continuous Trimix blending using 2 nitrox analyzers

[stepfen]

Hello!
First, I would like to emphasize that I am not trimix nor gas blending certified (nor I plan to become either of them any time soon). I don't plan to dive or to blend any such mixes myself. I just want to help the owner of the local dive store.

He recently learnt about a new (to him) trimix blending method using continuous flow and 2 nitrox/O2 analyzers. He thinks that the method sounds favorable for "medium" mixes (I think he mentioned as examples 10/45 and 10/55). More He than these it will require too much He causing too much stress to the compressor. Less that that and it can be easily done with partial pressure. His basic problem is how to utilize He tanks with low(er) pressure.

He has fount an excel file given here: Dive Shop Utility Software - Scuba Engineer (4th icon in the first row) doing the required calculations. It takes as input the old and the desired mix and it calculates the readings the user should get from the O2 analyzers (one measures O2 levels after He is added to air, the other O2 levels after any extra O2 has been added to the mix).
For example to fill in an empty tank to 10/55 the user needs to add He to air until the first analyzer shows 9.3% O2 and then add oxygen until the second analyzer shows 10%O2.

I offered to help him by making an (arduino based) device measuring these two O2 levels. This sounds pretty easy as I have already done an arduino based O2 analyzer (based on this: DIY Nitrox Analyzer (Arduino based) ) and now I am working on a Trimix one (based on this: Nitrox/Trimix & CO analyzer ).

But before me and this guy start working on this I would like to ask the opinions of the Gurus here in SB about the feasibility of the method.

I searched a bit and yesterday fount an old post (oddly enough today I can't find it!!) discussing the method. There somebody said that the method doesn't really work in practice because of the compressor's "blowback" effect. Basically the compressor sucks He easier than O2 (or was it the other way around?) hence the input percentages are disturbed by the compressor and the output mix ends up being wrong.
But I think in this post: Trimix continuous blender that the same method is described suggesting a correction for this blowback effect.

Any thoughts or updates on this? Should we bother investigating this method or??? Of course any related resources will be greatly appreciated.

Thanks a lot!!

 

[rjack321]

The dual O2 sensor method is a proven technique. Most people adjust the helium and O2 additions by hand as they will continually change slightly depending on the pressure you're pumping to.

You do have to do some trial and error due to the differences in compressibility of gases, especially helium.

 

[stepfen]

The dual O2 sensor method is a proven technique. Most people adjust the helium and O2 additions by hand as they will continually change slightly depending on the pressure you're pumping to.

You do have to do some trial and error due to the differences in compressibility of gases, especially helium.

Thanks a lot. That was the kind of answer I was hoping for. Yeap I am pretty aware that there will be quite a bit of trial and error to get the desired results.

For now we are looking only for monitoring. Assuming this goes well we might consider automatic control of the procedure such as gas control with needle valves etc.

 

[InColdBlood]

Can i ask you why the first % i mean helium it's different? with the second analyzer i will check the oxygen percentage for the trimix how can calculate the first?
Example why on a trimix 10/70 i will read 5.546 on the first oxygen sensor and 10 on second? which is the operation that i have to do?

and how to revert? so starting from first % after helium injection how can take the correct %?



Thank you

 

[sprockjohnson]

Can i ask you why the first % i mean helium it's different? with the second analyzer i will check the oxygen percentage for the trimix how can calculate the first?
Example why on a trimix 10/70 i will read 5.546 on the first oxygen sensor and 10 on second? which is the operation that i have to do?

and how to revert? so starting from first % after helium injection how can take the correct %?



Thank you

Typically the first o2 cell in the line is measuring o2% after HE has been added to air (20.9% o2) and based on the drop in % you can extrapolate the HE content of the mix and then the second cell after the o2 injection shows the final o2% for the mix. This can be done in reverse as well with slightly different equations.

 

[InColdBlood]

Typically the first o2 cell in the line is measuring o2% after HE has been added to air (20.9% o2) and based on the drop in % you can extrapolate the HE content of the mix and then the second cell after the o2 injection shows the final o2% for the mix. This can be done in reverse as well with slightly different equations.

Thank you, i'm looking of the equation if you know.

 

[sprockjohnson]

The reverse method (blending o2 first and then HE second) is o2/(o2 + N2). So for something theoretical like TX 21/30, which is .21 o2 + .30 He + .49 N2 (nitrogen) it would be- .21/(.21+.49)= .30 or 30% nitrox. You would then blend the nitrox 30% with the He until the second o2 cell reads .21 and theoretically it would contain the 30% He.

 

[inquis]

i'm looking of the equation if you know.

The Ideal Gas Law is one, but not accurate enough at typical pressures. To do it more accurately is far more complicated than one equation, unfortunately. I suggest you check out the Gas Blender Toolkit.

 

[inquis]

You would then blend the nitrox 30% with the He until the second o2 cell reads .21 and theoretically it would contain the 30% He.

The trick is how much nitrox to start with so you finish with 21/30 AND a full tank.

 

[InColdBlood]

Thank you for your help.
my question was about the percentage of the first oxy cell after helium injection.
From the first percentage example 5.546% how can calculate the % of helium which i'm uploading?

direct calculation should be like this
round((21/(21+79x(1+(70/20)))x100),1)