LHC: First Magnet Failure

By John Conway | September 19, 2008 3:09 pm

Well, no one said this would be easy…and it isn’t. The LHC has had its first major incident, the failure of one of the 1230 main superconducting dipole magnets. This was apparently due to a “quench” in which the magnet goes rapidly from the superconducting to the normal conducting state, which then means that the tremendous electric current in the magnet suddenly starts heating it up, causing huge internal mechanical stresses. As pointed out elsewhere, quenches are expected to happen quite often in the LHC. The magnets are designed to withstand these forces, in principle, and were tested extensively. What happened here is not clear yet, and I have not seen an official statement from CERN. Probably best to wait for that.

As for the effect on the LHC commissioning, to repair or replace the magnet requires warming up the relevant sector, then cooling back down after the repair. This takes several weeks (I am being deliberately vague here) and in the mean time, no tests with beam are possible.

This is a setback, for sure! People were getting quite excited about the possibility that the LHC could collide protons at high energy, 10 TeV in the center of mass, by late October or November, before shutting down for the winter in December. Could this still happen? My guess is that it is unlikely given this failure, but we’ll know for sure soon. Stay tuned…

  • http://www.cgoakley.demon.co.uk/qft/ Chris Oakley

    Thanks god! A reprieve for a few weeks before we destroy our planet with a man-made black hole.

  • Daniel Bezerra

    Nothing too serious, one hopes. I was looking forward to the first real collision tests.

  • Stephen Webb

    As an aspiring accelerator physicist, I can assure you that my peers wake up screaming in the middle of the night, drenched in cold sweat, from nightmares about magnet quenches.

  • Sili

    Is this gonna be like ENIAC where ‘up-time’ depends on how long it takes before the next valve burns out?

    How often is “often”?

    Ah well – teething troubles are to be expected. Presumably they happened with LEP too. We’ve just ‘forgotten’.

  • John R Ramsden

    John wrote “As for the effect on the LHC commissioning, to repair or replace the magnet requires warming up the relevant sector, then cooling back down after the repair.”

    Bearing in mind that there are no stupid questions (only inquisitive idiots), wouldn’t be quicker and easier not to heat up the sector and for the repair team to work in something like a space suit?

  • http://uslhc.us/blogs/?author=9 Seth Zenz

    Hi John R. I don’t think the problem is the cold in the tunnel, it’s that a magnet is impossible to service while it’s filled with liquid helium. And removing that helium safely means warming it up, and weeks…

  • Count Iblis

    Why would the LHC be shut down for the winter?

  • Garbage

    “Bearing in mind that there are no stupid questions (only inquisitive idiots), wouldn’t be quicker and easier not to heat up the sector and for the repair team to work in something like a space suit?”

    Now…that was funny 😀

    Look at the magnet’s temperature, it seems to be cooling back down already…


  • omh

    But the plot in the lower right hand corner is missing data points. They simply disappeared earlier today. Furthermore, I don’t think the temperature shown in the upper plot could have been lowered as quickly as it seems to indicate. I think that some sensors or software is malfunctioning, removing data points from the warmed magnets, and so I wouldn’t trust what you see there…

  • Christopher M

    To the winch, quench!

  • Adam A.

    At the tevatron (the accelerator near Chicago), ‘catastrophic quenches’ happen reasonably often. One happened while I was working on shift (a three month period). Regular non-catastrophic quenches happen all the time – a few times a month. In the case of the Tevatron, that means downtime to make extra anti-matter. However, the LHC only uses protons, not anti-protons, so after a regular quench, you could restart as soon as the diagnostics finish running.

    When a magnet quenches catastrophically, as Seth says, the liquid helium must be removed from the magnet and that sector has to be decoupled from the cryogenics system so that the magnet can be removed. In addition, the superconducting cables that connect to the newly installed magnet must be ‘baked’. That is, superconducting cables start life as a powder filled metal sheath. To become superconducting, it has to be heated to a specific temperature for the powder to fuse and form a special type of ceramic. That process takes some time.

    Usually quenches of this type happen for some external reason – the one that happened during my shift was due to a rodent of some type chewing on the interface between magnets – the Tevatron techs didn’t find much left of that rodent. Other things than can lead to bad quenches are power sub-station glitches, lightning strikes nearby (that can cause the beam to go off center slightly and hit a magnet), and even earthquakes on the other side of the world. This isn’t going to happen every month, but if it happened once a year, I wouldn’t be surprised.

  • Tim M.

    May I ask what it means to say “10 TeV in the center of mass”? Is that the energy of all the particles taken together?

  • BJM

    Re: Why would the LHC be shut down for the winter?

    Not an insider, this just from general reading: I see that winter heating needs of the surrounding communities push electricity demand and cost up.

  • http://chetsnicker.com chet snicker

    i shall hold myself back from joining the chorus of eschatological humor….

  • Sili

    Thank you, Adam A. That was really interesting.

    BJM, how fickle these big machines are. I guess the employees appreciate being able to go home for Crimbo, though.

  • http://www.nutcase.org H.M. Amir al-Mumenin al-Mutawakkil ‘Ala Allah Rab ul-Alamin Imam Yahya bin al-Mansur Bi’llah Muhammad Hamidaddin, Imam and Commander of the Faithful, and King of the Yemen.

    So Holger Nielsen was right.

  • http://blogs.discovermagazine.com/cosmicvariance/john John

    Tim M: when we talk about relativistic particle collisions, though it may at first sound counterintuitive, it turns out that the total energy available for creating new particles, etc. is most properly calculated in the reference frame where there is zero net momentum. We call this the “center of mass” or “center of momentum” frame. That’s the frame where the incoming protons have equal and opposite momentum.

    If you have a high energy particle hitting a stationary one, there is a lot less energy than the energy of the high energy one available in the c.m. frame…

  • Kletter

    It does make you wonder if the cost is worth it – although someone did point out the following:

    LHC : $10,000,000,000 over 15 years
    US Military : close to $400,000,000,000 in a single year!

    DOE’s yearly budget for nuclear weapon design etc. is about $10 billion as well – not like we need more advanced nukes, is it? DOE’s budget for solar energy research? Sub-$100 million, and that’s all to private industry. I’d rather see a billion-dollar a year solar energy research program than more overblown high-energy physics experiments, though.

    The cosmic ray detector systems and their results are also more interesting than CERN’s LHC: http://physicsworld.com/cws/article/news/18459

    Well, good luck anyway – looks like it will be needed.

  • http://www.gianfalco.it gianfalco

    I have a doubt

  • http://blogs.discovermagazine.com/cosmicvariance/joanne/ JoAnne

    The CERN press release about the incident can be found here. It appears that the problem was a faulty electrical connection that melted during the commissioning for 5 TeV beams. Downtime expectations for the repairs is currently 2 months.

  • Charles Tye

    A large helium leak really ought to be announced by a spokesman in a high-pitched squeaky voice.

  • Sad

    Talk about a bitter disappointment….. :(

  • http://www.myspace.com/danofx Dan

    I was disappointed by the news. I too am quite hopeful for the real expiramentation to begin; I’ve been following along since I read an article about the LHC in Scientific American, and have been quite excited about it. Oh well, keep spirits high. Drop by my myspace @ myspace.com/danofx, talk science or politics, or whatever.

  • Blake

    “However, the LHC only uses protons, not anti-protons, so after a regular quench, you….”

    Really? wow I guess I had always just assumed that it was a proton anti-proton collider. What’s the difference between a proton anti-proton collider and a proton proton collider? Is it just that you would have ~2GeV less center of mass energy at the collisions and at 14TeV who cares about that, or are there other implications? are p – p-bar collisions “cleaner”?

  • none of the above

    Blake on Sep 20th, 2008 at 8:16 pm writes:
    “What’s the difference between a proton anti-proton collider and a proton proton collider? ”

    Number of collisions. The more collisions that you have, the greater the fraction of the available centre of mass energy that you can effectifvely use [this is because a (anti)proton beam is effectively a broad band beam of quark and gluon constituents with a rapidly falling distribution as a function of fraction of proton momentum carried].

    Antiprotons are hard to make so anti-proton proton colliders have fewer collisions than comparable proton-proton colliders. For example the design luminosity [measure of number of collisions per second] at the LHC is two orders of magnitude larger than the TeVatron luminosity.

  • blake

    but then why did fermilab decide to make the tevatron a proton anti-proton collider at all? what’s the benefit?

  • none of the above

    blake on Sep 21st, 2008 at 9:53 am asks:
    “but then why did fermilab decide to make the tevatron a proton anti-proton collider at all? what’s the benefit?”

    Ease and cost. Because an antiproton-proton collider requires only one ring of magnets. If you already have a synchrotron capable of carrying protons in one direction, you can just inject antiprotons in the opposite direction. It is for this reason that the CERN pbar-p collider was just the old SPS used as a collider, and the Fermilab tevatron collider is just the tevatron accelerator used as a collider. You do this because you can do it quickly and cheaply, recycling the magnet ring and vacuum tube of an already existing accelerator.

    A p-p collider must be built from scratch and requires either two rings of magnets, or else a 2 in 1 magnet design [2 beamholes in 1 magnet yoke] as in the LHC. You do this because you’re interested in really doing it right, and getting the highest luminosity possible.

  • Blake

    I seeeeee!!!! I should’ve thought of that on my own! but…….didn’t! one final question then. How do you keep the beams from colliding all the time if they are counter rotating in the same magnetic field in the same beam tube and don’t their respective magnetic fields create messy weird interactions? (I guess that was 2 questions)

  • http://blogs.discovermagazine.com/cosmicvariance/john John

    Blake, the other advantage of a proton-antiproton collider is that then you have
    lots of high-energy antiquarks from the antiproton to annihilate the quarks in
    the proton. This is necessary, at Tevatron energies, to make significant numbers of top-anti-top quark events as it turns out; that’s what led to the top discovery in 1995.

    For the LHC, at “electroweak scales” of 100-200 GeV, there are plenty of antiquarks and quarks from the virtual “sea”. This is something I intend to post on in the near future.

  • Adam A.

    Hi Blake –

    It’s a good question about why the beams don’t collide all the time in the Tevatron. There are two mechanisms to prevents this:

    1) Electrostatic separators – periodically throughout the accelerator, there are charged plates to pull the two beams apart.

    2) Beam focus – in the non-colliding sections of the accelerator, the beams are much larger than at the interaction point. Because of the decreased focus, the beams can largely pass straight through each other without many interactions. On both ends of each interaction points, there are sets of quadrupole magnets to focus the beam significantly.

    I would encourage you to look up the design of the accelerator complex at Fermilab – It’s very cool. There are many accelerators in the complex which all work in concert to get counter rotating beams of protons and anti-protons into the Tevatron (the anti-proton side of the complex is very interesting – making antimatter is hard). Working in the CDF control room during shot setups (preparing the beam for entry into the Tevatron) was one of the coolest things I’ve done as a grad student, and one of the most informative.

  • Sili

    Thank you for the clarification on p-bars. I’d understood that the problem was luminousity, but the the details are interesting.

    If someone has a stroke of brilliance and figures out how to make lossa anti-protons, will it be possible to use the LHC as is or does the design not allow for that?

  • teadrinker

    I would accuse McCain-Palin of delaying the LHC so it would start up and destroy Earth right around election time so Obama could be blamed for this, but those two probably don’t know how their toasters work, let alone a giant particle accelerator.

    Get well soon, LHC. I want to see that Higgs Boson!

  • Adam A.

    @Sili –

    I would doubt that someone will come up with a better way to make anti-protons. The current way is to shoot protons as a spinning metal target and to sift through the spray on the other side. Some small fraction of that spray will be anti-protons, so it’s just a matter of keeping those and putting them into a storage ring. The energy of the incoming beam is tuned to maximize the yield of anti-protons. There are not convenient caches of anti-protons laying around, so you’ll always have to do something like this.

    The LHC magnets have a novel design – instead of having one beam pipe down the middle of the magnet, the magnets have two separate beam pipes in the, and the windings of the magnet are such that the magnetic field points in opposite directions for each beam pipe. With all of the effort put in to designing magnets that could contain counter-rotating proton beams, it would be a waste to use anti-protons. Anyway, you’ll always be able to make protons easier than anti-protons – protons only need hydrogen gas and some manner of ionizing system. Anti-protons always start off with protons, so you’ll never reach parity, even if you increase the yield. Finally, storing and cooling (reducing the momentum spread transverse to the beam) anti-protons is difficult.

  • Pingback: Metal-eating bacteria vs. the Tevatron, the next Hollywood blockbuster « Peculiar Velocity()

  • Pingback: Delay for the LHC | Cosmic Variance()

  • Pingback: Chicago Boyz » Blog Archive » Brain Rinse()

  • 12 year old wiz

    Hello, I’m just wondering about the ideologist who thought that the LHC could create a micro black hole, it can, but can it not suck up completely nothin due to the fact it is, once again, micro sized has little or no power what so ever to even have a gravitational pull? So it would just fade? Someone please tell me!

  • Pingback: LHC Restart | Cosmic Variance | Discover Magazine()

  • Phil


    Due to the differential nature of transversal matter waves, the gravitational pull can interact with ordinary flat space time only in a restricted range of frequencies.

  • Pingback: Endgame for the Tevatron | Cosmic Variance | Discover Magazine()

  • Pingback: Spooky Signals from the Future Telling Us to Cancel the LHC! | Cosmic Variance | Discover Magazine()


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Cosmic Variance

Random samplings from a universe of ideas.

See More

Collapse bottom bar