New Bug Reports/Feature Requests

[Linus Torvalds]

On Tue, Feb 23, 2016 at 5:20 PM, Thiago Macieira <thiago at macieira.org> wrote:
>
> You can do an isothermal expansion of air, or let the captured air warm back
> up to ambient temperature. Not that I expect people do it.

So that theoretical calculation is exactly what we do to turn the
imperial size into the metric size.

It's how we get 11.1l from 80 cuft.

But as mentioned, it's actually unphysical for various reasons. Not
only isn't air an ideal gas, but "isothermal" is just not realistic in
practice.

At 3000 psi, the ideal gas differences aren't big, but with
high-pressure cylinders (ie 300 bar, so the 4300psi ones) you not only
want to use DIN valves, but the difference between air compression and
the ideal gas compression is actually pretty big - even when doing
isothermal expansion in theory (and isothermal expansion is likely
even harder to get in practice, since the pressure drops are even
bigger).

See for example

    https://en.wikipedia.org/wiki/Compressibility_factor#/media/File:Compressibility_Factor_of_Air_250_-_1000_K.png

and follow the black 300K line ("warm water"): it's within a couple of
percent up to ~175 bar (so about 2500 psi), and is still just 3% off
at 3000 psi. For things like SAC rate, even 3% is still largely in the
noise.

But with high-pressure cylinders at 300 bar (~3400 psi) it's about
10%, and keeps rising after that. At that point the errors of not
taking compressibility into account start being noticeable.

              Linus