Error in average MPG display

[shopsis]

I was really pleased to see what high average mpg I was getting on the
display in my brand new OBW (2006 2.5i, manual trans). 28.5 mpg and up
on my regular commute of about 22 miles highway and 3 subarban/city.

Then I filled the tank. The display, which I had reset when I first
filled the tank, said 26.1. The number of miles divided by the output
of the gas pump came to 24.35.

This is a 6.7% difference. In the wrong direction.

It's only one data point, and I will of course, check after each fill
up for a while, but it leaves me wondering what other people are
getting when they compare the mpg display to real consumption, and what
people here, and Subaru, think are reasonable tolerances.

Charlie in de Bronx

 

[Mark T.B. Carroll]

It's only one data point, and I will of course, check after each fill
up for a while, but it leaves me wondering what other people are
getting when they compare the mpg display to real consumption, and what
people here, and Subaru, think are reasonable tolerances.

It might partly be a distance-measurement error: the speedometer and
trip counter in my WRX seem to be a bit off. (I should check the
odometer with my GPS to make sure I'm not being penalised for too much
mileage!)

-- Mark

 

[speednxs]

IIRC I8 has signs that precisely measure 3 miles. This is old school
compared to a GPS. Every odometer I've had read a couple tenths over 3
miles. I'd just love to be there when someone returns a lease with
over miles and demands a correction for an optomistic odometer. Let
the fireworks begin!

 

[Michael Pardee]

I was really pleased to see what high average mpg I was getting on the
display in my brand new OBW (2006 2.5i, manual trans). 28.5 mpg and up
on my regular commute of about 22 miles highway and 3 subarban/city.

Then I filled the tank. The display, which I had reset when I first
filled the tank, said 26.1. The number of miles divided by the output
of the gas pump came to 24.35.

This is a 6.7% difference. In the wrong direction.

It's only one data point, and I will of course, check after each fill
up for a while, but it leaves me wondering what other people are
getting when they compare the mpg display to real consumption, and what
people here, and Subaru, think are reasonable tolerances.

Charlie in de Bronx

Of course, the error from tank to tank will vary depending on the pump and
other things that change the amount the tank will hold. The true error can
be calculated by comparing over a handful of tanks.

My analysis is like this: the computer knows many things with great
accuracy. It knows how far you have gone, normally with the same sensor
output that drives your odometer so any error nulls out. It knows precisely
how often it fires the injectors. What it knows with less accuracy is how
much fuel is injected with each firing. It knows how much it asked for, but
it makes assumptions about the viscosity of the fuel (probably a good
assumption) and the fuel pressure.

I think most of the error is in minor inaccuracies in the fuel pressure
regulator.

Mike

 

[y_p_w]

Of course, the error from tank to tank will vary depending
on the pump and other things that change the amount the
tank will hold. The true error can be calculated by
comparing over a handful of tanks.

Don't forget gas station fuel pump accuracy.

<http://seattlepi.nwsource.com/local/264462_gaspumps27.html>

Some fail inspections. Fuel delivery speed sometimes
affects accuracy.

 

[Michael Pardee]

Don't forget gas station fuel pump accuracy.

<http://seattlepi.nwsource.com/local/264462_gaspumps27.html>

Some fail inspections. Fuel delivery speed sometimes
affects accuracy.

I looked at this a few months ago and decided the pumps were *supposed* to
be accurate enough that they wouldn't account for the several percent
discrepancies most of the time. Of course, there will be exceptions....

Mike

 

[Ian]

I have noticed the display on my car ('06 Tribeca) reads 0.5 to 1.0 MPG
higher than what I calculate from miles driven and gallons purchased. The
error seems higher in colder weather. Could it be the computer doesn't
register gas used for some period of time after each start?

 

[David Betts]

I have noticed the display on my car ('06 Tribeca) reads 0.5 to 1.0 MPG
higher than what I calculate from miles driven and gallons purchased. The
error seems higher in colder weather. Could it be the computer doesn't
register gas used for some period of time after each start?

These things are toys. They are not expected to be precise. They give
a general impression, that's all. Useful for comparison purposes - ie
seeing if you can improve your mileage by adjusting your driving style
- but don't ever take them seriously as a measure of consumption.

 

[Michael Pardee]

These things are toys. They are not expected to be precise. They give
a general impression, that's all. Useful for comparison purposes - ie
seeing if you can improve your mileage by adjusting your driving style
- but don't ever take them seriously as a measure of consumption.

The only one I have is on our 2002 Prius. It's pretty accurate, always 3% to
5% optimistic. Considering all the inputs to the calculation are digital and
the assumptions are typically good to better than 5% it's really no
surprise. Overall it is about as accurate as the speedometer.

The gas guage is notoriously inaccurate in the Prius (owners often call it
the "guess guage") and the MPG display, including the miles since we filled
the tank, is the best way of determining how much gas is actually in the
tank.

Mike

 

[Don Dunlap]

It might partly be a distance-measurement error: the speedometer and
trip counter in my WRX seem to be a bit off. (I should check the
odometer with my GPS to make sure I'm not being penalised for too much
mileage!)

-- Mark

I have a Tribeca and it is very consistent in a 1 MPG error. I always enter
all gas purchases in a spreadsheet and it gives very accurate measurement.
When the Tribeca shows 19.7 MPG for a tank, for instance, the spreadsheet
calculation will show 18.7, give or take one tenth.

Don Dunlap

 

[Bryan J. Lee]

I think it's off because it doesn't count time at idle with the average. My
wifes '05 buick show the average mileage and it start to go down when you
sit at idle. I think are cars only count when you are moving. Just a
thought.

 

[Rick Courtright]

sit at idle. I think are cars only count when you are moving. Just a
thought.

Hi,

The way one such system was described to me is that it "measures" the
fuel flow rate thru the lines and compares that to the speed of the
vehicle in making its calculations. If the car's sitting, there's a
"divide by zero" kind of situation, no? If so, then I think you're on
target with your thoughts!

Rick

 

[Victor Roberts]

Hi,

The way one such system was described to me is that it "measures" the
fuel flow rate thru the lines and compares that to the speed of the
vehicle in making its calculations. If the car's sitting, there's a
"divide by zero" kind of situation, no? If so, then I think you're on
target with your thoughts!

There is no division by zero. The calculation is miles
traveled divided by fuel used. The numerator is zero while
you are idling but the denominator is not.

When you are going down a hill, THEN the numerator is much
greater than the denominator and the instantaneous display
will go to its maximum value and then stop displaying, at
least on my Passat.

The time at idle should be correctly included in the average
MPG calculation.

 

[shopsis]

I noticed this morning, on two occasions while I was idling at a red
light with the car in neutral, that the "average mpg value" decreased,
suggesting that gas consumption while idling is included in the
average.

I'll do a more focused experiment, when I have a chance, by pulling
over and idling for several minutes shortly after the car has had a
long enough trip for an average trip mpg to register on the display.

Charlie in de Bronx

 

[Victor Roberts]

I noticed this morning, on two occasions while I was idling at a red
light with the car in neutral, that the "average mpg value" decreased,
suggesting that gas consumption while idling is included in the
average.

Obviously you are correct, but someone else here had that
same observation and came to the conclusion that the idle
time is not included.

I'll do a more focused experiment, when I have a chance, by pulling
over and idling for several minutes shortly after the car has had a
long enough trip for an average trip mpg to register on the display.

I have seen this many times in our Passat - as is to be
expected.

 

[Bonehenge]

I noticed this morning, on two occasions while I was idling at a red
light with the car in neutral, that the "average mpg value" decreased,
suggesting that gas consumption while idling is included in the
average.

Stopped idling time affects correct, real world MPG. While stopped
and idling, the car is getting zero MPG. On the same token, during a
long descent, the engine is using idle fuel at highway speeds,
creating a period of high MPG.

If only moving time is counted in the display, it's a gimmicky toy.

 

[Michael Pardee]

I noticed this morning, on two occasions while I was idling at a red
light with the car in neutral, that the "average mpg value" decreased,
suggesting that gas consumption while idling is included in the
average.

The idle consumption pretty much has to be included. MPG readings can't be
averaged, so the total miles has to be divided by total gallons.

Our Prius has to "dummy up" the readings when sitting - much of the time the
immediate consumption rate is actually zero, and even zero divided by zero
is undefined.... The interval readings are correct, though. At speeds over
11 mph, if the engine hasn't started, the display just shows 100 mpg rather
than trying to divide by zero.

Mike

 

[shopsis]

Mike:

With the Prius does the actual mpg (miles per
tankful)/(gallons/tankful) agree with the car's calculation?

Charlie

 

[Michael Pardee]

Mike:

With the Prius does the actual mpg (miles per
tankful)/(gallons/tankful) agree with the car's calculation?

Charlie

It's pretty close, running 3% to 5% optimistic. I normally think of the
error as 2 mpg; like 47 instead of 49. I don't check it much any more, but
did check it out the first few tanks. I've occasionally thought I could
calibrate it by tweaking the fuel pressure regulator about 1 psi lower, but
the risks seem higher than the value of calibrating the display...
especially since it's mainly my wife's car.

Mike

 

[Rick Courtright]

is undefined.... The interval readings are correct, though. At speeds > over 11 mph, if the engine hasn't started, the display just shows 100

Hi,

Just out of curiosity, does the engine start at a given speed (or pretty
close), or does it somehow figure acceleration rate, cruising speed or
other factors into its "decision" when to start up?

Rick