The magnet worked properly until it suffered two unexpected stops in August due to technical problems on the service infrastructure.

During the June Technical Stop, the CMS magnet was ramped down to zero tesla in order to allow a preventive regeneration of the cold box on the cryogenics to remove any possible trace of air pollution. The mechanical filter of the first turbine was replaced, and a full check of the cryoplant control and safety chains was performed. No faulty component was found that could explain the former cryoplant stops earlier this year.

The two spare compressors and their motors were delivered and are now stored at Point 5.

The magnet operation was smooth throughout July, until the fast dump of the magnet on August 10 caused by a sequence of events. First the cooling water temperature of the cryoplant compressor units increased. The water temperature was measured up to 37°C, and the compressor lubricant went up to 50°C, triggering an interlock that stopped the compressors and the cold box. The magnet was then cooled by the spare volume of liquid helium in the intermediate Dewar, as it has to be in such a case. The origin of the increasing water temperature was a faulty temperature sensor module on the Programmable Logic Controller (PLC) of the primary cooling water control, indicating a temperature of 4°C. This resulted in the automatic opening of a by-pass on the primary water cooling, which disconnected the primary cold water coming from the cooling towers.

During the reconnection of the cold box to the magnet, an interlock on the pressure value in the low-pressure helium return line prevented the operator from depressurizing the cold box from 5 bar to 1 bar in the downstream return line. The cryogenics team tried to release this pressure through another valve, but the pressure transient went up to 1.26 bar in the return line, affecting the helium flow circulating in the magnet current leads, also connected to the same helium return line. As the pressure rise was very fast, the cryo team couldn’t react against it. As a result, there was no flow of helium in the current leads for 12 seconds. The magnet safety system worked well and triggered a fast dump to protect the magnet and the current leads against the quench. The current leads didn’t suffer any damage as they had been designed to hold 20kA without helium for several minutes followed by a fast dump.

The interlock preventing the depressurization was triggered because the pressure value in the return line went 39 mbar above the defined threshold. This threshold has been slightly increased after this event. This might be a consequence of the increase of the pressure on the low pressure lines last year to limit the risk of air ingress in the helium circuitry, which proved to work well so far, as there has been no sign of blockage of the cold box with impurities since. The recent modification of the cryoplant process will be tested during the September Technical Stop with the magnet off.

Due to the fast dump, the magnet temperature rose to 70K, and four days were required for the temperature recovery. On August 14, the magnetic field was back to 3.8T.

Then a second magnet stop took place on August 21. This was caused by a fault on an uninterruptible power source (UPS), which went off either with a manual local command or an undetermined temporary material failure. Unfortunately, this UPS powers the cryoplant PLCs and also the command circuit of the dump resistor breakers. As a consequence, the compressors and the cold box stopped again, and the dump resistor breakers opened in the fast dump configuration. In such a case, the magnet safety system detects a fault on the configuration of the electrical circuit and triggers immediately a ramp down of the current.

Unfortunately, the cold box reconnection was delayed because of both the investigation on the UPS and another technical problem on the command circuit of a switch breaker on the power line of one compressor of the cryoplant. During this time, 87% of the volume in the spare liquid helium Dewar was used to keep the magnet cold at 4K. The Dewar had to be refilled before the magnet ramp up could resume. It took about 30 hours to get the level with the necessary safety margin, still below the operating level that was reached later on the evening of August 22 when the magnet was already at 3.8T.

All the safety chains of the magnet systems worked properly. After these two unexpected stops, the magnet is now in its fifth magnetic cycle this year.