2016 Odometer Error

I am curious about this myself.
I track the commute roads I do, and changing the winter to summer tires (16" vs 17" tires) does change the overall distance traveled.
I will try to do a test with the app with a loop with the Gen3 and whenever I will get my Gen4 repeat it.
So far my average underestimation of fuel consumption is, considering changes at the pump that can hardly be verified, 6.5%. How much of that is gas station pump inaccuracy, foam in the tank due to speed of fueling up, odometer wrongly calculating distance based on type of tires used and their wear, volume of fuel based on temperature (winter/summer), it is quite likely "unknown"...

 

That new odometer error would approximately match my post-odometer-lawsuit 2014 Forester, which is somewhere between 1.8 and 2% low.

My 2010 and 2012 Prii were/are underreading by just 0.2-0.3%. My pre-lawsuit Subaru was similarly close to 0%.

For calibration, I used many roadside mileposts on I90 from the Columbia River to the Idaho border (some posts are offset, but there is only a single busted segment, around MP220 at Ritzville) and a Garmin automotive GPS.

 

@krousdb: Thank you for starting this thread! If you get a chance, could you see how accurate SpeedTracker is around your 17 mile loop?

I agree it is important that people list their tires when posting data, because even with the same size designation the circumferences of tires from different manufacturers could be different.

BTW, there was a famous incident with AJ Foyt and the Ford GT40 being tested at Daytona where tires stamped the same actually had different circumferences. Dirt track racers like AJ knew this, and always measured their tire's circumferences so they could put them on the car according to size to create "stagger" so the car would turn better in the desired direction (usually left I think).

 

No odometer can be absolutely accurate. The consequences of an odometer reading too high are lawsuits, from owners claiming they were excluded from warranty prematurely. Consider too that as tires wear down and OD reduces, the odometer reports will get closer to accurate, perhaps eventually read slightly high.

 

I used Delorme and set start and finish on 100 mile route, my odometer reading was 99.6 miles, delorme had it being exactly 100.0 so I was .4 mile off, which gives me about the same reading you got. I have Gen 3, 15 inch wheels, have 30k miles on Michelin energy savers A/S. The Delorme is real close within a few feet.

 

There is an error in your reasoning. If your mileage is underestimated by 2%, mpg will also be underestimated by that same 2%, so MFD should be reporting less mpg than your really get.

 

% units affect using multiplication, not sumatory. To substract 2% from that 5% you should do this: 0.05*(1-0.02)=0.049 equal 4.9%.

So basically adding a 2% odometer error means that the corrected overstated mpg calculation is 4.9% above real instead of 5%.

Turns out it is negligible!

 

If this is true for the whole time, you'll have 1900km extra warranty.

But seriously, as your tyres wear, it will become less, and possibly reverse. Are you able to do the calculation over the life of your tyres and report back?

 

Are you sure about that? Thinking in terms of ratios, if miles/gallons = 1.05 of normal, and miles = 1x normal, that would means gallons was 1/1.05, or about .952 of the true gallons used (95.2%). If miles is also 2% low, wouldn't the ratio be .98/.952 = 1.029? In that case the displayed MPG would now be 2.9% high. Maybe my math is wrong, but that's how I'd calculate it.

 

Normally (with typical inflation pressures and working loads), the portion of the tire on the road is flat, not round, so the normal circle equations don't work well for figure changes in rolling circumference as the tread wears. The actual changes appears to be much less than one would expect from the circle equations.

With modern steel belted radial tires, the rolling circumference is very closely tied to the length of the steel belts underneath the tread. These belt lengths do not change much, or at all, as the tread thins.

Tire spec sheets usually include RPM: revolutions per mile. Use that figure.










Merged.

????????

(1 + 5%) * (1 - 2%) = (1.05)* (0.98) = 1.029 = 1 + 2.9%.

2.9% is more than a negligible change from 5%. The difference between the 3% that Krousbd expressed, and the 2.9% dropping out of this calculation, is negligible.

 

I'll have to check this out next week... I tend to overinflated my tires a bit too, and I wonder if there is a % difference because of that...

When folks post data, can they post tire size, psi, miles on the tires and the mileage and mpg results?


iPhone ?

 

With the old bias ply tires, this may have made a difference. But with the modern steel belted radials, I'm betting that you won't be able to see the odometer calibration difference without some high precision measurements, or some absurdly high pressures. (Or too-low pressures, or very lightly loaded tires.)

 

You are correct. The right way to do the math is taking into account the total value, 1.05, and not just the .05 as I did.

So taking this result as correct, we end with a corrected 2.9% error over refueling method.

I think that margin is about the same range of error we could get with refueling measurments, bBut the car's estimation error figure is consistent for everyone that calculates it. It should be a consistent error with some dispersion and measurements should be in a small range instead of that closely wrong all of them.

So I think it's possible than the the car's computer estimation is already corrected by the engineers that designed it, and we are making consistently the same mistake and getting an underestimation with the fillup method. We really don't have as much data and knowledge about the car as them.

After concluding a consistent 2.9% between refueling calculations and onboard computer figure, I think it means that it's highly probable that it is already correct and we are making wrong assumptions somewhere.

We should probably take the car's value as reliable.

 

Are you saying the Prius accurately reports mpg? Or just the odo is accurate?

The latter yeah, the former nyet.

 

It could depend on how it senses whether the tank is full at the start. Does anyone know how the vehicle senses how much petrol (or gas) has been added to the tank or if it has been filled.

I only part filled mine (a FORD) recently (I guess only 80-90% full), but the vehicle seemed to think that I'd filled it completely, causing false readings. I always fill to the first click of the bowser, but I can slowly get another couple of litres in if I wanted to, but either way, it thinks it's full.

 

I'm pretty sure it's measuring fuel flow, regardless of your fill level. and with the odometer readings it's got the two variables it needs, to calculate mpg, or any similar units.

It's only owners that need to rely on fillup amount to do their calculation. As long as you fill similarly each time your calcs will be reasonably accurate. And any variation cancels out, doesn't magnify, with subsequent fill ups and calcs.

What I see is a very reliable 7.5% discrepancy.

 

Odometers, fuel pump meters, computers, and other parts of any mileage tracking system will all contain variables and amounts of inconsistency. I'd say a 2% to 5% error is probably expected and acceptable. The equipment just isn't precise enough for better measurements.