Okay, I finally filled up my tank after spending the last one without the radio, CD player, A/C and heater... no gunning the engine, altered work route to take expressway instead of freeway (note: I did have to turn on the fan for about 15 min). So, my efforts with hypermiling produced the following numbers:

Manual calculation (total miles / gallons filled) = 26.17 mpg
Energy Monitor calculation = 27.5 mpg

=> Energy Monitor 5.0% higher than the "actual" mpg

So, how does this discrepancy compare to the tank prior to this where I had the CD player playing all the time, drove on the freeway for work commute, gunned it here and there? The numbers there are:

Manual calculation = 24.8 mpg
Energy Monitor calculation = 26.1 mpg

=> Energy Monitor 5.2% higher than the "actual" mpg

Namely, the discrepancies to within expected uncertainties are identical. I measured similar discrepancies for other tanks as well.

In other words, you should expect your Energy Monitor measurements of your MPG to be about 5% higher than the actual values.

(Would be interesting to know what percentage of the numbers stored in the mileage database are based off the EM versus manual calcs).

I do have one measurement that indicates that this discrepancy can be additionally about 2% higher when you've used the A/C and/or have a trip with lots of elevation changes from climbing/descending hilly streets. I'll keep track of this during the summer months to see if this is real.

Anyways, I'm going back to enjoying my HiHy; cranking up my CD-player, my butt-warmer and occasionally outgunning those obnoxious Suburbans. Hypermiling is fun... but only up to a point. Always holding back 268 hp is just plain cruel. As long as I know the EM is always about 5% higher, I can at least know where I really stand (i.e. keep the EM above 26.25 mpg (=> 25 mpg actual)...

 

The first manual calc I did on our HiHy came out at 26.1 MPG while the computer was showing 25.3 MPG. Don't forget that ultimately gas mileage is nothing more than the miles driven divided by the amount of gasoline used in gallons. This is true for 10 miles, a full tank, or a year's worth of driving (the longer you go, the better the average).

In order for these two numbers to match, the computer must be reset at the station and the tank must be filled up to exactly the same amount as the previous fill up. This is very hard to do consistently even with the autoshutoff feature on the gas pump. I would be curious to see how the computer and the gas station data compare after one year, or even six months (I'll never know, I always reset mine).

 

THIS IS A GOOD TOPIC. MY HIHY DID 24.8 ON THE SCREEN AND 24.4 ON MANUAL CALCULATION. tHIS IS THE FIRST TEST.i WILL COMPARE THE NEXT 10 TANKS TO SEE HOW MUCH IT CHANGES. JUST I MORE SQUEEZE ON THE GAS NOZZLE COULD CHANGE IT SOME.

 

The reason for the discrepency is because of a cummulative sampling error. When you drive in city/suburban conditions, frequent speed changes result in the sampling error. Your computed MPG will be different than your displayed MPG. If you do a long highway trip, at constant speed, you'll find your display matches your computed MPG, within a two or three percent.

 

I contend that the onboard FE display is going to be much more accurate from a tank to tank basis than hand calculation because the data source and the way the data is compiled for the vehicle's calculation of MPG is much more accurate than the source data available to us for calculation.

The car calculates MPG based upon fuel metered to the engine over distance traveled. We calculate FE based upon tank fill up over distance traveled. We and the car see the same distance traveled, however the onboard computer knows how much fuel was metered into the ICE. The only thing we know is how much fuel we were able to put in the tank. Bladder or not, that tank fill can vary from filling to filling. A few ounces one way or the another can have a huge impact on FE on vehicles which get exceptional economy. On the other hand, the metered fuel is a known value which is stored with the onboard computer system, compared to the known distance traveled and the vehicle comes up with an accurate average based on values that are fully known with no "fill-up variance" error.

Now, if all gas pumps shut off at the same place and time on a vehicle, or we had a trans parent filler indicator on the side of the vehicle so we could see the actual fuel level in the tank, then we could refill the tank to the same point each time we refilled and do a calculation. I would suspect that if we could do that, we would get an FE number that was quite similar to what the vehicle calculated.

 

Tank to tank you're correct. But if the onboard is accurate you should also see some tanks that calculate higher than onboard and some lower than onboard. It seems people are not see this high/low just too high from the onboard.

The fill-up variance disappears with more tanks in calculating your cumulative average. Unless the fuel pumps at the gas stations are inaccurate, the owner calculated accumulated FE (and GH database) should eventually match the onboard value.

I don't know for sure but I don't think the onboard actually measures fuel use with a precise fuel meter, but does a fuel use estimate based on several other control sensors like the o2 and MAF along with fuel injector pulse durations.

 

I agree with what a couple of other posts say. I also see substantial differences on a single tank between the calculated mpg and the mfd mpg. Over time though the actual gallons used will not lie. I have a little less than 15K on my 05 Prius and have noticed up to 10% or even more single tank differences. I also suspect the computer is missing negative data on the real time caclulations. I have another idea that cold weather is causing the differences. I have seen numbers varying less than 2% since it warmed up here in Dallas over the last 2 tanks (59.8 calculated vs 60.6 mfd and 59.5 calculated vs 60.6 mfd). That's about 850 miles of driving.

I've kept a tank by tank log since purchase (mfd figures added 1/1/06 forward) and plan to send it to Toyota after I have a full 6 months of the variance. Just to see what they say.

 

Did a google looking for info on the accuracy of the fuel pumps at gas stations and found this.

http://www.npr.org/templates/story/s...toryId=5304258

NPR All Things Considered talkd to a gas station inspector that actually checks accuracy. Bottom line, there is slight variation, which overall goes in the favor of the customer, though it is certainly possible to get significantly less or more on occasion.

One tip, fill at the slower notch to get maybe slgihtly more than at the fastest notch on the handle.

 

So, let's look at some data that I've collected over the last 4 tanks of gas, starting from the most recent first (this is a AWD HiHy):

Manual (mpg) / Energy Monitor (mpg) / %Difference
-------------------------------------------------------------
26.9 / 27.7 / 3.0%
26.2 / 27.5 / 5.0%
24.8 / 26.1 / 5.2%
25.2 / 27.1 / 7.5%

These tanks were filled up at different gas stations and different pumps. Each tank had different driving conditions (A/C, %Freeway, Gunning, etc). The most recent tanks had the least % of freeway driving, and is interestingly correlated to a lower %Difference. However, there is insufficient data to claim that the correlation is statistically significant.

As you can see all 4 tanks show that the EM is consistently higher (~5.1%) than the manual calculations. This data shows good evidence that the EM reports systematically higher mpg estimates than that calculated by the manual method (total miles/gallons of gas). The NPR report provided from the previous post indicated that they measured discrepancies of typically 1/2 cup per 5 gallon or about 0.6% discrepancy, which is far lower than the 5% discrepancies from these measurements.

I find it unlikely that the different gas stations represented by these measurements are all systematically calibrated to report 5% more gallons than actually delivered. (Although it would make an interesting conspiracy theory.) I don't see how that could go unnoticed over time...

I'll keep monitoring this to keep adding to the statistics.

 

Hi Gang.

We just bought our Toyota Highlander Limited 4x4 about a week ago. The whole thing we wanted was better environment, gas mileage, and of course the high resale value of toyota, amongst other pluses in quality. We paid $43k which included :

42000 MRSP -3k off = 39000. We got gap and extended warranty = 2k . We paid taxes title insu which was also 2k. So we paid 43k for the top end hybrid. Pretty pricy.

It said 31 and 27 mileage EPA. I already knew from our 2 other vehicles that EPA is full of it as a rule.

We only have 100 miles on it and the computer says 22mpg. Most has been small trips of about 2 miles. I took it out on the highway last night just to help break the engine in. From what i have read here, we could be doing better, but it isnt that important to alter our life to squeeze a few more mpg.

My jeep grand cherokee limited is 'supposed' to get 16 and 20 according to the EPA but in reality gets about 12 mpg around town. The 20 is fairly accurate on the open road under optimal driving conditions. The Jeep has an inline 6 , 4.0 litre and torques about 195hp.

we also have a little ford escort which is a total miser. She gets about 20mpg in the city real world, and at least 35mpg on the highway. Love that sucker.

We traded a beautiful and powerfull but economical disaster Chevrolet Trailblazer LTZ because of the high miles on it and the horrible 10mpg real world city driving. My wife loved her 'machine' and was sorry to give up her 'kick azz' monster truck. So we got her the Highlander hybrid to get the best of both worlds and appease her needs.

This highlander has better pickup and power than both of the suv's, and of course the gas mileage is much much better. We didnt buy the thing to squeeze every drop out of the EPA , or we would have bought a pirus. Better yet, bought a bicycle.

To think we can drive a 268 hp fully loaded SUV and get 22mpg real world out of the sucker in the city is a miracle in its own right. As much as i detest the bait and switch of the EPA of all vehicles, i already knew to expect 80% real world.

We are delighted that we are helping the enviroment and putting a whole pile of $$$ back into our pockets, instead of giving it to oil companies who artificially inflate gas prices whenever the hell they think they can get away with it , using excuses that are pathetic and intelligence insulting.