Serious Magnet Failure at CERN's New Accelerator

GrepNut writes "CERN is reporting that the giant magnets that steer the particle beam in the new and highly anticipated Large Hadron Collider have just failed catastrophically in a stress test, apparently due to a design oversight. It doesn't help that the magnets were designed and built by CERN's US competitor Fermilab." While safety precautions were followed, and no one was injured nor were any rifts in the space-time continuum opened, it's still a rather large setback for the project.

Fidgeting magnets...

[wakaranai]

Hmm.... sounds nasty.

Each of the ~1200 superconducting magnets is about 50 foot long. There's a photo here showing one being put in place (March 2005):
http://www.msnbc.msn.com/id/7119458/ [msn.com]

Comment removed

[account_deleted]

Comment removed based on user account deletion

Oh, crap.

[swschrad]

Fermilab has built electromagnets for many particle accelerators, including SLAC. They are apparently the only source. If you want something else, you have to go to TDK in Japan for fixed-intensity ceramic magnets.

According to an old neighborhood buddy of mine who is at SLAC, when he was in redesign of the linear accelerator in the 80s, those were the only two bids. For flexibility, they went with Fermi and electromagnets.

And they haven't failed yet.

While we're whining about cars, you can't keep headlamps and taillamps in a VW, wiring issues burn 'em out. nobody's perfect. that's why you negotiate warranties in the contracts for stuff.

no wonder you don't dare sign your name. which, BTW, is quite imperfect in itself. Can't stand on the courage of your convulsions, as a rabid right-wing wacko radio commenter used to say.

Anti US Slant

[laing]

The article seems to place the full blame on Fermilab's poor design. I will withhold judgement until all the facts are known. Did CERN provide specific requirements for asymetric load bearing capacity? If there were no requirements provided to Fermilab, then it would seem to me to be a problem at the CERN end.

Give me a break

[stox]

The forces induced in these magnets during a quench is obscene. Given the size of the LHC, I would guess that these are the largest such magnets ever fabricated. When pushing the envelope so hard, failures are going to happen. It amazes me that the public's quality expectations are so high for such work. If Windows was built to the same standards, it would have uptimes measured in centuries.

Re:There were no injuries

[WoLpH]

It would just make us start floating I assume, atleast they were able to let a number of things float because of an intense magnetic field over here: http://www.hfml.ru.nl/froglev.html [hfml.ru.nl]

Re:Got what they deserved

[Anonymous Coward]

Globalisation means that anyone with a big enough budget can do pretty much everything mankind is capable of.

It also means never having to say you're sorry.

Mainly because it opens up a whole new class of other people to blame when it goes bad

But to be serious; in this case it looks like a case of overlooking a possible engineering problem. It's quite understandable, as things of this nature present some unique and new problems and sometimes present gotchas that only sometimes gets caught in time. The Huygens Titan Probe almost wound up being a very expensive paperweight until a single engineer caught on that the doppler shift of the radio signals might need to be compensated for. Hubble's main mirror wound up being ground incorrectly due to some chipped coating on an end cap of the null corrector and lack of testing. Other space probes, and some very specialized machinery in other fields, weren't lucky enough to have an engineer catch the gotcha in time.

The problems become obvious with hindsight but are understandably hard to forsee, simply because these sorts of things have either never been done before or are at a whole new order of scale and complexity. Some turn out to be simple bone-headed mistakes but most are, "We never even thought of that!" That sort of thing is truly global.

Re:Give me a break

[ScrewMaster]

No, it's why software engineers aren't just programmers. The problem is that large mechanical engineering is done to standards that are well-understood and don't change that much. Software is much more of a moving target. Eventually, though, computing will become more mature, more stable ... and the job of engineer will take on more of its traditional meaning when applied to software development.

Of course, at about that time we'll have invented a true AI and people won't be programming anymore. Hopefully I'll be retired by then and can take up programming as a hobby.

Re:There were no injuries

[John Courtland]

Or you carry a loaded firearm into the chamber:

http://www.ajronline.org/cgi/content/full/178/5/10 92 [ajronline.org]