Are AFRs accurate when under no load?

I've read somewhere (some really old forum post) that AFRs are not accurate and are to be ignored if under no load. Is that correct? One example would be free-revving. This is of interest to me because I've been concerned as I've noticed when free revving I go super lean, sometimes 17:1, sometimes 18:1. I have done a bunch of tests and datalogged just driving around for 30 minutes while my ecu averages sample AFRs recorded along the way for each cell. I go and filter out any samples where TPS < 5% and AFR is less than 11 or greater than 19 to eliminate tip in and fuel cuts/decel and those same cells that go super lean free-revving aren't so bad, worst one is 5% over stoich.

Another example would be idle, I'm pretty sure they are valid there.

The part of the map for this discussion is the 30kpa column in my case, either for idle or free-revving.

When free-revving it's pretty easy to hold the rpm in any of the relevant cells and the MAP is stable as well so I can get it right in the center of my table cells. I had been looking at the 30kpa as basically the free-rev and transient column since otherwise it's very hard to stay in it under any sort of load (other than the low rpm idle cells).

Currently if I hold a free-rev in the 5,000 rpm 30kpa cell I'll get something like 17:1 - 18:1 for an AFR.

I had been considering adding some fuel in these areas but felt I was missing something. I think I need a better understanding of this area of the fuel map and the relationship between load and AFRs?

I don't have any concearns over AFRs in other areas of the map, just this one has me scratching my head.

I've asked a similar question before but I think I may be getting to the bottom of understanding it now (I hope).

Uh, I think you're getting hung up on thinking the AFR (lambda) 'has' to be a little rich everywhere. This isn't true for several reasons, such as.

Under light throttle there isn't the heat in the combustion chamber to keep spark plugs clean with over-rich mixtures.

Following on from that, the engine will run cleaner as all the fuel is burned and so there will also be less oil dilution and potential loss of oil film on the bore Wall, which should make a small improvement in wear resistance.

For fuel economy you want to run around 'best lean power' or 'best lean torque' - this is where the most energy is recovered from the fuel burn to propel the vehicle. With most modern engines this is around the 16-17:1 range and can make a BIG difference in mileage.

Depending on the vehicle's use, I would suggest using the 'lean best' for low load and/or rpm and progressing to 'best torque' closer to full load and/or rpm. Your wallet and spark plugs will thank you.

Point well taken! However still curious for the sake of knowing whether AFRs are accurate when not under load aside from idle.

It completely depends on engine configuration, and/or fitted smog components. For example high overlap cam/port profiles or air injection pre sensor such as a rotaries port air functionality from the air pump and ACV will both cause whacky readings while under light load and/or idle but not present under high RPM conditions (depends on the function of smog gear for one, port design on the other.)

Interesting, and noted. In my case I'm on a standard 4 cyl. My only emissions equipment are evap purge and cat, no EGR or air injection or anything too fancy.

The way we measure AFR by wideband O2 sensor can only give inaccurate readings is under the load since 02 sensor reading can be affected by back pressure... The higher the pressure the bigger the error can be.

So therefore under no load, the wideband AFR is more accurate since back-pressure will be less?

Maybe, maybe not. You can compensate for said pressures. I think you're overthinking things, a wideband is a pretty damned accurate tool but you'll need to listen/feel what your engine wants as well, its not the be all end all. Theres a reason why I still do things such as inspecting exhaust ports and reading plugs.