Following article is very good...
http://hybridcars.about.com/od/owner...seandglide.htm

Ken@Japan

It's slightly more accurate topping off every time, but that's bad for a few reasons, pollution being one of them. It's really best to just pump until automatic shut-off.

hmmm, can't speak to your personal experience but the objective data says something different. given that the combined mpg between hch.ii and prius ii are relatively the same (and it may turn out that ima in hch.ii surpasses the gen ii prius this year) it must either be:

1. hsd and ima get relatively the same mpg at both city and highway speeds.
2. hsd gets better city but ima gets better highway mpg.
3. hsd gets better highway but ima gets better city.

given that the above 3 scenarios define all possible result sets, i'm willing to place my bet on #2.

I'll bite. Why would it be slightly more accurate topping off every time?

They're driven by different drivers. Most of the forum posts I've read suggest that the Honda drivers, on average, drive more conservatively than the Prius drivers. But all I know for certain is I got better mileage driving a Prius (I rented one for a couple of days, twice) than I did with the HCH. And I get slightly worse mileage with the TCH than the HCH, yet this car is 700 lbs more and I tend to drive a bit quicker in it too (and it's like 4-5 seconds quicker to 60 miles per hour).

well let's see, you certainly didn't help the hch.ii database with your mileage!

seriously, there are hypermilers in both databases. there are lead footed drivers in both databases and there are lots of "normal" drivers in both databases. given the sample sizes, the drivers in both databases have been averaged out. i don't see the drivers as the reason for the hch.ii as having great mpg. as i see it, the only factors at this point that could be affecting the data are (and these all affect the hch.ii data in a positive way):

1. summer/winter fuel. the hch.ii is just starting to benefit from the summer fuel while the prius gen ii has at least a year or more of this benefit already built in. expect the mpg in the hch.ii database to go up b/c of this.

2. break-in time. the hch.ii data may improve somewhat at the cars in the sample break in. the prius already has this data built in for most of the cars in the sample. there is no definitive data that shows that break-in actually affects mpg. so, i think this is almost a moot point.

3. driver learning curve - the hch.ii data may improve somewhat as drivers learn how to driver their hybrids. again, prius already has this built in for most of the cars in the sample. more than break-in, in think this factor can and will have a significant impact on the data in the hch.ii sample. i'm still learning how to maximize my mpg and i've owned my hybrid for 4 months.

4. early adopters - here the hch.ii may have more early adopters in its sample than the prius. of course this is less and less true as months go by. however, we don't know anything about how early adopters relate to mpg. and, in the end the sample is large enough and enough time has gone by that any effects of early adopters is pretty minimal.

my conclusion is that the data has not yet stabilized for the hch.ii. that we will see an increase in the mpg for the hch.ii but that this increase is not due to a different driver set but rather to other external factors.

i'd be interested in hearing others' comments/analysis on the data.

cheers,
neil

Just a guess - topping off is less sensitive to differences in the "kickoff" threshhold of various gas pumps. Some will shut off earlier than others, while topping off produces a visual representation that the tank is full (via spillage, or near-spillage).

I meant to answer that question but forgot. In all cars except the Prius, the tank is topped off at basically the same point, but the gas pump automatic shut-off does not happen at the exact same point every time. This can result in some tanks' calculated mileage being a bit too high or low, but it will be balanced after the next tank.

So you might drive 500 miles over 15 gallons but the pump shuts off after pumping 14.9, thus making your mileage calculation a bit higher than it really was (33.6 instead of 33.3). In other words you used 15 gallons but you pump 14.9 so you mistakenly think you only used 14.9. But this will be balanced on the next tank since there's less gas in the tank than usual, so you might drive 497 miles over those 14.9 gallons, then pump 15 gallons before shut-off, making it look like you did 497/15 (33.1) when you really did 497/14.9 (33.3) so it balances out.

To sum up, your lifetime mileage will be perfectly accurate but mileage for individual tanks might be off a bit -- slightly high on one tank, then slightly low on the next, balancing out.

But remember: topping off is bad for the environment. It causes unburnt gas to vaporize or even spill (which is REALLY bad). There is a reason both the pump and the car's gas cap contain warnings not to top off.

Soo...I guess that begs the next question . If I do the math by hand as in I set the trip computer and then just divide that number by hand with the amount of gas I pump, that number is accurate regardless of whether I top off or not right?

If that is true, then it sounds like the car thinks its full when its not or maybe it is when calculating. That sort of sounds like the car doesn't actually calculate how many gallons is actually burned but uses the gas tank display? Using the display sounds even more inaccurate if you fill the tank say always somewhere between E and 1/4. How would the car know if you burned say x.1 or x.2 gallons at the point?

I have a fuel economy/trip computer in my 02 Trailblazer, and I am fairly certain that there is a sensor that is monitoring the amout of fuel being used (instantaneously). I don't think that it is based on the fuel gage - I think this would be far too inaccurate.

It then keeps track of the cumulative instantaneous values, and uses those to calculate the average MPG.