitor the cat's efficiency, but this inaccuracy is unacceptable for AFR control. This is typically why engine computers continuously vary their fuel output between slightly rich and slightly lean, in order to keep the O2 sensor signal from going to equilibrium.

http://users.du.se/~h02johed/img/lambda.jpg

Now, my question is - If an EFIE device is supposed to trick the engine computer into running leaner than before, by modifying the O2 sensor signal to make it appear that the engine's running more richly than it really is, how can you maintain control of the air fuel mixture if the O2 sensor output only goes to 15:1? Because it seems to me that if you're going to go beyond 15:1 (where real fuel economy gains would be realized), you're going to need something other than a device that merely changes the signal coming from a sensor that is limited to 15:1 on the lean end. Otherwise, you have no idea what the AFR is, and you'll eventually end up burning something up inside the engine due to an unanticipated over-lean condition.

A wideband LAFtype O2 sensor will read up to a 22:1 A/F ratio.

Here's an idea.

You may be able to install one in addition to the one your trying to trick.
Then monitor voltage separately to get a idea of how lean your running.


Honda uses these type O2 sensors in their lean burn engines.

>

Redneck beat me to it. Yes, you will need a wideband in order to accurately monitor the AFR outside the typical "narrow" band range. Something like this one from Dynojet: AirRam - Cold Air Induction System

Last summer, I installed an ancient old PLX M-250, and it takes the place of the pre-cat O2 sensor. It has a wideband output suitable for an aftermarket engine computer or a gauge, and a narrowband output suitable for the stock engine computer. It enables me to determine when my truck goes out of closed loop operation when I accelerate, and is pretty neat to watch.

I could build a circuit that could take the wideband output, and generate a narrowband output in the desired AFR range, and that would only take effect at 20% throttle opening or less. Above 20%, the circuit would revert back to an unmodified signal. I'd still have to replace the post-cat O2 sensor with my other M-250 that I have sitting around, and have this circuit modify that signal as well, so as to successfully fool the stock engine computer.

However, that is something that we would do. I imagine that most people who buy these EFIE things do not actually take the effort to ensure an accurate AFR by buying a $300 or so wideband O2 sensor to monitor their output. What do they do, instead?

they cause damage they can not repair ...
 

Standard narrowband O2 can detect wider than 14-15 AFR . while that range is were it is most accurate I have seen it able to output values wider than that graph (something like low-mid 13's for 0.9v and probably same on high end,close to around 16 AFR but i never checked that end )

As above poster running lean has many side affects , I wouldn't do it alone .

mwebb -

Does this agree with your statement? :

Automotive electronics handbook - Google Books
http://ecomodder.com/forum/member-cf...-emissions.jpg

CarloSW2

low resolution 02 senSor waveform showing MISFIRE
 

[QUOTE=t vago;218060]I have a question about them, but I'll get to it at the end.


http://www.plxdevices.com/products/w...utputGraph.jpg

http://users.du.se/~h02johed/img/lambda.jpg

QUOTE]THIS LOW RESOLUTION 02 SENSOR WAVEFORM
is showing pronounced consistant MISFIRE
====================
yes
that NOX graph agrees with my SWAG
but
it is true that combustion temperatures are HIGH during hi load with slightly rich conditions , when NOX output is very low, perhaps because there is no 02 availible to combine with the N
to form
NOX , during rich condition.

This is the book (rather, two-volume set) that I reference whenever I start researching something or other about gasoline engines

It would appear that my question was answered. I think it would be a good idea to put up a sticky on the potential drawbacks of using an EFIE (particularly EFIEs that only rely on the stock narrowband O2 sensor), perhaps as part of listing the potential drawbacks of running lean.

So how much less O2 is there after the catalytic converter? Can this sensor be used to trick the ECU into running leaner?


http://ecomodder.com/forum/member-cf...-emissions.jpg

What? (can you draw a picture?)

That's how an O2 sensor works. It works off of the diffusion of oxygen from one side of the sensor to the other. If there is a condition where extra oxygen is demanded and there is none present in the exhaust stream (such as for a rich mixture), oxygen flows from the air outside of the sensor into the exhaust stream. If there's an excess of oxygen in the exhaust stream, that flow slows down, or stops. Higher oxygen diffusion rates cause higher voltages to be generated by the O2 sensor, up to a point.

We're talking about minuscule amounts of oxygen, though.

t vago -

Do you mean the outside air is "diffusing" through the membrane (shown below)? :

Oxygen sensor - Wikipedia, the free encyclopedia
http://upload.wikimedia.org/wikipedi...Sensor.svg.png
http://upload.wikimedia.org/wikipedi...Sensor.svg.png
CarloSW2

02 sensors detect combustibles not 02
 

I read that an EFIE adds a bit of voltage to the O2 sensor signal (like 0.2v for example), to basically skew the median to a leaner value. I agree, just installing this gizmo, dialing in a guesstimate voltage, and not having something to accurately monitor the result is not a good idea. However, I expect that is what several people do...and will be surprised when something goes ka-boom.

Well, I've been thinking, on and off, of using a WBO2 controller to control a WBO2 sensor, and passing the output through a microcontroller to generate a simulated narrowband O2 signal for the stock engine computer. The circuit would modify the WBO2 signal as required, taking throttle position and MAP into consideration, to fool the engine computer into running lean.

After thinking and thinking about it, I've come to the conclusion that I well and truly hate this idea. Why? Because decent off-the-shelf WBO2 controllers cost at least $150 apiece, and I'd need to install 4 of them into the Magnum (2 pre-cat and 2 post-cat). That's on top of the 4 each WBO2 sensors themselves, which I think are about $50 a pop. That's at least $800, all told.

Alternately, to test the WBO2-controller-as-EFIE idea, I could go install my other ancient PLX WBO2 controller into the truck (for the post-cat O2 sensor), but... why? I don't drive it that much any more.

the AFR "cycles" for two very important reasons....

#1 being that it's how the narrowband sensors are used to correct for fueling. if the sensor is constantly sitting near any one value, it's indicating that the sensing element is dead, it's not up to operating temperature or an extremely rich or lean condition is present. by forcing the AFR to "swing" around stoich, the ECM will cause the AFR to average stoich by attempting to make the time above and below the reference voltage equal.

#2 is the cat itself... supposedly they operate more efficiently when they are put into situations of slighty rich followed by slightly lean.... i really don't understand why, but it's been suggested that it needs some extra oxygen inside it to help the reactions. if this were the case, an AIR system that pumps air into the cat nearly all of the time would probably work quite well. or at least get it hot and keep it that way very quickly, though it wouldn't hurt to estimate cat temp so not to overheat it.

This is primarily why modern engine computers have AFR swing around stoich as you described. It's to get rid of air injector pumps that would be driven as an engine accessory - one less drain off the engine's output. The extra oxygen is pumped into the cat substrate during a lean cycle, and is then used to reduce the slight amounts of NOx and CO in the exhaust stream. This stored oxygen does not normally get depleted under normal operating conditions, but will get depleted when AFR goes rich (as with hard acceleration).

I did find the Zeitronix ZT-2 has a wideband O2 sensor that has an adjustable narrowband output. So you could skew it lean as you like. But as you indicated, with a Dodge V8 you'd need at least two for the pre-cats. I asked Zeitronix and that's what they told me. Said I might throw a code from the post-cat O2s when running leaner. The truck doesn't get inspections (OK licensed) so that wouldn't be an issue, just annoying. At $300 each, I am inclined to just bite the bullet and blow my money on other things...like a lighter dampner or lighter wheels or something.

Dust, dust... *cough* *cough*

I found this interesting thing... It's an open-source wideband O2 sensor controller, which uses an AtMega8 as the brains. They even provided the source code! Provisions for heater circuit and Nernst cell PID controllers are provided, as well as input signal processing and narrowband output simulation.

I may be able to adapt this controller to control 4 different WBO2 sensors. If this is possible, then I may be able to fool my Magnum's engine computer into going to 16:1 AFR for cruising.

Looking forward to hearing the progress. I'm guessing you'd try to stick to the 20% upper limit you mentioned earlier...