Theory about why MPG indicator reads high

I believe the fuel usage is measured by fuel injector on period, so any fuel of one gallon is one gallon..

I use non ethanol and my offset reading is 6.4% higher.

Following is copy of my post.
-----
scientific numbers...
gasoline energy: 34.6 MJ/liter
ethanol energy: 21.2 MJ/liter

-> E10 energy = 34.6 X 0.9 + 21.2 X 0.1 = 33.3 MJ/liter
-> 33.3 / 34.6 = 0.96

The E10 has 4% less energy than gasoline.
We will see 4% less mpg number.

Ken@Japan

 

Not necessarily. If the viscosities of the two fuels are not identical, the volume that flows thru the injector may differ even if the injection time and feed pressure are the same.

This calculation assumes that the E10 actually contains 10% ethanol. The labeling on the pump usually reads "May contain up to 10% ethanol", so unfortunately we really don't know what we are actually getting. So the mpg penalty based on energy content could be anywhere from 0 to 4%.

There is also the miscibility issue. Gas and ethanol are miscible, so when you add 10 mls of ethanol to 90 mls of gasoline, you get less than 100 mls total volume.

 

The Prius MPG calculation counts injector pulses to determine the amount of injected fuel. It's a reasonably accurate and precise method, but there are rounding errors. I suspect Toyota deliberately rounded in their favor, as it looks better to unsuspecting customers.

Tom

 

The indicated distance travelled and road speed are based on a tire rolling-radius which may not account for slippage between the tire and the road, and certainly not for the effects of wear and tire pressure. So even with new tires, the indicated distance travelled may be more than the actual, and as the tires wear the indicated distance becomes more optimistic.

 

I don't know how the Insight-I does mpg calculations. Whatever method Honda chose clearly works quite well, much more accurately (at least in mine) than what people are reporting for the Gen-III Prius. I don't have good data for my 2011 Prius yet.

For my Insight-I, I haven't seen any increased discrepancy between the FCD mpg and hand-calculated mpg over the six years since I bought the car in 2006. I don't know if I was getting non-ethanol gas at the beginning or not, but of course now all gas sold in CT is up-to-10% ethanol.

 

Unfortunately, both the hand-calculated mpg and the mpg on the HSI display rely on the odometer readings. So any differences between them arise from differences in the measurement of the amount of fuel used.

 

Guys, attached is a chart I made plotting the calculated error of the MFD per each tank on my Prius. Current average is 6.8%. You can also notice a lot of variability. The thin line is the linear trend. I don't know why it is increasing, rather than being constant. I also don't know how I got a negative number.... and why sometimes I am getting an error which is well below average (maybe because I stopped pumping immediately after the first click ).

I have not noticed any appreciable difference in fuel consumption since using E10 (since Feb. 2011 sold in Germany, before it was E5) and in any case this would (should!) not affect MFD estimation (the car is designed to work with E10...).

These I think are the external factors (i.e. not related directly to the MFD calculation) that can affect the calculation of FE:

• distance travelled vs distance shown in MFD (affected also if you change tire size, i.e. summer vs winter tires (in my case 17" vs 16") - actually I have a question - is tire diameter always the same independent of wheel size?)
• fuel pumped vs fuel *actually* pumped (likely less, not more...)
• when tank is "full" vs when tank is *actually* full (when does one stop pumping, how much air has gone in due to the speed of the fuel going in affecting when the pump will detect a "full" and hence click)
• temperature that can affect fuel density (albeit I guess to a very small degree, and my chart shows a variability that is more frequent than temperature changes between tanks; as you see some tanks are only a few days apart)
• changes in ethanol content between different gas stations of the same company and changes between different companies

How each of these items affect the error calculation is hard to assess, IMHO. Therefore, the average over several tanks is what counts. And from what I see, and considering the wide differences in calculated vs displayed L/100km, I am of the opinion that:

1. either the MFD does a good job, but the factors above are affecting heavily (e.g. the pumps are stealing a lot of fuel by pumping much less than displayed on the counters at the gas station , or I have a different style of fuelling my car each time , or....whatever )
2. or the MFD does a very bad job (6.8% bad, in my case) because Toyota decided, and I wouldn't be surprised and it makes economical sense, to use a cheaper method to estimate FE, that is "good enough" for most drivers. Obviously they used a method that shows better numbers than worse...I would have done the same!

I am more inclined to think it is number 2, affected even more by 1, i.e. MFD is maybe 4-5% off and the measuring/hand-calculation adds the extra 1-2%.
What does your error chart look like? do you see the same variability I am seeing?
Does anybody know how accurate the distance (km/mi) in the MFD is?

 

Thanks for the data!!

To minimize variability, it's important to fill the same way every time. It's easy to just stop at the first click, and it's probably a good idea to slow down the pump as you approach full to avoid a "false" first click. Ideally one would fill at the same pump at the same station every time, but of course this isn't possible for most people.

Not necessarily. It depends on the two tire sizes. Some of the online tire stores have comparison charts that give you correction factors for different tire sizes.
But any change in tire diameter will affect your tank-to-tank mileage, not the difference between MFD vs. hand calculation. The reason is that both the MFD and the hand calculation rely on the same odometer data, so they will be affected equally by an error in the odometer.

Yes, this will make the hand calculation lower than the MFD number, which is what is observed. It wouldn't surprise me that gas stations cheat us by 6-7% on average.

I had noted the MFD data for my Insight-I over 60,000 miles is only 0.6% higher than the hand calculation. But, thanks to your post, I suddenly realized that this could mean either that:
(1) the Prius MFD is optimistic by 6-7%, the Insight-I MFD is accurate, and gas stations on average overcharge us by 0.6%
OR
(2) the Prius MFD is accurate, Honda offset the Insight-I MFD by minus 6-7% to cover overestimates of volume pumped, and gas stations overcharge by 6-7%.

 

I do not believe that most gas stations are dispensing an under amount of gas. Gas stations are regulated and besides it being unethical.

If someone has the ability to use 0 ethanol gas and compare the guage reading with a tank of 10% ethanol gas that will answer that question.

Here in Oregon its is a 10% Ethanol state all year long, otherwise I would do it.

alfon