Is There an Optimal Size for Magrid Casings?

[tombo]

You called for 10T and 16e6 amp-turns
Indrek's equation shows force going up as current squared.

A 1 T field is a LOT easier.
And, 2 meter diameter coils are a LOT harder.

So how many amp-turns are we going to design for?
It makes all the difference between merely more money and UnObtainium.

(That is right next to UnUnObtainium on the periodic table.)
(It kinda feels like the money is UnObtaininum too.)

[tombo]

Dropping the field by a factor of 10 to 1 T
160,000 amp turns gives 65,000 N or 14,000 pounds.
That's easy. I can hold that with a 3/4" (18mm) dia cheap steel rod.
With good steel I can go a lot smaller.
Say one 7/16" grade 8 bolt (10mm grade 10.9 in most of the world.)

Actually we would want to use non-magnetic materials.
But you get the picture: Please specify a field no bigger than we truly need.

[KitemanSA]

It makes all the difference between merely more money and UnObtainium. (That is right next to UnUnObtainium on the periodic table.)

Nope, look again. UnUnObtainium is next to UnNilObtainium. UnObtainium is ten spaces lower!

[KitemanSA]

YES. This is almost perfect except the sides should be aligned a bit more

They are aligned exactly on the cube edges. It is hard to tell unless you are looking exactly along the plane. But that view does not show the overall shape well.

You know, when I look along the equitorial line, I think I can see what you mean. At least along that line, the centerlines of the pipes look aligned. Other angles look off. Oh well, I'll take your word.
The other drawing you have of the "square" plan form independant coils look to have the outside edges of the coils aligned, not the centerlines.

Please bow them, but not make them circular.

Yes that was what I was trying to describe. Words are so hard to use when describing geometry. (I have not yet because that shape is harder to draw.)

I can't even do 3D, so you a miles ahead of me!

I think it will improve the funny cusp performance.

I've heard that before. But why? or maybe more to the point, how?

Not being a physicist, I can only guess as to his reasons, but to me, I think the length of the quasi-linear cusp gets MUCH shorter, and the higher curvature of the magnet causes the field to run more uniformly between real and virtual coils. But then what do I know?

Also, if we can place thin tensile members across the vertexes connecting the circular arc segments then most of the forces resolve to simple hoop stresses.

The problem of course is the plasma. But, I think that is the lowest plasma density highest B field region in the whole sphere.

My aim was to provide support around, but not directly in line with, the funny cusp which has been intentionally made as small as possible.

funkiest looking tire I've ever seen

Yes, I was just picking up the term as a zeroth order approximation to set the scene before getting on to what I really wanted to say.

Gotcha.

WB-6 failed due to movement of the coil windings during pulses abrading the insulation of the coils leading to a catastrophic defect.

All the more reason to think hard and early about the stresses on those SC's.

True, and I expect the best way to limit the stresses is to make as much of them axial (hoop) as is possible.

[Billy Catringer]

Chief Simon? I'm gonna wanna cast these pieces. Right now that looks like the cheapest way to get the job done.

Will that cause problems for the plasma or magnetic fields?

[tombo]

Billy, You're going to cast multiple layers right onto the superconductor coils without overheating them? I suppose the first layer IS insulation but it still sounds dicey. Thermal control will be critical. Do you have a work around in mind? Do you have enough casting experience to tell me than I'm jumping at shadows?

Re my previous post:
With only 14,000 pounds we might not even need corner ties in the one-pass bent tube configuration.
We really really need to know the field strength needed.

[tombo]

I think it would be quicker and give more confidence to mock up a coil set and measure the forces on a coil than to do a full up computer simulation.

I think a 0.15 M (5.91 inch) radius set of 5 fixed coils to match Indrek's configuration.
Then place 1 identical coil on the top cube face suspended on a simple balance beam with adjustable weights and constraints (thread?).
Or the far end of the balance beam would push down on a triple beam balance or other suitable (digital?) scale. (Unfortunately I don't have these, or space to set it up.)
I would use a car battery and a set of ballast resistors to get data from several different currents spanning at least an order of magnitude.
Wire all 6 in series to balance the currents.

We can confirm Indrek's equations at http://www.mare.ee/indrek/ephi/force/
Obviously we use much lower currents but we can see if his curve fits the data.

Also if the top movable coil is suspended to swing & twist freely (say from its centerpoint) it will give information about lateral and torque stability/instability.

I would welcome any suggestions about tooling.
I'm thinking corrugated cardboard (apple box, duct tape) and zip ties with a few dozen turns of transformer wire.
But, I would really like to find a cheap/free round form to wind them on and to hold them properly round during the measurements.

JohnShearing did some nice mockups similar in style to what I have in mind but much more complicated. viewtopic.php?p=11252&highlight=tooling#11252
John are you still with us?

The next step would be to model square planform coils arranged as a truncated cube.
That one will, for sure, need a measurement of the torque around the coil axis, maybe using a pair of "L" shaped balance beams.
It will also need a stiff axle and a protractor. Later it will need a flexible axle run and ways to measure more degrees of freedom.
Left to their own devices those corners pointed together will push away from each other. They will all want to turn 45 deg to form a non-truncated cube IMO. They may well also want to flip over.
This one is complicated by the desire to explore the effect of the corner radius on the forces. And, that requires multiple coil configurations.

The questions that need answering here are
1. How strong will we really need to make the supports?
2. How close to the perfect alignment will we need to keep the coils to keep support forces within reason.

A later experiment could repeat some or all of these with a conductive sphere simulating the plasma ball. Say, aluminum foil glued over a toy ball the right size.

It is amazing how a simple proposal snowballs into a much bigger one, even before it starts.

[Billy Catringer]

Billy, You're going to cast multiple layers right onto the superconductor coils without overheating them?

Nope. Two castings for each jacket. 316SS or some other austenitic. We could, maybe, bend pipe and split it after bending for the vacuum jackets, but even the vacuum jackets will see a lot of stress.

With only 14,000 pounds we might not even need corner ties in the one-pass bent tube configuration.
We really really need to know the field strength needed.

As much as I would LOVE to have tension members in this assembly, how are we to cool them? The plumbing is already complicated and very iffy.

[tombo]

Nope. Two castings for each jacket. 316SS or some other austenitic. We could, maybe, bend pipe and split it after bending for the vacuum jackets, but even the vacuum jackets will see a lot of stress.

Then what? Weld the 2 halves together the long way?
Possible but tough.
Welding on castings is always an iffy proposition.
Also, this has to be done for 4 or 5 layers.
The outer layer also has to be leak tested to what? 10e-9 or 10e-12?
I guess there are welders who can do that kind of work, but not many.

As much as I would LOVE to have tension members in this assembly, how are we to cool them? The plumbing is already complicated and very iffy.

Yes, I know.
Somewhere I found an equilibrium blackbody temperature for the environment and it was not impossible although very tough.
It was in a thread about carbon fiber or CNT grids.
I recall that the upshot was that it was possible for carbon fiber and IIRC maybe even moly or titainium. (still need to watch out for radiation)

Also the tie only needs the outer layer of water cooling.
So, a piece of 316 tubing welded onto the outer jackets should do it.
It would get the same coating of in-situ vacuum-evaporated Boron that the rest of the interior gets for neutron protection.

Another thought to reduce piping complexity is to have each coil mounted directly to the outer wall and fed independently. The control people might like this option too. Maybe on a tripod of radial struts with only 1 needing to be fat enough to hold the superconductor. It only needs to be fat enough to hold 2 sc wires and not all the turns of sc wire.

Back to casting:
The sc is sintered at 900C in the factory.
Also, some sc's are supposed to be annealed or something after forming.
So, they can take considerable heat during fabrication.
The inner layers could be made from material that melts considerably lower than that.
By the time you get to the outside layer that has to be 316 or copper (1050C) or something else hot, some of the inner cooling channels could be made operational.
I myself would rather try that route. If casting is to be used at all, we should use its full potential.
I'm liking this idea more all the time.

Does anyone out there have casting experience?

[MSimon]

But you get the picture: Please specify a field no bigger than we truly need.

OK: as much as we can get. Remember B^4.

[MSimon]

As much as I would LOVE to have tension members in this assembly, how are we to cool them? The plumbing is already complicated and very iffy

.

Build a cage - either internal or external. Eventually it should be designed into the chamber walls. We don't have to start there.

[Billy Catringer]

Then what? Weld the 2 halves together the long way?
Possible but tough. Welding on castings is always an iffy proposition.

Welding on cast iron is tough and iffy, so much so that is often better to braze cast iron pieces rather than weld them. Welding on cast steel or cast stainless is a lot easier. The long seam welds will be necessary no matter how the torii shells are formed. Casting accurate half-shells for the torii will be a much simpler prospect than forging them or forming them out of sheet and a lot cheaper.

Also, this has to be done for 4 or 5 layers.

And, in-the-tube supports must be designed as well. I know.

The outer layer also has to be leak tested to what? 10e-9 or 10e-12?

Er, not sure what values you are talking about here. All of this will have to be built in accordance with ASME B31.1 (power piping code), so yes, all of it will undergo 100% UT examination during and after welding, dye penetrant between each weld passes and pressure testing up to one-and-one-half times the working pressure held for a minimum for one hour with certified charts for temperature and pressure variations. I would have also said radiograpic examination as well, but there is no way x-rays will tell you anything with these geometries.

I guess there are welders who can do that kind of work, but not many.

I got welders available who can weld up the crack of dawn. You want welders? Get out your checkbook. That's the least of our worries.

Somewhere I found an equilibrium blackbody temperature for the environment and it was not impossible although very tough. It was in a thread about carbon fiber or CNT grids.
I recall that the upshot was that it was possible for carbon fiber and IIRC maybe even moly or titainium. (still need to watch out for radiation)

Assume, just for the sake of smirks and grins, that Doctor Neble makes his results public tomorrow morning. Will these super-ohmigod materials be available? If so, great. Otherwise, plan on building a structure stiff enough to do without 'em. Oh, that reminds me. I'm thinking we might do well to use 9% chrome-1/2% moly alloy for the cool water jacket. I need to check on its magnetic properties.

Also the tie only needs the outer layer of water cooling.
So, a piece of 316 tubing welded onto the outer jackets should do it.
It would get the same coating of in-situ vacuum-evaporated Boron that the rest of the interior gets for neutron protection.

I'll 'splain my problems with the plumbing before too very long. For the moment, believe me when I say that it is going to be damnably difficult to cool tension supports.

Another thought to reduce piping complexity is to have each coil mounted directly to the outer wall and fed independently.

This is the only workable solution that I can see. I am planning to do exactly this in the Blender model of this beast.

It only needs to be fat enough to hold 2 sc wires and not all the turns of sc wire.

Think about this some more. You missed one or two of the demons lurking in the details.

The sc is sintered at 900C in the factory.
Also, some sc's are supposed to be annealed or something after forming.

They are heat treated during and after forming. Copper and nickel, both of which are in SC jackets, work harden and must be annealed during the process. I would not be suprised to learn that they must also be annealed after or during the winding phase. This is another argument in favor of Bitter plates over windings.

By the time you get to the outside layer that has to be 316 or copper (1050C) or something else hot, some of the inner cooling channels could be made operational. I myself would rather try that route. If casting is to be used at all, we should use its full potential.
I'm liking this idea more all the time.[

Ah, yeah, we'd have to run water around the SC's, probably in the inner vacuum and LN2 jackets while welding up the cool water jacket. The cool water jacket will be thick enough to require pre-weld and post-weld heat treatment. After everthing has cooled down we could use an alcohol and a nitrogen blowdown to dry out the inner jackets. Yes, this last step has its hazards, but we are talking about industrial grade stuff here. It has to be done right anyway.

[MSimon]

Helium leak rate testing will be a requirement.

[Billy Catringer]

Helium leak rate testing will be a requirement.

And bellows sealed valves in the water loops. Gotcha. Will do the cage. I can't see a good way to avoid it.

[MSimon]

Billy. Blogged your "weld the crack of dawn comment":

http://powerandcontrol.blogspot.com/200 ... lders.html