Been tracking the difference since I got the car, results are interesting... they seem to track ok, but there is a significant delta between the two, with the computer estimate being overly optimistic most of the time. We already know this, but thought some hard data would be interesting. See attached picture.
Computer vs. actual mileage
1 - 16 of 16 Posts
I plan on doing the same, I just got my car but it'll be interesting to see how it compares!
nice chart, I've been comparing mine over about the last 10 fill ups and the AVERAGE reading is sometimes low and sometimes high vs the manual calculation - bottom line over that many fills is very little difference, so to me the AVERAGE ready is fairly accurate, enough that I'm not going to worry about it.
The calculation of distance on the trip computer divided by the number of gallons required to fill the tank is flawed. You are using the total distance driven on a previous quantity of gas with an unknown energy content divided by the amount of gas required to "replace" that gas. Unless you know how much gas you put into the car previously, and what it's total volume was, the calculation most people are using is nothing more than a rough estimate.
FrugalDriver
Hey LordHelmet Thanks - That's a cool web site that uses the words green and performance in the same sentence! Lots of good (real world) stuff on that site...
Again, I'll point out, this calculation is flawed.
You put in a quantity of gasoline when you filled up the car, and you drove a certain number of miles on that quantity. You then refilled the car with another quantity, and used the mileage from the previous quantity and the volume of the current quantity to calculate the MPG.
A given quantity of gas can have different energy content depending on the manufacturer, temperature, and lots of other factors. You are only coming up with an estimate when you use that type of calculation, because you are assuming that each gallon of gas you pump has exactly the same energy content, which is impossible.
you are talking about very small differences here. energy content of the same brand of gas (assuming you stick to one brand) a week apart is minute at best. but you are correct, it is an estimate, albeit a fairly good one....
I'd be willing to wager that the differences are more than minute. Unless you are using the same pump, at the same time of day from the same station, you are going to have variances in both density of gas per gallon, as well as energy content per gallon.
I don't see any problem comparing average MPG on the trip computer with the actual MPG calculated at the pump. Provided you reset the trip computer after each fill, you will get a relevant comparison. Same distance, same tank of fuel; what is the problem? Energy differences just don't come into it. LordHelmet's chart is mostly in sync, so he is comparing the appropriate readings.
I've checked this quite exhaustively on a recent holiday and the trip computer was consistently 1 MPG optimistic. Another car was spot-on over a 7000 km trip (Average = 11.5 km/litre, calculated 11.59 km/litre; rounding error since display had only 3 digits).
The average readout can also be instructive; on one stretch of road, I had achieved 38 MPG with careful driving, but when my wife took over the wheel, it rapidly dropped to 35 MPG. -Although she is a good driver and was maintaining the same speed, she tended to accelerate/brake more than necessary. There was also some initial confusion over setting the cruise control and the repeated acceleration made the average take a nose-dive.
From a purely scientific standpoint you are correct, however the differences will be minute for most measurement needs. The volume of liquid gas chanages about 0.12% with a 1degree Celsius change. That is roughly 1% with a 10 degree C change. Not a significant amount for practical purposes. Average temperature change from one fill to the next is not likely to be more than 10 degrees Celsius in most cases.
Some years ago, the Australian Institute of Petroleum did extensive field tests to measure the effect of delivery temperatures. Mainly due to the insulated underground tanks, the extreme density variation was only plus or minus 1% , and the conclusion of the study was to recommend staying with the volumetric basis of sale.
http://www.docstoc.com/docs/91337141/THE-TEMPERATURE-CORRECTION-OF-PETROL
If you have been reading the posts, you will see that this has nothing to do with the comparison of trip and actual calculations. One parameter that needs to be considered is the accuracy of the odometer, if that is used for the manual calculation. Calibrate against GPS or known distances.
Major error sources would be odometer and consistency of pump shutoff at same amount of fuel in vehicle. Averaging over several tank fillups would average out the pump shutoff differences. Odometer error would be constant, except for tire wear decreasing tire diameter over time. As noted, volume is variable, but not to a great extent. Energy content can vary significantly between pure gas and gas/ethanol mix. However, it doesn't matter what mix you put in when calculating mileage for the previous tank, what matters is what volume is used, which you measure by the amount of fuel needed to replace previous tankful.
That's a good summary.
On the matter of tyre wear. I checked my odometer when the car was new and it was 99.7% accurate. That is, for every 100 km travelled, the odometer registered 99.7 km. After 15,000 km, the tyres have lost 1 mm of tread, so the wheels are rotating a little faster and odometer is 99.999% accurate!
99.7 X 668 / 666 = 99.999 (not a devilishly hard calculation).
For the first time I caught my odom real time average consumption meter lying. It was stuck on the same number for hours - no way this could be true. While displaying the real time average, I reset it by holding down the trip button for 5 seconds. It blinked and reset. Then it displayed some ridiculously high number and as I drove it tapered all the way down to well below my 'norms' for my model.
Bottom line - do not rely on trip counter or on board computer as it is not reliable.
Bottom line - do not rely on trip counter or on board computer as it is not reliable.
here are the figures from my Rio for the last 4000KM or so. The graph is a little cluttered but the grey line indicates what mileage the trip computer thinks the car got, and the white line is the economy as determined by comparing pump readings to the odometer. Red line is a running average, and the white and yellow line up the top shows the price fluctuation. The wild variation in economy is basically down to road trips up the freeway every month. As you can see, the trip computer has been overly enthusiastic on all but 3 of the 15 tankfuls so far. I wasn't entering the trip computer readings for the first three tanks though.
![]()
1 - 16 of 16 Posts
Join the discussion
