NB does NOT give you AFR....

AFR isn't constant. Narrow band sensors are very narrow band in the electronic sense of the term. They tell you if you're rich or if you're lean. Think a ball balanced on a knife, it'll fall one way or the other; very hard to keep it in the middle. The only way to use them is to vary the intended AFR by a percent and watch for it to swing high->low or low->high, then adjust as appropriate (Short term/long term fuel trims relate to this as to what adjustments over/under base are required to cause the swing). As Fred mentioned, it also makes Californians happy as modern catalytic converters are more efficient under an oscillating oxygen rich/learn environment, which was likely the primary push for narrowband use in this manner.

Fine, Fred, let me rephrase that; if the AFR is CONSTANT at 14.7:1... why would the [sensor] OUTPUT also not be constant??

OEM narrowband output oscillates because the fueling is oscillating.

14.7:1 is NOT an "AFR"?

Mike carefully explained this to you above. Read his detailed explanation again. He's obviously got too much time on his hands, good for you.

I would say if you had some very high accuracy combustion apparatus that could give you exhaust gas with a constant "lambda" held to with within about 3 decimal places and that lambda that you want to measure was somewhere between about 0.900 and 0.950 or 1.050 and 1.100 and the sensor temperature was kept constant to within about 20°C, then yes, your narrowband sensor would probably give a constant output voltage with reasonable accuracy.
If any one of those variables are outside of the examples given above then you can only use a narrowband sensor as a "switch" - i.e on/off, rich/lean, 0/1...

Redlines show where the curve slope is possibly useable:
[/url]

**Later edited picture**

That's an optimistic graph that relies on the sensor being up to temperature, and varies, a lot, with temperature. Probably varies a lot between one an the next too. Also over a narrow Voltage range, so errors are large, etc.

He was specifying 14.7, at which point, no, no way, can you use it to hold stable. You could hold a stableish value away from that, but you couldn't say what value it was that you were using. The only way would be to reach steady state, then find stoich, and use math to find other AFRs assuming your calibrations were all perfect, or close to it.

Isn't that just saying what I said but in a different way?

So a NB can be used as a "semi-wideband"??. Could you use a NB to read from 11.78 to 14.7 ??.

Multiple knowledgeable people just explained that no, you cannot.
You're either very good at ignoring any information that doesnt suit you, or you're doing some expert level trolling.

If you're convinced that a narrow band can be used as a wideband, go out there, prove your point, amaze the world
with your engineering prowess. All those evil companies selling us widebands all this time while we could have settled
for marginally cheaper narrowband sensors, it's time someone put a stop to them.
You go work out the details, I'll go put on my tinfoil hat and chant some mantras in support.