More on MAF tuning
I took some more data for Iteration 3 (meaning I did not apply the MAF tune implied by Iter 3 in my recent post). The resulting tune has a bit higher peaks than the Iter 3 graph I posted. I'm not surprised. I decided the additional data was well advised, considering that most of the data for Iter 2 as well as the first data set for Iter 3 were taken when it was "cold out." (By "cold," I mean 65-75 degrees.)
I took two more datasets on my way home when it was above 90 outdoors. The result was a bit more aggressive for iteration 3. I'll be applying this tonight, and taking data tomorrow for iteration 4. I may be driving my car up to WinStar tomorrow night (TBD), in which case I'll get some data for a 2hr highway drive mostly cruising.
Still, I'm pretty happy with the fairly strong convergence even over a wide range of ambient temperatures. Behind the cut is an explanation of why this is difficult.
The Mass Air Flow sensor works on a simple principle. It heats a wire by applying a fixed voltage. Because the resistance across the wire is roughly constant, a fixed current flows through the wire. Therefore, under ideal circumstances, the wire dissipates a fixed wattage.
If the wire were sitting in a vacuum, the energy imparted on the wire would need to be radiated away. Incandescent light bulbs work by this principle: Put a filament in a vacuum or near vacuum, run electricity through it and it glows.
If the wire were sitting in a non-vacuum, energy would be carried away by two methods: Radiation and convection. Convection happens when the fluid in contact with the filament (air in this case) gets heated and changes in density. Less dense fluid rises relative to more dense fluid. Hence the concept that "heat rises." Heat doesn't rise, but hot air does relative to cold air. That's how hot air baloons work, BTW.
There's a third method by which the filament can dissipate energy, though, and that's conduction. That is, whatever mass it comes in contact with it can engage in the transfer of heat energy. If that mass is colder, energy transfers from the filament to the mass. If the mass is warmer, energy transfers from the mass to the filament. It's this third method that's integral to the functioning of a MAF sensor.
Suppose you hold the ambient temperature constant and have perfectly still air. You'll lose energy only due to radiation and convection. The filament will reach an equilibrium temperature determined by these properties. (I could explain why, using Newton's Law of Cooling here, but already this is getting pretty deep.)
Anyway, we can establish a baseline from this. Given this baseline, now suppose we start air flowing around the filament. To a first order, what you'd find is additional heat transfer away from the filament as a function of how much air moves past it. It's this principle that a MAF works upon.
So how do we measure it?
The filament in the MAF sensor is thermally coupled to a thermistor. By thermally coupled, I mean that they're glued together in such a way that they should be close to the same temperature. For now assume the coupling's perfect.
A thermistor is a resistor whose resistance varies with temperature. By measuring the resistance of this thermistor, we can determine what the temperature of the thermistor is, and by proxy, the temperature of the filament.
How does that help us?
Ideally the filament is hotter than the incoming air. So, the less incoming air there is, the hotter the filament gets. The more incoming air there is, the cooler it gets. It's a direct, monotonic relationship. So, the PCM measures the resistance of the thermistor and throws that into its calculation of incoming air mass.
But, as I alluded to above, there's other frustrating factors. The temperature of the incoming air determines how much heat energy it pulls away. Thankfully, a different sensor, the Inlet Air Temperature (IAT) sensor measures that. So, between IAT and MAF, the computer has a pretty good chance of estimating the mass of air entering the engine.
So what gets in the way?
Recall the other two mechanisms for dissipating heat energy: Radiation and conduction. It turns out that the MAF isn't the only show in down when it comes to heating air. The air intake and especially the throttle body both heat the incoming air. Furthermore, the they radiate their own heat, which can in turn heat up the MAF directly, separately of heating the incoming air. All of this serves to throw off the measurement.
So, in the end, the MAF reads the Mass Air Flow to a first order, but gets second order effects from the intake and throttle body heating up. At idle, the intake path alone is inefficient enough that it can start breathing air as hot as 165°F. The throttle body heats up even more-it's a mere 3" from the exhaust cross-over.
This is why it's important not only to shield the intake, but also tune the sensors for a wide range of operating conditions. Those second order effects become significant and throw off the readings.
I realize I've glossed over many of the deeper details-those of you who know your thermo deeply will recognize that. I hope, though, I've touched on enough of the basics to highlight the reasons why MAFs aren't perfect and tuning over multiple temperature bands is important if you want a robust tune. The alternative is a per-temperature-band tune. That might provide better peak performance at that temperature, but they tend to be more brittle.
I took two more datasets on my way home when it was above 90 outdoors. The result was a bit more aggressive for iteration 3. I'll be applying this tonight, and taking data tomorrow for iteration 4. I may be driving my car up to WinStar tomorrow night (TBD), in which case I'll get some data for a 2hr highway drive mostly cruising.
Still, I'm pretty happy with the fairly strong convergence even over a wide range of ambient temperatures. Behind the cut is an explanation of why this is difficult.
The Mass Air Flow sensor works on a simple principle. It heats a wire by applying a fixed voltage. Because the resistance across the wire is roughly constant, a fixed current flows through the wire. Therefore, under ideal circumstances, the wire dissipates a fixed wattage.
If the wire were sitting in a vacuum, the energy imparted on the wire would need to be radiated away. Incandescent light bulbs work by this principle: Put a filament in a vacuum or near vacuum, run electricity through it and it glows.
If the wire were sitting in a non-vacuum, energy would be carried away by two methods: Radiation and convection. Convection happens when the fluid in contact with the filament (air in this case) gets heated and changes in density. Less dense fluid rises relative to more dense fluid. Hence the concept that "heat rises." Heat doesn't rise, but hot air does relative to cold air. That's how hot air baloons work, BTW.
There's a third method by which the filament can dissipate energy, though, and that's conduction. That is, whatever mass it comes in contact with it can engage in the transfer of heat energy. If that mass is colder, energy transfers from the filament to the mass. If the mass is warmer, energy transfers from the mass to the filament. It's this third method that's integral to the functioning of a MAF sensor.
Suppose you hold the ambient temperature constant and have perfectly still air. You'll lose energy only due to radiation and convection. The filament will reach an equilibrium temperature determined by these properties. (I could explain why, using Newton's Law of Cooling here, but already this is getting pretty deep.)
Anyway, we can establish a baseline from this. Given this baseline, now suppose we start air flowing around the filament. To a first order, what you'd find is additional heat transfer away from the filament as a function of how much air moves past it. It's this principle that a MAF works upon.
So how do we measure it?
The filament in the MAF sensor is thermally coupled to a thermistor. By thermally coupled, I mean that they're glued together in such a way that they should be close to the same temperature. For now assume the coupling's perfect.
A thermistor is a resistor whose resistance varies with temperature. By measuring the resistance of this thermistor, we can determine what the temperature of the thermistor is, and by proxy, the temperature of the filament.
How does that help us?
Ideally the filament is hotter than the incoming air. So, the less incoming air there is, the hotter the filament gets. The more incoming air there is, the cooler it gets. It's a direct, monotonic relationship. So, the PCM measures the resistance of the thermistor and throws that into its calculation of incoming air mass.
But, as I alluded to above, there's other frustrating factors. The temperature of the incoming air determines how much heat energy it pulls away. Thankfully, a different sensor, the Inlet Air Temperature (IAT) sensor measures that. So, between IAT and MAF, the computer has a pretty good chance of estimating the mass of air entering the engine.
So what gets in the way?
Recall the other two mechanisms for dissipating heat energy: Radiation and conduction. It turns out that the MAF isn't the only show in down when it comes to heating air. The air intake and especially the throttle body both heat the incoming air. Furthermore, the they radiate their own heat, which can in turn heat up the MAF directly, separately of heating the incoming air. All of this serves to throw off the measurement.
So, in the end, the MAF reads the Mass Air Flow to a first order, but gets second order effects from the intake and throttle body heating up. At idle, the intake path alone is inefficient enough that it can start breathing air as hot as 165°F. The throttle body heats up even more-it's a mere 3" from the exhaust cross-over.
This is why it's important not only to shield the intake, but also tune the sensors for a wide range of operating conditions. Those second order effects become significant and throw off the readings.
I realize I've glossed over many of the deeper details-those of you who know your thermo deeply will recognize that. I hope, though, I've touched on enough of the basics to highlight the reasons why MAFs aren't perfect and tuning over multiple temperature bands is important if you want a robust tune. The alternative is a per-temperature-band tune. That might provide better peak performance at that temperature, but they tend to be more brittle.