Better Mileage with a Half-Empty Gas Tank?

I think it has to do with the fact that the fuel pump must work harder when the tank nears empty. The required vapour pressure in the tank is harder to maintain with less gas in the tank.

Also, as the tank nears empty, any sludge or water at the bottom of the tank will mix with the gas, and make it less efficient by both; taxing the fuel filter, and burning less clean.

You may have been misled by an old wives tale. The pick up in the fuel tank is always on the bottom of the tank. What ever is in the bottom of the tank is coming out first. And I don't think there is anything floating on top of the gasoline to cause the poor running.

And now for the reason the op gets better mileage when full then when almost empty.

The OP is using a fuel with alcohol. The fuel pump in the tank has a bad spot in the wire feeding power to the pump. The electricity is energizing the fuel pick up tube. The flow of electrons in contact with the alcohol is giving off hydrogen gas. the hydrogen gas is getting mixed with the gasoline going into the engine. And we all know that hydrogen gives better mileage. As the tank level lowers, there is less alcohol in contact with the fuel pick up tube that is functioning as an anode. That causes less hydrogen to be ingested by the engine, lowering the gas mileage.

That was so simple, I can't believe none of you thought of it. You all just kept quiet so as not embarrass the OP right?

I regret to inform you that we have sent someone out to break your legs.

What. Is it something I said?

You have not convinced me, sorry!!!!

Please let me know what pills you are taking, I need a few myself!!!

If you are using the car's gas gauge to determine when you're at half a tank, and empty, that's going to introduce a substantial error. They are just not that accurate.

I think your getting the same MPG ether way. One MPG doesn't prove anything. You'd have to do this experiment many times to really prove it.

Has it got something to do with the amount of fumes given off between full and empty, and in the case of empty, the greater space for them to evaporate into?

There are many factors that can influence mileage. One of them is the time of year and how much butane is in the fuel. Urban legend has it that the oil companies fill the fuel up with butane in the winter months so as to get rid of it from the refinery, and to sell a lower-grade fuel that improves cold starting.

Remember that a gallon weighs more when it's cold. Also, air is more viscous when cold.

At a given temperature, the vapor pressure is constant regardless of the volume. This is true of all liquids. For example, the vapor pressure of water is 23.76 mmHg at 25° C (77° F). This figure makes no reference to volume.

Bricktop is correct in saying that the accuracy of the fuel gauge is not constant, but varies depending on the fuel level. A more accurate fuel gauge (such as used in aircraft) measures the volumetric rate of flow through the fuel line. The rate of flow does not depend in any way on the fuel level in the tank.

One MPG difference in mileage is statistically insignificant for a standard automotive fuel gauge because it is within the error variance of the fuel-level sensor.

Everything you said is urban myths or very very old things that use to happen.

Post 1975 cars have pressureized fuel tanks, vapors don't leave. When you fill a tank, the vapors are absorbed into the fuel.

The composition of gasoline is set by LAW. Butanes can be added more in the winter to assure there will be vapors in the tank, this is called the vapor pressure. Iso-butane is added to increase octane and the vapor pressure, iso Butane cost more than regular butane and is 99% the cost of gasoline, so refineries are making higher grade fuel.

First, the viscosity of air at 90F is .019342 Cp and at 50 F its .018256 Cp, it thinner. and secondly, what does air viscosity have to with any thing about fuel milage or are your trying to say that cold air is more dense and that increase friction on the car as you drive down the road?

Bottom line, if anything you'll get better mileage by running the car dry, then filling it up 1/2 way. BECAUSE of the reduced weight.

YEEEESSSSS! At last!

Good lad.

Cheers,

Stu.

I think it has to do with fumes but they are not in the tank...

Very funny, took me a second to catch on (might have some fumes of my own to blame!)

Keep 'em coming, there is alway a place for humor on CR4.

Doh!

Quick! How much less does a half a tank weigh compared to a full tank?

@ Anonymous Hero

I think you perhaps misunderstood? What he's saying is that when he refills his tank at no less than half empty he gets better mileage then if he waits until the tank is nearly empty to refill. So it seems he gets better mileage when he's lugging around > 1/2 tank.

So in theory if he could attach a large enough tank to his car then it wouldn't use any gas at all.

But that silly thought aside...here's another one: Could it actually be a weight distribution issue?

Ah. I don't know, then.

I would first examine his testing setup for error. He is only looking at less than a 3% difference and we have no idea how the experiment is set up and executed for the experimental and control studies.

D'uh..

Think about it...fuel has weight...
Check your mileage with the boot (trunk) full of sand bags...then remove half of them and try again.

Not rocket science...hmm well it is actually 'cos rocket fuel is heavy too .

Del

"Check your mileage with the boot (trunk) full of sand bags...then remove half of them and try again."

But which half? Or does it have to be every other bag? Be precise man! We take our science seriously here at CR4. Apollo would never have made it to the moon if the astronauts had just pitched all their sandbags overboard at random!

Hello Guest,

The answer to your question should be self-evident.

The sandbags removed are the ones underneath the others.

The scientific and precise reason for that, is the lower sandbags are closer to the mass of the Planet Earth, thus actually weigh more than the bags of sand (further away from the mass of the Planet Earth), immediately above them.

You should always act carefully, and choose scientifically, when removing sandbags from the sandbag pile.

Kind Regards....

I missunderstood too...*
I think it's bollocks...

Best is to drive between empty and half full.

Del

* But only 'cos the thread title is missleading...it should read Better millage with full tank.

Nah!

Best is to carry around several one pints of gas in the trunk, so's he could run the tank always near empty, then he could just add them as needed to keep going. Save all of that fuel mass in the tank causing a problem, ie inertia on accelleration and momentum at slowing. (LOL)

Cheers,

S.

My calculation say thats about 3 and half tablespoons(.05 liters). I don't think they have made a gas gage in a car that could read accurately enough to distinguish that amount.

Time and temperature of the day when the filling was done would corrupt the results.

The shut off on the gas pump nozzle is not that accurate. Just the position of the nozzle in the fill tube off a few degrees could change when the pump nozzle shuts off.

I have been testing my mileage for quite some time now. Testing all sorts of ridiculous internet claims and uban myths on fuel economy. That being said, I have to do many base tests to keep a good record. One thing I have noticed using a scan gauge is that my tank average changes all the way to the end. Obviously at the end it changes much less but the average is indeed affected. Many people say they do a mileage test based off of about 50 miles of driving, then they top off the tank and divide the amount of fuel put in by the miles driven. This is a crap test because there is a huge difference in a 50 mile test and a full tank test. Over a full tank I think the average is just a better (more accurate) number). Also, one test and 1 mpg diference is hardly enough to say it is so. It is pretty interesting though. Keep in mind that every little thing effects mileage... EVERYTHING

My car gets substantially better mileage when the tank is half-empty than when half-full

Seriously, I would assume even though the tank is pressure sealed there have to be some vapor losses. Hey were only talking 3% here. It could be as simple as the time period difference between fill-up's.

As a control, test your mileage from half-empty (or even half-full!) to near-empty. But the trick would be how to make the test accurate. Try filling a known volume of gas and running until the car dies. (strongly recommended to keep a gas can in the trunk for this experiment).

Might it have something to do with the temperature of the fuel?

Many fuel injected cars have a pressure pump located inside the fuel tank. The design used the liquid fuel as a coolant for the pump. When the tank is more than half full there is more liquid and thus more cooling than when the tank is less than half full. Pump produces same amount of heat regardless of volume of fuel in tank. So a less volume (mass) of liquid results in less cooling. End result is warmer fuel going to the combustion chamber.

Hot rodders among the readership will likely remember the ice cans for fuel used on dragsters in hot climates. The idea being to deliver cold fuel to the engine because this gave slightly better power than if running with warm fuel. The ice can was simply a coil of fuel tubing inside a can to hold ice cubes ( or dry ice)

Actually, Delmar has the best suggestion for what will still be wild inaccuracies and conditions.

If you work, as I once did, in the industry dealing with dispensing gasoline there are cans called gauge cans.

They are used for instance, where such things are inspected, to pump a gallon of gas as measured by the gauge can so that the amount showing dispensed on the dispenser pump can be checked against.

Get a couple of gauge cans. Run your car dry. Pour an exact amount of fuel into the tank. Read your odometer. Start your car and then run it dry. That will give you accurate mileage per gallon.

Except that it won't. It will only give you accurate MPG over that particular piece of road and circumstances, temperature, humidity, etc.

There is a way to do it but it will be expensive. There are computerized test rigs that when installed will measure flow through the gas line and then integrate that with all the other factors.

This is of course about figuring out how to get the best mileage.

The best way to do that is to leave your tank dry.

j.

My car has a very non-linear scale on the fuel gauge. So the fuel level seems to drop rapidly when full, and then it slows down below half-way mark. This has nothing to do with the rate of fuel consumption. I think it is just that when the tank is full the float arm must be near horizontal and so a small change in level would cause a large change in the arm angle and when the tank is low the arm must be hanging down. The resistance potential divider which is used to sense the arm angle is unfortunately linear.

GA

Blink wrote: Elnav's theory on fuel temp is interesting, but the fuel has such a long dwell time in the fuel lines within the hot engine compartment, (and there are other things, such as the exhaust system, that heat the fuel tank) that I don't think additional heat from the fuel pump could degrade efficiency in a measurable way.

REPLY

?? long dwell time? Most of the fuel is returned to tank PDQ as part of the circulation. Only a small portion gets injected at any given time. My comment was based on observation. My F150 has dual tanks. On some hot summer days I get engine stumbling when tank goes below halfway mark. I switch over to other tank which up to that point has not been used during that trip.

Quite often the amount of fuel is very similar (somewhere midscale) on the fuel gauge. Switching to the other tank generally smooths out the engine. The effect is more noticable when in-use tank is lower than half way mark. I once checked the tank with my hand and found the in-use tank much warmer than the not-used tank. After replacing my fuel pump on another vehicle my mechanic advised me to never let the gauge get below half since this contributed to pump over heating and premature failure. Not surprising since most pump parts except armature and magnet is plastic.

Good post.

Stumbling, I'd think, should be independent of fuel temperature -- unless the temp is sooo high that you are getting vaporisation in the lines, despite the high pressure. F150's normally operate just fine in Texas in 110 degree weather, so I'd think that the stumbling might be from some other cause (maybe the less full tank has more condensation -- maybe you are peering in the tank too often, letting in warm, humid air?? )

I have not had a vehicle in which the fuel tank gets particularly warm, but the theory is interesting. Aside from the heat given off from fuel pump motor inefficiency (hopefully slight) there is the heat generated from endlessly stirring the gas via recirculation. I don't have figures in front of me for changes in efficiency with fuel temperature, but it would be interesting to research.

In any case, it's an interesting theory and great observation.

I guess the OP has not returned. When he does, we will all be surprised if he says something like "Oh, ya mean I can't use the fuel gauge as an accurate measure of fuel consumed?"

Then for reasons of longevity of the pump and safety, it sounds as though some US cars need to shield the fuel piping from too much heat!!!

By the way, it sounds as though you have a pump or filter problem if when the tanks get to half full the engine does not run smoothly......surely that could be fixed?

It's no wonder there is a shortage of feul as all these crazy tests are bing performed. Get a bicycle and cycle - that way you will save some of the fuel you have purchased and actually get some exercise

Hello elnav

<"....There is no one alive today who can remember the stench of horse s*** on the urban streets of yesteryear.....">

Excuse me, I can remember streets covered with steaming horse ordure, and the aroma on a hot day.

I'll be sure I'm not the only person here, who remembers that.

The method of transport is still used in some locations. (Not my present location, I hasten to add).

Concurrent with the horse transport here, were electric trams.

Kind Regards....

I'd love to do a run chart on the data to see it is statistically in control or out of control! (SPC)

[Edit: Boy, am I a nerd!]

Sure, but in a good way! GA.

Is your AX petrol or diesel?

On our Petrol Volvo 940 estate, we only get 25MPG (UK)

and on our V70 diesel, we get 46.7MPG (UK)

Petrol. Just did a couple of return trips to St. Davids (Pembs), some with only me, some with car loaded to the roof. One refill after 292 miles came out as 72 mph! Think the cutout on the pump must've tripped early. Can't tell from fuel gauge - it stays on FULL for the first 80-odd miles.

You will find the same thing if you travel with a passenger and then throw him (or her) out. If you would only put a gallon of gas in you would increase the mileage my two mpg!

The weight of the gas is the only cause!

If you read the OP carefully, you'll see the claim is for it to work the other way - i.e. better milage when the tank is, on average, containing more fuel.

The answer lies in realization of pressure, The more fuel in the tank the greater the pressure. the greater the pressure the better atomization, the better atomization the greater the mileage. experimental cars have been getting 100 plus MPG using greater fuel pressures to fully vaporize the fuel. The problem with this is catastrophic failure results in the bomb effect. We cant have cars exploding because of a backfire now can we?

For fairly obvious reasons, fuel injection pressure is regulated, and the fuel pressure delivered to the engine is not related to the fuel level (head pressure) in the tank. Increases in fuel injection pressure have little to no effect on engine performance in spark ignition engines like that in the OP's Mazda. Therefore, injection is at very low pressures in comparison to the very high pressures used in a diesel.

There are no experimental spark ignition cars (comparable in size to the OP's car) that are getting 100 mpg by virtue of higher injection pressures. High pressure injection has existed for many decades: imagine how profoundly dumb combustion engineers would need to be to retain low pressure injection in SI engines if HP injection could triple fuel efficiency.

In today's engines, essentially all the fuel that is introduced into the cylinder burns in the cylinder. Inadequate vaporization is simply not an issue. Thus, the many "better vaporization" devices (that claim to improve fuel efficiency) are all scams.

Many years ago, cars went from carburetion to fuel injection. There was a remarkably tiny efficiency improvement in going from 1. introducing the fuel at nearly atmospheric pressure to 2. injecting the fuel at small multiples of atmospheric pressure.

Good common sense Ken, thanks.

Ahhh, could it be that you are hauling around less weight?

By this, I should be getting peak milege in my Dodge 1500 van because I can hardly afford more than 1/4 of a tank at a time.

Travis

Its nothing complicated, but as many have already mentioned, using the tank gauge is patently wrong as they are simply not accurate enough. The gauge is just there to give you an approximate indication, no more no less.

I have had cars that used the "first half" slowly and others that used the "first half" quicker than the second half.........mileage wise of course!

I always let a car run out of fuel once (with a full spare can handy. By the way, its not even a problem to restart modern diesels now) just to see exactly where the needle stops. My present car ran for almost 70 Km saying that the tank was already empty!!!

OK, 1 MILE PER GALLON BETTER? 36.5 VS 37.4?

I would suggest that your driving up hill the get the fuel and down hill for your test.

The extra trip to the station cost you the most money!

I read all the responses, and I gave a GA vote to each one that mentioned the fact that your gauge is nowhere near close enough calibration to make this judgement. I check my mileage to and from work on a weekly basis, and I have done so over the course of a year using two different vehicles. One is a 4-cyl Buick, 12 gal (US) tank, the other an 8-cyl Dodge truck with a 25 gal (US) tank. They both exhibit the same syndrome. The "top" half of the tank goes in three days, the "bottom" half in two. The distance appears to remain constant. The gas is usually got at one station, usually the same easy-to-reach pump. This leaves the gauge as the sole remaining random factor, so far as I can see.

I've heard that auto makers intentionally design gas gauges to go slower through the "top" half to give the impression of greater gas mileage. That's why the "top" half lasts 3 days and the "bottom" half only 2.

Boy are you in trouble now! Where the heck is my GA vote? In my answer at post 21, I wrote: "It sounds as if you know not to use your fuel gauge as an accurate indicator of the amount of fuel used."

I think I should get 1.5 points, at least, because my phrasing is more tactful than accusing the OP of not knowing how to perform a fuel mileage test. From his two other CR4 posts, I've deduced that the OP is a process engineer, and probably knows better than to rely on a fuel gauge reading to come even remotely close to being able differentiate between 36.5 and 37.4 mph.

I've driven many vehicles in which a full tank is shown as noticeably beyond the F mark (true "Free Energy", yes?) and many in which the fuel gauge remains pegged at full for the first 10-20 miles, implying essentially infinite mpg during that period, if the gauge were to be trusted.

I imagine that many of the "gains" described in testimonials for all the many fuel economy "improvers" available come from people who measure mileage based on the first 1/8 or 1/4 tank used (down from full) as shown on the fuel gauge. I have driven several trucks (about moving van size) that appear to get 20 some miles per gallon on the first 1/8 of a tank (using the gauge as an indicator) while actually achieving about 6 or 7 mpg. One can see how someone who has just spent several hours installing a $1000 "HHO" device, could easily deceive himself. He fills up, and in a hurry to see what great gains are to be had, he uses the (apparent) first 1/8 or 1/4 tank to come up with a 200% improvement in fuel efficiency -- even though he may have actual historical records over several tanks indicating the "previous" 6-7 mpg. Very likely, utterly thrilled with the improvement, he gives up even bothering to record actual fuel use versus mileage -- and instead immediately comes to CR4 to post how his HHO device is causing phenomenal gains.

But I don't think that the OP falls into this class. I think he is probably recording actual gallons consumed and actual mileages, but erroneously concluding that a .9 mpg difference is somewhere close to statistically significant. Given all the variability involved, as clearly shown in JohnDG's excellent post, it is not.

Good point Ken. #21 was a very good answer. I have rewarded you so.

Umm…I don't see where gas gauge errors have to do with anything with the MPG calculation.

You calculate mileage when you refill the tank and take a mileage reading. You drive for some random time period and fill the tank again, hopefully to the same full level. Divide the miles travelled by the gallons pumped and you get MPG. The variable of the tank being at 5/8 or 3/8 full before refilling does not affect the calculation. No gauge reading required except for determining when to stop & fill up.

The original question was why the calculated MPG appeared to be less when refilling an empty tank than when refilling a half full tank.

Delmar, actually, you make a very good point here. However, the OP did make it sound to me that he's using the gas gauge as an analytical tool. After a reread or 2, perhaps I misunderstood.

The big picture, to me anyway, is that the mileage is statistically the same.

You appear to not understand the nitty gritty part, which is:- "How can you tell when the tank is EXACTLY half full (or empty if you are a pessimist!), when the fuel gauges are so inaccurate??????"

I hope this helps you in fully understanding the question!!!!

I wonder if you read the OP's post correctly: he is saying that with more weight he is getting better mileage -- which he correctly describes as counter-intuitive. I have a computer model I use for such purposes, (and which correlates well with real world figures from my own vehicles,) and in steady state driving, the additional weight of half a tank of fuel (about 40 lb for the Mazda) should reduce the mpg figure by .2 mpg (or in the perhaps more sensible European standard, increase consumption by about the comparable amount). In real world stop and start driving, the effect would be more pronounced -- but realistically, virtually impossible to measure reliably and repeatably.

GeeZZ!

It looks like I missed the point too.

Sorry Chaps. Must be this dammned gammy leg acting up distracting me.

CHeers,

Stu.

Have you ever checked the shape of your gastank? maybe the lower half is just smaller than the upper half.

When I fill my tank up I am so pissed concerned about the cost of gasoline that I am afraid to step on the gas pedal. By the time the tank is low I have gotten over the screwing transaction at the filling station, and I drive like an idiot again. Your results of course may vary.

it's really simple if you think of gasoline as extra weight. If you fill up a 5 gallon gas tank and walk with it a mile it starts to become heavy to carry after a while, but if only fill the tank with a gallon of gas it's much lighter.