Question re: CO2 & Leaks

There's nothing that will tweak your anal retentiveness faster than an empty CO2 tank.

No, you're not overreacting, IMO. A leak can occur at any moving part (also at non-moving parts, but that's not very common). Crank up the pressure, then turn off the gas. If the pressure in the line stays constant for a day or so, then you're probably good (of course, the beer's now over carbed )

Any time I touch anything related to gas - even just moving a keg - I test for leaks by turning it up to 50 psi and spraying everything with StarSan or soapy water. Even insignificant leaks will show up like this - fix those, too.

 

I just went through the same issue this week. After reading up on it we have nothing to worry about. There are some calculators online somewhere that will measure PSI compared to temp. It's possible you have a leak but I highly doubt it, so don't worry.

 

The pressure of your tank is directly proportional to the temperature. If you raise the temp 10%, the pressure will rise 10% . . . cut the temp in half and pressure is halved. You still have the same amount of gas, you just have a different pressure. If your tank went from 1000 to 500 it's only because you cut the temp in half or you are leaking.

Sorry for the stupid question, but what do I use to fix a line leak

If I understand the question, you normally troubleshoot the connections. Don't know your exact set-up but in general you will have three type connections:

Anything listed as a npt (national pipe thread) is a tapered threaded fitting where the seal is made by compressing the threads of the connection. Brass is ideal because it compresses and re-compresses easily, although you somethings see steel or alum. Usually this type fitting is used on a shutoff valve coming out of your regulator. Some people will tell you to not use pipe tape here, I am not one of them. I use gas pipe tape (yellow). It takes a moderate crush to compress the threads to stop leaks. You don't want to crank till you kill it, but it takes some torque to make the seal.

Flare fittings are threaded connections where a gasket/washer makes the seal. The threads are straight and you compress the gasket (usually nylon). This is common where your regulator is mated to the tank valve. The seal is made by compressing and expanding the gasket. I would say a little less crush (torque) on your part here. Of course the gasket must be clean and seated properly and not flattened out from excessive use.

You may have fitting that are sealed by o-rings, usually associated with quick disconnects. The pressure of a spring or the gas pressure itself forces the o-rings to expand and make the seal. Of course no torque required, but it is vital to have the correct size o-rings and it must be seated correctly. A good practice is to replace at the slightest doubt, some recommend replacing whenever it is opened.

Soapy water is excellent for finding leaks, but be careful if the connection touches your beer. You really don't want to serve Bier de Palmolive. Spare gaskets/o-rings are a must. Good luck.

 

While you're correct that temperature affects pressure, a 10% increase or a halving of temperature is meaningless. If the temp is 0F, for example, doubling, quadrupling, or even increasing it by a factor of 1,000 would have zero effect on pressure. I'm not picking nits, I'm simply trying to prevent someone from concluding that something is wrong when the proportions don't work out.

Tape or dope is always used on NPT threads. Anybody who says otherwise is simply wrong. That's not an opinion.

 

When one speaks of proportional changes in temperature, one is referring to temperature measured from absolute zero. That is −273° Celsius or −460° Fahrenheit.

 

True, but still incorrect. The OP said that the gauge read 1000psi warm (presumably, room temp, let's say 75F, or 297K). If the change in pressure was proportional to the change in temp, then 500 psi would occur at about 149K, or -191F. I'm going out on a limb here, but I suspect the OP's kegerator is incapable of chilling that low

 

PV = nrT is physical law. What I said is completely correct. Note that I made no claims at all about the source of OPs problem, only that your understanding of what is meant by a 10% change in temperature is incorrect.

 

Actually, my understanding of what is meant by a 10% change in temperature is sound . I was simply illustrating that a 10% change or a halving of the temperature is a meaningless concept when dealing with an arbitrary scale such as Fahrenheit. And that using such a scale in this way to accurately predict a change in pressure would be incorrect. Reducing the temp from 75-ish to 38-ish (roughly by 'half') was observed to reduce the pressure by half. The fact that both numbers were reduced exactly by half is coincidence, and would only occur at these temps and pressures (I didn't do the Math, but I think that's a safe statement).

 

I had a new regulator that was leaking through the gauge. I did not realize it until I lost the entire tank. They replaced the regulator.

 

The ideal gas law has little to do with directly determining the vapor pressure for a liquid (which is the phase of the carbon dioxide inside your tank). In either case, it was developed for a theoretical gas with zero molecular size and no intermolecular attractions... it's especially inaccurate for gases at low temperatures and high pressures (the conditions you're discussing). What you really need to determine is the change in vapor pressure of the liquid carbon dioxide on account of the change in temperature. Decreasing the temperature of the liquid changes the pressure roughly logarithmically (using the Clausius–Clapeyron relation). Reducing the temperature from 290K to 275.5K (~95% of 290K) equates to a drop of vapor pressure by approximately 56%. The OP is experiencing a drop in pressure that is theoretically predicted and quite normal.

 

True. But when using the Fahrenheit or Celsius scale, 'halving' the temp from 75 to 37.5 is meaningless without first doing some conversions. Consider that 'halving' -20F is actually increasing the temperature. That was my point.

 

I had to go to the local welding store and purchase a quality pressure regulator setup for this purpose. I went from gas every month or two to gas twice a year. I do get new nylon bushings for standard connections and tape NPT where applicable when reconnecting a setup. Fixed the leaks!

 

I agree; it's just a kind of happy coincidence that halving the temperature on the Fahrenheit scale approximately halves the pressure. That's why I specified Kelvin units. Sorry, I was more of responding to the general discussion:

I was just explaining that your and kscaldef's and PortLargo's understanding of the thermodynamics of the carbon dioxide in your tanks is incorrect. The gauge pressure (i.e., the vapor pressure of the liquid in the tank) is independent of the quantity of liquid carbon dioxide in the tank (until it's basically empty of liquid); the vapor pressure is dependent on the temperature, but logarithmically so. If you don't care for theory (not to make any assumptions about your backgrounds), Wikipedia has an easily accessible liquid/vapor equilibrium data table for carbon dioxide that's pretty much directly taken from the CRC Handbook of Chemistry and Physics (found via the Supplementary Data page on the Carbon Dioxide main page). These data indicate that at 2.78 degrees C (37 F, refrigerator temperature), the vapor pressure of liquid carbon dioxide is 3,735 kPa, or 542 psi (i.e., the OP's pressure is exactly what it should be). There could still be a leak... you just won't find out based on the gauge pressure until you're empty.

 

Thanks for the corrections. I was definitely overlooking that the CO2 in the tank is actually in the liquid phase.

 

Thanks for your detailed response, and yes, my response to the OP in regards to tank pressure was incorrect.

For the OP: A more practical method of comparing tank pressures to temperature is illustrated by this chart (as explained by ne0m00re):



My reference to pressure moving proportionally to changes in temperature was meant to refer to the line side (output) of the tank, which is gas only. But the way I stated it was misleading unless you are dealing with Kelvin temps as kscaldef kindly pointed out. While common in the scientific community, Kelvin temps don't really roll off the tongues of homebrewers.

So, apologies to the OP for muddying the waters and thanks for the informed replies. Oh yeah . . . did you find any leaks?

EDIT: For some reason the chart would not display full size. Here is link to a more readable chart:
http://www.co2info.com/co2.html