Tank Pressure Tolerance

[gone2long]

Folks,

I have a 1970s vintage tank. Specs stamped on the top of the tank are as follows:
DOT - 3AA 2250
52162
SSGP
IC75 +PST
8 c6 94
20

Please forgive me if some of the stamps posted above are not exact, as it is hard to tell letters from numbers on the stamps.

Moving onward, I hadn't filled the tank more than a few times from the time I bought it, until 1995. Then, in 1995 I had it tested and filled it again at a US navy hydro-testing facility; so I did not take notice of the details. I never had problems using it a few times afterwards. Then, I did not use it for many years.

Recently, I had the tank successfully hydro tested and tried to fill it in SE Asia. The American (ex Army) facility manager who was responsible for filling it, did not take notice of the tank details, and he had one of his workers fill it without further instructions. The worker filled it to 3,000 PSI or so, and the valve on the tank blew, almost causing serious injury to his staff.

We then both read the inscription on the tank, which rated it (and or) the valve at only 2,200 PSI. As a result, I had him put on a new valve, which is rated at 3,000 PSI or 240 bar. By the way, nobody seems to understand PSI out here : (

After all this, the manager continues to insist that the tank can now be filled to 3,000 PSI. I have disregarded his opinion, due to my own ignorance and his relatively careless procedures, and I continue to fill the tank with no more than 2,250 psi. Consequently, I am now requesting other opinions.

Is the manager correct, and can this 50 year old, rarely used, recently hydro tested tank (with a new, PSI 3,000 rated valve) now safely be filled to 3,000 PSI? Or, am I risking my life taking his, perhaps not so reliable advice?

All replies appreciated,

G2L

[captain]

Legally no, the rating is not based on the valve, it is based on the tank construction. When you say the valve blew I am assuming you are referring to the burst disc on the valve and not the valve it self. From the sequence of events you describe the burst disc probably had not been replaced since 1970 and age took it toll causing it to fail, not an unusual event for a disc that old. A new burst disc is designed to blow at a pressure slightly below the test pressure of 3750 psi.
The old valve is easily capable of holding 3000 psi with a new burst disc.
That tank is tested at 3750 psi and rated at 2250 psi or 2475 psi if plus rated at the time of the latest hydro if it meets the required conditions to be plus rated. Unfortunately very few shops are knowledgeable enough or willing to do the simple math required to plus rate the tank.
So is filling it to 3000 psi legal, no. Is it safe, with a tank that has recently passed the hydro test, yes but you have to realize that you are using up a very small portion of the safety factor that is built into the tank. It is done all the time in Florida cave country. The decision is up to the person filling the tank not the owner.

[ScubaLawyer]

I agree with Captain.

[antique diver]

No doubt that the 2250 psi allows for some safety margin since the hydro test is to 3750. But I kinda like having that safety margin, and wouldn't recommend filling to 3000, which would narrow your safety margin.

How about us starting a forum discussion on pressure ratings and temperature. Isn't the rating based on something like the rated pressure at 68 to 72 degree ambient temperature? Rough approximate rule of thumb from way back stated that pressure change in the cylinder could change about 5 psi for every one degree F. change. So if your tank reached 110 degrees during filling on a warm day, you would have to over-fill by about 200 psi to "about" 2450 to have 2250 when it cooled back down to the "normal" temp of about 70.

It would make sense to fill to the higher pressure in that situation. Seems logical to me, and that's similar to what I do.
That's also what a local AIR-GAS facility does when filling commercial cylinders. They use a chart to correlate stamped pressure ratings with a given ambient temperature at the time of filling. Their customers don't like getting short fills, and neither do divers.

[Bryan]

Luis has done a lot of work on the subject and has some great information I'm sure he will share when time allows.....You can only imagine the amount of data an engineer can extract when he does his own hydrotesting...

[luis]

There are a couple of interesting subjects here and I will try to share what I have found when I have more time, but here are a few thoughts.

The steel 72 is stamped to 2250, but every tank that I have tested has passed hydro and the plus stamping requirements. If it passed hydro, I consider them a 2475 psi cylinder (2500 psi in round numbers).

I have no hesitation filling mine to 2700 or 2800 psi even when cold, but I don’t live in hot climate like Florida were tanks can get very hot in a car (and blow the burst disc).

I also like to have some safety margin. A 3000 psi hot fill is not too bad since it will normally drop 200 or 300 psi when it cools down, which would bring it to a comfortable level when it is cool.

A correct full steel 72 cylinder tank as mentioned is 2500 psi at 70 degrees Fahrenheit (per CFR 49 section 173). In the CFR (Code of Federal Regulations), it also mentions that the maximum allowed temperature is 130 F.

Doing the math using ideal gas law (and the proper absolute temperature, 70 Fahrenheit is 530 Rankine), the 2500 at 70 F will show as 2783 psi at 130 degrees Fahrenheit. That is not an overfill since it will go back down to 2500 psi at 70 Fahrenheit.

One interesting point that I should make. I am referring to the air temperature inside the cylinder. That temperature can be very different to outside temperature and also very different to the tank surface temperature. Air has a very low heat conduction coefficient (that is why is used as an insulator), therefore during the fill process, the air temperature inside the cylinder can be much warmer than the actual cylinder wall.

When filling a cylinder I always prefer to fill from a cascade bank than directly from a compressor, if the air in the cascade back has had a chance to cool down. The air directly out of the compressor is about as hot as it will be in the filling process.

I don’t bother using a water bath. The water bath will cool the scuba cylinder wall, but the air temperature inside still need time to cool down. Yes, cooling the cylinder wall helps, but only time will cool the air and the risk of injecting water with a wet filling connection into the cylinder is not worth the risk. I can be very careful when connecting my cylinder and make sure everything is dry, but I still use commercial fill stations.

Well this was longer than what I planned.

[swimjim]

I have a couple 72 twin sets rigged modern with double iso manifolds I use for tech diving. I routinely fill the in the 2800 to 3000 range. So far no hydro issues. I really prefer them to my LP 108 double set. The 72's are much easier on what is left of my back. Getting back up the ladder with the 108's can be a real adventure. Much of it felt hours and days after the dive.

[gone2long]

Legally no, the rating is not based on the valve, it is based on the tank construction. When you say the valve blew I am assuming you are referring to the burst disc on the valve and not the valve it self. From the sequence of events you describe the burst disc probably had not been replaced since 1970 and age took it toll causing it to fail, not an unusual event for a disc that old. A new burst disc is designed to blow at a pressure slightly below the test pressure of 3750 psi.
The old valve is easily capable of holding 3000 psi with a new burst disc.
That tank is tested at 3750 psi and rated at 2250 psi or 2475 psi if plus rated at the time of the latest hydro if it meets the required conditions to be plus rated. Unfortunately very few shops are knowledgeable enough or willing to do the simple math required to plus rate the tank.
So is filling it to 3000 psi legal, no. Is it safe, with a tank that has recently passed the hydro test, yes but you have to realize that you are using up a very small portion of the safety factor that is built into the tank. It is done all the time in Florida cave country. The decision is up to the person filling the tank not the owner.

Thanks Cpt.,

Truly appreciate your help.

Regards,

G2L

[gone2long]

There are a couple of interesting subjects here and I will try to share what I have found when I have more time, but here are a few thoughts.

The steel 72 is stamped to 2250, but every tank that I have tested has passed hydro and the plus stamping requirements. If it passed hydro, I consider them a 2475 psi cylinder (2500 psi in round numbers).

I have no hesitation filling mine to 2700 or 2800 psi even when cold, but I don’t live in hot climate like Florida were tanks can get very hot in a car (and blow the burst disc).

I also like to have some safety margin. A 3000 psi hot fill is not too bad since it will normally drop 200 or 300 psi when it cools down, which would bring it to a comfortable level when it is cool.

A correct full steel 72 cylinder tank as mentioned is 2500 psi at 70 degrees Fahrenheit (per CFR 49 section 173). In the CFR (Code of Federal Regulations), it also mentions that the maximum allowed temperature is 130 F.

Doing the math using ideal gas law (and the proper absolute temperature, 70 Fahrenheit is 530 Rankine), the 2500 at 70 F will show as 2783 psi at 130 degrees Fahrenheit. That is not an overfill since it will go back down to 2500 psi at 70 Fahrenheit.

One interesting point that I should make. I am referring to the air temperature inside the cylinder. That temperature can be very different to outside temperature and also very different to the tank surface temperature. Air has a very low heat conduction coefficient (that is why is used as an insulator), therefore during the fill process, the air temperature inside the cylinder can be much warmer than the actual cylinder wall.

When filling a cylinder I always prefer to fill from a cascade bank than directly from a compressor, if the air in the cascade back has had a chance to cool down. The air directly out of the compressor is about as hot as it will be in the filling process.

I don’t bother using a water bath. The water bath will cool the scuba cylinder wall, but the air temperature inside still need time to cool down. Yes, cooling the cylinder wall helps, but only time will cool the air and the risk of injecting water with a wet filling connection into the cylinder is not worth the risk. I can be very careful when connecting my cylinder and make sure everything is dry, but I still use commercial fill stations.

Well this was longer than what I planned.

Thanks Luis,

Lots to learn on my end about these technicalities. For instance, I have not exactly sure about the relationship of PSI to Bar, and when I talk to folks out here about such stuff, I get these blank stares. I was told by the manager noted in my original post that PSI 2,250 equates to 150 Bar and that, if I wanted to keep the pressure at that, he could do a 155 Bar "hot fill" like the one you noted above, apparently figuring that the tank would cool and pressure would drop to 2,250. Doing the math, each "Bar" seems to equate to approximately 15lbs PSI. Is that correct? Also, 1 "Bar" reflects average atmospheric pressure at sea level; so 150 Bar generally means 150x atmospheric pressure, right?

Also, everyone seems to think my tank is a "72". Is that a cubic inch measurement, or what? And what, exactly does the "94" on my tank represent?

Pardon my ignorance, but I am just getting back into the "sport" after many years away, due to the need to routinely clean the bottom of my recently purchased boat.

Truly appreciate your insights and assistance,

G2L

[georgeaustin]

Check the link regarding how to read crown markings on your cylinders. Others will chime in but I simplify by using 1 BAR = 1 ATM (15psi)
Burst discs wear out but a properly maintained steel cylinder, with rare exceptions should last for many decades - I have a few steel bottles that are immaculate inside and are 50 + years old and will last another 50 probably


http://www.catalinacylinders.com/suppor ... -cylinder/

http://www.scubaengineer.com/tank_servicingx.htm

[luis]

The conversion is:
1 Bar = 14.5 psi
1 Atmosphere (Atm) = 14.7 psi

Therefore:
2250 psi = 155 Bar
2475 psi = 171 Bar
2500 psi = 172 Bar
3000 psi = 207 Bar

150 Bar = 2176 psi


The steel 72 cylinder means that it has approximately 72 cubic feet of standard air when it is filled to its designed working pressure.

The designed working pressure of a steel 72 is 2475 psi, not 2250 psi. The 2250 psi is the stamped working pressure, but next to the original hydro date it has a plus “+ “ stamp. The plus stamp means that this cylinder was actually designed to work with a 10% fill above the stamped pressure.

The 10% extra is where the 2475 psi comes from.

At 2250psi this cylinder only has about 65 cubic feet of standard air.

To get the 10% overfill, technically the latest hydro date is supposed to have the “+” stamp next to it. Most hydro facilities do not have the necessary information to evaluate the plus stamp requirement so they do not stamp it.

On the other hand, many reasonable dive shops realize that most steel 72 would pass the “+” requirements and they go ahead and fill them to approximately 2500 psi.


Note: standard air is defined as the air in 1 atmosphere (at 70F).

[antique diver]

All conjecture and past experience aside, I wanted to accurately test the difference it makes to fill a given steel cylinder in water vs. dry? I conducted a carefully controlled and recorded test to see what really happens.

The following test was done outdoors in the shade, on a 95 degree day:
Using two fill whips on my 5 cfm Rix, I filled two individual Faber steel 45's from empty to 3100 psi. (took 20 minutes)
Cylinder #1 was dry, and in the shade. Cylinder #2 was immersed in 87 degree water which was kept at a constant temperature by letting additional tap water flow into the container and letting it slowly but continuously overflow.

At the time of reaching the fill pressure of 3100 psi, #1 had reached a metal surface temperature of 122 F. The wet cylinder had maintained a metal surface temperature of 87 F... the same as the water bath.

As expected, there was a significant difference in pressures between the two cylinders after allowing both cylinders to cool down for 3 hours to 82 degrees (measured on the steel surface of each cylinder).

At the end of cool down, Cylinder 1 had 2800 psi. Cylinder 2 had 3050 psi. A significant difference of 250 psi.


The same test performed on two aluminum 50' cylinders resulted in final pressures of 2900 and 3050 psi.

Of course an over-fill of the right proportion could be used to get the same ending pressures without the water mess.

[captain]

I had found the same regarding hot and cold fills. If you fill hot it means you have fill to a higher pressure that will cool to what you want or the extra hassle of letting the tank cool and topping it it off. I water fill, I have a 26" high piece of 8" PVC capped on one end. I installed a male garden hose fitting on the bottom. The tank sits in it with the valve above above the top of the pipe so it never gets wet. I trickle water through it and let it over flow through the three holes which also serve as finger holes to carry it. That way there is always cool water flowing up the sides of the tank.

[URL=http://s49.photobucket.com/user/bu ... .jpg[/img][/url]

[antique diver]

That's a nice water bath, Tom. We seem to think alike on air fills and water cooling, and you have given me some great ideas on improving my setup! Thanks for posting the photo.

[luis]

A key factor is that you are both doing very slow fills with small compressors. In a 20 minute fill (or anything similar) the heat transfer from water cooling will make a big difference.

The the commercial dive shops that I normally deal with are capable of simultaneously filling several steel 72 (or AL 80) in about minute or less. A slow fill (which they normally do) means that they can throttle the fill rate to take about 5 minutes.

The recommended fill rate for air is 600 psi/ minute (or less). At that rate, an AL 80 from 0 psi to 3000 psi would take 5 minutes to fill, and will be very hot.

The effectiveness of the water cooling is proportional to the amount of time it spends in it. In a one minute fill, the water cooling is basically meaningless. In the 5 minutes fill the water cooling starts to make a slightly measurable difference, but unless I am the one connecting my cylinder, I prefer the area stays dry. In the past, I have seen too many filling whips dropped in the water bath and the dive shop employee not bothering blowing the whip dry.

I always blow the valve and the whip dry even for a dry fill.

When getting a fill I often ask them to top it off and my best hope is that I am getting the cool air that has been in the cascade banks and had time to cool down.

If I had my own small compressor, I would probably wet fill. I would also add a few fill whips to fill several cylinders at the same time, to slow down the individual cylinder fill rate.

Someday I may have my own compressor… I look at them on a regular bases.