Lean bias narrowband

[serialk11r]

I just realized there's a very simple and cheap solution to bias a narrowband sensor lean by 1.5% or so. This is not enough to meaningfully change best timing, EGT, or low load efficiency but it'll make a small dent in WOT operation and add ~1% peak power to most cars without the expense and complexity of fitting wideband controllers. On cars with a wideband, you only need to bias the rear sensor which provides correction to the wideband.

A voltage divider between a 1.2V reference and the original output, with a 7:1 resistance ratio will scale the voltage from 0.15 to 0.28, 0.35 to 0.456, 0.8 to 0.85, 1.0 to 1.03 etc. The reference should be close to 1.1 so that you don't get a fault code.



Since car alternator output is quite noisy, to get a reference voltage you can first find some cheap 3.3 or 5V step down converter then toss on a capacitor for smoothing and use a voltage divider. This reference voltage can be shared across all O2 sensors if you have multiple sensors.

The actual resistor values need some experimentation to play nice with the sensor. It may be the case that the voltage ratio has to be more like 10:1 so that the sensor still read lean without having to go way way lean.

[Piotrsko]

Might be wrong here, but I was understanding that it was a frequency and not a voltage since the data input was digital on or off only. The resistor would cause the leading edge to peak sooner and decrease off frequency and if it was actually a sine wave, tend to increase the waveform towards clamping which would be the same result.

If the off frquency goes below a target value it means one state probably rich, if it goes above it's lean. Easey peasey to do digitally.

[serialk11r]

Originally Posted by Piotrsko

Might be wrong here, but I was understanding that it was a frequency and not a voltage since the data input was digital on or off only. The resistor would cause the leading edge to peak sooner and decrease off frequency and if it was actually a sine wave, tend to increase the waveform towards clamping which would be the same result.

If the off frquency goes below a target value it means one state probably rich, if it goes above it's lean. Easey peasey to do digitally.

I think you might be referring to the catalyst monitor function. Typically wideband cars also use the rear narrowband to calibrate the wideband (since the wideband has no fixed stoich reference, as the measured current at stoich can drift over time). I had this experience attempting to force my Scion FR-S into lean burn, where the car started recalibrating the wideband O2 on the freeway.

Bumping the voltage up will move the switching point just a little bit to the lean side. On some cars that run VERY rich, you could probably increase heavy acceleration efficiency by over 2.5% doing this, which is pretty significant.

[mpgmike]

Have you tried it yet?

[rmay635703]

That’s a golden oldie

There are dual efies that use a similar design but it’s fully adjustable to set how much lean you want.

[serialk11r]

Originally Posted by mpgmike

Have you tried it yet?

No, I haven't been able to buy a car to try it on yet. I think the market is supposed to cool a little because I see dealers dropping their prices a bit, but the good deals get snatched up very fast. I have some time during the summer to find a spare O2 sensor and do a little soldering.

I don't think I would do this modification to my V8 Vantage because I already have the ability to change the sensor setpoint in the ECU and have already leaned out the AFRs as much as possible. The typical load is so low that going a few % lean will really not help much.

Originally Posted by rmay635703

There are dual efies that use a similar design but it’s fully adjustable to set how much lean you want.

Yes it is actually better to just change the switching voltage at the ECU, but that would involve flashing the ECU which is a big upfront cost and added hassle in California.

Of course, this modification will reduce O2 sensor signal amplitude by a bit so in expected value terms you will have to replace the sensor a bit sooner.

A wideband sensor would be better but if you have a wideband, you might as well go much leaner and tune the car for it.

[mpgmike]

The reason I asked is NB O2s are extremely high impedance, it doesn't take much to short them out. An analog DVOM will read 0 volts because the meter is too low impedance.

WB O2s have a NB in the sample chamber, but use Ytrium Oxide to pump ambient oxygen in or out to maintain a fixed voltage. That trick wouldn't work at all on a WB, and I'm not sure how well it'd work on a NB.

[serialk11r]

Originally Posted by mpgmike

The reason I asked is NB O2s are extremely high impedance, it doesn't take much to short them out. An analog DVOM will read 0 volts because the meter is too low impedance.

WB O2s have a NB in the sample chamber, but use Ytrium Oxide to pump ambient oxygen in or out to maintain a fixed voltage. That trick wouldn't work at all on a WB, and I'm not sure how well it'd work on a NB.

Looking on the internet I see that a hot O2 sensor is supposedly about 200ohms impedance (no idea if that's right). Presumably the ECU measures the voltage against ground and draws some miniscule current, and since the reference voltage source used is higher voltage than the sensor voltage, current should be flowing from there to the ECU, so the top resistor on the voltage divider just needs to be a suitable value right? (switching frequency is constant and fairly low, I think)

To replace with a WB, you replace it with a WB controller and use the narrowband emulation output. They all support modifying the output through software.

[mpgmike]

A narrow band O2 sensor uses 2 dissimilar metals that chemically react with each other, when above 660 degrees F, and in the lack of oxygen. They are capable of outputting well over 2 volts (personally observed). As the sample chamber is diluted with oxygen, the output voltage decreases. In other words, they work similarly to a battery.

Starting rich, the voltage is around 1 volt. As the AFR goes lean, the output doesn't change that much. Then around the target AFR (14.7:1) the voltage begins to drop rapidly, and then tapers off. A new sensor may have enough output current to be able to use something like posted. As it ages, the output current decreases, even if the voltage remains reliable (again think in terms of a variable battery).

[serialk11r]

I decided I needed to double check the O2 sensor voltage calculations that use the Nernst equation instead of trusting some internet graphic, because I remember thinking this wouldn't work in the past...with complete combustion, the sensor would only be at 0.1v if you're just 1% lean.

So the issue is incomplete combustion and how much O2 is left in the exhaust stream. In this paper, it cites this graph from a 1987 paper that I can't find on Google, which seems to suggest 0.5% at stoich!: https://www.researchgate.net/figure/..._fig3_34286046


Of course, exhaust is hot so some of that is probably consumed by the time it hits the O2 sensor.

So narrowband closed loop fueling doesn't target stoich, it targets the point where the engine is running rich enough to consume that excess oxygen by the time the exhaust reaches the sensor. So all narrowband controlled engines run just slightly rich, but changing the switching voltage doesn't represent a meaningful change in oxygen concentration until you drop it past the point where the sensor is useful.