Hacker News Reader

Comment by xt00 | Hacker News Reader

For SPARC the big innovation sounds like it is being able to use some special superconducting electromagnets to generate the massive magnetic field in the tokamak -- since that part is such a critical part, it would seem like in the next say 5-10 years that it takes to get SPARC up and running there should be a parallel group just trying to improve that magnet design by say 2X, which would likely have a massive impact.. seems like a smart thing for National science foundation or darpa or whatever to be funding that.. (maybe already happening) but it would be a bummer to be talking about this in 2030 and them saying, ok now we just need to design a more powerful magnet..

View on HN

That's basically what's going on with ITER, it needs to be so massive because the technology available at the time of conception was not capable of generating more intense magnetic field. There is a very strong relationship between the strength of the field and the required size of the tokamak (hence time and cost of construction). By the time you're done better technology will be available, but that's how progress goes.

Unfortunately, it's not so simple. SPARC is just the latest attempt to benefit from stronger magnet technology and thus smaller designs, but another big problem remains unsolved: turbulence. Since smaller designs and stronger magnets directly imply higher field strength gradients, the plasma becomes harder to control for the necessary burn time to achieve Q>1. Lockheed also thought they could simply use stronger magnets to build smaller reactors faster, but in the end they discovered that smaller is not automatically better. I wouldn't be surprised if we also see the eventual SPARC redesign that's 100 times more massive. Unless we get a major theoretical breakthrough in the near future, ITER will probably remain the most likely design candidate for near term net energy production.

Can you provide a source for that? Because here's the head of MIT's fusion program, claiming the exact opposite: https://www.youtube.com/watch?v=KkpqA8yG9T4&t=2221s

See here for example: https://www-pub.iaea.org/mtcd/publications/pdf/csp_019c/pdf/... - but note that this stuff is not trivial. The intricacies of turbulence in MHD are incredibly complex and hard to study (near-impossible analytically and very hard numerically). Nevertheless, these scaling behaviours (see the figures in section 2) have been known for more than 20 years by now.

Also, beware that the head of MIT's fusion program (your source) is the head of SPARC. Since he siphoned a huge amount of money from gullible venture capitalists outside his field by now, you definitely shouldn't rely on him for an obective analysis of SPARC's fundamental design.

Well that paper is above my pay grade so I'll have to trust you on it.

I will push back on your second paragraph though. The presentation I linked was from two years before the founding of CFS. It seems likely to me that Whyte founded CFS because he believed in the physics he presented, rather than the reverse.

I see this sort of causation reversal and associated accusations all the time in internet discussions. For example, yesterday someone told me that Musk only likes electric cars because he owns lithium mining rights. Tesla bought those rights in 2020, because they produce lithium batteries.

SPARC is in fact a heavily marketed continuation of ARC, which has been around since 2014 (https://arxiv.org/abs/1409.3540 - note the same authors). So the commercialization ideas would have started way earlier. But a project lead believing in the project shouldn't matter in any objective analysis anyways. The guys over at Lockheed also certainly believed in their compact fusion reactor design when they started working on it (which was, incidentally, one year before ARC), but they too had to learn by now that belief alone is not enough when it comes down to the laws of physics.

Don't get me wrong, I'd love for SPARC to be a success. I just haven't seen any objective, external review that would confirm (or at least try to show) how they solved the problem of controlling turbulent plasma in a high gradient field. It's not even fully clear that ITER will be able to do that, and they'll have a much easier time due to the larger reactor radius.

The guys over at Lockheed weren't in charge of one of the leading academic fusion programs in the US. ARC in 2014 was just a design created by students in one of Whyte's classes.

At this point, as a woefully ignorant layman, I'm forced to weigh two contrary claims. One is by the head of MIT's fusion program. The other is by a commenter on HN, helpfully linking an impressive-looking paper I don't understand, which does not appear to say with layman-friendly clarity that higher magnetic fields make fusion plasmas less stable.

Hey, I'm not trying to convince you of anything. In fact I couldn't care less what people think of this specific project. It's not my money that these VC funds are spraying. I'm just trying to convey a modicum of common sense when it comes to evaluating sources for particularly extraordinary claims. I see Elon Musk get torn to shreds regulalry around here whenever he says something slightly optimistic about electric cars or rocket science, so I frankly find it astounding how easily people eat stuff that comes from guys behind private fusion companies. I guess plasma physics and continuum mechaniccs are less impugnable than rockets and cars for most people.

Well I was hoping you were trying to convince me, and would have an easier-to-digest source handy, or a simple explanation of the relevant part of the paper. You certainly don't owe me that, and I don't mean to doubt you, and for all I know you're a professional fusion researcher. Plenty of people like that are on HN. But I don't know, so I'm just muddling through as best I can, trying to get things right.

Personally, the claim that stronger fields make plasma more stable doesn't seem all that extraordinary to me, and certainly not one that's contrary to simple common sense. And when it comes to evaluating sources, Dennis Whyte seems like a fairly reputable one.

Well SPARC and the planned successor ARC can successfully generate power without needing more powerful magnets. Assuming all goes to plan (lol) ARC will be commercially viable with existing magnet tech. They have said that as you go towards higher field strengths, the physical forces pushing and pulling between magnets gets so extreme that it is difficult to build a machine that doesn't tear itself apart. So stronger magnets are only useful in conjunction with an entire machine design that can handle them. It looks to me like a good path is to get commercially viable machines with existing magnets, and then iterate on the entire machine once they have learned the basics of operating commercial fusion.

But I am just a robotics engineer who has watched a lot of the SPARC lectures, so I have no domain expertise here.

This goes for permanent magnets as well. I've had a ceramic magnet and a neo explode on me and to say that I was surprised at the incredible forces unleashed would be an understatement. I'm a sucker for working with proper safety gear and if not for wearing very good safety glasses I'd have been in serious trouble, multiple magnet fragments sat embedded in the glasses and a couple in my skin where it wasn't covered by the glasses (easily removed with another magnet :) ).

How large of a magnet are we talking, and what were you doing with it? So I know what to be terrified of, for future reference.

The ceramic one was a 2x2" square 1/4" thick one that had one stuck corner, trying to remove it shattered it, the other was a neo that accidentally jumped up when I put down a heavy piece of steel and that in turn allowed the magnet to start flying towards it, it shattered on impact. That one was a lot larger (2"x2"x1"), and neos are a lot more powerful for the same size as ceramics. In both cases it was to test sets of magnets under controlled circumstances to see which ones I would order in bulk for the rotor of a windmill.

Lessons learned: make no assumptions on whether or not a magnet really is free to be released, do not place magnets loose on a workbench even if they are far away from your work area because when the friction unexpectedly disappears you are going to be in for a surprise.

Don't mess around with very large neos unless you know what you are doing, they can really hurt you. (And when you do know what you are doing: observe safety precautions religiously and always wear safety glasses until your whole rig is assembled and the magnets are safely enclosed).

Something like this just scares me:

https://www.apexmagnets.com/magnets/6-x-2-disc-neodymium-mag...

I saw a PPPL group propose the use of permanent magnets to assist in stellarator geometry. My first and continued thought has been "this is completely insane". Permanent magnets are fragile, violent, and cannot be turned off. All engineering challenges aside, they propose making an industrially sized death machine that cannot be turned off.

You actually can turn them off. You can use mu-metal to short circuit the magnetic field.

https://en.wikipedia.org/wiki/Mu-metal

That can provide shielding, but how does that turn off the permanent magnetic field when a human considers working on (maintenance or decommissioning) the machine?

Well, during decommissioning that would be a tricky thing to do, but during maintenance I could image a way to slide the mu metal in front of the magnets shunting the field away in a safe way. But you are right, working on such gear requires strict precautions and attention to safety detail.

Let me give you one example of how hard it is to work with these: while building the windmill I had decided on an 'inside out' model where the rotor rotates around the stator, this has a bunch interesting properties, such as pushing the magnets onto the drump while they rotate instead of trying to eject them from the rotor in the normal designs. And because magnets get hot any kind of glue or other adhesive would be subject to all kinds of thermal stresses, might let go or melt. So that was a pretty neat solution. My rig for installing the drum was a half ton crane over a very heavy steel workbench that had the stator mounted to it.

As I lowered the rotor suddenly the whole assembly, table and all lifted off the ground and perfectly centered itself around the rotor. I swear I never ever saw that one coming and it really scared the crap out of me because only minutes before then I was still fiddling with pushing some wires in place inside the stator while the rotor was already floating above it. A bit earlier and I could have easily gotten pinched, in spite of all my precautions.

Reply on news.ycombinator.com