To Empty or not to Empty my Tins?

[Andy the Coastie]

<font color='#0000FF'>Jules,

YES , stored air does go OFF, and in taste over a period of time ,,,,, it's in the commercial divers manual, stating never to leave air in a cylinder for longer than 2 years. &nbsp;,,, if i can find my manual, i'll scan it for you &nbsp;

[BrianG]

There has been a recorded fatality due to stale air in the past. The incident occurred in a canal when a diver attemted to clear the prop of a friends boat. It transpired that his cylinder had been filled the previous season. The air when analysed contained less than 10% O2. The cylinder was found to contain water and in excess of one pound of ferous oxide.
Storing cylinders for a long period full is not a good idea, they are under maximum mechanical stress and high PPO2 promotes corrosion. This is even worse with Nitrox fills.
If storing cylinders for an extended period of time empty high O2 percentage gases and fill with air to approximately 50 bar. Store upright and ensure they cannot be knocked over.
In all the years I have been diving I am only aware of this one incident. The question is for the cost of a fill who whants to be the second?

[Woz]

<font color='#000080'>I am very, very sorry. I grossly underestimated how pedantic divers can be. &nbsp;Let me rephrase the question:

Let x=date of air fill of cylinder
Ley y=date of last breath of that fill from the cylinder
Let z=time elapsed

Now then, z=y-x

Please use any units you feel appropriate (SI seconds would be ideal) then express z as the time in years, rounded to the nearest year. &nbsp;And that's a f*cking Earth year, not lunar, solar or Alpha Centaurean.

I also feel that has been negelected recently so here's a couple to catch up.

[Jules]

it wasnt the cost of the fill, i was just curious ...cos i hadnt read it anywhere and i knew some one would have done.

[tom]

Originally Posted by [b

Quote[/b] (Steve W @ Feb. 26 2004,15:44)]Yes but we'll first have to establish acceptable parameters of +/- X% of oxygen, and of course we'll need at least 40 data points if we're going to correctly analyse the statistical significance of the variance, what do you think, parametric or non-parametric  is  p= &lt;0.05 acceptable or higher ?

See what you've started Jules ? someones gonna be demanding an apology any minute...  

Not quite sure what analysis you're thinking of here, but I'd suggest that calculating a 95% confidence interval for the mean across your data points would probably be appropriate (given that p&lt;.05 is the convention for hypothesis testing). If you want to do any kind of analyses involving variance then you are certainly talking about parametric statistics.

Of course, if you're interested in change across time (which I suspect you may be) then time series or trend analyses may well be the order of the day.

I apologise in advance, just in case.

[dlegros]

Originally Posted by [b

Quote[/b] (Steve W @ Feb. 26 2004,15:44)]See what you've started Jules ? someones gonna be demanding an apology any minute...  

I demand an apology - I was cr*p at stats and this makes my head hurt worse than the water at Stoney did! &nbsp;

Dom

[tom]

Does it hurt *significantly* more than it did in the water at Stoney?

I would suggest we need more data, more trials are required.

By the way, a friend of mine did his PhD on perceptions of pain, I'm sure he can help out if required...

[Mary]

<font color='#F52887'>OMG the scientists have started again, please save us.......

I know......

[Digger]

<font color='#000080'>I can see the ethical review committee just putting their big NO stamp on the front of an application for that! He must have had one hell of a proposal...

As for old air, can't see it being a major problem, unless of course it's been filled from a dodogy source, but then it's probably about as risky as if you'd got a dodgy fill anyway.

[BJ]

The air in your tank is less likely to have gone off than the air in your office, especially if you live in the Thames Valley! Your tank might have gone off though!