Monitoring Factory-Fitted Oxygen Sensors

Posted on September 16th, 2008 in AutoSpeed,Economy,Engine Management,Hybrid Power,testing,Turbocharging by Julian Edgar

This week we have the first in a two-part series, one that I am very pleased with.

The series is on how to use cheap and simple electronic kits to monitor the output of the oxygen sensor.

The first story I did on this, way back in the mid 1990s, resulted in the development of the ‘Mixture Meter’ kit – thousands have since been sold.

Now we both re-introduce the narrow band sensor display, updating the story to additionally discuss what many people want from such a display (and that’s improving fuel economy) and also, in Part 2, look at how a similarly cheap and easy-to-build display can be used with wideband sensors.

The latter is especially significant: while there are plenty of aftermarket air/fuel ratio meters that use wideband sensors, we’ve never seen a description of how to tap into the standard wideband sensor fitted to many of today’s cars.

And doing this is especially worthwhile – again, not so much for performance outcomes, but to improve real-world fuel economy. That’s primarily because, as its name suggests, a wide band sensor works over a much wider range of air/fuel ratios than a narrow band sensor.

So, if your car runs a lean cruise of 16 or 17:1 (or 24:1!), you can see on the wideband sensor display when it is actually in this mode. (And of course if you can see the conditions in which the car does this, you can replicate those driving conditions more frequently and with greater ease.)

Both the narrow band and wideband monitors are ‘relative’ displays: they don’t show air/fuel ratios in numbers but instead show how rich (or lean) the air/fuel ratio is compared with stoichiometric (ie 14.7:1 for normal petrol).

We could have developed a numerical display that would work with factory wideband oxy sensors, but we think that to guarantee accuracy on a given car and with a given factory-fitted sensor, each meter would have to be individually calibrated with a professional level instrument.

On the other hand, the ‘relative’ wideband sensor display is immediately useful, especially if it’s fitted before any vehicle modifications are undertaken. That way, the benchmark for air/fuel ratios in different driving conditions becomes the standard vehicle.

I have been monitoring the wideband oxy sensor fitted to my Honda Insight. I can see at a glance if it is running rich, stoichiometric or its ultra-lean cruise mixtures.  By driving in a style that causes the car to run lean cruise more often, and avoiding rich mixtures (something sometimes as simple as a gearbox down-change when climbing a hill), I can easily improve fuel economy.

And of course if the car develops a problem with fuel pump, or fuel pressure regulator, or MAP sensor – I’ll also be able to see that a problem is present.

Pretty good for a very simple kit that costs just twenty bucks…

10 Responses to 'Monitoring Factory-Fitted Oxygen Sensors'

Subscribe to comments with RSS

  1. milesinfront said,

    on September 18th, 2008 at 11:26 am

    I’d be interested to know about what is required from an engine to be able to run ultra-lean afr’s. I have some experience with a 2.3l 6cyl (M20) BMW running on a programmable ECU with WB02 input. I’ve found that once AFR’s drop below 16.5:1 the engine begins to miss-fire. I am aware that lean mixtures burn slower, so more ignition advance may be required. I’ve also read that lean mixtures are harder to ignite, so perhaps some ignition upgrades might make an improvement. (even multi-strike CDI ignition may be the answer.) Any knowledge you, or other readers, may have would be appreciated, and hopefully a good read!

  2. Franki said,

    on October 13th, 2008 at 1:03 am

    A good idea for an extension to this would be how to make your own wideband AFR meter using parts from existing vehicles (IE how to get the oxy sensor for an Insite to provide useful info in the exhaust of a EF/EL Falcon) for cars with narrowband sensors..

    It isn’t hard to drill a hole in an exhaust and weld a nut to it, so it must be doable to mount a wideband sensor from some other vehicle and use it to get AFR state change readings…



  3. Julian Edgar said,

    on October 13th, 2008 at 6:47 am

    Franki, if you look at the article on wideband sensors you will see that they are not ‘dumb’ sensors like narrow band sensors. That is, wide band sensors need a whole control system to work. So the ‘parts from existing vehicles’ would include the whole ECU! Easier and probably simpler to buy an aftermarket wideband air/fuel ratio meter.

  4. AJ said,

    on October 15th, 2008 at 11:16 pm

    Julian, I disagree. I have an aftermarket wideband sensor, they output 0-5v. Breaking them out to a lambda or a/f given the type of fuel burned is relatively easy. It’s the accuracy and speed of the conversion that is somewhat hard.

    Despite already having an aftermarket meter an article on the subject would be interesting, imho.

  5. Julian Edgar said,

    on October 16th, 2008 at 6:50 am

    AJ, are you saying that when the bare sensor is connected only to 12V and ground, and put in the exhaust, it outputs a signal of 0-5V that corresponds to AFR? I don’t believe that is the case – it needs pump cell control logic to work. The pump cell control is built into the meter, not the sensor.

  6. AJ said,

    on October 16th, 2008 at 7:08 am

    Julian, I wrote up some great stuff and then lost it with a click. Anyhow, i’m pretty sure that all wideband sensors deliver 0-5v. If i’m not mistaken they have a set coupler too. There’s a pin-out guide here:

    The trouble i see is that your daughter circuit could change the resistance causing your sensor readings to be off and as a result run lean or rich as well as having inaccurate gauge readings.

    I forgot this on my first post- greetings from the US and thanks for the great information, keep doing what you’re doing.

  7. Julian Edgar said,

    on October 16th, 2008 at 9:15 am

    1) Have a look at the circuits in the link you posted. The sensor certainly won’t work on its own.

    2) The suggested measuring technique covered in AutoSpeed uses a module with a very high input impedance so no change in the original signal will occur.

  8. AJ said,

    on October 16th, 2008 at 9:35 am

    1) oh of course not! I was talking about a daughter board that would read it. I think we’ve miscommunicated- i’m talking about a small controller that will break out a/f from a wb O2 sensor cannibalized from a car with a factory wideband. Or, accessing the wideband readings from within the car with the factory wideband sensor.

    My afx and the bosch sensor output 0-5v, 0v is 9:1 and 5v is 16:1, the math is (1.4*output voltage)+9. If your controller can calculate that it can break out wideband. It would probably be more simple than that if your output was a series of LED’s like the narrowband project.

    2) it’s been some time since i’ve calculated impedance.

  9. Julian Edgar said,

    on October 16th, 2008 at 9:42 am

    You said: “Julian, I disagree. I have an aftermarket wideband sensor, they output 0-5v.”

    They do not. The CONTROLLER outputs a 0-5V signal. Without the controller, the sensor won’t work.

    We do not intend at the current time to develop a wideband sensor controller.

  10. AJ said,

    on October 16th, 2008 at 12:49 pm

    “Without the controller, the sensor won’t work.”
    Yes, I know. I was saying that the controller need not be sophisticated.

    “We do not intend at the current time to develop a wideband sensor controller.”
    Okay, sorry to bug you.