I'm a guy who has hand calculated (trip odometer and full tank gallons used) MPG probably a thousand times in my life.

Usually I run my tanks near Empty and then fill up. (Total miles per tank is a "fun" measurement too. It's cool to break 400 miles with my V10.)

Anyway, I've noticed a trend over the last couple decades of doing this. Whenever I fill up at half tank (or quarter: whatever) I regularly calculate a solid .5 to 1mpg increase.

Anyone else noticed that?

Thoughts? Explanations?

Also there's a regular on here that said he gets lower mpg every time he gets new tires... Obviously a totally separate discussion, but man that is sure strange.

 

Correlation ≠ causation.

Jason

 

^^

Perhaps I'm the rare guy here who does not calculate or even worry what my MPG's are or are not. I need to get from here to there for work so whatever it is it is, as the kids say. I do have a rough idea how much I've used relating to trips made and because I tend to work in the same locations repeatedly this is somewhat easy. Besides I don't wanna work that hard for something mostly out of my control anyway.

What I will offer for newbies looking in is NEVER let your fuel tanks go below 1/4 tank----its premature death to the in-tank fuel pump. Well, maybe not never but doing so regularly is a known cause of this failure.

Replacing the pump itself isn't a huge challenge or cost but everything else that breaks while doing this is the real aggravation. Keeping the 1/4 full tank has served me well over the years for three different E250's, all with the 5.4 motor.

 

Right, but I am seeking possible causes for the correlation. I thought that was clear in my post.

BTW, how did you make that "not equals" sign?

 

It's not hard at all (to calculate fuel economy). It literally takes a few seconds.

I presently fill my van 2 to 3 times per week. That's about 10 times per month. Based on the 1/4tank fill- I'd have to do it 3 more times every month. I don't want to increase my time spent at gas stations.

I do mostly agree with the old adage about running low causing premature pump failure. But not fully agree: Even at its lowest I can only squeeze in <33gallons of fuel into that 35 gallon tank. That's 2.5gallons of fuel that remains in place (in the worst case scenario). Imagine a 5 gallon bucket that is just barely over half full of gasoline. That's a lot of liquid. More than enough to dilute the small amount of contaminents, and more than enough to prevent the pump from running dry.

 

In 23+ years of driving, I pretty much have ALWAYS run my tanks until they're almost bone dry (when I had dual tanks, I would run one dry, then switch to the other). I've replaced a grand total of 2 pumps, I think, and one of those was in a single-tank vehicle that never gets run dry (but often well below 1/4 tank).

Anyway, the not-equals sign was a copy-and-paste. I was looking for the ASCII code to type it in (Alt-2260), but I couldn't make it work, so I just did the copy and paste. I may not have the right character set loaded in my OS and/or browser. The reason I said what I did is that 0.5 to 1.0 MPG isn't really significant--there are many variables that can affect mileage by a total of that much and more, from cleanliness of the vehicle, to temperature of the air, to amount of weight in the vehicle, tire pressure, and so on. Without controlling for all those variables, what you have is a hypothesis, and no more.

Jason

 

A hypothesis implies some sort of answer. I have not remotely implied an answer. Contrarily, I have repeatedly asked others for a hypothesis.

I fully understand the other variables.

I dont understand the repeated results that I described. You can deny the repeated results if desired, but my records and personal experience trump that.

In the last 50,000 miles of driving my van it ranges from 12.1 to 13.1mpg. Just as you stated: varying .5 to 1mpg is normal.
However it has gotten lower MPG when bigger variables apply (towing, extensive idling, higher octane fuels, etc) but it's ONLY increases have come from filling when less than empty, and from using ethanol-free fuel.

 

Two possible reasons come to mind, vapors and fuel pump.

Vapors, when the tank is low there is more area in the tank for fuel vapors and they are more sensitive to heat and pressure fluctuations. Thus when the tank has low fuel level, then heats up as the day warms, more vapors are pushed out of the tank via the vapor/breather system, these vapors are not used to power the vehicle and are wasted.

Fuel pump variation, it shouldn't be but may be that since I assume you use the number on the fuel pump to calculate how much fuel has gone into the tank, that the pump is more or less accurate when pumping small or large amounts. This may be more true on pumps with a single hose and 3 selectors. Not only cause of variation within the pump but cause if for example the guy before you bought premium and you buy regular then the volume in the pump and hose between the selector valve and nozzle is still full of premium, or vise versa of course.

 

Sorry, but your hypothesis is that there is an increase in mileage when filling a tank that is not empty--it needs testing to control for variables such as the ones I mentioned. I'm saying that an answer is pointless without controlling for those variables which would require much more detailed data than simply calculating mileage for each tankful (air temperature, tire pressure, vehicle load, atmospheric pressure, etc.). Correlation means they happen at the same time. Causation means one causes the other, which may not be the case here. It could simply be coincidence.

Vapors are controlled and burned via the charcoal canister, purge valve and vacuum lines. That's why a loose gas cap can light up the Service Engine light--it's an emissions issue.

Jason

 

To some extent yes, but enough to efficiently utilize all vapors as well as liquid fuel(to the injectors) is utilized, no.

OP, jroehl does have a point, though he has made it badly. For example if times when you take the tank close to empty are also times with more fuel demands, compared to times when you don't, then that will have an impact on MPG. Further example, commuting is hard, and your likely to take the tank close to empty during the work week. But a chill weekend is easier on MPG and your more likely to do a preemptive fill up.

 

Well, we shall continue to disagree. The vapor system keeps the fuel tank from pressurizing when there is a rise in temperature by absorbing the vapors, not by venting them. They are drawn off the charcoal canister later and burned when they are sucked through the vacuum lines into the intake manifold. The oxygen sensors detect the extra fuel and back off the injectors appropriately. It's a very effective system, described here:

HowStuffWorks "How Evaporative Emission Control Systems Work"

Those are also POSSIBLE variables, of which there are many (that's why I tacked that "etc" in there). I was merely pointing out the OP's poor application/understanding of the scientific method. He was asking for conclusions based on incomplete or possibly even irrelevant data. Again, only extensive testing would reveal whether there is a relationship between the fill level of the tank at refill and the mileage of the vehicle. Anything we say is a cause based on his claim of an increase in mileage of 0.5 to 1.0 MPG with a half-full tank is merely wild-@$$ guessing because the variables are not controlled.

Jason

 

" I was merely pointing out the OP's poor application/understanding of the scientific method. He was asking for conclusions based on incomplete or possibly even irrelevant data."
- It's not scientific Bro. We are Ford truck owners talking on a forum.

" Again, only extensive testing would reveal whether there is a relationship between the fill level of the tank at refill and the mileage of the vehicle."
- Extensive testing is exactly what I am claiming. Hundreds of fillups calculated over many years and 100,000's of miles. There is a direct correlation of partial tank fillups and higher than normal MPG. Maybe it's complete chance. Unlikely though.

"merely wild-@$$ guessing because the variables are not controlled."
- The variables are relatively stable. I get 12 to 13 mpg under normal circumstances. Under those same circumstances, and filling up a partial tank, I have repeatedly gotten higher than normal MPG. This ?phenomenon? is with my current van and with other vehicles I've owned.

"Sorry, but your hypothesis is that there is an increase in mileage when filling a tank that is not empty"
- nope you're completely wrong here. a hypothesis would indicate WHY this happens. What you described is the Conclusion (of the Scientific Method- look it up). What I did was Observe (look it up) a phenomenon and ask some other Econoline drivers what they thought.

"Correlation means they happen at the same time."
- True. And that's exactly what I've experienced (measured) repeatedly in recent years.

"Causation means one causes the other, which may not be the case here. It could simply be coincidence."
- Also true. But it doesn't apply in this conversation because I never claimed causation. I ASKED others for input... And yes, it could be coincidence, but the volume of evidence implies otherwise.

 

I'm saying your evidence is thoroughly incomplete and absolutely meaningless by itself. It doesn't matter what your data says, it leaves out way too many variables. Without additional data, you may as well ask a cheerleader how to change spark plugs in a diesel. I'm WELL versed on the scientific method. You are the one who really needs to be looking this stuff up. One person with mileage data (and only mileage) is NOT "extensive" testing. It doesn't matter whether we're shooting the breeze on a forum or not, if you want a realistic answer to your original question, it goes way beyond what anyone hear could provide, unless they have a bunch of additional data on all those other variables.

Jason

 

YOUR evidence is thoroughly incomplete. you are just a guy commenting on another guy's ACTUAL experience.

I am the guy who has done the work and noticed the evidence.

The normal variables (including everything you mention) provide my van with a range of 12 to 13mpg. This has been measured and recorded over the last 15 months and 57,000 miles. Upon filling up a partial tank, it shoots above that range on repeated occurences. Any scientist (like yourself) would want to study that correlation.

The thing you seem afraid to say is that you don't believe my claim... Dont talk to me about evidence and data and variables when those have already been covered.

The definition of "extensive" is as follows: covering or affecting a large area.
Most people would agree that 50,000miles is an extensive sample size.

Allow me to reiterate: "YOUR evidence is thoroughly incomplete. you are just a guy commenting on another guy's ACTUAL experience. "

 

For that to be meaningful data, those other variables need to be independently tracked, which you have not done. You have not tracked tire pressure, air temperature and all that. It doesn't matter that it gives you a normal range of 12-13 MPG. What combination of those variables gives you 12 MPG, and which combination gives you 13 MPG? You don't know, because that data does not exist. Your experience is purely anecdotal because your data is incomplete. Any scientist would quickly agree with my assessment.

Jason