My car says it has gas but I don't think it does. It won't let me put more in. What to do?

March 15, 2008 7:36 PM

My car, which is rarely driven, is unexpectedly acting like it's full of gas and will not accept more. I have no recollection of filling it with gas, though I do recall that the last time I drove it, it was hope-i-make-it-home almost empty. Am I just super forgetful? Or is there another explanation?

I don't drive my car on a regular basis. Maybe once per week or once every two weeks. About three weeks ago, I was driving the car quite a bit, to work, to a home, to another home, back to work, every day, for about a week and a half. At the end of that time, I was running really low on gas, and the last day, I made it home but felt I was really pushing it. I made a mental note that I *really needed* to fill the car with gas next time I drove it.

The car then sat for about two weeks undriven. In the ensuing time, it did snow quite a bit (between 6-12 inches) and I never even bothered to clean the car off. No other interesting events.

The next time I wanted to drive it, I took it to the pump (about 5 city blocks away), went to fill it. It wouldn't accept more gas (the pump turned off immediately). I thought the pump was broken, went to a different pump, tried more, and ended up with the gas spilling out all over the ground. Not knowing what to do, I drove the car back home, and at that time I noticed that the needle was past the F.

Still, I have absolutely no recollection of filling my car with gas. I do not remember driving it or any car since the day I made a note to fill it next time I drove it. My fiance also says he didn't do it (and he wouldn't have access to the keys, nor does he ever want to drive it). I don't know what's going on.

Is it possible that the gas gauge is wrong? And is it possible this might cause the car not to accept more gas? The car is somewhat old ('95 Corsica) and not in really the greatest shape. I personally wouldn't rule out multiple things going wrong with it at once.

The alternative is that I took the car out and filled it with gas and can't remember it at all. Honestly that would be pretty weird. I'd never take the car out *just to* put gas in it (I'm too lazy!), so that means I went somewhere and fogot that, too. Also, I don't drive another car, I simply don't drive very often. So it'd mean I forgot driving, forgot going somewhere, forgot why I went there, forgot stopping at the gas pump and putting gas in it while I was out.

Anyhow, hoping for an explanation that doesn't involve my idiocy.

Also, is there any way to figure out how much gas is in the car besides the fuel indicator on the gas? I don't want deal with siphoning it, I tried pushing on the back end and couldn't hear any sloshing. Any other way?

I'm really nervous to drive it, I am afraid it's totally out of gas and I'll end up stranded someplace.

posted by FortyT-wo to travel & transportation (42 comments total)

Did you try inserting a wire or thin stick down the filler hole? A dipstick like. How about rocking the car and listening for gas sloshing? By the way, thanks for the smile.
posted by telstar at 7:48 PM on March 15

Get a gas can, put some gas in it, keep the gas can in your car.

And stop worrying so much.
posted by pants tent at 7:49 PM on March 15

If the tank is so full that petrol comes out of the filler, *and* the gauge reads F, there's gas in the car. But carry a fuel can for a while, if it helps.
posted by bonaldi at 7:53 PM on March 15

You either filled it up and forgot, or somebody else who may have access to the car did for you.

Check your online banking balances for the transaction.

Really, there's no other explanation for this.
posted by davey_darling at 7:54 PM on March 15

If you know where the gas tank is, you can try to thump it with your finger, or to move it ever so slightly.

But yeah, what pants tent said. But don't buy the gas can at the gas station--gas station gas cans are always way overpriced.
posted by box at 7:55 PM on March 15

I don't know that a full gas can is the best thing to carry around in your car, fire-hazard-wise.

I know it isn't actually. But the dipstick idea sounds good.
posted by drjimmy11 at 8:05 PM on March 15

If the tank is so full that petrol comes out of the filler, *and* the gauge reads F, there's gas in the car.

Occam's razor says this is the right answer.
posted by number9dream at 8:09 PM on March 15

Oh, yeah, don't drive around with a full gas can in the car. Unless you're in the Paris-Dakar rally or something.

Also, it seems possible that something in the fuel-gauge system (in crappy old cars like the ones I've owned, it's basically just a floaty deal that moves around in the tank) is failing intermittently, and so, at some point in the past, the gauge reported the tank was empty even though it wasn't. The gauge being wrong wouldn't keep the car from accepting gas, though.
posted by box at 8:10 PM on March 15

The fuel actually spilling out suggests that there is something in there. If it were a mechanical problem the gauge might be off, but you could still put gas in. Perhaps you are not forgetting some random trip in the last two weeks, but just forgetting that you did stop on your last trip home after that busy week, thinking (as I often do) "I've got time now, but the next time I need to drive I might be in a hurry, since I have no idea when I'll need to drive next, so I'll stop now"?
posted by Rock Steady at 8:18 PM on March 15

So maybe 16 gallons for the tank = several containers of gas. Can the cap be removed without a key? How about a favor from (1) a practical joker, or (2) a secret admirer, or (1) and (2) combined?
posted by Kevin S at 8:19 PM on March 15

You have two independent confirmations that the gas tank really is full of gas. One, the fuel gauge. This is inside the fuel tank itself. The other is the gas pump, which measures air pressure to determine a full tank. Gas pumps aren't terrible accurate, but you mentioned that the gas overflowed when you tried to fill it further. The only way this could happen and the tank not be full is if there is a blockage in the line. But the gauge already says there is lots of gas in the tank itself. It's possible that there is a major blockage and that the gauge is defective, but that seems incredibly unlikely. It seems far more likely that you or your fiance simply forgot (or maybe you did it in your sleep).
posted by dirigibleman at 8:33 PM on March 15

Here's a distressing thought. Assuming that it is filled, maybe it's filled with something other than gasoline, like water from some kids having a little "fun." If subsequently you only have driven a few blocks, conceivably it was working off of gas in the fuel line. Additional driving and sucking water further into the fuel system wouldn't be so great for your engine. If this is a possibility (especially if cap doesn't need key or was forced off) then I think you need to take a sample by siphoning - or tow to a garage.

Hope it ain't so; maybe someone else can think of a good reason to debunk this scenario.
posted by Kevin S at 8:38 PM on March 15 [1 favorite]

If the tank was full of water, it would show full on the gas gauge.

I wonder if the filler tube to the tank is blocked, but it doesn't seem likely.
posted by jjj606 at 8:49 PM on March 15

I was going to suggest something similar to Kevin S. The reason being that several years ago some dumb fuck put sugar into the petrol tanks of our cars, despite that we lived off the end of a small dead end street, down a long driveway with the cars totally out of sight of the road. So whoever it was went to a fair bit of effort to do something so stupid. Fortunately they also left sugar on the ground so we had the cars towed and drained (cheaper than you'd think) rather than ran the sugar into the engine.

The first thing I'd do is look at my bank transactions, as davey_darling suggested, and maybe take a look at how much cash I have left. Try to rule out filling it yourself more definitively than just memory. Then I'd try the dipstick idea but using something like cardboard which is a bit absorbent. This will tell you if the tank really is full and hopefully the smell of whatever liquid you draw up will let you know if it really is petrol. And if you don't already have a key locked gas cap then look into buying one.
posted by shelleycat at 9:02 PM on March 15 [1 favorite]

Kevin S: Petrol floats on water, so your theory holds, er, nothing, being as all fuel take offs are at the base of the tank. You can't drive more than a few feet on the fuel in the lines. The rest of your 'answer' shows you have no knowledge of how cars work.

It is unlikely that the gauge reads full AND the tank won't take any more fuel without there being fuel in the tank. Chances are your memory is flawed with the amount there is in the tank. It blows my head up how this is possible, but hey. It takes all sorts...

Carry, as mentioned, a full can (5L) of fuel. Ignore the idiots that say it is 'dangerous'. They have no idea what they are talking about. Keep the can in the boot (trunk) and wait to see if the car runs out of fuel before the gauge gets to 'empty'. If it does, you have an issue with the car (which we can probably diagnose) if not, then you have an issue with your memory (bless).

It's really not that much of an issue. If you have enough fuel to get to the nearest fuel station in the car (in the form of a perfectly safe, sealed, can) then the risk is essentially zero beyond inconvenience. Fill it back up when it runs out and report the facts...
posted by Brockles at 9:04 PM on March 15 [1 favorite]

If the tank were full of water, it would have never made it to the gas station and back. Citing Occam's razor again: you filled the tank and forgot about it. It's been weeks, after all.
posted by The Light Fantastic at 9:05 PM on March 15

Occam's razor shows a simpler answer yet: you were simply mistaken about the amount of gas in the tank at the end of the week. Perhaps you read the temperature gauge as the gas gauge? (I've done that.)
posted by Netzapper at 9:36 PM on March 15

So it looks like I'm just goofy?

Regarding the amount in the car when last I drove it, I'd been driving over 25 miles per day for 1.5 weeks, I was out of gas and expected to be.

Had thought of the bank statement thing, but my fiance fills his car from our shared bank account, and at least once a week (we carpool, and he has a small gas tank, so normally *he* is driving the 25 miles + per day) so it'd be hard to pin down that any gasoline went into my car instead of his. If we drove more that week, he'd fill up more. We'd have a lot of trouble tracking it.

My car sits out on the street in an urban area. I park it in front of our home, but it doesn't move much, and so it's subject to some vandalism (It's been broken into and my CDs were stolen out of it before, and I got a dent in it a few months back that was way too high to get in traffic - looks like some kids either hit it with a baseball or a baseball bat up near the hood) and it's not a key-locked fuel cap, so I was kind of worried that someone filled it with water, or the melting snow might have done this. I won't say that anyone specifically hates me, but in my area, I have some neighbors that are a bit ... different. But I definitely drove it more than a few feet to the pump... was probably a mile round trip to the gas station. So it would appear I was running off gas from the tank. Which means there's gas in the tank and I'm just really forgetful.

I *wish* I had a secret admirer (and one so practical as to show his undying love for me by filling my car with gas!) but... no. Still, I'm pretty amazed at the fact that even knowing I must have filled the car up, I can't remember doing it at all. Weird.

Thanks all!
posted by FortyT-wo at 9:59 PM on March 15

One year I visited a friend far away. He had been battling with a condo neighbor. After my first night there, I couldn't fill the gas tank. I didn't understand what was wrong so I exchanged the rental car. The rental folks later found plastic grocery bags inside the tank.
posted by ick at 10:23 PM on March 15

Right after the gas thing, a bounty hunter hauled that neighbor away. He was wanted in another state for murder when he burned someones house down. Avoid crazy neighbors.
posted by ick at 10:36 PM on March 15

Could your fiance have filled it? Had to use your car for some reason, noticed it was empty, filled it up, and forgot to tell you about it?
posted by hattifattener at 12:09 AM on March 16

I don't have an answer for your direct question, but I just want to reiterate how unwise it is to carry a full gas can in your car. Several years ago I ran out of gas on a long drive. Following that, I thought it was a nifty idea to keep a gallon of gas in the trunk, just in case. I figured it was in an approved container, what could go wrong. As it turns out, I was very, very lucky.

Several weeks later, on the first warm day of the year, I finished work and went out to the parking garage where I had parked. I could smell the gas fumes from several hundred feet away. Even though the gas can had been closed tightly, the heat had caused a vapor leakage. All it would have taken was one small spark, like one co-worker having a quick smoke on the way back to his car, and that garage would have gone up like a bomb. In order to make my car usable again I had to hose out my trunk (which then got water into the padding of the rear seat in the passenger cabin, which I then had to pull out and have dried by an auto detailer).

Please, don't carry a container of gasoline around in your car for a prolonged period of time, it's just not a very good idea.
posted by Lokheed at 5:16 AM on March 16

The gas can might be a good idea if you suspect the fuel gauge is broken, but I would not carry a full can of gas as an everyday thing (unless it's one of those metal military ones with a bulletproof seal). A can of gas is not going to help if someone's poured water in your tank; any gas you add will just sit on top of the water, as Brockles pointed out.

There's another thing you should keep in mind: it's not a good idea to leave the car sitting with a near-empty tank for weeks on end. With changes in temperature, the tank may breathe in and out, and moisture can condense inside. That could make ice in your gas lines, causing the engine to quit (this usually happens after you've left home, thus maximizing the inconvenience.) Try to keep the tank nearly full. If you don't drive it for months, you might have to put a gas stabilizer in; gasoline degrades over time.
posted by Kirth Gerson at 5:45 AM on March 16

This is the type of gas can I'm talking about.

Tom and Ray of Car Talk disagree with what I wrote about moisture condensing in the fuel tank, but also disagree with each other about keeping it full to prevent rusting.
posted by Kirth Gerson at 5:57 AM on March 16

if all other causes for the 'full tank' has been ruled and it could be attributed to a faulty memory, then maybe a visit to a doctor for further investiagations is called for.

it seems quite unlikely that one could forget filling up a gas tank, especially since it entails several different actions that need to be done sequentialy. drive into a station, get out of car, slot in credit card, take off nozzle, put into tank etc etc. the only explanation could be that you had a mental blackout ( an amnesiac episode ). this would possibly indicate that some health-related event happened that caused the amnesia. a seizure , maybe ? or something else?

Amnesia in wikipedia.
posted by kryptos at 7:02 AM on March 16

I have to side with the car talk people about the 'fuel line icing' being myth, too. A tiny amount of air breathes in and out of the tank. Teeny tiny. And only a proportion of that will have moisture in, certainly not enough to cause an issue in any reasonable time.

That could make ice in your gas lines, causing the engine to quit (this usually happens after you've left home, thus maximizing the inconvenience.)

And how does it get past the fuel pump to displace the fuel already in the lines? I think this 'issue' perhaps relates to much, much older cars in days of poor tank design.

There is still a chance (albeit small) that the gauge and the breather (for instance) is broken, and there is not as much fuel in the car as it appears. So carrying the fuel can until this is established has zero risk. Carrying a full can of fuel for 'several weeks' is not so good, but for the duration of the tank dropping far enough to establish the issue is perfectly fine:

Drive the car until the gauge says half full (or three quarters) and put some fuel in - if the same amount goes in as is representative of half (or a quarter) of a tank? Problem solved.

If no fuel goes in? Still a problem.

If the car runs out of fuel before this point, with the gauge still reading full - put the can of fuel in and restart and take the car to be fixed.

Simple.
posted by Brockles at 9:16 AM on March 16

When you figure out the answer to the current quandary I suggest you adopt a habit I've used for years: every time you fill your tank, zero out the trip odometer. This accomplishes a number of things. You will soon figure out exactly how many miles you can go on a full tank of gas, thereby giving yourself a secondary indication (besides the fuel guage) of how much fuel you have left. In my case, I know that in everyday driving I can go 400 miles on a tankful. It also makes me acutely aware of my driving habits and helps me conserve fuel…if I see that the odometer is lagging behind the fuel guage I know that I'm developing a heavy foot and back off a little. Of course, in your current situation a glance at the odometer would confirm that you absent-mindedly stopped at the gas station and simply forgot about it.

One last hint: develop the habit of zeroing the odometer before you get out of the car to pump the gas. It's all to easy to forget once you've filled up and are ready to drive away. Good luck.
posted by dinger at 10:30 AM on March 16

There's another way that a near-empty gas tank can refuse to accept gas and yet the gas gauge reads "full." It's a pretty rare occurrence, so it probably isn't your car, but I mention it for completeness.

There's a tube that admits air to the gas tank as the gas level drops. If this tube is clogged tight, no air can flow into the tank. As the gas pump continues to draw gas from the tank, the inside air pressure drops. The gas tank, being made of sheet metal, eventually crumples from the pressure of the atmosphere on the outside. The gas gauge, inside a crumpled tank, no longer accurately reports the gas level.

There's an easy way to check this out: Look under the car, between the rear wheels. The gas tank is that rectangular metal thing a few feet on a side, and should have smooth, flat, sides.

And lest you doubt that this is a real possibility, it actually happened to a co-worker of mine.
posted by exphysicist345 at 10:45 AM on March 16

Along the same lines as kryptos' answer, if the tank is indeed full, you may have forgotten filling it because of an exhaust link into the passenger compartment.
posted by jamjam at 11:47 AM on March 16

The gas tank, being made of sheet metal, eventually crumples from the pressure of the atmosphere on the outside.

So a steel tank was crushed on the vacuum formed from an electric fuel pump? And 14psi of external pressure was enough to deform steel?

Utter rubbish. It may make the sides flex in a little, but they will pop back out when you take the cap off. Permanent, structural deformation of the tank is so unlikely an occurrence as to be absurd. It is much more likely that the tank was physically deformed (by driving over a kerb or a rock) than air pressure was anything at all to do with it. The blocking of the breather was either a result of the deformation or a coincidence. The idea that pressure did it from the breather being blocked is ridiculous. It would have stalled the pump well before even beginning to damage the tank.

Your example is flawed, or your co-worker was plain wrong.

Along the same lines as kryptos' answer, if the tank is indeed full, you may have forgotten filling it because of an exhaust link into the passenger compartment.

Nice hysterical leap. There is no evidence at all to support this, and I think that is panic mongering. Carbon monoxide poisoning has many more symptoms than pure memory loss - headaches, sickness, drowsiness etc - and so leaping to that (as opposed to getting the days mixed up or just not being as aware of the fuel level as they thought) is not helpful.

I think the OP needs to take the 'reset odometer' advice (I use that myself) and try to use that and the fuel gauge as a more effective method of knowing fuel quantity than purely trying to remember. The vast majority of people should be watching your gauges more than they do anyway.
posted by Brockles at 12:00 PM on March 16

The theory of the crushed fuel tank may be a bit far-fetched, but the idea of air pressure crushing a metal tank is entirely plausible, and it doesn't take anywhere near 14 psi to do it. A differential of only 8 psi acting on a surface area of two square feet generates nearly one ton of total pressure, and a fuel tank is not built like a scuba tank. Let a small pressure act on a large enough surface area, and you can create tremendous force…why, you might even be able to raise a car or truck up on a garage lift. Naw…on second thought that could never happen. Hydraulics would never work.
posted by dinger at 2:31 PM on March 16

Correction…it only takes 7 psi over a two square foot area to generate a ton of force. My bad.
posted by dinger at 2:40 PM on March 16

Drive the car until the gauge says half full (or three quarters) and put some fuel in - if the same amount goes in as is representative of half (or a quarter) of a tank? Problem solved.

You have a lot more confidence in the uniform accuracy of fuel tank gauges over their span (E to F) than my experience will allow me. Every car I've owned has had a gas gauge that went down at a radically different rate when nearing empty, compared to near full. Mostly, they dawdled down to somewhere around 1/2, then approached Empty at an increasing rate. From that, I conclude that they are not useful indicators of the actual contents of a tank, or of what percentage of a full tank is remaining or used up.


Utter rubbish. It may make the sides flex in a little, but they will pop back out when you take the cap off. Permanent, structural deformation of the tank is so unlikely an occurrence as to be absurd.

Maybe not. Coincidentally, today on MetaFilter.
posted by Kirth Gerson at 2:40 PM on March 16

Heh -not today.
posted by Kirth Gerson at 2:51 PM on March 16

My prediction: you're going to have a head-slap moment when you remember that yes, you did drive the car 25 mi/day for a week and a half and yes, you were nearing empty, and yes, you were thinking that the next time you drive, you have to remember to fill up, but then, maybe the last day you drove, on the way to where you were going so you had a whole day's worth of activity to drown out the memory, you decided to assuage your anxiety over it and just fill 'er up. You spent so much time worrying about how low you were getting that you forgot you fixed the situation at the eleventh hour. :)

The moment will happen 2 minutes before you would otherwise fall asleep, in one of those suddenly bolt upright situations. And then the realization will keep you up for an extra hour and a half.
posted by iguanapolitico at 5:43 PM on March 16

You have a lot more confidence in the uniform accuracy of fuel tank gauges over their span (E to F) than my experience will allow me.

Accuracy, not necessarily, which was why I used 'representative'. Rarely does a linear rate be used - possibly deliberately, all the fuel gauges I have used seem to have the halfway mark somewhere nearer the 3/8 full (as you say). Consequently, the second 'half' makes the gauge move faster than the first. I always imagined this was so that the needle was more likely to attract your eye when you were running out. However, they were all most certainly repeatable, and the graduation marks still lined up pretty damn closely (needle parallax notwithstanding). My regular drivers have been mostly either German, or using German gauges (plus some Smiths gauges), but even so I'm surprised that you have experience of random ones - I have driven a LOT of different cars and not noticed a random element at all - guessing the amount to fill up hire cars has always been easy, for instance. Even my 1991 Crown Vic (a perfect example of a piece of shit with cheap-arse gauges) takes 50litres +/-1 litre every time when the light comes on. It will certainly be 'roughly' the right amount by enough to work out if the gauge is working correctly.

but the idea of air pressure crushing a metal tank is entirely plausible, and it doesn't take anywhere near 14 psi to do it.

14 psi is the absolute most that could theoretically be produced. A car fuel pump is incapable of creating a perfect vacuum (or anywhere close), so it is unlikely that a sizeable pressure differential could be created. Also, the fuel tank is a fairly small box, with many strengthening elements (pressed into the flat faces) so there isn't a sizeable flat face weak enough to act on. They're pretty damn strong. I've had fuel cells with blocked breathers that have popped the cap out of my hand (after filling) with a pop, but it has never damaged anything or even deformed it permanently.

jamjam: This has nothing to do with my job, but more to do with my experience of road cars. Also, my experience with Carbon Monoxide - it is unlikely that purely memory loss will occur - as your example clearly states (and as mentioned in my original post) the more likely symptoms are headache and sickness. Jumping from 'I don't remember filling my car up' to "carbon Monoxide poisoning" is a step too far and would easily scare someone. It is important that they know a bit more of what to look for than "Yeah! It can cause random memory black outs!!". If they know it also causes other things, they may be better armed for spotting any potential likelihood of the issue.
posted by Brockles at 7:26 PM on March 16

...the graduation marks still lined up pretty damn closely (needle parallax notwithstanding).

Not my experience at all. The "1/4 tank" just below Full disappears much more slowly than the "1/4 tank" just above Empty. Nor am I saying they're "random," if I understand what you mean by that. They are repeatably inaccurate, in the sense that the indicated 1/4-tanks are not even roughly equal. I have had a couple of German cars, a bunch of American ones, and a couple of Japanese ones. None of them, so far as I can recall, had a fuel gauge that accurately showed the last 1/4-tank as producing as many miles as the first1/4-tank. I've always thought it was a product of the geometry of the float arm arc. They use very different systems on, for instance, oil-tanks and tankers, which aren't constrained vertically the way car fuel tanks are.

...is unlikely that a sizeable pressure differential could be created.

It doesn't take a sizable pressure difference. Please look at the Mefi thread I linked to, or at least this picture of the subject of that thread. "These ... are rated to a test pressure of 100 psig and a minimum burst pressure of 500 psig. Minimum carbon steel plate thickness is 7/16 inch [11.1 mm]." The pressure differential was produced by steam-cleaning it, then closing all the vents to atmosphere. As you say, the maximum possible pressure difference is about 15 psi. There are a lot of si's in a fuel tank.
posted by Kirth Gerson at 8:40 AM on March 17

Not my experience at all.

Really? That's weird.

None of them, so far as I can recall, had a fuel gauge that accurately showed the last 1/4-tank as producing as many miles as the first1/4-tank.

Ah, maybe we're getting somewhere here with the difference in perception/experience - but did the last 1/4 tank produce the same amount of miles as the previous 'last 1/4 tanks'? So rather than being repeatable across a tank full (4 parts of 100 miles, say), are they repeatable for the same range across tanks (does that make sense? The last 1/4 tank is always '75 miles', even if the first one is 120).

They use very different systems on, for instance, oil-tanks and tankers, which aren't constrained vertically the way car fuel tanks are.

The size/shape of the tank is pretty much irrelevant, though - they are either calibrated or they're not.

As you say, the maximum possible pressure difference is about 15 psi.

(I read the thread and the linked text). Actually it isn't necessarily so in that example, as the tank will have been heated through cleaning, expanded, and then closed and contracted overnight. So part of that failure (perhaps the initial start?) may be due to the heat cycle, too, and simply evacuating the contents is not necessarily the sole source of pressure/loading.

Regardless, that is a massive volume of air/steam. Fuel tanks in a car really aren't that big, and the only thing that can generate anything approaching a vacuum is the fuel pump (by removing the contents) but they just aren't strong enough to produce much in the way of negative pressure- this was my definition of sizeable (rather than 'a big number'). If the pump can't generate the force required, it can't happen as the pump will stall.

I've pressure tested aluminium oil tanks that could take 15psi without deformation. I've also had people fill fuel tanks (from empty) without connecting a breather (the exact opposite of the example) with no ill effects to the tank at all. In all my years of dicking around with various cars, I simply can't see an instance where a: this vacuum can be created by the system, and b: it being enough to damage such a small tank. That train car has a large, unsupported area compared to a small, squared off fuel tank with many strengthening ribs.

It is possible, perhaps, that attaching a vacuum pump to a fuel tank may cause it to buckle or deform, but the system doesn't involve a pump strong to do that in a car.
posted by Brockles at 3:35 PM on March 17

...but did the last 1/4 tank produce the same amount of miles as the previous 'last 1/4 tanks'? So rather than being repeatable across a tank full (4 parts of 100 miles, say), are they repeatable for the same range across tanks (does that make sense? The last 1/4 tank is always '75 miles', even if the first one is 120).

Yes, that's exactly what I mean. Given that, you cannot say, "The fuel gauge's indicates the tank is down by 1/4, so if I put 1/4 of the tank's capacity in, it will be full." If that's not what you meant by the "Drive the car until the gauge says half full" part of this comment, then I'm not following you at all. Half full as indicated by the gauge is unlikely to be half the full capacity of the tank.

The size/shape of the tank is pretty much irrelevant, though - they are either calibrated or they're not.

No, it's not irrelevant. Deep tanks use a linear gauge that measures straight up and down; shallow automotive tanks use a float on the end of an arm that pivots. Since the base of the pivot is at the top of the tank, the float's movement is more vertical near Full than near Empty, where it has some horizontal travel. That's where I think the difference between the first 1/4 and the last 1/4 comes from.


(I read the thread and the linked text). Actually it isn't necessarily so in that example [that the maximum pressure differential is 15 psi], as the tank will have been heated through cleaning, expanded, and then closed and contracted overnight. So part of that failure (perhaps the initial start?) may be due to the heat cycle, too, and simply evacuating the contents is not necessarily the sole source of pressure/loading.

Yes, it is necessarily so, as you said in the first place. Absolute, perfect, zero Torr vacuum in a vessel on Earth is going to produce maximum one atmosphere of pressure on the outside. Your experience with pressure-testing fuel tanks is precisely congruent to the tank car being tested to 100 psi. A vessel that can stand internal pressure is not necessarily going to withstand external pressure, even of a much smaller magnitude.

Try this: Take a one-liter plastic Coke bottle and pump it full of air at 80 psi. It will hold that easily. Now let that air out and suck as much of the remaining air out as you can, using just your mouth and lungs. The bottle will collapse like that tank car. Now find a vacuum gauge and see how much vacuum you can apply to it using your mouth. If you're very strong, you might see 3 psi. Vessels that can withstand external pressure, like deep-sea submersibles, are massively reinforced, compared to vessels designed to hold in internal pressure. Auto fuel tanks are not reinforced against external pressure.

I suggest that if you have a spare fuel pump and an old gas tank, that you fill the tank half full with water, seal it up, and pump the water out. You may be surprised.
posted by Kirth Gerson at 5:57 PM on March 17

No, it's not irrelevant.

Yes it is, either they are calibrated, or they are not. I'm not sure how your post doesn't agree with that statement. It's not about the shape at all - it's about whether thay have been calibrated. And being as most fuel gauges (despite the massive, rather than a function of the tank. I can't find any sources for that, though. Regardless, it is either calibrated (ie accurate) or not.


"Drive the car until the gauge says half full" part of this comment, then I'm not following you at all.

Again, you are completely ignoring the fact that I said representative of half a tank - as in 'representative of whatever topping up the tank is in your car from the half full mark'. I thought that was pretty clear. It only needs to be very rough to ascertain if the gage is working and +/-5 litres would be fine anyway.

I suggest that if you have a spare fuel pump and an old gas tank, that you fill the tank half full with water, seal it up, and pump the water out. You may be surprised.
posted by Kirth Gerson at 8:57 PM

Based on one, possibly flawed, example that could have many other explanations (And hasn't been defended, I maintain on my experience of the design and capabilities of automotive fuel pumps that it cannot create a negative pressure sufficient to damage a fuel tank. "I think it could" won't sway me based on zero evidence to support it.

Absolute, perfect, zero Torr vacuum in a vessel on Earth is going to produce maximum one atmosphere of pressure on the outside.

No. The volume of the tank changed due to the temperature. There is more here than a simple evacuation of the tank. It may be less, it may be more. Because the volume of the container changed by a tangible amount, it is not correct to assume that the maximum pressure is one atmosphere. As I said, the evacuation of the tank may not have been the sole source of pressure change or loading.
posted by Brockles at 7:05 PM on March 17

No, it's not irrelevant.
Yes it is, either they are calibrated, or they are not. I'm not sure how your post doesn't agree with that statement. It's not about the shape at all - it's about whether thay have been calibrated.

It's because the vertical gauges used in deep tanks are linear, and the ones used in cars are not. Linear gauges have consistent increments; pendulum types like those used in cars don't. I also don't believe automotive gauges are calibrated to read consistent increments over their range, at least not in consumer vehicles.

No. The volume of the tank changed due to the temperature. There is more here than a simple evacuation of the tank. It may be less, it may be more. Because the volume of the container changed by a tangible amount, it is not correct to assume that the maximum pressure is one atmosphere. As I said, the evacuation of the tank may not have been the sole source of pressure change or loading.

Please tell me just what in any of that is going to produce more than one atmosphere. Are you claiming that somebody dropped a big rock on the tank, or put it in a hyperbaric chamber, or what?
posted by Kirth Gerson at 8:44 AM on March 18

Please tell me just what in any of that is going to produce more than one atmosphere.

Because the volume of the tank changed. The system changed. So, as I said, there may be stresses other than that of purely the pressure. Pressure relates to volume, and the system wasn't necessarily constant. I said it may go up or it may go down, but my point was that we cant' just say ' max force relates to 1 atmosphere' as there are other considerations. The expansion and contraction of the tank (outside it's usual expected temp range, perhaps) may have induced stresses. It's not as simple as 'air inside vs air outside'.

I also don't believe automotive gauges are calibrated to read consistent increments over their range, at least not in consumer vehicles.


So you are agreeing that they are either calibrated or not (as I was saying) but that you just don't believe that they are. I very much suspect that they will be, just not within any fine degree of accuracy. In fact, I'd bet large sums that they have been checked in production to reasonably demonstrate a given volume (=/- a nominal amount) of the contents of the tank.

pendulum types like those used in cars don't. (have consistent increments)

Yes they do. They are entirely consistent, just not linear. One does not equal the other.
posted by Brockles at 9:16 AM on March 18