Chamber Construction/Strength

[David Kunkle]

Hi all,

Trying to finalize plans for the chamber dimensions. It will be a cylinder18" in diameter and 18" long made from 1/8" thick SS. Archives stated that 1/8" is even overkill for chambers up to about 18". But, is it OK if the ends are flat (not curved or domed) or could they deform under vacuum? One end will have a 10" flange which should add a lot of support. The other end will be solid. Thanks.

[Rich Feldman]

Of course they will deform under vacuum.
The question is how much, and will it be enough to create a problem.
An 18" diameter plate of 1/8" SS is not going to burst,
but it will deflect inward enough to see next to a straight edge.
There will be lots of bending stress at the joint where it meets the cylinder.
Will it be welded to the cylinder, or welded to thick, bolted flanges?

Here is one online calculator with which you can figure the deflection,
even if you aren't particularly handy with strength of materials.
http://www.engineersedge.com/beam_bendi ... ending.htm

[David Kunkle]

The 18" end plate will be welded to the 18" diameter cylinder. The one end plate will have an ISO250 half nipple welded into the center.

Looking at your link to the calculator, I used 238 mm radius, 3 mm thickness, modulus of elasticity = 200,000 (wouldn't let me change it, but I found 193,000 for SS- close enough for now), Poisson ratio .29, load pressure of .1014 N/mm2.
I may not be understanding what my "Inner radius limit of loading, ro" is for my planned chamber. I thought that would just be my total radius. If I put it 238 mm, it gives me 0 center displacement, and 234 gives .036 mm??

Thanks.

[Rich Feldman]

Good work!

Try it with ro=0.
That means the load pressure is applied from the outside edge all the way to the center.
When you say ro=234, that means the load pressure is only applied to the outermost 4mm.

It will be interesting to see the difference between the
"free edge" formula, which would apply if you had a rubber gasket between disk and cylinder,
and the "clamped edge" formula, which is closer to the welded-joint condition.
In one case there's a slope but no bending moment at the outer edge.
In the other case there's bending moment but no slope at the outer edge.

[David Kunkle]

ro=0 gives a response of "NaN".
ro=1 gives 42 mm center displacement for simply supported, and 10 mm for a fixed support. A whole cm doesn't seem acceptable if not a little scary.

Going to 3/16" SS plate (4.76 mm) gives only 2.6 mm displacement, and 1/4" plate gives only 1.1 mm displacement. I may try to go with 3/16" for the ends then- unless 1/4" isn't as outrageous in cost as I'm guessing it's going to be.
Do you have any idea how close the "fixed support" calculation is to a welded joint?

Maybe I should talk to the shop that'll be welding it together about constructing partial domes for the ends and then use 1/8" SS for everything? Might not be that bad compared to the cost of going with thicker endplate material?

Thanks for all the help.

[Rich Feldman]

Do you have any idea how close the "fixed support" calculation is to a welded joint?

The welded-joint case would be intermediate between simple and fixed support, if the cylinder and end plate have the same material thickness. Assuming that the weld is strong enough for the bending moment, a longitudinal section of the tank would still have 90 degree corners. But that 90 degrees would be rotated slightly. The end plate would be "dished" all the way to its edge, but not as much as if the edge were simply supported. The end of the cylinder will have a corresponding taper, creating a bulge as we move away from the end.

It's time for empirical knowledge or finite element analysis. Good luck. You might confirm the online calculator result with another calculator or formula -- I just googled deflection of circular plate. Also it was fun to google vacuum vessel heads and look at the images.

Large metal vessels not strong enough to withstand vacuum can buckle and collapse violently, as in the railroad tanker car demonstration at http://www.youtube.com/watch?v=Zz95_VvTxZM

[David Kunkle]

Between the thickness, cost and weight of going to something like a 1/2" to 1 " thick end plate, I started looking at other options. A google search hit on "tank heads", and I wound up talking to these guys: http://www.tankcomponents.com/tank_head ... _heads.htm

I got a quote from them for parts to make a 20" diameter and 20" long chamber when assembled.
2 tank heads- $194 each
1 Cylinder- $237
SS304, 3/16" thick, mill finish (#180 grit polish inside only for an extra $150) Think I'll polish it myself- most of it should be easier if it's in 3 pieces to start.
The dish of the heads is 4-9/16" deep. Anyone know if the dish of the head is deep enough to counteract a vacuum? I told him specifically that this is for a vacuum chamber.

1/8" is no doubt cheaper, but I'm borderline needing 3/16". I figure an imploded chamber would be a lesson learned the hard way.
I'll still need my local shop to weld it all together along with all the flanges. No doubt they can't stamp dished heads out of sheet metal, and I figure this company can make the cylinders cheaper than a local shop can make one custom cylinder.

Searching here for tank heads turned up nothing. I wanted to run this by everyone in case someone might know any reason this might be a bad idea, and if a good idea, it may be a good option for others.

[Richard Hull]

Assumption here is that you are absolutely fixed on the idea of an 18" diameter cylindrical chamber. If so, go for the quoted prices. You must have some special series of experiments planned. With so large a chamber, are you planning on changing out grid structures or running tests where the large volume will be needed for materials inside to be altered regularly?

This is a very large chamber and will require a plus ultra vacuum system. No simple two stage pump of under 5CFM and a 4 inch diff pump or decent sized turbo will be demanded if you look to work with usably short pump down times. The pumped deuterium gas bill will be significant as well with this rather massive volume.

I had such big ideas only rather recently. I bought a fabulous chamber with many ports rather inexpensively for so large an ensemble. Reality set in, thankfully, and I abandoned the idea as a wasteful exercise involving a bit of megalomania on my part. Fortunately I sold the chamber without loss to my purse. Pumping issues and deuterium consumption were the sticking points.

Good luck

Richard Hull

[David Kunkle]

Actually, I do have some "special" experiments planned. I'm pretty well set at an 18 to 20" chamber at this point. Probably will be changing some grids, but the size of components makes it counterproductive to scale it down much further. I'd make it bigger, but I think I'm now at a good compromise that'll make this a satisfactory testbed.

For pumps, I have an ebay Trivac D4A that pulled down to about 3 microns right off with a bunch of NPT fittings, Teflon tape, a piece of soft clear tubing and some hose clamps to attach it to a TC. It was shipped dry, and the initial oil change wasn't road tar, but not terribly pretty either. Successive oil changes, including flushing fluid, didn't seem to lower the ultimate pressure. And I have Andrew Seltzman's rebuilt 200 l/s turbo. Hopefully, that doesn't put me into the multiple hour pump down territory?

Not worried about pumping out deuterium. Initially, it'll be run in demo mode before moving on to other uses that shouldn't involve pumping out deuterium.

Might be too bad I missed the sale of your chamber. I've been watching ebay for some months and still haven't run across anything close to suitable. I've decided it'll be just as well to start from scratch- almost anything used would still need a bunch of hacking and welding.

[Richard Hull]

Sounds like you are pretty well setup for the large chamber. Good luck in your efforts.

Richard Hull

[Rich Feldman]

Hi David.

Good luck, and please tell us what you learn about your tank head thickness requirement.
I did a little more investigation of design calculations.
Much of the following may be redundant with your own searches.

The external pressure case is, of course, more complicated than that for internal pressure.

There are a handful of industry standard head shapes, such as 2:1 semi-ellipsoidal.

I think most "vacuum tanks" are designed for only partial vacuum, such as we might find in a honey wagon application.

Fabricators eagerly provide data on volumes and weights, but seem to assiduously leave all questions of strength to the customer's engineer.

Most leads I have found have the goal of "building to code". In most of the world, the code is called ASME VIII, more specifically ASME VIII-1 UG-28,
which is probably hard to find for free on the Internet. You can buy it at https://www.asme.org/shop/standards/new ... -code-2013
One handbook with formulas is Pressure Vessel Handbook by Eugene Megyesy.
There must be some fusor forum readers with easy access to resources like those.

On other forums I saw suggestions to use pressure vessel design software with demo versions or trial periods.

Here are a couple of my favorite instructional links.
http://www.pveng.com/ASME/ASMEComment/E ... essure.php
http://www.eng-tips.com/viewthread.cfm?qid=84873
http://www.eng-tips.com/viewthread.cfm?qid=196371

[David Kunkle]

Thanks for the links, Rich. The first one was very helpful. At first I was worried, but after getting to this part, I think I am in good shape.

"A typical vessel 48" diameter with a straight shell 96" long needs a shell 0.056" thick for an internal pressure of 30 psi, but needs to be 0.225" thick for a 15 psig external pressure (full vacuum) per VIII-1 UG-28. The F&D head on the left end needs to be 0.082" thick for the interior pressure but 0.142" for the vacuum. A SE head needs to be 0.046" thick for the interior pressure but 0.127" for the vacuum. The F&D and SE heads are both calculated as if they are part of a sphere, but the two heads are given different equivalent radiuses resulting in different required thicknesses."

Before that, he stated how increasing diameter and length, and decreasing thickness weaken a vessel.

For a comparative whopper of a tank at 4' diameter and 8 ' long, it only need be .225" thick and .142" thick on the heads. What I already ordered is 3/16" or .1875" for my 20" chamber. The heads are F&D- about 2.2:1. Only about .04" thinner on the cylinder, but .05" thicker on the head. The author goes on to say this large tank amounts to 4x the collapse strength, and 3x is the usual safety factor.

Even if he is using some metal stronger than SS304 on this 4' X 8' tank, my less than 2' X 2' cylinder with .188" thickness should be way overkill based on my smaller size- at least I would think. Cost is $625 for the cylinder and 2 tank heads, and weight is about 90 lbs. Might be somewhat cheaper and lighter to go thinner, but I'd rather not worry if it's strong enough. I found one manufacturer of SS vacuum hemispheres that goes from 1/8" to 3/16" at 18 to 20 inches in diameter.

Thanks for the input.