From what I heard from a friend who contacted Ford, there are to two sets of data used to calculate the average MPG. It is my belief from comparing it to the SG, that at least one of the two sets is dropped or reset completely at around 200 miles. The reason I say this is because I seen my SG tank average continue to read correct while the nav sys average (both reset at fill-up) started climbing to 4-6mpg higher. It was a good mileage day and my today average was about 6mpg higher than my tank average on the SG. When I switched my nav screen from battery level to MPG, I saw the average tracking my today average almost tenth to tenth MPG up and down. I was on my favorite road on Jupiter Island and getting ~60mph on the current trip data on the SG while watching the Today and the nav average data match. That day I had traveled over 100 miles and the nav MPG average was within 2 tenths of the today reading. The nav sys MPG average was now 5 MPG higher than the SG tank average, and they were almost the same that morning.

Folks, I do some extreme driving changes where one minute I'm getting 36mpg and the next minute I find a road to exceed 60mpg. Watching all the MPG averages on the SG and comparing them to the nav sys average for two years, I think know when things suddenly change. On top of that, my tank averages exceed 600 miles at fill-up, so I see a change in the nav sys MPG average 3 times during one tank of gas.

My advise if you want know your average MPG by the nav sys, fill-up every two hundred miles and reset. If you don't want to do that, get a SG.

GaryG

 

GaryG -

I beg to differ with your assertion that the MPG computer resets automatically after approx every 200 miles. If this were the case, the avg. MPG readout would swing wildly at the point of reset - just as it does when I manually reset it. As noted previously, I reset the avg. MPG calculation field each time I fill up (while sitting in the gas station prior to moving my MMH). Initially as I drive away from the station the avg. MPG readout swings rather widely as the computer digests the variations in throttle position (MPG going down as I accelerate, going up as I back off the throttle and reach crusing speed). As I accumulate more miles on the tank (currently averaging about 430 miles per tank) the readout becomes less and less volitile moving only 0.1 MPG at a time as it begins to settle into what will prove to be the actual average for that tank (assuming road conditons, speed and driving terrain remain the same). Obviously, major changes in any of these factors can move the overall tank average significantly up or down even if one has already accumulated several hundred miles on that tank.

I have never observed any wide swings of the average MPG mid tank which would indicate an automatic reset!

Doug

 

I don't use that to calculate my MPG. The math is always more accuate. Although mtberman does have a point that to be "accuate" you would have to always fill up to exactly the same level in the tank.

 

The BEST MPG reading, hands down, is to do what I do.

Enter 100 tanks of fuel into the GreenHybrid database!
My ave MPG is as true as it gets.

Otherwise, consider ANY electronic display, a very close estimate.
The stock Ford system is usually generous by about 1 MPG on average.

On something short, say, 25 miles, the estimate will be +/- 5 MPG from true. On something long, say a 500 mile tank, the estimate is probabably +/- 1 MPG from true.

Over 50,000 miles ( by data table, or by hand )? Probably within 0.1 MPG of true.

Doing the math at the pump for an individual tank is, hands down, the WORST possible estimate of your fuel use. Math "by hand" at the pump can, and will be off by more than 10% each time. It is impossible to fill your tank more accurately than the electronic gauges report. You just can't. There is about 3 gallons of "air pocket" in your tank. And based on individual conditions each time, that air pocket will vary in size every time.

My advice from nearly 3 years of driving one?
Trust your electronic readings, and reset them at every fill.
The electronics will be far more consistent than any pump or human effort.
If you don't reset, the memory buffer will get full, and then you don't know what you are measuring, since data will be dropped at a rate you won't know about. I've seen no evidence of anything special happening at 200 miles. I have seen evidence that recent driving ( last 10 miles this tank ) carries more weight than the first 10 miles on that tank 2 weeks ago.

To me, it is unreasonable to assume Ford would have built a system that holds less than 1 tank of data. But at the same time, I know from experience, it does not hold much more than 1 tank worth.

-John

 

And take the temperature of the gas you get each time...

And ensure the energy content of the fuel is unvaried...

From wikipedia: http://en.wikipedia.org/wiki/Gasoline

Gasoline contains about 34.6 megajoules per litre (MJ/l) or 131 MJ/US gallon. This is an average, gasoline blends differ, therefore actual energy content varies from season to season and from batch to batch, as much as 4% more or less than the average, according to the US EPA.

I'll stick to the electronic readout.

Cris

 

Doug, I didn't state that all the data is reset at 200 miles, I said at least half the data was reset. Of course if all the data is reset, we would be seeing the same thing with the big variations like we do when we manually do a reset.

The point your making about the swing is what I was trying to point out. Once you have ~200 miles on a tank, the gauges don't change up or down near as fast. However, the nav sys did begin a quicker variation at some point at around 200 miles. This variation started matching my "today" average on my SG. This is why I think a large amount of data was dropped from my nav sys MPG average. I've known for a long time that the nav sys made this change at about 200 miles, but I didn't know it dropped as much data to cause it to match my "today" SG average. That's a lot of data at one time. At the end of the day, all the gauges had averaged pretty close to each other and that's why I thing you beg to differ from my observations. If you average say 34mpg, chances are you will end up not that far off at the end of a tank.

Now, what I've said here was documented during an entire tank at Hybridcars.com by me. It just so happens I was posting and reporting a change in my speed setting on my SG when I noticed this event. Check out post 15 and the others associated. You can see I documented my Dash, nav sys, SG "today" average and my tank averages.
http://www.hybridcars.com/forums/mil...t-t1578p2.html

GaryG

 

You just gotta love all this copious analysis over 1 or 2mpg.

 

It would be very odd for a programer to drop all of the data. Most likely the data goes into a queue with the oldest data point falling off the back to keep the queue at a pre-determined length.

A monitoring application I run keeps a data point for every second for 5 minutes. At 5 minutes is calculates the average for that minute and drops all of the data points for that minute and just keeps the average. When an hour has passed all the data points are used to calculate and average, all of the data points are dropped and the average is kept for 30 days. At 30 days the average for that hour is dropped.

I would not be surprised if Ford is doing something similar.

Also Gary and mtberman are comparing 05 and 08 navigation systems which I'm sure have undergone a code change.

 

Will, now that makes sense. Don't know if that's what Ford uses, but I was thinking along those lines.

It is obvious after a reset, and you drive away, it is recording samples about once per second. You get very rapid adjustments.

So, just a theoretical example here, tell me if this is possibly true...

The FEH may have two banks of 1k ( 1000 FDP )

FDP = Fuel Data Point
ADP = Averaged Data Point
WDP = Weighted Data Point
DDP = Displayed Data Point ( your running average on the screen )

During the first 1000 seconds, the FEH records 1000 Fuel Data Points.
During this time, you see an ADP on the the screen.
ADP = (FDP1+FDP2+FDP3+FDP4+FDP5.... / n)

After after 1000 seconds, this "bank" is full, and it saves a WDP to bank 2.
WDP = ( ADP x n )
Bank one is cleared, and starts over.
During this time, your DDP is a combination of the WDP and ADP.
DDP = ( WDP1 + ADP ) / (nw+n)
After 1000 more FDP, a second WDP is sent to bank 2, and the process continues.

Thus, you have "space" to save 1000 WDP and 1000 current FDP.

Each WDP can have a weight of 1000 seconds, or fraction there of.
Each FDP always has a weight of 1 second of drive data.

Overall, the computer could store 1000 x 1000 seconds of data with just
"2000 data points" or only 2k of memory.
This would be about 275 hours of drive time data.
This is approx 10,000 miles for an average driver.
Less miles if you do mostly city, more miles if you do mostly highway.

Is this possible with clever software and crude hardware?

Example:

First 1000 FDP recorded in first 1000 seconds of driving ave. to 35 MPG.
35 x 1000 or 35,000 gets recorded into register 1 of memory bank 2.

You only drive for 44 more seconds and turn the car off. Those 44 seconds ave. to 20 MPG.
20 x 44 or 880 gets recorded into register 2 of memory bank 2.

Your DDP or number displayed is 35,000 + 880 / 1044 = 34.4 MPG

You have recorded 1044 seconds, or 17.4 minutes of fuel economy, at a resolution of 1 second, but only used up 3 data points. Two for WDP and one for the runtime counter.

I'm no computer programmer, but that sounds BASIC to me.
-John

 

Sounds plausible but I don't think you would ever out right dump the first bank you would just right an algorithm that calculates the average in such a way that all data points in the first bank make it into one of the data points in the 2nd bank before falling out the back of the queue. You then shift all data points back one slot as a new one comes in. This way you always have a full first bank to do what ever you please with.

If I were writing it I would take the data points from the first bank average them out the same way I do when I'm adding a data point to the 2nd bank. I'd then take that value add it and all the values from the 2nd bank and divide to get the displayed average MPG. The first slot in the first bank would always be the current MPG,

But really with out documentation from Ford this is all wild speculation.