High Vacuum Differential Pressure 3

[weed78]

We are trying to bring back to life an old perlite insulated liquid O2 tank.
We have been pumping down the vacuum in the insulation space but can not get down below about 400 microns.
With the tank isolated from the pump we pull down to less than 10 microns so the pump appears to be capable of doing the job. Helium leak testing has not found any external leaks. Any ideas as to what may be happening there?
Another weird observation is that when we close the valve at the pump, we are reading a differential pressure between the tank and the vacuum hose to the pump of about 100 microns. We expected to see the two pressures equalized after a few hours but this is just not happening. We know that there is some kind of screeing device behind the vacuum connection to the tank but we still expected that after some period of time the pressures would equalize.
Is there some weird phenomenon happening at this high vacuum level that is preventing the tank and piping to equalize in pressure?
Thank you for you thoughts.

 

[iainuts]

Just a few thoughts, most tanks I've seen that loose vacuum have defects in the internal piping. External leak tests obviously won't pick this up. I believe you need to introduce helium inside the inner vessel and look for helium coming through to your pump down port.

Also, perlite will be crushed and broken up, eventually turning into powder as the inner vessel expands and contracts. This powder settles to the bottom and I've heard it described as 'cement'. This can add additional stress to the inner lines, causing cracks to develop.

As you say, there's a screen inside the vessel to keep the perlite from being sucked into the pump down port. I don't know how that could get blocked, but I'd think you might also see indications of perlite dust coming out, an indication of a problem with the perlite.

 

[handleman]

What does a measurement of microns mean in relation to vacuum? Sorry for the ignorance...

 

[quark]

1 milli meter of Hg = 1000 micrometer of Hg (or 1000 microns)

 

[handleman]

That was my suspicion, but I hadn't heard that measurement used before. It seems to me to be an odd way to express, as I'm sure nobody's measuring the height of a mercury column to a few micron accuracy.

 

[MintJulep]

It actually is a mercury colum height measured in microns. Not actually measured that way of course, but it "could" be.

 

[weeeds]

We use this terminology for deep vacuum as follows:

1 atm = 760 torr
1 torr = 1,000 microns

 

[wfn217]

You may be getting molecular flow.

 

[zeusfaber]

Showing my total ignorance here, but do you get a lot of gas or vapour adsorbtion onto perlite?

A.

 

[griffengm]

Weed78,
Has the perite chamber been allowed an opening to atmosphere? A valve left open while not in use? Been stored in a damp area? Any situation that increases the moisture exposure to moisture is going to increase your drawdown time. Moisture will need to outgas and, if memory serves, this occurs usually at about 400-500 microns.
Web definitions for Perlite
"Perlite = A mineral, which when expanded by a heating process, forms light granules. Perlite is a good addition to container potting mixes, to promote moisture retention while allowing good drainage." from

http://www.enviro-explorers.com/outdoor_classroom_curriculum/glossary.htm

Depending on the volume drying the insulation could take a while.
I agree with IAINUTS that some sort of helium leak test would be in order befroe investing much in this.

Griffy

 

[weed78]

Thank you all for your comments, they are very helpful.

iainuts: What type of problem with the perlite would develop if there was evidence of perlite dust? We are seeing some dust in the hose from the tank to the pump.

wfn217: What do you mean by "molecular flow" and why would that cause a problem.

griffengm: We do not know the history of this tank so I can't comment on whether it would have had any contact with atmosphere. Any idea what happens at the outgas level of 400 - 500 microns you refer to? Is this related to the boiling point of water under this level of vacuum, or is there something else happening?
Thank you

 

[rmw]

If it is moisture in the pearlite, then the 400-500 microns is related to the temperature of the moisture embedded in the pearlite.

Is there any way to elevate the temperature of the tank at all, even just a few degrees above atmospheric? This will raise the saturation temperature of the moisture and help it "boil off" faster. Otherwise, it is a slow process to boil it all out.

I've seen the same thing with refrigeration systems left open to atmosphere too long and when a little heat is applied to the places where moisture can be trapped, the moisture comes off a lot qucker.

rmw

 

[iainuts]

Good comments from griffengm and rmw. Helped touch off a few more thoughts.

Regarding the history, do you at least know the age of the tank? Should be on the plate. Who's the mfg? Who did you buy it from? The original manufacturer may be able to shed some light on the situation, have you given them a call? Have you talked to the company you purchased it from? If it's a very old tank (ie: more than about 30 years) that could introduce other issues. If the tank was used in a highly cyclic nature or allowed to warm up with any regularity, that could indicate problems with the inner lines.

We have a huge number of tanks rehabed and they all go to a mfg where warming the vessel is standard fare. They even have ovens, and they can cycle hot air/nitrogen through the inner vessel as well. If there's any significant amount of water in the vacuum jacket, it could freeze as you pull vacuum even though the temperature outside is above freezing. This might happen as the water boils and looses heat. At 70 F, water boils at .36 psia (19000 micron).

At low pressure the heat transfer is reduced significantly so it's tougher to get the heat to the water. Sounds like you don't have an oven though, so how about heating the inside? Set up a closed loop system that circulates hot, dry nitrogen or air through the inner vessel. You might try heating it externally also with any means you have available. Throw blankets over it and use electric heating elements on low points.

Also, vacuum is only broken using dry nitrogen or argon. You might consider doing this a few times to help get heat to any water inside. Use hot nitrogen and let it sit for a day or two.

Regarding the possibility of a leak from the inside, I think if the vacuum problem doesn't clear up you need to do the leak test by putting helium on the inside and do a mass spec leak check through the pumpout port. Hopefully that isn't an issue. If there's an inner line leak, the tank can be repaired at a proper facility.

Regarding the use of perlite, it does have a tendency to crush which is the dust you're seeing on the evacuation line. The problem is it creates a cement like pile on the bottom of the tank which can create extra stress on the lines. Have you considered removing and replacing it?

Another problem with the crushed perlite is that after it settles, it creates void areas at the top. This creates patches where there is no perlite between the inner and outer vessels which allows increased radiation heat transfer between the inner and outer vessels. You could remove the blow-off plate (break vacuum with dry nitrogen or argon first) and add some perlite to the top, but all that extra perlite is in the form of powder at the bottom. Better to just have the tank emptied of perlite, then refilled with fresh.

 

[weed78]

Thanks guys, your thoughts are very helpful.
The tank manufacturer is no longer in business and the tank is almost 40 years old. Is it reasonable to think that it may have outlived its usefulness?
We did a Helium Leak test but found no indications of problems with leaks into the vacuum space. We also pressurized the inner vessel with Nitrogen and did a stand check, with no evidence of problems.
Based on your comments we are focusing now on the perlite and will try to get some heat to it.
The tank was open to atmosphere for a short while during a perlite "top-up" so some moisture could have gotten in. Also, the fact that we needed to top it up indicates that a lot of the old stuff got crushed.
Does the crushed perlite still have acceptable insulating properties?
Thank you.

 

[wfn217]

Molecular flow: Gas-flow phenomenon at low pressures or in small channels when the mean free path is of the same order of magnitude as the channel diameter; a gas molecule thus migrates along the channel independent of other gas molecules present.

I meant it can take a long time to pump down if you have molecular flow.

 

[griffengm]

Weed78,
Rate of pumpdown is affected by surface tension of the materials. The feature that permits perlite to retain moisture in potting soil is what makes it a pain to dry as an insulator. When you get the vacumm to around 400-500 microns, the moisture begins to boil off but in small crevices, joints, and lattices in perlite the surface tension is stronger than just surface moisture. It takes a lot more energy to move it to vapor. This is why heating the system can help but the very property of perite making it a good insulator makes the additional heat less effective than it might be on an all metal joint.
We have found that pulling a vacuum, closing the system overnight, and renewing the level in the morning is somewhat effective in situations like yours. You might need to do it several times to get where you want to be.

Griffy

 

[WoodyPE]

Sounds like you guys have it figured it out. When I was an engineer in the refrigeration compressor industry I worked with some processes that removed all moisture from the compressor housings. This involved an oven baking process to boil it off from inside the compressor housing (typically 2 to 4 hours). We also had an advanced cryogenic vacuum system to pull off remaining moisture. It used extreme cold temperatures to "draw" the moisture and other gas molecules to a cold sink. The gas molecules naturally seek the cold sink. Pretty neat idea in physics.

http://www.angelfire.com/planet/gadgetforum/index.html

 

[PRESSEDSTEEL]

If understand correctly, you are trying to pump out a chamber that contains perlite. If that is so you could pump on that for days and never get down to a good vacuum suitable for leak testing. The perlite is acting as a sponge for air molecules as well as other gasses. These constitute a 'virtual' leak. Small cracks and crevices or screw threads (unless gthey are vented properly) will wreak havoc when trying to achieve vacuum. A method for testing this tank would be to pressurize the tank with two to five PSIG of helium and use the leak detector in 'sniff' mode. What type of leak detector do you have?

 

[MintJulep]

I don't know if the configuration would allow this, but you could try flowing heated desicated air or nitrogen through the space.