How much liquid co2 in a tank?

[Wat]

Over at docsmachine we're trying to figure out, in a full 20oz co2 tank, what percentage by volume is liquid vs gas. I checked my thermo engineering tables for the data (wow, dust). But i don't believe my numbers. Can someone proof?

According to my math

Volume of liquid = (Total Mass - Total volume * density of gas)/(Density of liquid - density of gas)

Density at 25C of
Liquid CO2 = 709.7kg/m^3
Gaseous CO2 = 236.3kg/m^3

Total mass = 20oz = 0.56699kg
Total Volume = 850ml = 0.00085m^3 (doc measured this by filling the tank with water its about 51ci excluding the neck).

Plugging this all in
Volume of liquid = 773ml
or
91% of total tank volume.

That somehow doesn't seem right. It seems rather high. For you math geeks, i derived the top equation from these two:

Volume liquid + Volume gas = Total tank volume
Volume liquid*density liquid + volume gas*density gas = total mass

[Pretzel Boy]

I am not sure here but i think that the "warmer" a tank is the more gas there is, as opposed to a freshly filled "cold" tank. I know that my mag doesn't operate as well(at all) on a freshly filled tank. I think it is because there is more liquid in there and less gas. And as the tank gets closer to room temperature the ratio changes.

[314159]

if you have a co2 tank without a anti-siphon or siphon tube, some clear macro line, and a thermomiter.

you could weigh the tank when empty, and after the fill, and do some math to get the weight of the co2.

measure the temperature.

sloly tilt the tank, see what angel you need it at to get the liquid co2 in the clear macroline.

use the angle that the liquid co2 starts to enter the macroline, and the geometry of the inside of the tank. to find the volume of the co2 at a given temp.

[314159]

you could also find the amount of co2 that can exist in a bottle with no liquid. maby this will help you with your above calculations.

weigh empty tank.
use a gun with clear macro line.
don't fire rapidly (don't chill the tank too much)
fire the gun till no liquid appears in the macroline.
weigh the tank

[hitech]

I think your "math" is wrong somewhere. As i remember it the "average" CO2 tank is filled with ~60% liquid.

[Wat]

not to be rude but

The point of this excercise was to find out if all those rumor numbers floating around of it being 2/3's full was right or not.

Merely saying the math is wrong because it doesn't support the hypothesis of 2/3 full isn't very helpful.

If the math is wrong, show me where as i've gone over the algebra a couple of times and haven't found a problem. I also went over my thermo engineering tables and haven't found a problem either. If my data points are right (i'm certain they are) and my math is right (again, i can't find a problem) than my only conclusion is that tanks hold a lot more liquid CO2 than originally thought on a volume basis or that 20oz co2 tanks are rarely filled with 20oz of co2.

[bjjb99]

We're dealing with saturated carbon dioxide gas (a liquid and gas phase in which the gas phase has an effective "CO2 humidity" of 100%). Perry's Handbook for Chemical Engineers lists the following data for saturated carbon dioxide:

At 300K (an 80 degree day),
P = 67.10 bar (973 psia)
vf = 1.470E-3 m^3/kg (fluid density 680.3 kg/m^3)
vg = 0.0037 m^3/kg (gas density 270.3 kg/m^3)

Your algebra looks sound to me. Your numbers are pretty darn close to mine as well, and you're using a slightly different temperature.

Using the same formula you used and the above values for the gas and fluid densities, I get a fluid volume of 823 ml and a gas volume of 27 ml when the tank is completely full. This puts my figures around 97% liquid volume for a complete fill, with the remaining 3% volume as a gas... certainly a far cry from the "2/3, 1/3" rule of thumb that's tossed around. Somehow I think the method used to generate the 2/3 full figure is flawed. Since I've never seen a description of the actual method that generated the 2/3 figure, I can't identify what might be the problem.

BJJB

[bjjb99]

Something else to consider...

The fraction of liquid contained in a full tank is going to vary depending on the tank's internal volume. Some 20 ounce tanks might be more than 850 ml on the inside, which would reduce the liquid fraction. The 2/3 figure might have been true for the old 7 ounce tanks used back when the rule of thumb came into practice.

It might be worthwhile to measure the internal volumes of several different tanks (3.5, 7, 12, 16, and 20 ounce) from several different manufacturers to get an idea how the internal volume varies with respect to how much CO2 the tanks is rated to contain. For that matter, toss in an internal volume measurement for the old 12-gram CO2 powerlets as well.

BJJB

[mongrel]

i fill a 20 oz tank with 16-18 oz of liquid co2.
liquid co2 will expand 30 times its size from liquid to gas.
when a tank gets warm (IE setting in the sun)
some of the liquid changes into gas the result is more gas in the tank a higher pressure. when a tank get cold the gas some of the gas changes back to liquid so a lower pressure in the tank. that is why the tanks arent totally full or the blowout disk would blow as soon as the tank gets warm.
hope this helps.

[Doc Nickel]

Thank you, Mongrel, but that's not what we're after.

Simply put, a "full" 20-ounce tank- as in one that is filled to weigh twenty ounces more than it's empty weight- is not totally full of liquid.

This "vapor bubble" is there to allow the CO2, a very dynamic gas, to expand and contract with temperature changes. If the tank were actually "full" of liquid, a slight rise in temperature would cause the liquid to try to expand. Without the room provided by the vapor bubble, the expanding liquid would increase the tank pressure dangerously.

All that is a given.

The question here is, just how much is liquid, and how much is gas, or vapor, in a "full" tank.

Now, in an article I read, written by Bud Orr, in Paintcheck Magazine back in 1988, he stated the tanks were filled to roughly 3/4 capacity with liquid, and the rest was gas. He further stated that a 7-ounce tank could theoretically hold 10.3 ounces, assuming no vapor bubble.

I've read the same "two-thirds to three-quarters full" statement many other times and from many other sources. I never had a reason to question otherwise.

However, Warpig's article on CO2 dynamics, which I'd read but not fully comprehended several years ago (and never really went back) states- and gives at least some of the math- that the tank is in fact only about ONE-third full of liquid, about 34%.

I'm quite bad at math, at least anything above simple arithmetic, so Wat and a couple of other Guild regulars have been checking and recalculating.

So far, what they've come up with is a little off from both the old "foklore" number, AND the Warpig article numbers... So we have 34%, 66-to-75%, or around 90%... who's right, and how do we verify it, short of having a clear tank made?

I my ham-handed way, I've been thinking.... I have a stainless steel tank, which used to be a "sample bomb" for in-line sampling of anhydrous ammonia. Anyway, it's a thickwall stainless cylinder, with a pipe-thread fitting at both ends. If I can dig up a length of the aforementioned clear microline, I can possibly rig up the tank with a microline fitting at both ends, and have the clear line in between as a "sight glass".

I'd have to determine the volume of the tank, first, plus that of the various fittings (there'd be a few) and come up with some way to actually get CO2 into the tank safely (perhaps a custom fill nipple?) and a way to bleed it out afterward (the tank has pipe fittings, it won't take a conventional pinvalve) but it might work.

All I'd need to do is get a reasonable level idea- X number of ounces equals approximately Y level. The differences between 33% and 90% will presumably be quite dramatic. :)

If the "bomb" works, it might be interesting to see the level change with temperature, too... it's still getting down below zero F here at night, so I could give a quickie check between awfuldamncold and room-temp... Throw on a guage... more fittings.... Getting complicated.

Stay tuned.

Doc.

[314159]

as far as volume goes, you could shove a plug in one end.

put some type of float in the tank that would fit in the opening, with a graduated portion "ruler" sticking out of the top, or you could just measure it.

this way when you add a given amount of watter, you can see how high the float rises (minus the volume of the float below the liquid, but that should be a small volume and could be ignored for a fairly good rough estimate) , and get the volume at that point.

[314159]

why plug up the bottom when you could leave the macroline and fittings hooked up to the bottom, add a given amont of water, and see where the water rises to on the macroline and mark it.

[cphilip]

Originally posted by Doc Nickel
...but it might work...If the "bomb" works...
Doc.

Been nice knowing you Doc. Should be a Blast!

At the risk of sounding like a gussher...seriously this is so cool of an idea to do. But please be carefull. But I just gotta know the answer to this all now. So I am willing to let YOU take the risk.

[CameraGuy]

Why not check the liquid CO2 level the same way that the liquid propane level is checked in propane tanks? (in the hopes that it this might prolong Doc's life...)

This is how it works: you buy a temperature sensitive adhesive strip (of the same type used for BBQ tanks, I've got one on mine) and attach it to the side of your CO2 tank, running from the valve to the bottom of the tank. Set the tank upright (probably easiest with a flat bottomed tank) in a sink. Then simply pour hot water over the strip (according to the instructions that come with it), wait a moment, then watch for 2 different colours appearing - one over the portion of the tank holding a liquid and another over the portion of the tank holding a gas. An explanation of which colour represents which medium should come with the temperature strip.

The physics behind this? Gases and liquids can/will absorb different amounts of heat energy, therefore a temperature differential will form between the portions of the tank holding each. As the strip turns different colours at different temperatures, some indication of how much liquid is in the tank should be given. This method isn't all that precise, but it ought to show whether or not any of the numbers floating around right now are correct.

[Wat]

To dig up a really old thread, the answer to how much of a Co2 tank is liquid has finally been answered...

http://www.network54.com/Forum/message?forumid=9013&messageid=1092071082