Real support for multiple concurrent sensors..

Linus Torvalds torvalds at linux-foundation.org
Thu Aug 3 11:06:14 PDT 2017


On Wed, Aug 2, 2017 at 11:44 PM, Anton Lundin <glance at acc.umu.se> wrote:
>
> On the subject of SAC rate, when we interpolate the SAC between a start
> and end pressure, we currently do that on _pressure_, as pressure_time.
>
> Since you made our gas volume calculations more correct, Ie factor for
> the non linearity of gases, we now get a pretty odd momentary SAC in the
> info box.

Ahh, yes. I was wondering why the SAC rate creeps up, but you are
obviously correct: it's because we calculate the real SAC rate using
the real volume, but the momentary one as if it's an ideal gas.

> If you open dives/sac-test.xml , you will see that the SAC value in the
> info box rises during the dive, as the pressure drop in the cylinder. I
> understand why, but this would be really confusing for a end user.

It was confusing to me too. I for some reason thought we always used
the proper gas volume, but you are obviously correct that we do not.

I'll fix it. I worry that I will have to make sure to use floating
point throughout to avoid rounding errors when turning milli-bars
pressure differences into milli-liter volume differences, but I'll
just have to see. At worst, I'll just have to do a "gas_volume_fp()"
the complex case, and move the "rint()" part outside.

          Linus


More information about the subsurface mailing list