I have no idea what you're trying to say. Everybody who uses the "by hand" method of calculating FE keeps track of how much they put in, and though I use "by dash" exclusively, I also keep track of how much the spinny-dials tell me I've put in. The FCD error is there, it's widespread, and it's consistent. How large a sample size do you need to admit something isn't right?

 

Hehe. Yes I can see you don't get what I am saying. There isnt anything to admit ( what does that mean?).



What I am saying is that the difference may be more noticeable on tanks that are more empty than others. Is that simple enough?

What I am saying is that the method the car uses to calculate dash MPG is flawed. I believe this may be due to the inability of the car to balance miles driven as tanks are emptied vs. how start up and quick accelration affects the MPG. You cant budge the dash MPG much at the end of a tank by driving ( even by hitting the 100MPG mark) but you can decrease it instantly by start up and accelarating at a stop. Its an imbalance between how MPG-dash is affected early on in a tank vs later.



SO-- X-Y graph. Gallons used on one axis. MPG/Dash vs Calculated number on the other. Plot it. This would help define whether or not what I am suggesting is true or false.

My guess is a flat graph until about 5-6 gallons of use.

mmhmm- whataver; we are just discussing reasons behind a known consistent discrepancy-- not debating whether one exists; hope this makes sense to you.

 

I think this behavior is just a function of the math. It is much easier to dump a lot of gas without traveling far (you can do this at any time just by stepping on the pedal). However, it is very hard to travel far without using gas. If you are towards the end of a tank, let's say 10 gallons at 50mpg, cruising in EV mode for a full mile would only increase your average to 50.1mpg.

If you want to test your theory, you could start filling up after 2 or 3 gallons for about 10 or so fillups. My guess is that you would see more noise in the difference between calculated and the dash, but the average would still be a difference of 1 to 1.5 mpg with the calculated being higher.

 

I'm with Kite... The whole "miles per gallon" calculation depends on two digits... absolute distance travelled (going in reverse counts as a positive number), divided by gallons of gas consumed. Since both numbers can only increase during the course of a tank, the only way you can possibly increase the mpg figure is by driving at a higher FE than everything that has come before (very difficult when you're already up in the 60s). Decreasing the number, though, requires nothing more than sitting at an idle without auto-stop, or performing an action that requires a significant amount of fuel (like a jackrabbit start). I would expect the miles traveled figure to be very accurate (and easily verified by driving a known distance, like between a couple of mile posts on the freeway), but the gallons consumed figure depends on a sensor we don't have access to, and may not be entirely accurate. It's this flow meter that I (and probably others) am pointing my finger at. Measuring fluid flow isn't the easiest thing to do, and it might actually come down to the engineer having spec'd "1VDC for every gallon per hour," and the sensor that was actually used was "1VDC for every liter per 15 minutes." The actual difference between one gallon and four liters isn't all that great when you're chugging drinks at the pool, but would certainly throw off a measurement like this.

 

I am actually thinking of trying this out. Just enjoying not having to visit the gas stations very frequently anymore though....

 

Actually, it's not a sensor, it is the engine control unit (ECU), counting the time that it keeps the fuel injectors open. The injectors are very accurate, at least on a new car. The only factors that normally affect this measurement are the voltage applied on the injectors, and the temperature of the fuel. Most newer models compensate for both.

I am too puzzled that soo many people report a consistent underestimation of the mpg. If anything, you would have expected a random error of perhaps one percent - when comparing different cars.

It would be interesting to know if the error is the same if you use metric units. From other discussions, such as about the 63-mile bug (100 km) and about displayed mpg, it seems that the unit uses metric units internally, then converts to US for the display. Maybe only the conversion factor is wrong?

I will track this for my car (always using metric display, for obvious reasons), but it will take a while to get data from several tanks.

 

Does not seem to be a metric conversion problem. Using the converters at http://www.easysurf.cc/cnver4.htm, I plugged in my trip b (lifetime) numbers of 8174.5 @ 4.7, and ended up with the same mpg/consumption as standard (4996 @ 50.1). The bug is elsewhere