Posted by mark b on September 27, 2006, 1:02 pm
With a 0.1 mile accuracy on the odometer, you would only have to
measure 10 miles to have 0.1% accuracy on the measurement. This would
also be enough distance to over-ride any error from taking the
measurement while moving and having to judge the stop and start points
on the fly. If you are off by 4% after a 10 mile trip, I would agree
that the odometer is off - or the mileage markers are off.
MarkB
Carpe Diem wrote:
> > A digital odometer only implies more precision, but its precision is
> > actually lower than a mechanical odometer because you can't estimate the
> > fractions of a mile like you can with a rolling digit (this applies
> > whether you are talking about an integer mile or a 0.1 mile type display).
> >
> > I assume, based on the number of significant figures in your measurement,
> > that you did this over a fairly large distance (like 1000 miles,
> > otherwise, the last few digits are meaningless). Are you converting from
> > kilometers as well? That could introduce a fixed error into the
> > calculation depending on how many digits you use in the calculation.
> >
> The Prius has a digital odometer & "trip" meter ; the trip meter has a 0.1
> km display.
> Miles or kilometers : I did not convert anything. When 100 km are in fact
> 104 km, the 100 miles will be 104 real miles...
> Measurements are done over rather short distances, but next week I have a
> trip of about 550 miles (885 km) and I'll try to do a new measurement...
Posted by Carpe Diem on September 27, 2006, 1:33 pm
> With a 0.1 mile accuracy on the odometer, you would only have to
> measure 10 miles to have 0.1% accuracy on the measurement. This would
> also be enough distance to over-ride any error from taking the
> measurement while moving and having to judge the stop and start points
> on the fly. If you are off by 4% after a 10 mile trip, I would agree
> that the odometer is off - or the mileage markers are off.
I agree, and as I cannot imagine that ALL the mileage markers (along
different highways) are off, I guess it's the odometer which is...
Posted by Bill on September 27, 2006, 5:41 pm
>> With a 0.1 mile accuracy on the odometer, you would only have to
>> measure 10 miles to have 0.1% accuracy on the measurement. This would
>> also be enough distance to over-ride any error from taking the
>> measurement while moving and having to judge the stop and start points
>> on the fly. If you are off by 4% after a 10 mile trip, I would agree
>> that the odometer is off - or the mileage markers are off.
>>
> I agree, and as I cannot imagine that ALL the mileage markers (along
> different highways) are off, I guess it's the odometer which is...
We know this much: Since you have new, OEM tires, it should be off 2%.
When they are worn out, it should be off 2% in the other direction.
Posted by Bill on September 27, 2006, 5:40 pm
> With a 0.1 mile accuracy on the odometer, you would only have to
> measure 10 miles to have 0.1% accuracy on the measurement. This would
> also be enough distance to over-ride any error from taking the
> measurement while moving and having to judge the stop and start points
> on the fly. If you are off by 4% after a 10 mile trip, I would agree
> that the odometer is off - or the mileage markers are off.
> MarkB
But Mark, aren't you assuming less than a 1% error in mile marker placement
and ignoring tire wear and the attendant +/- 2% error? If the tires are new
(2%) and the markers are off (1%) and if readings are taken on the fly, 1%
after a 10 mile trip is meaningless. Tire diameter is significant to such
an extent the Nav system includes a recalibration procedure to be performed
in conjunction with installing new tires.
Posted by NeoPhyte_Rep on September 28, 2006, 2:52 pm
Bill wrote:
> Tire diameter is significant to such
> an extent the Nav system includes a recalibration procedure to be performed
> in conjunction with installing new tires.
>
Could this also affect the MultiFunction Display of Miles Per Gallon?
> > actually lower than a mechanical odometer because you can't estimate the
> > fractions of a mile like you can with a rolling digit (this applies
> > whether you are talking about an integer mile or a 0.1 mile type display).
> >
> > I assume, based on the number of significant figures in your measurement,
> > that you did this over a fairly large distance (like 1000 miles,
> > otherwise, the last few digits are meaningless). Are you converting from
> > kilometers as well? That could introduce a fixed error into the
> > calculation depending on how many digits you use in the calculation.
> >
> The Prius has a digital odometer & "trip" meter ; the trip meter has a 0.1
> km display.
> Miles or kilometers : I did not convert anything. When 100 km are in fact
> 104 km, the 100 miles will be 104 real miles...
> Measurements are done over rather short distances, but next week I have a
> trip of about 550 miles (885 km) and I'll try to do a new measurement...