I waited, and periodically tried to get more from the cylinder. Did get a small amount, but not more than 10 percent after the first batch. I also did not try warming the cylinder beforehand.
I have tried manipulating the input values on the spreadsheet to see what errors could predict the measured results. Taken one at a time they are pretty extreme. Volume 165ci, Temperature -70F, or a 13 psi error in the pressure gauge at 49 psi. Individually these are beyond liklihood but some combination is possible. I wonder what the gas temperature inside actually is. I don't think the volume is that far off, but the gas temperature could be pretty wrong. And the gauges are definitely not great. I have not seen a 13 psi difference between gauges at such a low pressure, however.
Of course perhaps there is an error in the spreadsheet or in the assumptions, but I have not found that yet.
I am buildng a launcher with a 50ci chamber. That should be done soon. It will provide another data point. I may also try some 16gm cylinders though due to cost I would not use them often.. I also have some other volume chambers on launchers that I can try.
It would also be pretty trivial to make a test chamber. One problem with testing in launchers is they often slowly leak down so waiting for the temperature to equilibrate doesn't work.
Looking at the Spudtech web the TBL250LP is about 46 ci and gets about 100 psi. The TBL500LP is about 90 ci and gets about 50 psi. The spreadsheet shows 126/65 for these.
One other issue is the way the Ultraflate handles the schrader valve. It does not open the valve but overpressures it. I would not expect this to be much of an error for CO<sub>2</sub>. In the case of the QE19 there is a one way valve and I took the center out of the schrader so we have only a psi or two loss from the one way. In the case of the CSV19 there is a normal schrader. If the schrader takes 15 psi to overcome then one extra atmosphere of CO<sub>2</sub> would be 'stuck' in the cylinder. It is such a small volume that should not make much difference to the main chamber's pressure.
Clide, do you recall the pressure, temperature and volume of the tests you ran some time ago? Are they consistent with any of these measurements?
Another set of tests on the 3x7 chamber 3/13/05 12 gram 70 degrees, carts warmed to about 80 degrees in the sun. 62 psi first charge, 25 psi a few minutes later, second 12 gram takes it to 83 psi, a few minutes later the second recharges the chamber to 23 psi.
Edit - 3/13 8:30 pm:
Reviewing my spreadsheet and notes, comparing the new data, looking at the web:
1) I cannot find any errors in the calculations. I find other calculations that are similar.
2) The experimental data do not agree with the calculations.
3) The measurements are closer to the 4200/ci of the originator of this thread.
So, how can this be?
1) Approximately 25% of the liquid CO2 in the cylinder goes to Dry Ice at 70 degrees ambient. Thus only about 9 grams is available to develop pressure. This is clearly the case since waiting awhile produces another 25% pressure.
This STILL does not compensate fully for the errors.
So I played with the temperature. Clearly the temperature of the gas is NOT 70 degrees. To compensate for the measured pressure it has to be quite cold. -50 F works for my measurements, but this may not be a terribly accurate number since the error bars on it are considerable.
The bottom line: 4200/ci is not a bad estimate, but you'll only get about 75% of that pressure right away, the rest will have to wait for the dry ice to sublime.
It would be interesting to do a test where the gas is allowed to warm after filling a chamber. It will require a very low leakage rate, perhaps a copper chamber to get better heat transfer to the gas. This would show if this hypothesis is correct, or there is some other effect.
Edit 3/16/05 7:31pm
I just calculated the temperature for which the 4200/chamber volume is correct for 12 grams of CO<sub>2</sub>. It comes out -64 deg C or -82 deg F. So it is right in the neighborhood of my previous prediction. I've calculated the factor for temperatures from -110 F to +100 F and it varies from 3900 at -110 to 6230 at 100 F.
We should still verify this hypothesis by more testing, but it is starting to agree in several ways.
If this is all accurate, two things are important to remember - The 4200 rule may be good for an approximation, but if the CO<sub>2</sub> is allowed to warm in the chamber, the pressure can increase by 50% or so. Make sure this is within safety limits. It could certainly be a surprise if used in paintball and warming created a 'hot' velocity.
Secondly, we need to use a lower temperature for GGDT simulations for CO<sub>2</sub> unless we warm it up before launching. I just compared CO<sub>2</sub> at -40 to air at 60 in my CSV17 tennis ball launcher. It makes a considerable improvement to use the warmer air. I have not verified this with measurements yet, but I will check it.