Not sure, but I think even 2 (3 at the most) tanks would get it very accurate. If your pump happened to shut off early one time, it should even out the next time. The dash display is at least very consistantly inaccurate, in my experience around 6.5% optimistic.
Hand calculated. I like the suggestion of 10 or more tanks, and use total mileage/total gallons. If you average the mpg for each tank for the 10 tanks, I'm not sure that would be as accurate (as I recall I figured it both ways with another car and the numbers were different, though not hugely). Having figured mpg for a very thirsty vehicle, I know there's a number of factors that can affect one tank, including differences with how a pump shuts off. I wonder if 2 or 3 tanks is enough to even out those differences. My car's per tank seems to be just over 5% on the optimistic side compared to the calculated, the longer term trip odometer seems to be closer to what the overall average I'm seeing.
We received our car in May, 2009. I believe manufactured date was April, 2009. An early one to say the least. We now have 60,000 miles on the odometer. The MPG guage has been consistently off by 2 to 4 mpg. I believe this was done with purpose by Toyota. As most people would rely only on the guage and they would be seeing 2-4 mpg than the were actually receiving. I know for a fact with the 2006 Honda Civic Hybrid the MPG guage was just the opposite. When it read 50 MPG it would actually calculate to 51 to 52 mpg consistently. Toyota can correct this; but will they? alfon
Hand calculation, of course, but with the proviso each fillup must be to the top of the filler cap. Why? Because gas station pump shut-off valves aren't neccessarilly calibrated to shut off exactly as others. Personal experience with my Prius shows I can usually put in ~1.8 gallons more after the initial pump shut-off. However, there is a danger to this method. Due to EPA requirements, the fuel filler venting systems on most cars may take some of that topped off fuel and direct it to the gas vapor canister in the engine compartment. If that canister become saturated with liquid fuel (as opposed to vapor), it can really screw up the emission system. (That happened to me with another vehicle.) I believe you can safely top off the tank without incurring this problem as long as you drive your car immediately, and not let it sit for several days or more. PS: My calculations consitently show ~2 mph over the onboard display.
Yes. I set a spreadsheet that does the calc tankful-by-tankful. This is good for tracking variations, seasonal effects, and so on. But then I also set up a function off to the side that just tallies all the gas, and the total distance, and comes up with a number based on that. It's a little different than averaging the per tank mileage values.
The Tank MPG calculation is fairly precise but not accurate. There is a bias on mine of 5 - 6%. Hand calculation will be the most accurate, especially as your number of tank fill ups increases. After that, the Tank MPG value becomes predictable because you can add a fudge factor. For me and my typical driving, I can subtract 3.5 mpg and be within 0.2 mpg of the hand calculated value. Knowing this makes it a whole lot easier to estimate the range of the vehicle, assuming a complete fill up. It also clues me in if there is a problem with the pump, because I know, typically within 0.1 gallons of where the pump will shut off. This prevents underfilling AND overfilling.
My Scangauge cumulative average MPG is almost exactly the same as the Prius' number if they are both reset at fillup. I assume it's because they both use the same input data. Whenever I calculate it from the gas pump and the trip odometer miles, it comes out lower by 2-3 MPG. Maybe that's the bladder issue.
I've been using pencil & paper from the beginning and have been getting 3-4% under the dash readout. Let me throw this out there, does anyone think that just maybe the readout is higher than actual consumption because of the use of ethanol in the gas?
The error in mine is too large for this, and I seem to remember from previous threads that straight gas doesn't fix the Gen3 error. And Gen2s don't have this error with ethanol, at least not on a similar scale.
I've posted some of this in other threads but it looks like it'd help this discussion as well with credit this info taken from a 2002 article from http://www.caranddriver.com/features...andal!-feature If you want your display more accurate all you need to do is buy different tires to offset the inaccuracy. We don't know how much variation there is but the general recommendations have been discussed thoroughly in http://priuschat.com/forums/gen-iii...esistance-replacement-tires-current-list.html The number to watch is revolutions per mile (RPM) and my current assumption is Gen II RPM of 845 and Gen III RPM of 835 are good targets. It's slightly complicated because the odometer and speedometer are calibrated differently. As mentioned in http://priuschat.com/forums/gen-iii-2010-prius-main-forum/98662-how-fast-have-you-taken-prius.html the Gen III Prius speedometer goes non linear at 40+ MPH. We can probably assume similar nonlinear behavior for the Gen II. Swapping tires can improve the accuracy but since there are three targets no single change will make it all perfectly accurate (low speed, high speed, and odometer). Still if you pick the right RPM you could get the accuracy closer. Ironically depending on the width of the tire going from and to making the display more accurate could reduce your real world MPGs by increasing rolling resistance/weight/drag though you'll gain better handling/traction in the exchange so I wouldn't worry about the minor loss in fuel economy.
When I had my Cadillac Seville and the VW TDI, the consumpsion readouts were usually within 1/10 mpg either way, so I stopped bothering with the pencil & paper. I'll probably keep checking with the Prius for about 6 months or so to see if the variance stays about the same. The main thing is that my wallet notices the difference.
My MFD error has not stayed the same. It was 5.7% optimistic in the first year, and 7.6% optimistic the second year as driving style changed and mpg improved.
Averaging over 10 tanks won't work well unless you want a yearly average. In that case use 12 tanks. Pearl takes a month to go through one tank. Over 10 months the temperature varies a LOT. Temperature affects mileage a lot. Flooding the carbon canister with liquid fuel won't hurt it but I suppose it -could- upset the emission control system for a short time. I've never seen that though. With GII Prius it's a bad idea to totally fill the fuel system, as if the fuel then warms up and expands it -can- pop off hoses inside the fuel tank. That will result in eventually having to replace the tank. Not sure what the consequences are with the GIII. Probably none. Pearl's display has been high and low compared to hand calculation, depending on who knows what. I do fill the tank as full as I can, consistent with leaving -some- expansion room (or none when it's colder outside than in the ground where the fuel comes from, provided I will then drive far enough to use a few litres). If you do accidentally overfill a GII leave the fuel cap loose for a few hours to prevent pressure buildup. I haven't seen codes thrown from this practice, but don't leave the cap loose for days.
You are assuming that the fuel pump and the odometer are accurate Assuming they're accurate enough, if you simply tracked how many gallons of gas you put in your vehicle for 10,000 miles, it would not matter whether you filled the tank each time or not, if you've put 200 gallons of gas in the tank, you'd have 50 MPG. 10,000 miles is a lot, but if you were off by just 2 gallons over that period of time, your MPG rating will change by 0.5 MPG. With that in mind, a tank-to-tank deviation of 1 or 2 MPG is pretty much a rounding error. When I fill my tank, I usually use the same gas station, the low octane gas, fill to the first click-off, and I reset the trip odometer. Example: 500 Miles Traveled / 10 Gallons Added = 50 Miles per Gallon
