Fooling the Computer for Better MPG

[xjbubba]

The catalytic converter requires the mixture to vary between lean and rich--extended running above or below 14.7:1 fuel/air ratio will eventually kill the cat. Any economic gains achieved by running lean would be lost in the cat's replacement. You can also re-program the fuel tables in the ECM to do what you want. With Renix, you could bias the O2 output to run lean all the time--no codes or check engine light to worry about--but I hope you have ears that hear better than mine, and you might as well gut or remove the cat; of course, you won't be doing the environment any good:nono:

 

[RaccoonJoe]

........Sounds like an interesting project. Keep in mind that the Renix O2 sensors are different. The ECU supplies 5 volts to the O2 sensor and the O2 sensor acts as a variable resistor. The post 90 O2 sensors produce a 0-1 volt signal that is proportional to the O2 concentration, so matching the correct current/voltage and impedance matching may be real tricky with them............

OK, next question. I have a O2 sensor eliminator kit (basically a resistor in a weatherproof box), that was designed to replace the downstream sensor in ODBII vehicles. It was purchased when I was having some cat issues with a 98 Blazer, but never used. Would this work to fool the single sensor in my 89 XJ??

 

[87manche]

OK, next question. I have a O2 sensor eliminator kit (basically a resistor in a weatherproof box), that was designed to replace the downstream sensor in ODBII vehicles. It was purchased when I was having some cat issues with a 98 Blazer, but never used. Would this work to fool the single sensor in my 89 XJ??

no

firstly, renix boxes used a diffeent kind of O2 sensor.
Secondly, the upstream sensor is what the ECU uses to determine mixture. The Renix ECU doesn't have a preset fuel map like an OBD system, it uses realtime data gathered from it's sensors to run the motor.

 

[Ecomike]

no

firstly, renix boxes used a diffeent kind of O2 sensor.
Secondly, the upstream sensor is what the ECU uses to determine mixture. The Renix ECU doesn't have a preset fuel map like an OBD system, it uses realtime data gathered from it's sensors to run the motor.

If the Renix does not have somekind of a preset fuel map, how does it start, and how does it run (idle at least) with out TPS and O2 sensor attached? I have started and idled the Renix with out the O2 sensor and TPS sensor attached.

A resistor will not work on OBD I or II to replace the O2 sensor because the O2 sensor produces a voltage when it is working, it is not a resistor, but it can be used in series to lower the voltage output to the computer thus changing the computers perception of the A/F ratio (making it think it is running leaner and forcing the computer to run it richer) But you need the right Ohm rating, low enough to work and not so high that the computer decides to ignor the O2 sensor, and not to high so that you burn up the engine. It will not make it run leaner, only richer (I am pretty sure).

I would think the Renix would see a fixed resistor for what is, fixed, and ignor it when it did not see the resitance change with changes in the A/F mix, just like it would ignor a bad O2 sensor that started displaying a fixed or off resistance.

However the right size (ohms) resistor in parallel or series with the renix O2 sensor ( which is a variable resistor) might be used to change the preset computer A/F ratio!

Not sure which direction the A/F ratio moves on the Renix, but adding a small, say 1 to 10 ohm resistor to the ground leg of the Renix sensor should shift the A/F ratio a bit too. If the Renix follows the same direction as OBD-I and II, namely that lower voltage means it thinks it is running leaner, thus forcing the computer to run it richer, a small resistor in parallel might shift it the other way, making it think it is running richer and forcing it to run leaner.

 

[BC-Handler]

Hell, LIFE is risky. Risky was when I was flying fighters in USAF in the Vietnam War.

Hey, Bob;

Welcome home.

BC-Handler, US Army, RVN '70-'71

 

[87manche]

If the Renix does not have somekind of a preset fuel map, how does it start, and how does it run (idle at least) with out TPS and O2 sensor attached? I have started and idled the Renix with out the O2 sensor and TPS sensor attached.

A resistor will not work on OBD I or II to replace the O2 sensor because the O2 sensor produces a voltage when it is working, it is not a resistor, but it can be used in series to lower the voltage output to the computer thus changing the computers perception of the A/F ratio (making it think it is running leaner and forcing the computer to run it richer) But you need the right Ohm rating, low enough to work and not so high that the computer decides to ignor the O2 sensor, and not to high so that you burn up the engine. It will not make it run leaner, only richer (I am pretty sure).

I would think the Renix would see a fixed resistor for what is, fixed, and ignor it when it did not see the resitance change with changes in the A/F mix, just like it would ignor a bad O2 sensor that started displaying a fixed or off resistance.

However the right size (ohms) resistor in parallel or series with the renix O2 sensor ( which is a variable resistor) might be used to change the preset computer A/F ratio!

Not sure which direction the A/F ratio moves on the Renix, but adding a small, say 1 to 10 ohm resistor to the ground leg of the Renix sensor should shift the A/F ratio a bit too. If the Renix follows the same direction as OBD-I and II, namely that lower voltage means it thinks it is running leaner, thus forcing the computer to run it richer, a small resistor in parallel might shift it the other way, making it think it is running richer and forcing it to run leaner.

I will clarify, it has a preset for starup and open loop. Closed loop it runs what the sensors tell it to run.
you could very well use an inline pot to vary the A/F ratio, and the renix would probably be better suited because of it's flexible programming, and the fact that it has a knock sensor.
I would for sute install a wideband O2 sensor and guage though.

 

[Ecomike]

A gauge! What a cool idea:yelclap:, why not 2 gauges, one for the true A/F and one for the lying A/F going to the computer????

I don't disagree, but wondering why you suggest the wideband O2 sensor? Do they even make a wide band that will work with a Renix????

 

[8Mud]

Just throwing a couple of things out there. But IMO you have to keep the RPM band in mind (primarily) along with different loads and vacuum.
Some techniques may have more successes in different RPM bands. If any adjustment you make throws the sensor (say O2) out of the envelope, the computer will likely default and switch to closed loop (much like WOT). I can picture most any O2 sensor tweak causing trouble near peak torque, in the 2-2400 RPM range.
Some small tweaks, that add up may produced better results. The tweaks could even be seasonal. There are some things that may work better in the summer.
The engine temperature sensor would be easy to fool, a percentage resistor from the input wire to the sensor bundle ground splice. Same with the MAT sensor. You'd be fooling the program into running a warmer, leaner, curve. That the programming is likely to have answers too through the entire RPM band.
It may only be a few percent mileage improvement (if any). I've also had some luck tweaking my basic TPS setting a little.
A little here and a little there, it all adds up.

 

[bobinyelm]

The catalytic converter requires the mixture to vary between lean and rich--extended running above or below 14.7:1 fuel/air ratio will eventually kill the cat. Any economic gains achieved by running lean would be lost in the cat's replacement. You can also re-program the fuel tables in the ECM to do what you want. With Renix, you could bias the O2 output to run lean all the time--no codes or check engine light to worry about--but I hope you have ears that hear better than mine, and you might as well gut or remove the cat; of course, you won't be doing the environment any good:nono:

If one were to bias the O2 sensor with a summing OP AMP (i.e., the O2 sensor would work normally, but the output of our OP AMP, what the PCU actually "sees," would be the sum of what it reads from the O2 sensor plus perhaps 0.05-0.10 volts), the computer would STILL see what is supposed to see (i.e., a rich/lean pattern) except that instead of varying between (say) and A/F of 13.5-15.2 (centering around 14.7), it would be varying from perhaps 14.7-16.2, varying about a somewhat leaner value.

I seriously doubt the cat would care (actually, with a bit less raw fuel to burn, it should be quite happy), and the PCU would be none-the-wiser. It would simply lean the mixture so as to return ITS input to 450mv (what it thinks is producing an A/F of 14.7).

Naturally, the vehicle's driver would need access to adjust the center A/F mixture (a potentiometer that controlled the voltage fed to on leg of the summing OP AMP) to compensate for load, altitude, etc. One would probably want to return the additional reference input to zero (assuming the OP AMP is set to a gain of 1.00) when towing or under load (to reset the actual mixture to 14.7), and be able to increase it to perhaps 0.10v when one desired to run slightly leaner.

If you do a "Google" on OP AMP Circuits, you will see numerous examples of simple summing OP AMPS using 99 cent ua741 Integrated Circuits. The only other components required would be a 5 volt regulating IC ($2 at Radio Shack) and a few resistors.

I am not presently set up to bread-board such a circuit, but it should be pretty easy.

Just cut the O2 output wire somewhere up the O2 harness from the O2 inline connector (since the O2 sensor gets its reference through its own leads and connector we don't want to mess with it between the sensor and the its connector) and insert the OP AMP there. The actual amp could be anywhere (in the cockpit probably) to make access to the potentiometer easy.

Installing a male and female in the O2 output wire in the harness would facilitate removal of the device should one wish to return the circuit to original, yet permit insertion of the device, which would have identical male/female connectors.

Other than the access to the O2 output and PCU O2 input (both provided by severing the O2 line), one would only require a 12v source to power the 5v regulator, and a ground reference, of course.

I wouldn't even mess with any other sensors, since when in closed loop, the O2 sensor is in effect the "Master Control" for the whole system.

Bob
Bob

 

[87manche]

A gauge! What a cool idea:yelclap:, why not 2 gauges, one for the true A/F and one for the lying A/F going to the computer????

I don't disagree, but wondering why you suggest the wideband O2 sensor? Do they even make a wide band that will work with a Renix????

a wideband O2 sensor wired to a A/F ratio guage is what I meant.
at least then you could see what the actual mixture is looking like.
friend had one in an rx7 he turboed, no idea who made it or what the proper name of it is.
a stoich meter perhaps?

I googled stoich meter and came up with this:
http://www.egauges.com/pdf/AutoMeter/837j.pdf
they even have a matching O2 kit with the weld in bung for the pipe.
I don't believe that the renix O2 sensor would be compatible with this guage.

 

[DaveKerwin]

what if you got one of those manual aw4 shifters ( http://awshifting.com/) and controlled when it shifted to keep rpm's low, that may be more practical, easier, and has added offroad benefit. just .02

 

[Bored400]

This is seriously heady stuff! I've read all of the responses to this thread and I've gotta say... if you guys arent in the biz, you should be. I wander into threads like this and tread DEEP water! LOL! Dummies like me will have to depend on superior minds to run the maze, find the cheese, and plot the course! LMAO!!!!
Godspeed, Gentlemen!

 

[Rick Anderson]

There are wide-band and switching O2 sensors. I think some of the Renix used some sort of variation of the wide-band, not sure, some of the Renix experts will probably explain. But the wide-band O2 sensor is very very expensive.

While the switching O2 sensor is very cheap. Why? Because the switching O2 sensor is very simple and wildly inaccurate. The only thing it is accurate for, is to consistently switch over at the same voltage right at the 14.7:1 AF ratio, in all conditions. Roughly 0.5 volts. Any other point for AF its an extremely steep curve that moves greater than the slope with temperature. i.e. 16:1 AF ratio when the sensor is at 600°F is the same voltage for 14.8:1 when the sensor is at 650°F.

The computer really can't tell what the actual AF ratio using a switching sensor. It can only tell when the AF ratio has swung thru the 14.7:1 AF ratio. So thats how the computer maintains the 14.7:1 AF ratio, by just bracketing it constantly, constantly moving back and forth between slightly rich and slightly lean. The computer shoots to hit an ideal target frequency that the O2 sensor switches back and forth, then it knows it has the best fuel settings.

So, you want to make this kind of system run lean? Not sure if it can easily be done. If you do something to skew the voltage drop across the O2 sensor, "MAYBE" it will continue to switch back and forth at the proper frequency just the switch will occur at a richer or leaner AF ratio OR MAYBE NOT. I wouldn't be surprised either way.

You don't necessarily need an OP AMP to change the reference voltage of a sensor. A simple resistor may do it, maybe even better. Remember, a Resistor just creates a voltage drop in a circuit proportional to the current. So, depending on where you put the resistor, its own voltage drop will create a different voltage drop in the sensor, either raise it or lower it. It may or may NOT work, you just have to try.

 

[Ecomike]

There are wide-band and switching O2 sensors. I think some of the Renix used some sort of variation of the wide-band, not sure, some of the Renix experts will probably explain. But the wide-band O2 sensor is very very expensive.

While the switching O2 sensor is very cheap. Why? Because the switching O2 sensor is very simple and wildly inaccurate. The only thing it is accurate for, is to consistently switch over at the same voltage right at the 14.7:1 AF ratio, in all conditions. Roughly 0.5 volts. Any other point for AF its an extremely steep curve that moves greater than the slope with temperature. i.e. 16:1 AF ratio when the sensor is at 600°F is the same voltage for 14.8:1 when the sensor is at 650°F.

The computer really can't tell what the actual AF ratio using a switching sensor. It can only tell when the AF ratio has swung thru the 14.7:1 AF ratio. So thats how the computer maintains the 14.7:1 AF ratio, by just bracketing it constantly, constantly moving back and forth between slightly rich and slightly lean. The computer shoots to hit an ideal target frequency that the O2 sensor switches back and forth, then it knows it has the best fuel settings.

So, you want to make this kind of system run lean? Not sure if it can easily be done. If you do something to skew the voltage drop across the O2 sensor, "MAYBE" it will continue to switch back and forth at the proper frequency just the switch will occur at a richer or leaner AF ratio OR MAYBE NOT. I wouldn't be surprised either way.

You don't necessarily need an OP AMP to change the reference voltage of a sensor. A simple resistor may do it, maybe even better. Remember, a Resistor just creates a voltage drop in a circuit proportional to the current. So, depending on where you put the resistor, its own voltage drop will create a different voltage drop in the sensor, either raise it or lower it. It may or may NOT work, you just have to try.

You have me thinking about how steep the slope is on the narrow band O2 sensors, which includes Renix, OBD-I and OBD-II. Tricking the computer may be real tricky with that steep slope to deal with. May be able to get away with it for a very small change then you may suddenly hit a point where it losses control and drops back to open loop mode, due to that step slope and the need for it reach 14.7 to switch back and forth. The Renix uses a resistive O2 sensor and runs between 0 and 5 volts, while the rest ( OBD-I and OBD-II ) run back & forth between 0 & 1 volt.

The existing sensors (Renix, OBD-I, & OBD-II) can all be gauged very easily with a high impedance $10 multi meter (voltage) wired from the dash to the O2 sensor output wire (just tap into the wire between the O2 sensor and the computer) and then just ground the meters ground wire!!!!!

 

[Ecomike]

Just throwing a couple of things out there. But IMO you have to keep the RPM band in mind (primarily) along with different loads and vacuum.
Some techniques may have more successes in different RPM bands. If any adjustment you make throws the sensor (say O2) out of the envelope, the computer will likely default and switch to closed loop (much like WOT). I can picture most any O2 sensor tweak causing trouble near peak torque, in the 2-2400 RPM range.
Some small tweaks, that add up may produced better results. The tweaks could even be seasonal. There are some things that may work better in the summer.
The engine temperature sensor would be easy to fool, a percentage resistor from the input wire to the sensor bundle ground splice. Same with the MAT sensor. You'd be fooling the program into running a warmer, leaner, curve. That the programming is likely to have answers too through the entire RPM band.
It may only be a few percent mileage improvement (if any). I've also had some luck tweaking my basic TPS setting a little.
A little here and a little there, it all adds up.

I was thinking along the same lines about the 2 temperature sensors, but then I realized that a working O2 sensor is probably compensating for any error produced by small systematic errors in the readings coming from the temperature sensors. I think the MAP + MAT + CTS sensor inputs are used to make an estimate which is then corrected by the O2 sensor if the O2 sensor is working. If this true, and I think it is, a biased temperature or MAP sensor should get corrected by the O2 sensor.

I know that others have used adjustable MAP sensors to run theirs richer. But did they still have working O2 sensors on those rigs?

 

[8Mud]

I was thinking along the same lines about the 2 temperature sensors, but then I realized that a working O2 sensor is probably compensating for any error produced by small systematic errors in the readings coming from the temperature sensors. I think the MAP + MAT + CTS sensor inputs are used to make an estimate which is then corrected by the O2 sensor if the O2 sensor is working. If this true, and I think it is, a biased temperature or MAP sensor should get corrected by the O2 sensor.

I know that others have used adjustable MAP sensors to run theirs richer. But did they still have working O2 sensors on those rigs?

I was thinking of pulse width. The hotter the motor gets, the narrower the pulse width. And/or my butt dyno tells me that the hotter the motor gets, it's likely that the fuel curve programming changes, it gets leaner at lower RPM's for sure and likely gets less lean at mid range and likely richens up again under low vacuum and/or higher RPM's. Tricking the MAT or the engine temp. sensor may not affect all RPM bands or vacuum conditions, but may help mileage in certain RPM bands. A small percentage gain in mileage could be possible, though it may be lost again by an inefficient burn.
The O2 sensor does compensate for mixture, using other sensor inputs, but it's doubtful the Renix computer has the technology to be infinitely variable. Much more likely it is a programmed curve, that is selected when certain criteria or sensor inputs are received. Messing with the O2 sensor seems likely to cause mid range bucking and/or act as an RPM limiter or even low vacuum detonation.
The trouble with an O2 sensor adjustment that is *static*, is that motor runs lean near peak torque anyway, a little richer at idle, low and high RPM's. When the O2 sensor fails it most often fails to a, default closed loop or at mid RPM range when it runs the motor lean anyway, it causes detonation.
The O2 senors second function is to make sure enough oxygen is present in the exhaust for a good catalytic reaction. It may even be the primary programing.
This is most all conjecture, your test equipement, could sort it out pretty quickly.
Many years ago I got a look at the Renix ECU programming, in a graph format (remember this is a primitive computer, it was old stuff before production) with a simple yes/no switching setup. I recognized it for what it was from some similar technology of the era, for other types of controllers.
IMO, massaging the sensors a little, makes much more sense than trying to outbrain the original programmers. A quicker more responsive O2 sensor could likely have some benefits, as could improved injectors. Fooling the computer into using the next higher heat program curve, may have some benefits, it may also make for a very rough idle.
OBD 1 and 2 series computers likely have a lot more possibilities for tweaking. Though much of what they do may also be yes/no preprogrammed responses to inputs, there are likely more options and/or finer delineations..

 

[Ecomike]

I recently tested my Renix O2 sensor while it was connected and while it was in park at various speeds. I used a high impedance old style FET analog meter.

I noticed the width of the oscilation and time on each side of the mid point (14.7 A/F ratio is about 2.5 volts on the Renix) between cross overs was at a maximum at idle (1 to 4 volts), and got very tight and narrow at about 2000 rpm (2.3 to 2.7 volts at one point). It seems the A/F ratio control gets tighter and better as the RPM increases on the Renix. This was no load condition, but it did not seem to care about deceleration or acceleration, only absolute RPM. Not quite sure why, but it was quite noticable.

 

[8Mud]

I recently tested my Renix O2 sensor while it was connected and while it was in park at various speeds. I used a high impedance old style FET analog meter.

I noticed the width of the oscilation and time on each side of the mid point (14.7 A/F ratio is about 2.5 volts on the Renix) between cross overs was at a maximum at idle (1 to 4 volts), and got very tight and narrow at about 2000 rpm (2.3 to 2.7 volts at one point). It seems the A/F ratio control gets tighter and better as the RPM increases on the Renix. This was no load condition, but it did not seem to care about deceleration or acceleration, only absolute RPM. Not quite sure why, but it was quite noticable.

About what I figured, if you add a vacuum curve to the same calculation (load) it will likely richen up again (some) at higher RPM's.
You could probably hook it up for a moving test or maybe a load test, wheels blocked and parking brake on.
Fooling the temp. sensors into thinking the engine and manifold are warmer than they are, may narrow the under 2000 RPM ratio a bit, without affecting much of anything else markedly. Likely not a large gas savings, but noticeable.
I'm not trying to sell the temp. sensors as *the* answer, but as a possibility that likely wouldn't affect overall performance to any extreme extent. Not something I'd recommend year round anyway, tweaking the temp. sensors would likely have some noticeable (bad) affects in cold weather..
Another possibility would be to fool the MAP at startup, when it takes the reference atmospheric reading.

 

[xjbubba]

I don't think the Renix ECM is as "dumb" as some would think. I beleive the Renix, as well as most other ECM's of the time, used similar algorithms and technology to achieve the same ends. I think the two primary differences between, let say an GM ECM, and the Renix were the type of O2 sensors and the ability to store and display trouble codes; however, programing wise, they work pretty much the same. The one thing that both computers assume, is that fuel pressure and volume meet specific specifications. That's why the first thing you do to trouble shoot a drive ability problem with any FI vehickle, is check fuel pressure. You run too much pressure and the system runs rich. To little fuel pressure, and the system will run lean. Remember, the ECM is using sensor inputs to control injector pulse width and engine spark advance. The O2 sensor merely fine tunes the air/fuel (A/F) mixture, during the times the computer is programed to use it. There are a number of ECM routines that require ignoring the O2 output, such as wide open throttle, or deceleration en-leanment.
If you try to run the A/F mixture too lean, you will probably run out side of the linearity curve of the common O2 sensor, meaning it would be difficult to predict what the voltage out put signified; you wouldn't know how lean or rich the engine was running.
It's probably possible to build an oscillator to simulate the O2 sensor, fooling the ECM into thinking the A/F mixture is perfect at all times. Now the only inputs controlling A/F and timing advance would be air temp sensors , MAP, and hopefully, the driver.
The driver could control fuel pump pressure, thereby biasing the A/F mixture to what ever level they wanted; however, they would need a wide band O2 sensor to tell them the condition of the current A/F mixture, and the means to manually change the fuel pump pressure from the driver's seat in real-time.

 

[bobinyelm]

There are wide-band and switching O2 sensors. I think some of the Renix used some sort of variation of the wide-band, not sure, some of the Renix experts will probably explain. But the wide-band O2 sensor is very very expensive.

While the switching O2 sensor is very cheap. Why? Because the switching O2 sensor is very simple and wildly inaccurate. The only thing it is accurate for, is to consistently switch over at the same voltage right at the 14.7:1 AF ratio, in all conditions. Roughly 0.5 volts. Any other point for AF its an extremely steep curve that moves greater than the slope with temperature. i.e. 16:1 AF ratio when the sensor is at 600°F is the same voltage for 14.8:1 when the sensor is at 650°F.

The computer really can't tell what the actual AF ratio using a switching sensor. It can only tell when the AF ratio has swung thru the 14.7:1 AF ratio. So thats how the computer maintains the 14.7:1 AF ratio, by just bracketing it constantly, constantly moving back and forth between slightly rich and slightly lean. The computer shoots to hit an ideal target frequency that the O2 sensor switches back and forth, then it knows it has the best fuel settings.

So, you want to make this kind of system run lean? Not sure if it can easily be done. If you do something to skew the voltage drop across the O2 sensor, "MAYBE" it will continue to switch back and forth at the proper frequency just the switch will occur at a richer or leaner AF ratio OR MAYBE NOT. I wouldn't be surprised either way.

You don't necessarily need an OP AMP to change the reference voltage of a sensor. A simple resistor may do it, maybe even better. Remember, a Resistor just creates a voltage drop in a circuit proportional to the current. So, depending on where you put the resistor, its own voltage drop will create a different voltage drop in the sensor, either raise it or lower it. It may or may NOT work, you just have to try.

What you say is TRUE, but reducing the O2 voltage output to the PCU (by adding a resistor in series) tricks the computer into thinking that the current mixture is TOO LEAN, so it will richen the mixture so the bracketing voltage returns to the reference 0.45 to 0.50 volts. That's not what we are after.

One would want to INCREASE the output of the O2 sensor as I see it, tricking the PCU to believe it is too RICH, so it will lean the mixture to return the bracketing voltage to the 0.45-0.50 volts that it THINKS is producing its 14.7 Stoichmetric A/F ratio.

As I see it, to INCREASE the O2 output (without actually richening the mixture), one would have to ADD voltage to it ( a summing OP AMP perhaps) or amplify it (a non-inverting OP AMP circuit).

Of course, one COULD wire up a test circuit with just a battery and a couple of resistors, and ADD it's voltage to the O2 output (i.e., wire it in SERIES) so the PCU sees the SUM of the battery voltage PLUS the actual O2 output.

Below is a hand drawn simple example of this idea. The 2.7k resistor is in series w/ the 1k potentiometer because it limits how much voltage can be added (to prevent lossible damage to the PCU). As it's shown, the circuit should be able to add 0-0.40 volts to the O2 output. The PCU should then try to lean the mixture enough to return its O2 input voltage to the reference of 0.45-0.50 volts.

If we are, say, adding 0.10 volts w/ our battery circuit, the PCU would lean the mixture until the actual output were 0.35-0.40 volts (0.45-0.50v - 0.10 v = 0.35-0.40v). In order for the O2 probe to put out 0.35-0.40v it would actually be seeing a leaner than 14.7 A/F mixture, which is what we are TRYING to do.

One could cut the PCU input lead anywhere between the O2 sensor connector and the PCU itself (we do NOT want to mess w/ the O2 sensor or its connector because it gets its reference from its own harness), and insert our "device" in series w/ that lead. We could mount our battery and resistors inside the car so we could adjust the mixture with a knob attached to the potentiometer.

Obviously, the battery circuit is a temporary training aid because we would have to replace the "AA" battery yearly (the current draw is negligible at 300 microamps), and it is not an elegant solution, but it SHOULD do the job well enough to test the theory I propose.

Bob
NOTE: Since this forum doesn't allow one to post images from one's computer, please see the circuit in Posting #34 at:
http://jeepsunlimited.com/forums/showthread.php?p=7571241&posted=1#post7571241