superconducting magrid design

[mvanwink5]

Support each coil separately from the wall, not like shown on the WB-6, that way you are just making one magnet X 10 for a dodecahedron (or 6 for a cube). In fact that is how folks here think WB-8 is built, with 6 separate coils. Nubs are thought to be subject to too much heat and abandoned in WB-7. (I think I am understanding your objection).

Look at the picture of WB-7.1, there are no nubs.
http://www.emc2fusion.org/

[mvanwink5]

The SC would not need to be driven by HV, but the case would be at a high voltage and so the SC would need to be insulated from the outer case. At least that is what I am thinking.

[jon117]

gotcha. okay, that way it'll be then. i'll get on re-designing it later on tonight. glad you understand what i meant about the welding, i could illustrate it, but that'd be overkill.

by outer case, do you mean the outer grid? the part that's outside of the magrid and held at a high -ve voltage to push electrons in towards the center?

thanks btw, this is all very insightful and helpful!

[mvanwink5]

Jon117,
The magnets are held at +HV with the vacuum chamber at ground, so the SC inside of the water jacket would probably be best to be insulated. Hopefully that makes sense.

[jon117]

updates:

as you can see shrinking the size further is limited by the diameter of the nuts and bolts holding it together...

Do you see what I mean now about the feed-thrus?

I still think there will be massive issues separating them. I mean yes, you could use L-joints with holes cut in them, but whatever you do to one feed-thru you have to repeat at least TWENTY-FOUR times... and you've got at least five pipes! 120 separate parts JUST for the cooling? and all of them need to be welded to an extremely high quality? No, I think that's getting way out of hand and is setting yourself up for failure.


There must be a more elegant solution. any ideas?

[mvanwink5]

Jon117,
Just brainstorming here, why does the back side of the coil need a water jacket? Further the chamber itself is vacuum. That should simplify things a lot, no? There still needs to be structural material, but the nesting is cut way down I think.

Also, isn't the LN only needed between the water jacket and the LHe? So, is nesting LN around the LHe needed for the feed throughs? Of course outside of the vacuum chamber things change, but if it is too much trouble, maybe that can be the point to go to LN cooled copper, or water cooled copper, or just copper?

Another thought, what about putting a circular copper bar on the backside for loss of SC state and current collapse, to keep the B field from collapsing? Isn't that standard SC magnet precaution?

[hanelyp]

Thinking there must be a more compact way to fit shells together than those flanges. I'll be back when I have an idea illustrated.

[mvanwink5]

Flanges are because the upper and lower shells are bolted as opposed to welding them as was done by Bussard in WB-6. Is that right? Still, on the previous drawings, there were no flanges, but the halves were bolted.

[hanelyp]

[mvanwink5]

Based on what I know about helium, it might be best to seal weld the tube that encloses it. As far as I know, the rest could use bolts...JMDAO. BTW, we used helium to leak check vacuum leaks. It might be important to take care not to have helium leaks as that might make troubleshooting vacuum gasket leaks difficult. Just saying.

[Teahive]

I suppose any kind of interlocking seam that self-seals via the pressure of the liquid inside the tube would be too leaky?

[mvanwink5]

I'm thinking construction would be in three parts:
1. bottom (plasma facing) assembly with water jacket, LN jacket and insulation (vacuum zones)
2. Sc and helium shell assembly (I'm thinking seal welded)
3. top assembly, just vacuum, maybe inner shells for strength, maybe not.

As far as I can see the top and bottom assemblies could be bolted, or welded together.

[mvanwink5]

Helium is monatomic, unlike H2, and is known to leak out of canned, sealed vacuum tubes, causing the cans to collapse. We used to apply helium to seams to find vacuum leaks, so I can say a very large seam (3 meters circumference) will be nasty to seal right. That is just my thinking, but some engineers may have sealing systems that would work...

Just remember, most vacuum seals just have to work against air and might not be so good against helium.

[jon117]

sooo... I just realized something. cooling the LHe tube to ~4K will effectively create a cryopump. Now, you wouldn't want any water vapor trapped in the vac chamber (obviously), but please someone correct me if I'm wrong, but we don't NEED the thermal insulation vacuum chambers to be at ultra high vacuum. I know using a roughing pump before-hand would be preferred, but anyone have an idea if we could get away with JUST using the cryocooling properties of the pipe to pull a decent enough vacuum?

If that's the case, we wouldn't NEED feed-thrus for them. and we just saved ourselves a massive set of headaches.

also, great design hanleyp

mvanwink5: are you sure you're not thinking of superfluid helium? that shit's nigh on impossible to keep contained...

[mvanwink5]

Jon117,
No, just He. I suspect He that leaked into the space between the LHe sand LN would transfer heat and any air in the gap between the LN and water jacket, same thing. That is just my thinking with no numbers, so maybe I am wrong. So, I take it you weren't impressed with the idea of just using the water jacket and LN on the plasma side? It would make plumbing easier (at least it seems to me).

PS The vacuum zones in the coil just require venting as the chamber is vacuum.