Edmunds.com | Inside Line | Auto Observer Your Account | Help | Directory
You are here:
![[F]](http://a332.g.akamai.net/f/332/936/12h/www.edmunds.com/media/townhall/webxicons/site.gif){kind=small}
Forums
SUVs Mazda CX-9 Mazda CX-9 Real World MPG
140 messages, Last post on Feb 13, 2009 at 11:14 AM
You are in the Mazda CX-9 Forum. Your Hosts are steve_ & tidester
|
Just filled up the tank for the first time. 16 mpg in urban traffic, not bad. Will average out over the next fill ups using "odo B" to keep track of eventual improvements. I know it is heresy, but comparing it to the Sienna minivan it replaces, the engine needs its revs to move. |
|
|
Replying to: steve_ (Jul 07, 2008 6:15 pm) It dilutes the error rate. Let's say for example the pump you're using sometimes adds an extra 1/10 of a gallon of gas more or less. Adding that 1/10 of a gallon to a 2.5 gal fillup will create a 4% error (.1 / 2.5 x 100), while that 1/10 gallon pump error will only account for a 1% error on a 10 gal fillup (.1 / 10.0 x 100). Let's say your pump error is 1/2 gal because it's flowing faster or slower, more is released on the click, you're holding the pump slightly differently, etc. That 1/2 gal difference will create a 20% error on a 2.5 gal fillup vs only a 5% error on a 10 gal fillup. It's basic statistics. The larger the sample size (i.e. the quantity of the fillup) increases accuracy (eg MPG for that fillup). That's also why scan gauge really isn't accurate for snapshot MPG ratings, since MPG is Miles Per Gallon and if the car only moved a few feet and scan gauge tried to calculate a Miles Per Gallon number based on a few feet of driving, it wouldn't be very accurate in comparison with calulating a MPG over a hundred mile range. So if Scan Gauge calculates that it took 0.023 gal to travel 6ft that would equal 20.24MPG (0.023 x 5280 / 6), but if Scan Gauge was 1/100 of a gallon off and the actual amount used was 0.033 instead of 0.023, then the MPG would actually be 29.04MPG, so by being 1/100th of a gallon off, the MPG inaccuracy was 43% off.
|
|
|
My daughter is in 6th grade. Her science project this year is MPG vs PSI. My CX9 GT-AWD played an important role in this. Here is the prelim result. On a stretch of 2-mile local road w/o light/stop sign. Each data point was obtained by averaging four data points (back and forth twice to counter slop and wind effect). Vehicle was maintained at target speed with cruise control. 35 40 45 50 --------------------------------------------- 40 27.3 26.4 25.4 24.7 35 26.7 26.4 24.8 24.7 30 26.4 24.8 24.5 24.7 25 25.6 24.4 23.7 23.5 20 24.3 22.8 22.8 21.4 -------------------------------------------- ^ PSI (all four tires) - warm As you can see the CX9 is well-capable of doing 27mpg provided that - run tires at 40 psi (MAX=51psi), it is a bit bumpy... - maintain speed at 35-40 whenever possible on local roads. My daughter and I spent the entire afternoon for that data. BTW, the TPMS lights came on at 20psi (not 25psi or above). It went away when I pumped them up to 35psi (my typical) and CX9 reached 15mph from start. Also, a PSI drop of 40/35 down to 20 costs you ~10% drop in MPG. The MPG was based on my ScanGaugeII, which was +7% compensated. I have calibrated it for almost 1-year at the pump. It has been very accurate now (+/- 0.05 gallon error out of 17-18 gallons)
|
|
|
Replying to: bobw3 (Feb 10, 2009 1:02 pm) Yes, but basic statistics depends critically on using correct data. The departments of weights and measures in the various states regulate the accuracy of gas pumps to be about 6 cubic inches per 5 gallons. Since there are 231 cubic inches to the gallon, that works out to a relative error of just 0.5% per gallon pumped, not 20% If one is rounding off their numbers to the nearest 10th of a gallon then one is just being lazy. The price you actually paid for the gas and the posted price per gallon are sufficient to get an appropriately accurate measure of the volume of gasoline that you pumped. And, speaking of statistics, the "Law of Large Numbers" tells us that whatever errors you incur at each fill up will tend to the correct average over the course of time.  tidester, host SUVs and Smart Shopper
|
|
|
Replying to: ceric (Feb 11, 2009 11:13 pm) tidester, host SUVs and Smart Shopper |
|
|
Replying to: ceric (Feb 11, 2009 11:13 pm) It would be interesting (but very timeconsuming) to see measure the accuracy difference by only driving 0.2, 2, and 20 mile lengths. 40 27.3 26.4 25.4 24.7 35 26.7 26.4 24.8 24.7 30 26.4 24.8 24.5 24.7 25 25.6 24.4 23.7 23.5 20 24.3 22.8 22.8 21.4
|
|
|
Replying to: tidester (Feb 12, 2009 12:16 am) Correct, but that's only the "fuel actually pumped" vs "what's shown on the gas pump gauge accuracy" What I'm talking about is the accuracy of person's ability to fill up the tank to the same level every time when they fill up their tank and then calculate their MPG. Even if the pump is 100% accurate, the point at which it clicks and shuts off varies a lot, depending on the speed of the fuel flowing into the tank, position of the nozzle, etc. Your original question was why it made a difference if you calculated MPG on a 2 gallong vs 20 gallon fillup and that's what I was trying to explain. It's the inaccuracy of the fillup (not the pump) that creates the error, and that error (which relates to the Law Of Large Numbers) will create a larger MPG inaccuracy the less you pump. Another way to think about it is to imagine if you drove one mile, filled up the tank and calculated MPG. The drive 100 miles, fill up the tank and calculate MPG. Which do you think will provide a more accurate result? Yes, it's an extreme example, but it shows that when trying to increase the accuracy of an MPG calculation it's better to drive more miles. |
|
|
Replying to: bobw3 (Feb 12, 2009 10:57 am) The four numbers at the first row represents the MPG collected from various speed (35, 40, 45, 50 mph). Each data point is an average of four trials (back and forth twice). The ScanGaugeII computes MPG every 1 second or so. I reset the trip computer at the start and observe the average MPG at the 2-mile end. There are other factors that cause variations more than the ScanGaugeII such as the wind resistance. (on the day of our test, it was pretty windy) I also employed the cruise control to avoid human foot pressure variation on gas pedal. |
|
|
Replying to: bobw3 (Feb 12, 2009 11:10 am) Actually, that was Steve's question, not mine. What I'm talking about is the accuracy of person's ability to fill up the tank to the same level every time ... The information for doing it correctly is available based on what you pay and the price per gallon for your current fill up AND your previous one. Nevertheless, you have a point because most of us are not going to take the extra bookkeeping steps. tidester, host SUVs and Smart Shopper |
|
|
Replying to: bobw3 (Feb 12, 2009 11:10 am) Well, ok, but I live in a perfect world and Tides will vouch that my mpg spreadsheets are accurate out to 9 decimal points. 
|
|
You are here:
Forums
SUVs Mazda CX-9 Mazda CX-9 Real World MPG
New? Join Now!
Forum Tools
Search Forums
Browse by Vehicle
2010 Mazda CX-9
Browse by Board
Browse by Topic
