I came across a discussion in a small circle called “AFR 16,” where “冬夏” mentioned an idea about adjusting the air-fuel ratio (AFR) to achieve lower fuel consumption. I decided to verify this idea.
Disclaimer: This article is a collection of notes after studying AF (Air-Fuel) and related knowledge. As someone not involved in automotive engineering, some of the content may not be entirely accurate and is for reference only.
I always believe that in the chaotic world of vehicle modifications in China, it makes more sense for both the individual and the vehicle to grow together than blindly spending money on trendy modifications. So, I am recording all the modifications, their effects, and some insights here.
For gasoline engines, different AFRs determine different combustion characteristics, as can be seen from the curve below:
However, in general, we often hear about the “golden” AFR of 14.7:1, but as shown in the graph, 14.7 is not in the “Best fuel economy” range.
From the article “Experimental Study on Lean-Burn NOx Aftertreatment System for Gasoline Engines” (link in the appendix), we can find the following passage:
We can make two preliminary guesses regarding why manufacturers set the AFR to 14.7 instead of 15.4 or higher:
- To improve power performance.
- To improve emission performance.
We know that the quantity of nitrogen oxides (NOx) is one of the indicators of emission standards. With the full implementation of China 6B emissions standards in 2023, the NOx limit has been reduced from 60 mg/km to 35 mg/km compared to China 6A.
Let’s look at another graph showing the performance of different AFRs on emissions:
I couldn’t find the specific paper where this graph was used.
It can be seen that nitrogen oxides (green line) reach their peak between 14 and 16. This may confirm the earlier guess of “To improve emission performance,” especially considering that modern vehicle ECUs are the result of compromises between fuel consumption, emissions, and calibration for different markets within the same vehicle category.
So, based on the theoretical part above, we can arrive at a preliminary conclusion: if you don’t care about environmental performance (or you’re not an automobile manufacturer), and you’re aiming for better fuel economy, you should adjust the air-fuel ratio (AFR) (more accurately, the closed-loop AFR) to be close to 15.4.
To verify our theory, we can conduct experiments. First, let’s introduce the test vehicle:
- 10th generation two-door manual Honda Civic
- Fuel: 98 octane(RON) gasoline + HKS 0W-20 engine oil
- Tires: 235/40/R18 AD09 tires with a tire temperature of 25 degrees Celsius, front tire pressure of 2.2, and rear tire pressure of 2.3
- Test road: A section of the Shanghai Middle Ring Elevated Road
- Weather during the test: Intake temperature of 30 degrees Celsius, light rain
Since we have Hondata, we can easily modify the closed-loop AFR values of the vehicle. The original settings are as follows:
You can see that the factory calibration is at 14.7.
We will modify these values to 15.4, 15.5, and 16 in the ECU for testing:
For safety reasons, the “Closed-loop target lambda high load” value was not adjusted because running a lean AFR under high load conditions (such as suddenly applying throttle during stable cruising in a high gear at low RPMs, where load is high but AFR is lean) carries a higher risk of detonation.
The testing method involves using adaptive cruise control (ACC) to maintain a constant speed of 70 km/h and recording fuel consumption data after driving for more than 10 kilometers to obtain relatively accurate fuel consumption figures.
Fuel consumption: 4.8 L/100 km
Fuel consumption: 3.6 L/100 km
Fuel consumption: 3.9 L/100 km
Fuel consumption: 4.1 L/100 km
From the results, which may contain some errors, we can see that increasing the AFR from 14.7 to 15.5 reduces fuel consumption from 4.8 L/100 km to 3.9 L/100 km, aligning with the “Experimental Study on Lean-Burn NOx Aftertreatment System for Gasoline Engines,” which indicates a fuel consumption improvement of 3.1% to 10.1%.
|AFR||Fuel Consumption (L/100 km)|
Some Additional Points
What Does This Fuel Consumption Data Imply?
In my car, using 98 octane gasoline, and assuming a gasoline price of 9.93 CNY/L in Shanghai at the time of writing, driving at a constant speed on the highway allows for savings of (4.8 - 3.9) * 9.93 = 8.937 CNY per 100 kilometers.
Why don’t manufacturers do this? What other problems might arise from doing so?
Here are my personal thoughts on this:
- It would lead to higher nitrogen oxide (NOx) emissions.
- This emission is likely to prevent vehicles from meeting the emissions limits of National Standard 6B (leading to vehicles not being able to be sold), so it’s a compromise as a result.
- There would be higher exhaust temperatures.
In Hondata’s documentation at https://www.hondata.com/help/flashpro/index.html?closed_loop_parameters.htm, it is described as follows: Warning: Running leaner than stoichiometric (lambda 1, approx 14.6:1) will increase exhaust gas temperatures. For this reason, it is not recommended to change these parameters for vehicles with catalysts.
So, for cars without direct header sections, a leaner air-fuel mixture would result in higher exhaust temperatures, which could potentially damage the exhaust system (specifically, the three-way catalytic converter inside the header).
- It might cause a decrease in power.
- As shown in the following graph, a slightly richer air-fuel mixture than 14.7 could provide a slight increase in power, but a leaner mixture would cause power to drop rapidly.
As consumers, we can make informed decisions here after understanding the principles and potential side effects. After all, it’s our own cars, right? (Rather than vehicles that can be subject to manufacturer OTA updates and various unusual features after purchase.)
Interesting Findings While Tuning
I noticed that when using the Hondata public version program—meaning
the AFR table in the screenshot above had not been adjusted—the AFCMD (ECU-requested AFR) would constantly fluctuate.
However, after adjusting it, the AFCMD became much more stable.
I’m not sure if this is due to additional compensation in the stock ECU or if there’s a bug in the Hondata public program.
By combining a bit of theory with hands-on experience, you might gain a better understanding of your car and internal combustion engine characteristics.