SparkFun Forums

[victorf]

I would like to measure the fuel level in a Hot Air Balloon's propane tank much as the gas level is measured in a car's gas tank. The propane tanks are normally 10 to 15 gallon affairs and contain cryogenic liquid propane. There are usually 2 or 3 tanks in a typical balloon basket. I would like to have a fuel gauge(s) that tells me how much propane remains in each tank.

Any DOABLE ideas?

Any enlightenment will be appreciated.

Vic

[wiml]

Wow, they're cryogenic? I assumed the gas was simply liquid under pressure at room temp.

How much does the weight of the tank change when it's empty? Could you simply put a weight sensor under the tanks, and somehow arrange for the hoses not to exert vertical force on them? Is frost buildup significant?

A simpler method might be to have a device that records the amount of time the burner is on and estimates the amount of gas used based on that.

[victorf]

wimi wrote:

Wow, they're cryogenic? I assumed the gas was simply liquid under pressure at room temp.

Wellllll I guess I was a little misleading there. The cryo temps come when the propane is vented or expanded in the burner. In general the tanks are not frosty.

Weighing is not an option because as the balloon accelerates up/down the precieved weight will change. Need something better than that I'm thinking.

Vic

[upand_at_them]

How about a pressure sensor before the regulator?

Mike

[wiml]

oh, okay

re pressure regulator: I think the pressure would be roughly constant as long as the gas is partially liquified. As you let gas out, more liquid boils until the gas pressure is back to the pressure that will keep the remaining liquid liquid...

Re weight: you could weigh it and attach an accelerometer, and compare the results to get the tank's mass. I bet it wouldn't be very accurate though.

I guess the tanks are metal, right? no chance of using a through-the-side-of-the-tank capacitive sensor?

[eugen]

Hi,

From home experience with butane I remember that the area of the cylinder that contatins the gaseous phase is colder than the area with the liquid as long as the gas flows. The system takes heat from the ambient for the phase change. The difference in Temperature was so significant that there was water condensation on the cold top area of the cylinder and nothing on the bottom area where the gas was liquid.
I assume that propane, although it has a different equilibrium gas pressure as butane, should behave in a similar way. The energy necessary for the phase change is taken from the ambient heat.

So it might be possible to fix an array of thermistors from the bottom to the top of the cylinder and determine in which height the temperature gets colder. I assume that there is a gas flow more or less all the time so that the temperature will not even out eventually.

Just another idea, not tested or verified...

Regards,

Eugen

[GeekHollow]

Ever seen those stick-on level meters at big-box hardware stores? They're just liquid crystal thermometers doing just what you suggested.

And as you said, they only work while there is gas flow. In a balloon though, isn't the gas only on for short spurts a lot of the time? And since the tanks are metal, I'm not certain that a short burst would create a useable temperature differential.

I'm thinking the mass detection idea could use more looking into.

Jerry

[eugen]

Mass detection, also called 'weighing' , has probably a lot of challenges on the hardware side of the project: The cylinders have to be placed free standing unsecured on electronic scales or alternatively free hanging suspended on spring scales, or mounted in specially designed holders where piezo scale elements are incorporated. Quite a challenge...

What about sending an ultrasound signal into the bottom of the cylinder? The signal will be reflected by the top surface of the liquid gas and the running time will give the level. Condition: Cylinders must be placed quite accurately vertical... and waves on the surface will invalidate the result...

Aren't there LPG gas flow meters? Count the revs and know the consumption?

[Philba]

The problem with a flow meter is you have to know how much fuel was in the tank to start. That may not be a problem for just filled tanks but what about partially full ones?

I don't know how the ultrasonic transducers will work at those temps. If you can modify the tank to use one of those, a better idea is to use capacitance which is how most modern car gas gauges work.

If you have 2 pieces of metal (say 2 rods), they have a certain amount of capacitance. If you insert the rods in a liquid, the capacitance goes up. In fact it is linear with the depth of insertion. Now, I think you could use the tank as one metal surface and a stainless steel probe as the other (not sure if you can electrically insulate it). You will get a characteristic range of C from the probe. For a given value of R, RC will give you a time constant that you can measure. You could use this as the basis for an oscillator (via a 555) or even just time it directly with a microcontroller. I've done this for water tanks - works pretty well.

The trick is figuring out how to get the probe in the tank so it is a) insulated, b) doesn't violate any regs and c) doesn't leak. The voltage and current can be kept very low so there *shouldn't* be a spark hazard. Tank attitude won't cause significant inaccuracy if the inner probe runs exactly down the middle.

More pie in the sky - I wonder if it is possible to determine the resonant frequency of the tank. More LP, the lower the frequency. Just tap a full tank vs an empty to see what I mean. That wouldn't entail modifying the tank or valves. Use a "speaker" to generate a broad spectrum "ping" and a "mic" to pick up the result. Subtract the ping from the result and you should see the resonant peak (via an FFT I think). speaker and mic are probably piezo elements. Problem is that each tank needs seperate calibration. Maybe more processing power than you bargained for.

Phil

[Ben]

Looks like there is a commercial product that uses a hall effect. Doubtful this would be easy to implement....

http://wirelessmonitoring.centeron.net/ ... itors.html

[sparky]

Neat trick with the centeron sensor. Requires a ferrous float though (if I'm reading it right) which may be difficult to get into a propane tank.

I know Micromotion (branch of Emmerson these days) sells flow meters based on a coriolis effect. As fluid runs through a bent pipe, a sine wave is induced upon the pipe. Depending on the frequency response of the sine wave, you can tell how much fluid (extremely accurately actually) is running through the pipe. No moving parts. Very handy.

I imagine there is a similar trick by 'tapping' the side of the propane tank. A fully tank sounds different than an empty one, no? So attach a small inducer to the side of the tank and a pickup on the other side. You should be able to pickup distinct responses for different liquid levels.

-Nathan

[bobledoux]

Sorry to come in late on this question

Are you finding the existing tank gauges are not adequate?

One of my balloons is an Aurora 54K with a single 20 gallon laydown tank. Because there is only one tank, I can't afford to run it dry. This makes for very conservative fuel management. With multiple tanks, the issue is less critical.

It would be possible to calibrate each tank. Starting with an empty tank on a scale, add a fixed amount of propane and record the gauge reading. Create a table of readings for each tank. I know some tanks do not report the full range of fuel content. You'd have to live with that.

With vertical tanks I think weighing the tank would be the best idea. Unless you are accelerating up or down, something a balloon does very slowly, the difference between empty tank weight and current tank weight would be an accurate means of measurement. This is the best way to determine fuel content when refueling. It would also be a good way in the air.

When the balloon is at a constant velocity up, down, or level, the weight would be accurate. I would set up a scale under a tank, in flight conditions and just try it. I'll bet you would have to burn or vent hard to see much change in the weight due to acceleration.

http://www.proaxis.com/~bobledoux/