Two months after an electrical fault put CERN’s brand new Large Hadron Collider (LHC) out of action, the first damaged sections of the machine are making their way out of the tunnel for repair.

In the past week or so, seven of the LHC’s magnets (mostly 15 m long, 35 tonne “dipoles”) have been transported approximately 6 km through the 27 km LHC tunnel from the scene of the incident to a shaft on the main CERN site. From there, the magnets have been craned 50 m to the surface and taken to different locations for inspection. Some 50 magnets are expected to have to come to the surface in total, about 20 of which will not return, and the last one should be above ground before Christmas.

The incident in September was a major blow for us. But things are moving fast now and we can see a way ahead Roger Bailey, LHC operations leader

On 19 September, just nine days after protons were circulated in both directions of the €3bn LHC, an electrical connection between a dipole magnet (one of 1232 that bend the protons around the ring) and a neighbouring quadrupole magnet (one of 392 that focus the proton beam) failed during circuit tests in the last of the LHC’s eight sectors. At the time, a current of 8.7 kA was being pushed through superconducting cables the width of a stick of chewing gum to generate the enormous magnetic fields required to bend protons at high energies.

Magnets broke their anchors

Due to a bad connection, a splice linking cables between two magnets in “sector 34” suddenly developed a resistance and therefore disintegrated — producing an electrical arc which punctured the liquid helium plumbing that keeps the magnets (i.e. superconducting cables) at their 1.9 K operating temperature. Two tonnes of helium was released with such force that some magnets broke their anchors to the concrete floor, and a further four tonnes of helium was also discharged into the LHC tunnel.

Although some equipment was ready to be transported from the affected area within a couple of weeks of the incident, engineers have had to wait until two independent LHC sectors — sectors 23 and 12 — were purged of helium before it was safe to do so. That’s because only one of the LHC’s access shafts, located at the north end of the CERN site in the middle of sector 12, is wide enough to handle the dipoles. Indeed, with transport vehicles in the LHC tunnel travelling at 2 km/h, this is partly why it took two years for all 1232 dipoles to be installed underground.

In a presentation to the LHC experiments committee (LHCC) on Wednesday, LHC project director Lyn Evans stated that the repair was well underway with 100 people from CERN and contractors working on it. He expects that about 20 dipoles will be replaced with spares, and said that techniques have been developed to detect resistive splices at low currents to help prevent a similar incident. Although CERN has not yet finalized the costs of getting the LHC up and running again, it estimates the maximum cost of repairs and consolidation to be CHF15m (€10m) plus CHF10–20m to replace the spare magnets.

Testing times

We need to check for damage to the super insulation, which means de-cryostating, re-cryostating and then testing all over again Nick Chohan, CERN

Of the 30 or so magnets that will be repaired and reinstated in the tunnel next year, many will require a major refit. “It’s not just a case of removing the magnets and cleaning them [the electrical arc produced soot that contaminated the proton beam pipes in some magnets],” Nick Chohan, who spent five years testing each LHC dipole before it was installed underground, told “We need to check for damage to the super insulation, which means de-cryostating, re-cryostating and then testing all over again.”

Among many safety modifications, such as reinforcements to the magnet anchors, CERN is considering modifications to the spring-loaded relief discs on the magnets’ vacuum enclosures, which were unable to cope with the huge helium discharge on 19 September. Although the latter could be done in situ, it would require all 140 tonnes of the LHC's liquid helium to be taken out (raising the problem of where to store it) and all eight sectors to be warmed up to room temperature and cooled back down again — a process that would take months and possibly incur damage to the RF “fingers” between magnets.

CERN released an interim summary of the incident on 15 October, and plans to release a fuller report in early December outlining the repair schedule and plans for LHC operation in 2009. Despite frustration that the LHC did not provide even a few minutes of low-energy proton collisions, which machine operators were poised to deliver just days after the sector 34 incident, physicists working on the four LHC detectors are making the best of various data recorded during the start-up and of those from cosmic rays.

“The incident in September was a major blow for us,” LHC operations leader Roger Bailey told Physics World. “But things are moving fast now and we can see a way ahead.”