The last major piece to the puzzle! To be completely honest, AFR's are NOT hard to analyze. However, understanding WHY you want a particular AFR is a bit more complex, and that's where this article comes in. Today we hopefully answer questions like "What is a good AFR?" "What is lambda?" "How does ethanol effect AFR?"
What is AFR?
Likely, you're pretty familiar with the term "AFR" - air fuel ratio. Essentially, how many parts air to how many parts fuel. By the way, the "parts" are measured in mass, interestingly. Burning fuel is simply a chemical reaction, where we need so many parts air + so many parts fuel to get a burn. Remember high school chemistry?
Maybe you should have paid attention in class after all.
If we have exactly the right amount of fuel corresponding to a specific amount of air, we have what's called a "perfect burn". In this scenario, all the reactants (fuel and oxygen) are used up without any remaining fuel or oxygen. This perfect ratio is referred to as the stoichiometric ratio, and which is represented by the Greek letter lambda.
Okay...so why care about Lambda?
From a scientific point of view, lambda is actually better to work with than AFR. Lambda is always based off the stoichiometric ratio, so a perfect burn of methanol, ethanol and gasoline all have a perfect burn at a lambda of 1. However, they're respective AFR's can be quite different!
As you can see, AFR's change quite a bit as we alter fuel. However, lambda's stay very very similar. Luckily, our ECU measures and adjusts AFR with lambda, so it's actually quite adaptive to changes in fuel type.
What to look for in logs
The logs are pretty simple to analyze. We want to make sure of two things. Firstly, that our measured (real) AFR matches the target AFR. This signifies that car is operating consistently. Some loggers also support logging fuel trims, but this varies a lot logger-to-logger, so I won't cover that here. If it seems like your car isn't hitting AFR's, trying running the car to empty, getting a fresh tank of good top-tier fuel and driving it around for a while (50+ miles). It's likely that you've gotten some bad gas.
Fun fact: This graph comes from a 1980s Bosch Motronic Electronic Fuel Injection Manual.
Secondly, we want to make sure we aren't running overly rich or overly lean. Hopefully your logger spits out lambda directly, as we pointed out earlier it's much easier to work with. Under WOT, ideally we'll be in the peak power range close to ~0.82 lambda (reference the table for exact values). However, it's not uncommon to see lambda of 1.0 below 4500 RPM on Mk7's.
For whatever reason, the factory ECU keeps the system relatively lean through the low- and mid-range. If the car is running healthy, this won't cause any issue. However, if your ECU is doing adaptations, you can quickly get into dangerously lean conditions. It's not uncommon for a bad batch of gas to put fuel trims +/- 0.05 lambda. If you're already at 1.0 lambda, going to 1.05 puts you in dangerous lean condition, as the leaner the mixture, the more likely the engine is to knock. However, if you were at ~0.85 lambda, and the car ran 0.05 lambda lean, 0.9 lambda would still be safe (and powerful).
What about the other direction? Downsides of running too rich? Generally speaking, running rich results in a loss of power and increased soot (roll coal, right?). In the extreme and extended cases, running rich can foul plugs, destroy cats, and even piston rings. Generally speaking, these cases occur when the car is running overly rich all the time, not just on WOT.
The Real Take Away
Our ECU's work in a closed loop system that is based off lambda. If the ECU isn't hitting lambda, it's likely bad gas or a bad sensor (haven't seen this yet on a mk7). Unlike AFR, lambda doesn't change much with fuel type which allows Mk7's to work in "flex-fuel"-like manner. Take a look at your logs, if you're on a tune that runs in meat of the "Peak Power" AFR, try running some E and revisit the timing tables! Good luck and happy tuning!