New Bug Reports/Feature Requests

[Linus Torvalds]

On Wed, Feb 24, 2016 at 5:20 PM, Richard Houser <rick at divinesymphony.net> wrote:
> When you are looking into the tank editing, would you please try a couple
> tanks completely from the imperial specs we get from the manufacturers spec
> sheets?

I don't think that's reasonable, outside of some specialized cylinder
model editing thing.

Basically, I don't know what the manufacturer spec even *means*. The
numbers are meaningless.

> When I'm shopping for this stuff, I don't typically see water volume
> measurements.

Look for the metric version, you'll generally find it.

> Examples:
>
> Luxfer AL80 77.4cuft 3000psi (from manufacturer specs online)

Go to:

  http://www.luxfercylinders.com/products/scuba-cylinders/l6x-aluminium-cylinder

and pick specifications, and you'll see the actual wet capacity. There
it's called "Internal volume". It's specified as 11.1 liters.

Now, look at our AL80. That's exactly what you get in metric mode:
11.1 l and 207 bar.

And the way we get that? It's the 80 using an ideal gas.

Now, switch that thing to imperial, and you'll see that they say 77.4
cubic feet of air capacity.

And the important thing to take away is that if you enter "77.4"
*anywhere*, and expect people to do math on it, you're in for a world
of butt-hurt.

I have no idea how they get there. It's magic. With the best
approximation of air compression effects, I can get *close*, but I get
77.1 cuft.

In other words, there is some completely arbitrary thing going on.

It's not worth it, I tell you.

So you have two options:

 - use imperial measurements, and use the *nominal* capacity, not the
actual one.

 - use metric measurements

Those are your two options.

The third option, the one you seem to want, is not mathematically
possible. Even if you know that it's the actual size, there is some
magical constant that Luxor has used, and we don't know what it is.

It's *close* to the compressibility constant we now use, but it's not the same.

Of course, it's generally close enough that it probably doesn't
matter, but it really means that I think that if you use imperial
sizes, you should basically accept a "within 5%" factor.

> XS Scuba HP120 (Faber tank) 120.6cuft 3442psi (confirmed in the dive ship
> catalogs and online)

Grr. XS Scuba used to have much better spec sheets, including metric.
Now they seem to be skimping on it.

Or maybe it's just the Faber cylinders that don't have good spec sheets.

With some googling you can get the actual metric size online, although
Faber itself seems to make it hard to get.  But the online specs I
found seem to say 120 cuft and 15.3l.

And this time, if you actually put in 15.3 liter, with subsurface you
get a nominal size of 127 cuft, and a "actual" size of 119.6. So this
time the "120" in the "HP120" is the actual size.

But it's almost impossible to tell ahead of time. Is it the nominal or
the actual size? Who can tell.

> LP72 (forget the brand) 65cuft 2250psi (an optional + overfill hydro stamp
> gets it near 72, hence the name)

So that seems to be how LP cylinders are always done. They are sized
not according to their working pressure, but their "10% over"
pressure.

We actually compensate for it by just setting the working pressure to
the "+" pressure. So you'll see in our tank desciptors that we have
things like:

        { "LP85", .cuft = 85, .psi = 2640 },

where the "nominal" working pressure is 2400 psi, but to actually get
the rated size (85) you have to pressurize them to 2640 psi.

I don't think I've ever seen an LP tank at 2250. That sounds like a
*de*rated LP cylinder. But maybe it's just not one of the
manufacturers I didn't check (I checked a number of them when I did
our equipment tables originally).

> Catalina AL80 77.4cuft (from specs online)

The Catalina AL80 is the same size as the Luxfer one - 11.1 l.

> Also, when I look up the details on that worthington x8-119 tank, I show
> 3442 psi ~= 237bar and a volume of 14.8l (though that volume is reseller
> provided and does not show in the spec sheets over here).

Either way, with a wet volume of 14.8l, I don't see how you get 119 cuft.

>From my calculations, 14.8l at 3442 psi is either 122 cuft ("nominal")
or 116 psi ("actual").

I get the feeling that somebody decided to just split the difference,
and call it 119.

And I'm not even kidding. The imperial sizing really is *so* crazy
that I could see exactly that happening. Somebody couldn't decide
whether they should use the nominal or actual size, and decided to go
in between.

This is why I seriously am saying that if you insist on using imperial
sizes, you should just accept that "within 5%" thing. It's going to be
_close_. But you'll always wonder.. A cylinder that is called a HP120
from two different manufacturers could be slightly different.

And it's really not always trivial to even find out.

>    I don't suppose
> the 200/207 and 230/237 bar discrepancy is a factor either way?

I think it's just another one of those "metric sizes are sane" issues.

For metric people, the working pressure is about safety, not size. So
they'll just have 200 bar (with 230 bar for a "plus") and 300 bar for
a high-pressure thing, and a 3000 psi setup is just a 200 bar cylinder
in metric.

Even if 3000 psi is actually 207 bar if you care and do the math.

But metric people don't care, because the number is a safety limit
anyway, and those safety limits are _not_ about that kind of "a few
bars here or there".

So in metric specs, you'd likely just call that a 15 l cylinder (maybe
14.8 if you really care), and because the working pressure is a rough
range, they are just rounded off.

But then when you do imperial sizes, suddenly the *exact* working
pressure matters a lot, because it ends up impacting what you sell the
cylinder as.

You probably won't find those 300 bar cylinders in imperial versions
(you need DIN for them, and from what I understand the fills are hard
to find even in Europe - it's just often not worth it), so you'll find
the 200/230 bar thing for 3000/3442 psi.

"Close enough".

I _so_ wish that the US just did metric cylinder sizes. I don't mind
feet and Fahrenheit at all - those are just trivial conversion issues.
But the imperial cylinder sizes are just a source of constant pain and
confusion.

                 Linus