Yidney

How do our cars actually calculate the average mpg shown on the MFD? It is a calculated number from a bunch of input parameters, or is there a direct measurement of gas flow? If so, by what?

Anker

The ECU that controls the fuel injector pump knows how much fuel it uses at all settings.

I really like the "From Start" statistics, especially how they are reset automatically. Haven't seen that on any other car.

tw2

I would guess it uses the mass airflow and the injector pulse width to calculate it. The fuel pressure is always the same. Mine appears to correlate well with how many km I get from a 55L tank ie not many.

Rudeney

The ECU records how many times it has pulsed the injectors. The amount of fuel used by each injector pulse is a known value. It's simple math.

tw2

Rodney, would the amount of vacuum not factor into it as well since that changes the amount of fuel that actually passes with each pulse? Or would that be fairly negligible on average.

Glyn M Ruck

It is purely calculated from injector pulses & constant variation in injection pulses with RPM & flow demand. Effectively an in-line flow meter that measures total flow even in open loop conditions of enrichment vs mileage. That is why the consumption average starts so high (open loop) & then reduces as mileage increases from a cold open loop start & runs into closed loop fully warmed up. Any change in throttle rate will immediately reflect in the average reading. i.e more flow via injectors with greater TPS signal modified by airflow through the MAF.

%age of WOT TPS input plus airflow through MAF = injection rate trimmed by measurements from the O2 sensor.

Yidney

So then I guess my next question is whether the ECU "knows" how much gas is used with each injector pulse, or it just thinks it knows based on some rating of the injector. Like I could take a squirt gun that has a rating of 1 oz per pull of the trigger (assuming such a thing existed), and shoot it 100 times into a tub. I assume I shot 100 ounces, but will it really be 100 ounces when I measure the contents of the tub? One is a direct measurement, and one is a calculated measurement.

raymond g-

if the calculation relied solely on injector measure/sensor, it would have a
gaping hole in the data calculation, i opine, as that simply returns fuel used
and no one has yet mentioned (or take into consideration) miles traveled.
oh, I see that Anker implied mileage in his response vis a vis 'from start'

so, if the programmers are worth their salt, they would cross check the data
as well as utilize other sensor metrics not limited to MAF, intake pressure,
fuel remaining in tank, odometer, etc.

MPG is a moving target depending on a variety of conditions (cold, warm,
pressure, throttle position, current MPG, current tankful MPG, etc) that the
designers can program the results to average this based on a preset palette
of conditions.

Glyn M Ruck

No ~ Injectors vary in fuel flow (delivery) per pulse depending on parameters listed above. As long as the injectors are within spec the measurement is extremely accurate. Ring main pressure is maintained at > 3.8 bar or 55 psi. Flow back to tank in no way affects the calculation. All that is being measured is what the injectors actually inject for each pulse. Then summed vs mileage. A direct measurement of how much fuel the engine sees vs mileage.

The ECU knows exactly how much fuel is being injected for a specific throttle opening & fuel trim. (caveat no leaking or damaged injectors).

Rudeney

There seems to be some "fuzzy logic" programmed into the system because when you first start the engine and sit at 0 mph, you are getting some n on-zero MPG on the display. Obviously, if you at and ran then engine like that all day, you'd be getting 0 MPG, but it starts at some "base" value, assuming that you will idle for a brief time and start driving, thus building positive MPG.

Glyn M Ruck

The display at start makes more sense in l/100Km. In l/100Km it ramps up to pretty alarming consumption if you don't move off running open loop mixture enriched while not putting on distance.

But I'm with you.

xsever

One dimension to add to this is that the car's mileage calculator is pretty accurate. I use an Android app to track my gas and repairs.

When I fill from the exact gas station standing in the exact same spot, the mileage the app gives me and the "From Reset" dash display are exactly the same. (I reset the mileage at each fill-up). The accuracy is really perfect.

When I fill from a different gas station and/or with different inclinations, then there is some variance.

So this leads me to conclude that the programming on MB's side are pretty accurate.

tw2

My C32 used to start at 38L/100km as a maximum after startup even at standstill and then work its way down from there. I never use that screen on the CLK as I find I end up trying to save gas for some reason and not enjoying the car as much.

Glyn M Ruck

You are correct And especially if you do any hard acceleration under open loop conditions. But it does teach one when running downhill to take your foot right off the gas to shut off injection completely. Then the numbers come tumbling down.

sailorben

Yes I find I get obsessed with it if I view it. And the better I get the more obsessed I get about it.

On a 130mile journey, traffic was very congested, but managed to keep driving along at about 45-50 most of the way home. After about 60 miles, I achieved 60.2mpg! Not a typo! As I was coming closer to home, I got into the country lanes so had to get that screen off, after having 20 mins of fun it was down to 50mpg. Pretty outstanding.

Usually at 70mph I get 45mpg which is decent. My 45 minute commute to work has a mix of city and 15miles of 70mph, I average about 38mpg. I have to have the consumption screen on, as otherwise if I'm not careful I can quite easily get the consumption up to 30mpg without much issue. Which for a diesel is rubbish.