Fwd: Re: [PATCH] Fix up SAC calculations for ATM/bar confusion

[Linus Torvalds]

On Mon, Feb 25, 2013 at 5:10 PM, Dirk Hohndel <dirk at hohndel.org> wrote:
>
> Just for kicks and giggles, here is the comparison with our numbers -
> simply calculating gas_volume(1_L_Tank, pressure) for the given values.
> Second column is Suunto, third colum is Linus' function

I think a "percent difference" would have made it easier to read, but
from a quick look, we're generally within a percent or two of the
suunto numbers.

Not always, but the Suunto numbers have some odd jumps. Like here:

> 190  184.825  187.561
> 200  194.553  197.433
> 210  198.113  204.068
> 220  207.547  210.777
> 230  216.981  217.562

We have a noticeable jump at 200 bar (not surprising, that's our
"let's start correcting" point), where we go from roughly 10l / 10 bar
to roughly 7l / 10bar.

But the Suunto correction function is much worse. Look at the
differences: 10l, then 3.5l, then 9l, then 9l. It's even more
noticeable at around 300 bar:

> 290  264.357  259.846
> 300  273.473  267.155
> 310  273.369  274.540
> 320  282.187  282.000

We do that ~7l per 10bar thing consistently, but apparently Suunto
thinks that the gas volume *shrinks* when the pressure goes from 300
to 310 bar. So they seem to have two big discontinuities at 200 and
300 bar.

But on the whole, we seem to be in the same ballpark, with Suunto's
function just being somewhat odd. I think we're mostly ok.

                     Linus