Mixing Nitrox with a Partial pressure system

[DiveLvr]

I would like to know the general steps in mixing a 30% EAN from the point of connecting the tank to the system to final analysis. As i have looked this up it seems a tank filled in this fashion needs to "settle" before analysis. I have seen it suggested a tank needs to be "rolled" after at some point then analyzed. If either of these is the case if I let my 30% EAN filled tank sit in the closet a month or so is the mixture going to be different than when it was filled? ANother question, if the tank does neet to "settle" before analysis, how long does it need to settle? If the target mix is 32% and I analyze it 2 days after the fill and find the mix to be 26% what can the shop do to correct the problem? Thanks in advance.

DiveLvr

[Hickdive]

Different strokes for different folks but here's my procedure;

Make sure O2 cylinder, needle valve, dive cylinder and bleed screw are closed.

Open dive cylinder.

Open O2 cylinder.

Slowly open needle valve until oxygen flows.

Let oxygen flow into dive cylinder slowly. A maximum speed of 4 bar per minute.

When pressure reaches required level shut needle valve.

Shut all other valves, open needle valve and bleed valve to vent.

Connect dive cylinder to compressor and make up to required fill pressure.

Disconnect dive cylinder and give it a roll round/invert a couple of times (I've analysed fresh filled cylinders that were supposed to be 50% at 30 odd %; rolling/inverting for a minute or so homogenised the mix at the correct 50%)

Analyse (twice, ensuring analyser returns to calibration between analyses)

Record result, label cylinder.

Done.

[Hickdive]

To change the mix the shop needs to bleed it down, add more 02, then pump back to fill pressure.

[Dive-aholic]

I would like to know the general steps in mixing a 30% EAN from the point of connecting the tank to the system to final analysis. As i have looked this up it seems a tank filled in this fashion needs to "settle" before analysis. I have seen it suggested a tank needs to be "rolled" after at some point then analyzed. If either of these is the case if I let my 30% EAN filled tank sit in the closet a month or so is the mixture going to be different than when it was filled? ANother question, if the tank does neet to "settle" before analysis, how long does it need to settle? If the target mix is 32% and I analyze it 2 days after the fill and find the mix to be 26% what can the shop do to correct the problem? Thanks in advance.

DiveLvr

Nothing needs to settle and the tank doesn't need to be rolled. A basic understanding of gas laws makes this pretty obvious. There are gas blending classes that explain the process of partial pressure blending. As for your scenario of having 26% when it should be 32%, whoever at the shop mixed that has no idea what they are doing. A 6% difference in oxygen is unacceptable. The shop should refund your money and you should take your tank elsewhere.

[Hickdive]

I speak from experience. If a cylinder is filled slowly with O2 and then slowly with air and immediately analysed then it might not be a homogenous mix at that point and the analyses may appear to be wrong.

Leave it a few hours and it will homogenise, alternatively, give it a roll around and a couple of inversions to mix it. Some people put the air in at high speed to ensure an even mix but that causes too much heating IMO, putting the air fill pressure and thus the mix off.

O2 is very slightly denser than air, hence slow, careful filling of it and then air can cause a stratified fill. Of course brownian motion will mix it over time.

I find the secret to accurate pp blending is taking time and doing things slowly (particularly when blending with Helium) but I have the luxury of doing my own blending in my own time, something shops don't often have.

[windapp]

If you shop is giving you 26%, and their analysis is 32%, you should be complaining to the whoever gave them their blending license, and have it revoked. Make sure that there is nothing wrong with your analyzer, and that you know how to use it first though.

The shop analysis of the mix is not just a quality check. It is also a double check for your analysis. You should be agreeing with their analysis or something is wrong somewhere.

[OzGriffo]

Nothing needs to settle and the tank doesn't need to be rolled. A basic understanding of gas laws makes this pretty obvious. There are gas blending classes that explain the process of partial pressure blending. As for your scenario of having 26% when it should be 32%, whoever at the shop mixed that has no idea what they are doing. A 6% difference in oxygen is unacceptable. The shop should refund your money and you should take your tank elsewhere.

What gas law says that diffusion is instantaneous? If merely dumping gasses together immediately resulted in a homegenous mix, there would be no need for the baffles in a "stick" system.

I'm not saying a reading should ever be 6% out, but it will take some amount of time for the mix to settle.

[Jim Lapenta]

I speak from experience. If a cylinder is filled slowly with O2 and then slowly with air and immediately analysed then it might not be a homogenous mix at that point and the analyses may appear to be wrong.

Leave it a few hours and it will homogenise, alternatively, give it a roll around and a couple of inversions to mix it. Some people put the air in at high speed to ensure an even mix but that causes too much heating IMO, putting the air fill pressure and thus the mix off.

O2 is very slightly denser than air, hence slow, careful filling of it and then air can cause a stratified fill. Of course brownian motion will mix it over time.

I find the secret to accurate pp blending is taking time and doing things slowly (particularly when blending with Helium) but I have the luxury of doing my own blending in my own time, something shops don't often have.

But you're in Scotland. Everything moves at a more relaxed pace there. One reason I'd so love to visit. Just one dive in Loch Ness is all I ask. Just the chance to see would be worth it.

[elan]

I never was getting a fill changing the reading for more than .2% when blending nitrox. And that could be contributed by inacurate calibration.
6% difference looks like a screw up by the operator.

[randy88k5]

I have seen the "layering" effect with mixing nitrox and trimix. If you analyze a mixture right after the filling is complete, it can and will have a different mixture than if it is rolled around or let sit for several minutes. At the high pressures of a scuba tank, the gas acts much like a fluid, and can layer if filled slowly. But it eventually does mix.