Catch up on stories from the past week (and beyond) at the Slashdot story archive


Forgot your password?
Science Technology

Second Snag This Week Could Delay LHC for Weeks 160

sciencehabit writes "After a transformer failure earlier this week, the Large Hadron Collider has hit another snag — and this one is much more serious. As Science reports, 'At least one of the LHC's more than 1700 superconducting magnets failed, springing a leak and spewing helium gas into the subterranean tunnel that houses the collider ... How long [repairs take] will depend in part on how much of the LHC must be warmed to room temperature for servicing. If it's only a short section, the repair could be relatively quick. But the machine is built in octants, and if workers have to heat and cool an entire octant, then the cooling alone would take several weeks." Reader Simmeh contributes coverage from the BBC. We recently discussed the transformer malfunction at the LHC, which was a smaller problem and has already been fixed. Update - 9/20 at 12:52 by SS: CNN reports that the LHC will be out of commission for two months.
This discussion has been archived. No new comments can be posted.

Second Snag This Week Could Delay LHC for Weeks

Comments Filter:
  • by X0563511 ( 793323 ) on Saturday September 20, 2008 @12:44AM (#25082447) Homepage Journal

    Yes, because taking nearly 30 years to build this was rushing.

    Calm down, one of the magnets quenched. When that happens, it gets REALLY hot and things break. They knew it could happen.

  • by NixieBunny ( 859050 ) on Saturday September 20, 2008 @12:56AM (#25082527) Homepage
    I work on helium-cooled radio telescope receivers. They have trouble regularly - it sometimes takes five or six tries to get the thing cooling properly. These poor folks have over a thousand giant Dewars to keep cold! Give them a break.
  • by blueg3 ( 192743 ) on Saturday September 20, 2008 @01:09AM (#25082599)

    Having worked on one of these, two failures this early on is par for the course. There's a lot of work to be done even after the thing is build and initial testing is done before it's stable and working (and even then, most particular accelerators are only somewhat "stable" with very heavy maintainance).

  • by Animats ( 122034 ) on Saturday September 20, 2008 @01:39AM (#25082747) Homepage

    You can look at much LHC status online, including detailed cyro status. (I'm not giving the URL, so as not to Slashdot that server. You can find it if you really care.) Sector 34 of the LHC is at sector 34 at 4.5K-20K, instead of down below 4.5K where it should be. One of the magnets quenched and went normal, and much of the energy in the magnet is dumped as heat. Then the liquid helium boils to a gas and blows out through relief valves. But the sector hasn't been brought up to room temperature, so they apparently think they can fix the problem without major work on the magnet.

    Some of the cyrogenic magnets gave serious trouble last year, but apparently it's not as bad this time.

  • by Anonymous Coward on Saturday September 20, 2008 @02:51AM (#25083051)

    There is not just one line. There are 6 lines as far as I know. They transport superfluid helium as well as warm helium. Here is a paper about the cryo system:

    Anyways, they are now investigating with a remote inspection train that can travel in the LHC.

    Paper accessible here:

    Sorry but I am going to an anonymous coward -- but clearly, this post comes from CERN...

  • by perturbed1 ( 1086477 ) on Saturday September 20, 2008 @02:58AM (#25083073)
    About right... There are 1232 dipole magnets in total.
  • by imsabbel ( 611519 ) on Saturday September 20, 2008 @05:42AM (#25083537)

    I would expect at least a douzend of failures and faults of that magnitude until full power is reached.
    Its just too complex.
    And about the "expensive!!!1" aspect: A few months delay are so much cheaper than spending twice as much before so try to get everything 200% perfect. And even then things might go wrong.

    Even in a tiny normal synchrotron, shit happens. At the ALS in Berkelely they managed to detonante a main PSU because they only tested them one at a piece, and when build in they had bad crosstalk. Beam was down for several weeks.
    At the SLS in Villigen, even months after the full ramp-up beam instabilities or drops happened on rather regular basis.

    Such things happen.

  • by Anonymous Coward on Saturday September 20, 2008 @05:42AM (#25083541)
    Shit, you had better tell the US that. Fermilab seems to be rather under the impression that they a) built the quadrapole magnets for the LHC (which failed rather spectacularly a year or so ago and see the project back but thats another story) and b) that they are actively contributing to CMS commissioning, data taking and physics including setting up and maintaining useful systems like cmsmon (although that sadly is off due to greek hackers).

    The US has a huge role in the LHC and its experiments and contribute both man power, money and equipement. They are the biggest national group on CMS (although Europe as a whole is of course bigger) and have been involved in the construction of many subsystems.
  • by Richard Kirk ( 535523 ) on Saturday September 20, 2008 @06:57AM (#25083767)

    If you are not familiar with superconducting magnets, then some of these terms may seem a bit mysterious. So, here goes...

    A superconducting magnet is essentially a big coil of superconductor. Initially, you put current into the superconductor to build up the magnetic field. You then 'join the ends' of the superconducting loop, so the current circulates endlessly, and the middle has a constant magnetic field.

    There is a lot of energy in the magnetic field. An 11-tesla magnetic field has about the same energy per unit volume as TNT. Worse than TNT, there is no rest mass to the 'explosive' so all the magnetic field energy would be dumped straight to the surround. The surround is already under a lot of tension due to the magnetic field, so the magnet would blow apart spectacularly, if it wasn't properly designed.

    The magnet has a link in the superconductor which is heated to drive it 'normal': this is used when the magnetic field is being built up. This link usually has a great big conventional shunt resistor in parallel with it with great big heat sinks, and this arrangement is usually on the top of the magnet. If the helium level gets low or something else funny happens, the hope is that the coil superconductivity will go at this point rather than anywhere else. The magnetic energy, instead of getting dumped into the magnet's structure, gets dumped into this shunt resistor. It may glow yellow, and boil off lots of helium, but the magnetic field can collapse over a few seconds rather than instantly, and won't release an electromegnetic pulsed that might set off a chain reaction with the magnets next door.

    What has happened here is that the safety system has gone off in one of the magnets just as it ought to. I expect they will inspect the shunt assembly to check nothing has scorched when all the energy got dumped, and also to try and find out why it did. However, with luck they can get it all going again without interrupting the vacuum.

  • by Kjella ( 173770 ) on Saturday September 20, 2008 @07:28AM (#25083869) Homepage

    Pardon me, but have you or whoever modded that up any clue about that? It could be that way, or it could be that a missing magnet will cause the beam to veer off course, hit where it shouldn't and create a major fuck-up. These aren't exactly guide rails, they're the only thing keeping the particles in their place.

Heuristics are bug ridden by definition. If they didn't have bugs, then they'd be algorithms.