Filtration 101

[oxyhacker]

Whoops, I think I just pressed the "thank" key by mistake. Hopefully it will thank some deserving person.

The thing about calculating loads on alumuminum which always scares me, when it comes to highly stressed items like DIY filter canisters, is that they are much more likely to fail as a result of fatigue and stress concentrations at the machined ends, than have the sidewall just say the hell with it and let go. Since the stresses due to machining are very hard to predict mathematically, and since subtleties of how the machining is done can be very important, it is just about impossible to say with any reasonable level of assurance whether any one-off aluminum canister is safe.

Iaian. Using a 4 X 1 safety factor with such a rigid metal as aircraft aluminum seems appropriate.

 

[pescador775]

Does this scare you? It does me, aluminiuminmum

Separator Failure Report

 

[pescador775]

I kind of agree with that. I copied the Bristol filter, the type with end caps. Instead of 6351, I used 2024, a much tougher metal. Otherwise, exactly the same. Not worried, much.

The thing about calculating loads on alumuminum which always scares me, when it comes to highly stressed items like DIY filter canisters, is that they are much more likely to fail as a result of fatigue and stress concentrations at the machined ends, than have the sidewall just say the hell with it and let go. Since the stresses due to machining are very hard to predict mathematically, and since subtleties of how the machining is done can be very important, it is just about impossible to say with any reasonable level of assurance whether any one-off aluminum canister is safe.

 

[pescador775]

Guys,I think both of you are right. Basic physics: Higher temperature - more water in the air & higher energy of the vapor form water which BOTH contribute to the same result. Now, if you want to bash each other's heads I'll step aside.
Just my 9th grade physics 2 cents, no pun intended.

Thanks, I did not overlook your post. My sincere apologies to Craig. Somehow, I fixated on rate of uptake of moisture whereas it appears that Craig was talking about total potential versus actual capacity for moisture uptake of the dessicant. He is absolutely correct. For example, air which is at a temp of 35C can potentially hold about twice as much moisture as at 20C. This makes a big difference in total processing capacity of the compressor filter measured in cubic feet of free air.

I feel that some interesting stuff was provided in spite of the misunderstanding. I invite Craig to continue to provide such useful info and not feel that I am waiting to pounce. T'aint so. Learned my lesson.
Pesky

 

[pescador775]

I did a calc to see what Bauer and LF are talking about, and the numbers which they publish for filter capacity do seem to be about right if one is careful to read the fine print. For example, assume an average SCUBA cylinder holds about 2.5 m3 of air/ gas. This is about 87 cu ft. The amount of moisture separated out by the compressor condensator is related to pressure and temperature. At 30C and 150 bar the condensator should remove all but 0.5 gram from 2.5 cubic meters of air. Thus, the dessicant has to deal with that 1/2 gram of water in the course of filling the tank. My experiments showed that 13X can hold 24% by weight water. However, a practical limit is about 20% due to the steep downward curve of adsorption. Thus, I suggest that if one can estimate the weight of dessicant in the filter it is possible to retrace the logic which led to the original recommendations for the filter. If the dessicant in my DIY filter weighs six ozs (170 g) then the adsorption capacity is 170 X .20 = 34 g (water capacity). This is equal to 34/0.5 = 68 tanks = 5900 cf free air. Again, this is at 30C filter inlet temp.

I would guess that the standard triplex filter cartridge holds less than six ozs dessicant but don't know for sure. Maybe three or four ozs is closer than six so Bauer's capacity number would be less than 5900 cf.

 

[rjack321]

A P0 cartridge holds about 50gms (1.75 oz). Pretty skimpy IMHO.

Here are the Alkins P0 knockoff specs.

 

[pescador775]

The efficiency of 13X diminishes according to total run time. The ability of this dessicant continues until it reaches maximum capacity but at a diminishing rate which suggests a 1/2 life rule in relation to adsorption rate. I would have to plot the data to confirm this but it definitely does not look linear. Anyway, the question arises as to what point the dessicant is not removing virtually all the water before air enters the SCUBA cylinder and is this defined as the end of its life. Further, it raises a question as to the practical run time before depletion. I'm guessing that for purposes of specifying capacity that the mfgrs of the systems are using a simple set of assumptions as I furnished above. Yet, they are probably aware of the non linear nature of seive so perhaps they have looked at a graph of its performance and drawn a limit of somewhat arbitrary nature (CGA spec) as to how long they will say that the stuff lasts under specified conditions.
PS
Thanks, rjack for the info. It may come in handy.
Pesky

 

[veriqster]

Guys, you are priceless; this kind of in depth info for free - I wish I got comparable quality out of my college classes.
P.S. The more I read on this board the more i realize that my knowledge is limited and the wheel has already been invented.

 

[pescador775]

Verigster, we do what we can. Everytime I turn on the PBS channel (NOVA) or read Scientific American and see some wild stuff; I mean, a few years ago, some fellows actually reproduced the predicted Bose-Einstein condensate in the lab. I had heard about it but didn't know some of the details. Yeash, very humbling.

As to the present meanderings on my part, thank Lawrence Factor. At one time, they were claiming some fantastical performance figure for Vaporshell, I don't remember exactly, something in the neighborhood of 80% adsorption of water by weight. At the same time, I was trying to estimate the filtration life of a homebrew filter cartridge partially charged with that desiccant. Because their number looked a little high I ran some tests on new LF sourced Vaporshell. No matter what I did, the number came out to 24%. I measured it, I cooked some old and new samples, I measured it again, I rechecked everything, and this went on for a couple days until the number was verified. hmmmm. Since that time, I have been cautious in respect to the claims of vendors. Sharing such tidbits is fun and instructive; at least, feedback helps to check our claims and assumptions.

 

[rjack321]

Not so funny story...

I reproduced the table from my Alkins W31 manual in excel and posted it above my compressor. I have a 10/20/30% RH sensor between the stock P0 filter and my secondary filter. I log my hours. I have a soil thermometer stuck in the wall (measures air temp, partly the temp of the 2x4 its stuck through, lol). So I log my pumping temps.

No matter what I did I couldn't even come close to the published specs. Typically I was about 0.6x the published lifespan in hours when I hit CGA limits (~20%RH at 2000psi).

So I start asking around and discover that Alkins had copied the Bauer P0 filter specs. In the process substituted "compressor inlet temp" for "filter inlet temp".

So I write to Airtex, the US distributor and the US CPSC and tell them of the error and that a 'recall' should be undertaken for the manual. US CPSC tells me they have way too many defective products to investigate them all. Airtex tells me "thank you, we realize the published numbers are generous". The End.