O2 sensor questions.

Start by assuming I know absolutely nothing about computer controlled cars. That won't be far from the truth. I know just enough to be dangerous. I can trouble shoot and repair, but changing the system is beyond me. It's time to learn.

I'm installing a Innovate MTX-L AFR meter in the Samurai, in preparation for adding boost. It's in and wired and appears to be reading correctly.

I wanted to use one of the analog outputs to feed AFR information to the ECU so I don't have to run two O2 sensors. I assume it is a narrow band O2 sensor. It has 3 wires. Two for the heater and one for the output.

There are two outputs. Output One is 0V = 7.35 AFR and 5V 22.39 AFR. Output Two is 1.1V = 14 AFR and .1V = 15 AFR. Which do I need to use. I haven't been able to find anything online that tells me what the ECU is looking for. I guessed and used output 2. It idles at about 2500 rpms and is lean. Rather than trial and error and burning something up, I turn to the smarter guys in the room.

Thanks.

The ecu is looking for a 'narrowband' 0-1v signal so you would use the low volt output to feed into the stock o2 sensor signal wire.

In reply to Vigo:

That's what I thought so I wired it into output #2. Why would the idle be in the 2500 rpm range and the AFR bouncing between 14.5 and 16. The only thing changed was adding the AFR meter and using the output for the ECU input. My VOM shows the output shifting between .1V and 1V. What did I do wrong?

First off, put the original sensor back in and make sure you didn't berkeley up something else.

It should run on the second, but I've learned that not all narrow band o2 sensors are the same. I'd also check and make sure you are inputting the signal to the right wire. I presume you used the heater power and ground for the wideband? I don't think it will matter whether you have them reversed. How did you ground the signal portion of the new one, since you don't have a factory ground wire? The 3 wire would ground through the exhaust pipe...or does the heater ground there, and you are feeding spare amperage to some sort of ground monitor?

The MTX-L uses a sensor that plugs into the AFR meter. It provides heat, power and reads the sensor output.

Off of the AFR meter is a analog output to be fed to the ECU. It is supposed to simulate a narrow band sensor. It is one wire. I spliced it into the plug off the old sensor and plugged that into the harness.

When it didn't work, I wondered about the sensor ground. I grounded the sensor ground off the ECU back to the ground for the AFR meter and the exhaust manifold. There was no change.

I've unplugged the ECU sensor input. No change.

I haven't stuck the old sensor back in. I will do that tomorrow.

Obviously the ECU isn't happy with something. Could the AFR output be out of range for the ECU?

Worst case scenario, I'll weld another bung into the exhaust and run two sensors.

High idle means air is getting in somewhere, usually. Adding or removing fuel doesn't usually affect idle speed that badly.

"Output Two is 1.1V = 14 AFR and .1V = 15 AFR"

That doesn't make sense, I'm guessing it's a typo. A narrowband sensor basically just reads "leaner than stoich" or "richer than stoich", with the changeover being right around 0.45v. As noted, make sure you're feeding that into the correct signal wire.

This is a copy and past off the instruction manual.

"Optionally, the YELLOW (Analog out 1) and/or BROWN (Analog out 2) can be connected to the analog inputs of other devices such as data loggers or ECUs. If either one or both of these wires are not being used isolate and tape the wire(s) out of the way. The default analog outputs are as follows: Analog output one (yellow) is 0V = 7.35 AFR and 5V = 22.39 AFR. Analog output two (brown) is 1.1V = 14 AFR and .1V = 15 AFR."

According to the wiring diagram, the O2 sensor is three wires. Black, black and yellow/blue. The blacks are heater, the yellow/blue goes to the ECU. I wired output #2 from the AFR meter to the yellow/blue for the ECU.

When I do this, I wire the narrowband simulation output as close to the ecu as I can get. Grounds are very important.

You also may need to jack into the MTX-L and make sure that narrowband simnis active/programmed.

Toyman01 wrote: In reply to Vigo: That's what I thought so I wired it into output #2. Why would the idle be in the 2500 rpm range and the AFR bouncing between 14.5 and 16. The only thing changed was adding the AFR meter and using the output for the ECU input. My VOM shows the output shifting between .1V and 1V. What did I do wrong?

Based on your first note, the ECU is actually working as it's intended to run. All narrow band O2 sensors (which I will now call HEGO) operate between 0 and 1 V, and also like pointed out, stoich for whatever fuel you are using will be roughly 0.46v. Within all of that, they genreally will swtich on the lean side between 0.2 and 0.1v, and switch on the rich side between 0.8 and 0.9v. The fact the simulated output is what you posted, the engine ECU is working properly.

The real problem is that the simulated output really doesn't simulate anything.

(FWIW, all HEGOs are the same. they look different, and the theoretical switch point changes by 0.01v, but how one uses them is exactly the same. realistically, WB (UEGOS) are the same thing, except they add a pumping circuit, so that the reference is better. their only goal is to be a battery that uses the difference in O2 to generate a voltage. very simple device)

In reply to alfadriver:

"Analog output two (brown) is 1.1V = 14 AFR and .1V = 15 AFR."

Does that mean I need to reprogram the output. 14:1 and 15:1 are both lean. I would think 1.1V should be a rich condition.

Some of the posts above assume the simulated narrow band output is setup like a standard narrow band, which may not be the case (but why is another question). The AFR values listed from the instruction manual do not seem correct for what the ECU is expecting. You may want to call and get verification that that is correct or not. If the stated output is correct, then you will need to run your second O2 or find a way to adjust the narrow band simulated output.

Toyman01 wrote: In reply to alfadriver: "Analog output two (brown) is 1.1V = 14 AFR and .1V = 15 AFR." Does that mean I need to reprogram the output. 14:1 and 15:1 are both lean. I would think 1.1V should be a rich condition.

First thing is that you don't want to program it to mean anything above 1v. That doesn't happen.

The next thing is to know what 14:1 and 15:1 really means- in terms of the relative number that an O2 sensor really reads. Stoich for 100% gasoline is about 14.6:1, E10 is closer to 14.3:1, and I think E85 is as low at 9.X:1.

So what fuel are you running?

If it's gas- 14:1 is rich, 15:1 is lean, and it's pretty even.

This is roughly what you are trying simulate. And note the bottom line is Lambda, not a/f. Where lambda is (a/f)/(stoich a/f). But that's the more realistic output of the sensors. We then convert that to a/f.

The closed loop controllers that use a HEGO run it tight around stoich- whatever it is for the fuel.

Still, the ECU is correctly reacting to what you are feeding it.

Ok doing some quick calculations, the using the output you posted (and I read that 0.1v was 16:1, not 15, so my mistake).

That roughly means that 14.64 is 0.46V, which is roughly where computer expects stoich to be.

What's interesting is that you are measuring an output of 0.1 to 1v, which SHOULD be 14.1:1 to 15:1, but it's reading 14.5:1 up to 16:1. So the expected output isn't matching what it's telling you it should be. Could be a floating base- so make sure the innovate and the ECU are properly grounded. Still- if you read directly off the Innovate of 0.1 to 1v, and it's ALSO reading 14.5 to 16:1- it's not generating equal data.

In reply to alfadriver:

Got it. I'm burning gasoline. For some reason I had it in my head that 12-13 was where it needed to be. So, it might not be a lean condition causing the high idle. I wonder if a dirty signal might be confusing things. Does the input to the ECU from the O2 sensor need to be shielded? How sensitive are OEM ECUs to noise?

Toyman01 wrote: In reply to alfadriver: Got it. I'm burning gasoline. For some reason I had it in my head that 12-13 was where it needed to be. So, it might not be a lean condition causing the high idle. I wonder if a dirty signal might be confusing things. Does the input to the ECU from the O2 sensor need to be shielded? How sensitive are OEM ECUs to noise?

It's a very strong signal. No need to shield it. If the sensor isn't hot, it will just deliver less than 0.1v.

Ground noise, on the other hand, can be a problem. Which is why sensors went to 4 wires shortly after 3. Make sure the innovate box is grounded somewhere on the chassis. You don't want a floating ground, which can cause an offset in the voltage- which will be a problem.

but do make sure that the output you are getting matches with the a/f generated by the box.

I've only used the 4 wire variety. Is the ground path using the exhaust on a 3 wire system or is the heater return overloaded for both tasks? If it is - make sure the engine to chassis grounds are good and/or wire a copper braid to the exhaust somewhere. It wouldn't be too hard to introduce a 1v drop in that scenario.

In reply to Giant Purple Snorklewacker:

The signal ground for a 3 wire is just like with the single wire- through the exhaust-engine-chassis ground set up.

Does it run fine again if you put the stock o2 back in?

In reply to Swank Force One:

That will be this afternoon's project.

Interested to see what you come up with. I have an old JAW just running an AFR display, but have been meaning to wire the analog output in to the ECU to replace the stock narrow band. I already welded in a second bung in my downpipe for the wideband, but will lose it when I switch to headers.

Toyman01 wrote: In reply to Swank Force One: That will be this afternoon's project.

Cool i'll wait to see how this goes. I've used the LC-1/MTX-L to do narrowband sim on quite a few cars and never had any sort of problem like this.

You could always just megasquirt it now.

Kenny_McCormic wrote: You could always just megasquirt it now.

That sure seems like a good idea.

Until you start seeing how much development went into that OEM's computer. It's able to start from -20 to +120. It had actual spark data that relates to acutal best spark as well as when it starts to knock. It's has development on it so that when things get hot for whatever reason, it puts in the right amount of fuel to cool things back down and not break anything.

It's capable of running a rather wide range of fuel- we are not taling gasoline to E20, but fuel blends that don't like to evaporate- causing hesitation that is a pain do deal with... Let alone the fact that it actually starts and runs on the fringe fuels.

It has transient development on it- so that depending on how hard you tip in, it responds with the right amount of fuel- not too rich, not too lean.

Real dyno data was developed to determine what the actual best fuel is for power.

It's got countless hours on it to minimize the emissions without you even know it does that.

It can run below sea level and up to the highest roads you can think of.

It's auto trans can up shift and downshift and lock the clutch- all in reaction to what the driver wants, when they want it, the ambient conditions, and what the engine can deliver.

Basically, there 24 months of development from 8 engineers, 2-3 dynos running all the time, and a series of 20 or so prototypes (about 3 series of 8, but some are missing).

Granted, a lot of that is to make sure that the calibration will work for millions of cars. But there are many hours just getting the engine to run correctly.

Just sayn.

In reply to alfadriver:

And yes, one can get pretty close using an aftermarket comptuer. But it's not "just install X" easy. There is a LOT of development work needed to make it run well. Especially really well.