Wortgames,
I'll have to disagree with you there, a sodastream cylinder when filled with nothing but gas will only hold about 60grams of gas or so.
Calculating quickly P1V1=P2V2, and new soda stream cylinder has a nominal capacity of .46L and holds 330g (176 litres at 1 atomsphere pressure) of gas.
1atm x 176L = P2 x 0.46L
Pressure inside teh bottle = P2 = 382 atm = 38 Megapascals = 5600 psi (if we assume it remains as a gas)
However above 850psi (5800kpa) the CO2 will form a liquid. So there has to be liquid in a full soda stream bottle.
The pressure in the cylinder will always read a constant pressure as when the pressure rises above 850psi the gas turns to a liquid, thus dropping the pressure. It balances out so that you end up with a mix of gas an liquid in the cylinder. If you draw off any gas, thus dropping the pressure, more gas boils off to create gas to maintain the pressure at that magic 850psi. When you run out of liquid the pressure will drop as you remove any gas. For a big 9kg cylinder there may still be enough gas pressure to dispense a keg, but with small cylinders if will not last long at all. Rougly speaking, 850psi x .46L = 26L at 15psi (dispensing pressure) so you might just get a single keg dispensed out of a "gas only" soda stream bottle.
So the quick check is... if you have a constant pressure while carbonating/dispensing, you have liquid CO2 in the bottle.
That "constant" pressure of around 850psi is right for 20 degrees C. As temperature rises so will the gas pressure, until about 31 degrees C when any liquid WILL become gas (technically the gas and liquid states become indistinguishable). This results in a big increase in pressure (exact pressure will depend on the mass of gas inside).
In paintball systems this can result in burst safety valves which is a lot better than exploding a cylinder. 330g of CO2 takes up 0.206L, so soda stream cylinders are filled to about 45% capacity which compares to 68% for paintballing. By filling the soda stream bottles to less that 1/2 capacity it increases the safety margin in case the bottles do get left inside a hot car. Without doing the numbers I'm not sure if it would be possible in theory to explode a new soda stream bottle if it were allowed to get to say 50 degrees inside a hot car, but given that the cylinders are evidently rated to 250bar (3600psi) I dont think it would be likely. Could be a fun experiment though - just stand WAY back.
Having done the calculations now it looks like even on it's side the soda stream bottle would likely have the outlet above the liquid layer, even when "full" (aka 45%). So maybe it is something people can get away with. Personally I'd be looking to have the cylinder on a 45 degree angle, just to keep it well away from that potentially disasterous 'upside-down' regulator damaging position.
Anyway too much thinking, and my scottish amber ale calls.
Darryl