@veronica CFS is an *incredibly* well-funded ($2B) spinout from the Alcator C-Mod team at MIT with productized 20T high-temperature superconducting magnets.

https://cfs.energy/technology/#sparc-fusion-energy-demonstration

@elfprince13 20T, that's impressive. I was a fellow on the HiLumi LHC project at CERN, and when I left they were still working on the 11.4T magnets needed. Granted, that was in 2019.

@veronica I was in the DOE's National Undergraduate Fellowship program at PPPL back in 2012 and ended up deciding I'd rather do grad school in CS so I didn't have to spend the rest of my life implementing PDEs in Fortran, but one of the guys from the program who stuck with it is now their (CFS's) Head of Tokamak Operations.

@veronica and yeah, the 20T demo last year (and subsequent reveal they already have supply-chain in place to crank out more) was when I suddenly got optimistic about fusion again.

@elfprince13 If I recall correctly, a major challenge with our magnet design was the superconducting wire for the 11.4T. It was incredibly brittle compared to the type used for the current magnets in the LHC.

I mostly worked on simulations for the radiation in the magnets from the experiments, so I wasn't directly involved.

And yes, I worked with Fortran 😊

@veronica brittleness definitely seems like an annoyingly undesirable property for "wire" 😅

@elfprince13 Ah, the CSF one you mentioned is the one called SPARC. That one I've heard of.

@veronica yeah, SPARC is the (still intended to sustain net-energy-positive fusion) prototype, and then ARC is the intended commercialized follow-up

@veronica @elfprince13

I'm not really a magnet expert, however the geometry also matters in how easy it is to get a high field.

The LHC magnets are two-aperture bending magnets with a somewhat complicated winding and mechanics, and extremely tight tolerances for field quality at a wide range of currents, and not much room for adjustment once they are made.

@veronica @elfprince13

It could be that the high field magnets used in tokamaks (torroids, D-shaped solenoid that bites its own tail) are easier to get up high field?

@kyrsjo @veronica so this is related to one of the reasons I historically have favored stellarators - with a tokamak you're imposing an external magnetic field on the plasma (which really just wants to faff about and do its own thing).

@kyrsjo @veronica Moreover, the magnetic field is weaker on the outside of the torus due to relative increase in surface area with same number of coil windings, so as particles take random walks out through the field, you get one-sided error, and they begin to concentrate outward which is bad

@kyrsjo @veronica with a stellarator you physically bend the plasma into a shape that induces its own useful magnetic field, in such a way that you have a globally average good curvature minimizing the tendency to drift outward, but sadly the solutions to the differential equations letting you physically realize this really cool idea leads to shape require stupidly tight tolerances to manufacture and are thus much more expensive to build at any kind of scale.

@kyrsjo @veronica this had led to a bunch of political decisions wherein tokamaks get favored as the "easy" solution, and stellarators have a hard time getting funded despite being the "more correct" approach (PPPL literally has a stellarator in pieces in a garage because funding got cut before they were allowed to put them together)

@kyrsjo @veronica anyway, politics aside, the tokamak folks basically just need to make a "big enough" model so that the curvature isn't too crazy to get the confinement they want (ITER), or just build an insanely powerful magnet to do it at a reasonable scale (CFS).

@kyrsjo @veronica there are some other variations playing with the curvature of the basic toroidal design within this overall space, but that's the basic gist of the MCF game*.

*if you really want to deep dive there are also Field-Reverse Configurations (FRCs), which have a few passionate devotees, but are even more stalled (public-) funding wise than stellarators.