"James Sweet" wrote in message
news:OsOsj.375$ph.210@trnddc06...

I agree that the rated fuel consumption for motor vehicles is off - I
haven't averaged that low for at least 15 years. For instance, my car
is
rated 20/29 for city/highway - last week I averaged 33.2 mpg in about
440
miles of mixed driving according to the car's odometer and the reading
on
the pump when I gassed up on Friday evening. According the average
mileage
display on the dashboard of the car, I got 34.1 mpg, and I suppose the
difference could be due to the attendant (no self service gas in NJ)
filling
the tank right up to the gas cap and/or inaccuracies in the pump or
odometer, or even simply to the fact that I fill up the tank at the end
of
the day (when the car and the gas is at its warmest) but do about half
my
driving in the morning (when the gas is the coolest). Which means
simply
that a full measured gallon on a Friday afternoon is probably less than
a
gallon on Monday morning simply due to the expansion and contraction
that
go
with changes in temperature.



The expansion and contraction based on temperature for a volume as small
as
a tank of fuel in a car are so tiny that you'd never be able to measure
them
with anything around the house, and certainly not the odometer in your
car.
The fuel temperature varies over a range of perhaps 60F max, usually much
less.

Well, maybe I'm calculating wrong. There's an approximately 3% difference
between what I calculate as my miles per gallon for last week and what the
car calculated. The coefficient of expansion of gasoline is 0.069% per
Fahrenheit degree. Coincidentally, over a 30 degree temperature difference,
that's between a 2% and a 3% change in volume. For 15 gallons of gas, that
comes somewhere between 3 and 4 ounces of gas.

The meter on the pump reads out several digits to the right of the decimal
point - it appears that this level of accuracy is available at the gas
station. The odometer reads out only to the tenth of a mile, which means
that I don't have the accuracy at my end to calculate this by hand. I don't
know what the internal accuracy is when the car computes average miles per
gallon - I presume the fuel pump knows pretty precisely how much gas it's
pumped, and the odometer measures distance covered by counting revolutions
of something (one of the wheels?), and it seems reasonable that the internal
accuracy of the car's computation is more than adequate to notice a
difference of this magnitude.

I guess the other consideration is that the car is likely computing average
mpg using the gas burned (or at least, pumped to the engine) while any by
hand calculation is basing it on gas bought, and any difference the fill
level will throw the result off. Last week, the attendant took great pains
to fill the tank right up to the brim (he was evidently trying to get the
total to come out to a whole dollar amount), something that usually doesn't
happen. So I have no problem believing that I bought slightly more gas than
I burned.

Whichever figure is right and whatever the explanation, it still seems to me
that the mileage estimates published by the EPA are too low, and it's seemed
that way ever since I started paying attention (way too many years ago).