Hacker News Reader

Comment by gpm | Hacker News Reader

Is there a good reason we aren't seeing REBCO magnets in all new MRI machines? Seems like not needing liquid helium would be a pretty big win for hospitals?

View on HN

Yes: Commonwealth Fusion essentially is buying the world's supply of it.

MRIs are a more mature market, i.e. > 1 units, and need to justify component switches not just for feasibility, but for profitability. Swapping out a known-good for something in such short supply isn't feasible, it's more expensive than liquid helium.

Source: https://spectrum.ieee.org/fusion-2662267312 "Over two years, the team managed to buy up most of the world’s supply of 4-millimeter-wide HTS tape"

Commonwealth Fusion is such a Fallout name

Todd is in shambles that he can't use the name.

They probably still can't carry the necessary current without helium cooling. There are some NMR spectrometers (which work essentially by the same principles as MRI machines) with high-temperature superconductors, but those are still cooled by helium. The lower the temperature is the more current you can pack into those.

Could also be that it doesn't make economic sense yet, they're still pretty new compared to classical superconductors.

MRIs don't consume much helium. Keeping them topped off is a trivial expense compared to the rest of the costs of the machine and operation. If there were a new design that didn't require helium, it would also have to cost less than current machines to make sense.

There are already non-venting systems that use a reasonable amount of helium (7L) to manufacture and don't require refilling. It just was not a design consideration previously. https://www.philips.com.au/healthcare/resources/landing/the-...

While 7 liters is less than 1,500, why would they call this a helium-free system?

They say helium free operation, as they say you never need to deal with the helium (e.g. no refilling)

> Is there a good reason we aren't seeing REBCO magnets in all new MRI machines?

REBCO tapes are still very brittle, and MRI machines are notorious for the amount of banging they produce when active.

???

the gradient coils make the noise, and due to field homogeneity requirements, it is safe to assume the superconducting coil is not moved by the gradient coils

homogeneity is only optimized for a small volume within the bore. Regardless, the superconducting coils will still experience a change in Lorentz forces as sequences are run, and noise from other components will still be "heard" by the main magnet

It's still very much affected by the vibrations, and the electromagnetic forces from the gradient coils.

I would say it's because the technology has not been matured enough.

I would also say the reason CFS has been getting funding is not because of fusion (although, hey, why not take that longshot bet), but because of the value of being able to make these magnets for non-fusion applications.

Wait, isn't the byproduct of fusion helium anyway?

Yeah but you'd power a city for a year for 100 kg of helium.

A single MRI needs around 1000 liters, 1L is roughly 130g. For those who wonder prices vary between $15-35 per liter. Recent MRIs have recycling systems that reduce refill needs by 90% when they work properly.

So about $15k for 100 kg?

Fusion produces much, much less helium than that. D-T fusion that produces 100kg of He-4 would produce something like 25PJ of net primary energy. If you managed to turn 20% of that into electricity, you'd have 5PJ. This is within an order of magnitude of the yearly electricity consumption of the entire world.

I do actually have receipts. For example numbers I picked Washington DC, which has 700k people that each use an average of 600 kWh per month. That's 5 TWh / year, which I rounded up to 10 TWh. This yields 475 m^3 of helium, which is 85 kg. I rounded up to 100. DC is a modest city and all of humanity uses 15000 TWh / year (of heat). This "100 kg / city / year" estimate is even on the lower end since I started with electric consumption and never put in the factor of 3 to convert to thermal.

Your mass-energy conversion number is within error of mine, but your estimate of how much power we use is much lower.

https://www.wolframalpha.com/input?i=%2810+TWh%29+%2F+%2817....

5PJ is not that large, only 1.4 TWh.

In contrast, world primary energy consumption is at a rate of around 20 TW, or five orders of magnitude more than that per year.

For anyone else wondering about the five orders of magnitude:

https://ourworldindata.org/energy-production-consumption

Yearly total energy consumption 160 PWh (~18 TW of power)

114285 times more than 1.4 TWh

Reply on news.ycombinator.com