I've been tuning my 2015 Subaru WRX to run in lean burn and I finally got around to doing some proper A-B-A testing. This car has a direct-injection turbo engine that is pretty knock-limited at moderate loads, so I wasn't sure how much of an effect a lean burn tune would have, but I'm pleased with the results.
In order to tune it, I used engine management software (Accesstuner Race), and re-calibrated the stock wideband oxygen sensor to trick the ECU into running lean. I'm currently running 17.5:1 AFRs at lower loads, which transitions to 16.3:1 at higher loads, and switches to normal AFRs at loads above 1.0g/rev of airflow (~40% of maximum load). The engine runs in full-time closed loop. Ignition timing is advanced up to 7.5 degrees over stock. Lean burn is active from 1200-2800rpm, and activates as soon as the oxygen sensor comes up to temperature, unless coolant temperatures are under -10C. I go into a lot more detail in
This Post on NASIOC. I've also made a lot of changes that improve economy during engine warmup at the expense of emissions, which are detailed in that post (lower idle, disabled variable valve timing warmup settings, disabled stratified fueling idle mode, disabled post-start ignition timing tables).
I did some A-B-A testing to compare my tune with the stock map. Testing was done on a flat highway with the cruise set at 100km/hr in 6th gear with a warm car. There was a pretty steady cross wind that was more of a tail wind in the A-B direction, and more of a headwind in the B-A direction. Fuel consumption was determined by datalogging A/F ratio and mass airflow and then calculating the fuel flow in kg/hr. I also datalogged coolant, oil, and intake air temperatures to make sure they were consistent across all tests.
Average Fuel Consumption Results:
Stock Map:
A-B: 5.075 kg/hr
B-A: 5.587 kg/hr
Lean Burn: (17.0-17.1 AFRs)
A-B: 4.644 kg/hr
B-A: 5.274 kg/hr
Difference A-B: 8.5% less fuel with lean burn
Difference B-A: 5.6% less fuel with lean burn
The results were probably better on the A-B trip due to the lower engine load. (Edit: On the A-B trip, ignition timing was advanced an average of 6.1 degrees over stock. On the B-A trip, ignition timing was only advanced 3.2 degrees. At lower loads I can advance timing up to 7.5 degrees over stock). With a knock-limited engine like this, there is less to be gained by running lean at higher loads. I may do more testing at 40-80km/hr to determine if the results are even better at low engine loads.