[OregonRider]

I have a 2012 R1200RT with on board computer. I find that I ride slower if I am watching the MPG reading on the computer. (I tend to ride faster when watching the MPH reading!)
The MPG reading consistantly gives the same numbers:

48.0 49.0 50.0 51.1 52.2 53.4 54.7 56.0 57.4 58.8 60.3 61.9 ....

How are these figures calculated? There must be some lograrithmic equation used (although I suspect one would need to convert to kilometers and milliliters) Any mathematicians out there?

 

[11862]

Speculating, I would bet that they sum up the lengths of the fuel injector pulse lengths to calculate total fuel flow. A simple division by t he mileage since last reset would give a scale able value for net fuel mileage. There may be some turn-on or turn-off time constants included in the summation to accommodate the characteristics of the fuel injector(s).

This means if the fuel rail pressure goes down, or if an injector becomes partially plugged, that your mileage will mysteriously decrease.

Remember - processors are fast as hell and don't get bored........... :rotf:

 

[pbarney]

Interesting question. I may be wrong, but I believe mpg is calculated from fuel strip data. Reading posts from riders who have problems with fuel strips ( I have not had the pleasure...knock on wood), it seems the "Miles left" indicator becomes inaccurate. I doubt BMW would have two separate systems for fuel monitoring. It's correct that one could calculate fuel used from the injector pulse width, number of cylinders, engine RPM, etc. But I don't believe that the software guys would waste the computational overhead in the ECU just to give the rider an mpg value.

Just my $0.02.

 

[beech]

I suspect that miles per gallon and "range" miles are two different calculations. The mpg could be done via the fuel flow and distance/time information while the range includes the tank level reading.

 

[pbarney]

Beech, after a little digging on how other vehicle manufacturer's determine mpg, I have to agree with you. Check out the theory of operation at this URL: http://combustory.com/wiki/index.php/MPGuino

 

[OregonRider]

As I dont have a fuel strip on my 2012, the computer cant use that. Also, shortly after resetting the MPG screen, while sitting at a stoplight, I can watch the MPG drop rapidly. This makes me think it must use fuel flow. The distance to empy calculation, however, has to figure the amount of actual fuel in the tank.
I wonder if the distance to empy figure is based on the MPG since reset number or on current fuel flow and amount of fuel in the tank???

 

[Sotys]

...I wonder if the distance to empy figure is based on the MPG since reset number or on current fuel flow and amount of fuel in the tank???

The distance to empty is based on the current fuel consumption and availability in the tank. You can easily test this. Ride for 10 minutes with say 80 mph and see what is the indicated distance to empty. Then decrease the speed to 50 mph and you will see after couple of minutes your distance to empty increasing . The computer always compare current consyumption with remaining fuel.

 

[TheoM]

IIRC, BMW farms-out it's code to Siemens technicians. This seems to be true of BMW, BMW Motorrad, and MINI.

I know that the consumption/range behavior in my '00 328i and my '04 MINI S are similar. By that I mean the percent variance between hand-calculated and reported MPG seems consistent. I've found that in those vehicles, zeroing-out the average speed and consumption each time I fill up allow both to report more accurate data.

My '12 MINI Countryman S and my '09 RT report much more accurately. However, I can still improve the accuracy by zeroing out the averages. Currently the difference between reported and actual MPG data is 2.3% optimistic in the '12 MINI and 1.5% optimistic in RT.

Recently, I remember reading that MPG was calculated by fuel flow (injector pulse width, I am assuming) and distance travelled. I have no explanation for the improvement in accuracy by zeroing-out average speed! However, I believe there is a "rolling" calculation used for average MPG, so zeroing that data out starts a more current series of rolling calculations. The implication there is a situation where MPG will more accurate if the speed/conditions for the whole tank are close to the same. However, if you reset the average MPG at fill-up and drive the first 150 miles conservatively on the slab and the last 50 miles in stop-and-go conditions, the average mileage (due to the "rolling" nature of the calculation) may be less accurate (in this hypothetical case, MPG would report lower than the actual average).

Also, here is a bit of insight into the "cruising range" and "Average fuel consuption" calculation as offered-up in the MINI Countryman manual. Note that the average consumption (in the Countryman, at least) is calculated based on the time the engine has been running. We know it doesn't zero-out when the engine is turned off, so it is constantly adjusting an MPG number which, perhaps, was NEVER zeroed out; the "lifetime MPG, if you will. Personally, I want the average to reflect the time the engine has been running ONLY since I filled up the tank!

Of course, what Siemens programs for the Countryman may be TOTALLY different from what they do for the RT, but I would bet that similar algorithms are used by Siemens in the RT (sans two of the four cylinders, of course!).

 

[OregonRider]

So it is established then, that the computer monitors 1) current fuel flow, 2) Fuel flow over the past short interval and, 3) the amount of fuel remaining in the tank. What remains unexplained, is why the MPG readings jump in consistant, discreet steps i.e. from 58.8 to 60.3. Why do we never see 59.7 or 60.1? What alogrithm defines these numbers? And why should anyone care?

 

[TheoM]

So it is established then, that the computer monitors 1) current fuel flow, 2) Fuel flow over the past short interval and, 3) the amount of fuel remaining in the tank. What remains unexplained, is why the MPG readings jump in consistant, discreet steps i.e. from 58.8 to 60.3. Why do we never see 59.7 or 60.1? What alogrithm defines these numbers? And why should anyone care?

First, that we actually care probably condemns us to some sort of exile or at least ugly public ridicule!

Second, I'm guessing that the sensor sampling rate was coded with some constant value that is used to increment or decrement the current MPG value. Perhaps, in this case, it is 1.5 MPG. Further, this implies that the MPG is not updated unless the calculation shows at least an increase or decrease of 1.5 MPG.

Is the difference consistent, i.e.; is the adjustment delta always the same value?

:deadhorse