Fifteen Years of Operation of the Compact Muon Solenoid Detector Superconducting Magnet

The Compact Muon Solenoid (CMS) detector magnet has been in operation since 2008 at CERN's Large Hadron Collider (LHC). It will have to operate until the end of the High-Luminosity LHC run, beyond 2040. The CMS magnet comprises a large superconducting solenoid coil providing a magnetic field of 3.8 T with a free bore of 6 m in diameter and a length of 12.5 m. The coil is constructed with an aluminium stabilized Rutherford Nb-Ti/Cu cable and operates at 4 K with indirect conduction cooling in thermosiphon mode with boiling helium. The magnet reached 4 T and a record stored energy of 2.6 GJ when it was commissioned in 2006 in the surface hall at CERN Point 5. It was then transferred in 2007 to the underground experimental area, where it was recommissioned and successfully operated at a nominal field of 3.8 T since then. A summary of the magnet operating data is presented in this paper along with the observed progressive change of the Residual Resistivity Ratio (RRR) of the pure aluminium conductor stabilizer as a function of operating cycles and magnet warm-ups. The technical problems encountered, and the solutions implemented with the cryogenics and the vacuum pumping of the cryostat are described, as well as the upgrades carried out during the LHC shutdown periods on the control system, the cryogenics and the powering circuit where a freewheel thyristor system has been implemented.


I. INTRODUCTION
T HE Compact Muon Solenoid (CMS) Collaboration was founded in October 1992 [1].The CMS Collaboration brings together members of the worldwide particle physics community in a quest to advance knowledge of the laws of our Universe.The Collaboration has been in charge of designing, constructing, commissioning, and now operating since 2008, the CMS experiment at the CERN Large Hadron Collider (LHC) accelerator.
The CMS experiment is a highly integrated particle detection system, built with a large superconducting magnet [2].The CMS magnet is a 12.5-m long, and 220-ton superconducting solenoid operated at 4 K to provide a nominal magnetic field of 3.8 T.  The CMS magnet cryostat and its 12'000-ton iron yoke are the structural support of the CMS detector.
On 4th July 2012, the CMS collaboration, together with the ATLAS Collaboration [3] announced the discovery of a new particle corresponding to the Higgs boson [4].The Nobel Prize in Physics 2013 was jointly awarded to François Englert and Peter W. Higgs for this discovery [5].Since then, data taking continued and the total integrated recorded luminosity at CMS reached 216 fb −1 end 2022 [6].
This article presents the analysis of the main operating data of the CMS magnet from 2008 to August 2023.This period includes the preliminary phases of the LHC operation, the LHC physics run 1 (2010-2013) and run 2 (2015-2018), followed respectively by the Long technical Shutdown 1 (LS1) and Shutdown 2 (LS2).In 2022 the LHC physics run 3 has started.It will last until the end of 2025 when the Long Shutdown 3 (LS3) will begin to perform the final upgrades of both the LHC accelerator and the experiments for the High-Luminosity (HL) LHC run starting in 2029 [7].

II. THE SUPERCONDUCTING MAGNET
The CMS magnet conceptual design was proposed by CEA/Saclay-IRFU, France [8].Its final technical design and construction was achieved by a collaboration of seven Institutes [2].The CMS magnet design faced several technical challenges that were successfully solved [9].The main characteristics of the magnet are given in Table I.The coil is made of 5 modules, each with a 4-layer winding of a Nb-Ti/Cu Rutherford cable co-extruded in a sheath of high purity aluminum which was then electron-beam welded to aluminum alloy profiles acting as mechanical reinforcement to achieve a compound reinforced conductor [10].
The magnet was successfully commissioned at 4 T in 2006 at CERN on the surface facilities of the CMS site [11].It was transferred in 2007 into the underground experimental cavern UXC55 100 m below ground level (Fig. 1), together with its ancillary systems in the underground service cavern USC55 (refrigerator, power converter, switch breakers, cryostat pumps, control and safety systems).After a successful test in 2008 at 3.8 T, the magnet was ready for operation.

III. OPERATION OF THE CMS MAGNET
The operation is under the responsibility of the CMS Technical Coordination Team.Regular interventions, corrective and preventive maintenances, and upgrades are performed by several CERN groups that are providing 24/7 on-call services as well.These groups are from the Research and Computing Sector (CMS Experiment and Detector Technology Groups) and the Accelerators and Technology Sector (Cryogenics, Electrical Power Converter, Cooling and Ventilation, and Electrical Engineering Groups).

A. CMS Magnet Powering
1) Operation at 3.8 T: A 2-quadrant silicon-controlled rectifier based on four thyristor rectifier bridges allows a controlled ramp up and ramp down of the magnet electrical current.In case of faulty conditions, a discharge is performed on external air-cooled resistors.A slow discharge (SD) is performed in case of powering fault, and a fast discharge (FD) is triggered in case of magnet quench (Fig. 2).During operation, a partial ramp down to 9.5 kA (2 T) is implemented since 2012 in case of a cryogenics stop, to avoid the risk of a magnet fast discharge caused by helium flow fluctuation on the helium return circuit.
2) Free Wheel Thyristors System: The powering circuit was upgraded during the LS2, with the implementation of a Free Wheel Thyristors (FWT) system.This system contributes to increase the magnet's lifetime and the operational time at nominal field, preventing a full magnet power off in case of disruptions on the powering or cooling networks, therefore reducing the number of mechanical stress cycles on the coil, and avoiding 8 hours with the magnet unavailable corresponding to a full discharge followed by the ramp up of the magnet.It is based on a thyristor bridge developed for the converters of the Super Proton Synchrotron facility at CERN in operation since 2012.It is composed of 6 thyristors in parallel [12].The FWT by-passes the power converter in a closed loop when the converter goes off and allows to reconnect the converter to the magnet at any current value when the nominal operating conditions are back.It is always possible to open the breakers to safely discharge the magnet.This system has a large time constant, over 15 hours, with a low current decrease (|di/dt| < 0.33 A/s).Its cooling circuit is powered through an Uninterruptible Power Source (UPS) against power glitches.The FWT was successfully commissioned ahead of the LHC run 3 in 2021.
3) Magnet Warm up: To warm up the magnet to room temperature, a dedicated power converter is used.With the magnet above 10 K, the Joule heating ensures a uniform heat dissipation in the windings, limiting the temperature gradients inside the coil's structure, and preventing the risk of delamination of the fiberglass-reinforced epoxy insulation.It takes about 3 weeks to warm up the coil with a maximum temperature increase of 1 K/hour, and a maximum temperature gradient below 15 K.

B. CMS Magnet Control System
1) Process Control: Supervision and control of the magnet is performed through the Magnet Control System (MCS).The MCS interfaces with the control systems of the power converter, the cryogenics, the cryostat pumping.It is based on the CERN Unified Industrial Control System (UNICOS) for the industrial control applications [13].The Supervisory Controls And Data Acquisition (SCADA) software and the Programmable Logic Controllers (PLC) are commercial off-the-shelf products, that were upgraded during LS2, on the magnet and the cryoplant control systems.A dedicated PLC was also installed for the cryostat vacuum pumping system during LS2.
2) Instrumentation: The measurements are stable and identical to early publication [14].Of the 140 temperature sensors and 120 voltage taps installed, none are defective.The 4 full range vacuum gauges are shielded against the stray magnetic field, but they have been recalibrated or replaced every year.
The magnet ground insulation is regularly checked, in particular after each cool down to 4 K following a warm-up to room temperature.It is still good with a ground insulation resistance value above 10 MΩ, including the dump circuit.
Magnetic measurements have been performed using sensors installed on the magnet yoke [15].Two NMR probes are positioned inside the magnet free bore, on the outer radius of the hadron calorimeter, in a central transversal plane, and provide magnetic field reference values.NMR measurements are not continuously monitored as the signals cannot be locked because of a slow drift due to the local field gradient.The NMR measurements are done once or twice a year.The standard deviation of the NMR measurements is 1.710 −4 T.
The magnet current is measured by two redundant zero-flux current transducers (or direct-current current transformers -DCCTs), used for the current regulation loop.Their control electronics are calibrated each year.The DCCTs are class 3 (50 ppm) precision, with a resolution of 1 ppm.
The magnet instrumentation was initially split between the MCS and the Constructor Diagnostic System (CDS) that was implemented as a stand-alone system by the CEA-Saclay team for remote diagnostic.During LS2 the CDS was merged with the MCS, for direct access to the measurements and to use the CDS sensors as spare ones for the MCS in the future.

C. CMS Magnet Safety System
The Magnet Safety System (MSS) primary function is the magnet protection.It has two redundant electronics.It automatically triggers the magnet discharge from digital interlocks received from external control systems.The MSS is also interlocked to the LHC beam dump system.In case of FD, the LHC beam is dumped, because the magnet discharge has a slow but non-negligible effect on the LHC beams causing orbit deviations [16].The MSS detects a magnet quench on analog signals coming from the quench detection (QD) system.Fig. 3 gives the QD signals recorded during a magnet ramp up to 3.8 T. The spikes are observed at each ramp up and at high current.They may correspond to tinny coil movements.At 3 T, the ramp rate is reduced from 1.5 A/s to 1 A/s to reduce the spike amplitude well below the thresholds of QD system set between 1 V and 2 V for 1 s, depending on the QD channel.These spikes did not trigger the QD system.
The MSS system was successfully implemented in the early 2000's.After more than 10 years, critical components became obsolescent.Spares were still available but too limited in quantity for the many years of operation in the future.A new MSS, the MSS2, was designed, tested, and successfully deployed at CERN [17].It is based on a real-time embedded industrial controller.This system was configured and implemented on CMS during the technical stop at the beginning of 2018.

D. Cryogenics 1) Description and Performance:
The CMS magnet is cooled by a cryoplant, with oil-lubricated screw compressors in two stages, providing a helium mass flow of 200 g/s at 16 bar, and one refrigerator, for all loads down to 4.4 K.The compressors are installed in a dedicated surface building at P5, and the refrigerator is installed 100 m underground in the service cavern.The cooldown to 100 K is performed with a 30-kW precooler using liquid nitrogen (LN 2 ).A 6-m 3 Dewar provides liquid helium to the magnet in case of cryoplant stop to be able to safely discharge the magnet without risking a quench.The magnet is indirectly cooled in thermosiphon mode with liquid helium at 1.25 bar, with a phase separator connected to the coil heat exchanger circuit [18].
The cumulative operating time of the refrigerator turbines is 93000 hours, including all cooldown phases and stand-by at 15 K.The overhaul of the compressors was done during LS1.One turbine of the coldbox first stage broke and had to be replaced in 2009.The membrane of the third turbine broke in 2014.The causes of the breaks were either pollution or pressure fluctuations on the gas bearings.The third turbine had to be replaced in 2015 after 10 years since first operation, because of the degradation of its performance due to an erosion of the turbine blades.
2) Issues With Cryogenics: Several issues were encountered with the refrigerator that appeared following the LS1.
a) Refrigerator pollution: A clogging of the refrigerator was identified, due to a pollution of the coldbox heat exchangers and filters with Breox, the lubricant used for the compressors.Minor pollution effects were already observed in 2010 and the following years, but they could be solved during planned technical stops by changing the turbine filters or regenerating the entire cold box by warming it up to room temperature and Authorized licensed use limited to the terms of the applicable license agreement with IEEE.Restrictions apply.flushing it with warm helium.At that time of the first events, the clogging was explained by residual oil remaining in the helium circuits after a contamination of the cryoplant with 2 litres of Breox due to a problem with the final oil removal system, during the first commissioning in a surface building in 2004.A cleaning was done in 2004 with isopropyl alcohol.A fourth coalescer was added at that time to the oil removal system.
At the beginning of 2015 the clogging increased suddenly.To operate the magnet at 3.8 T for several short periods in 2015, many interventions were necessary, regenerating the coldbox, changing the filters and coalescers, and even using the pre-cooler to gain cooling power.In parallel the design and procurement of a cleaning system was launched.The coldbox internal circuitry had to be modified to connect a condenser and a purifier, to circulate the solvent in the contaminated parts, and to finally dry all the circuits [19].The solvent used for the cleaning had to meet the requirements for personnel safety in the CMS underground service cavern.The cleaning was followed by a drying period that lasted about one month.
Several consolidations were made.Two big filters were added in the coldbox: one filter in parallel to the 80 K adsorber outlet filter with a 30-μm mesh and a surface 40 times larger, and one filter in series and upstream the first turbine inlet filter with a 10-μm mesh (changed in 2021 to 5-μm mesh) and a surface 100 times larger.The 80-K and 20-K refrigerator adsorbers were changed.A new oil separator and new oil coalescers were designed and installed (Fig. 4).
Two new high pressure helium gas lines DN100/80 and DN25 were installed from the compressors at the surface to the refrigerator in the service cavern, as it was easier-faster-cheaper than cleaning the polluted lines over the length (200 m) and height (100 m).A helium drier was also installed to further reduce the residual humidity in the helium circuitry.
These repairs and cleaning were major achievements as they allowed CMS to be operational again for physics data taking in a reasonable amount of time.
b) Nitrogen pre-cooler leak: The nitrogen pre-cooler heat exchanger had a leak inside the coldbox.The repair was done during the same period as the cleaning of the coldbox in 2016.
The heat exchanger was damaged by pressure oscillations during its use.A possible cause was a bad insulation vacuum in the LN 2 transfer line.
c) Damaged valve: A pressure regulation valve on the intermediate Dewar was damaged, degrading the overall refrigeration performance.The valve was bent, and a crane hook impact in the experimental cavern was suspected.Also in this case, the repair was performed in parallel to the other interventions on the refrigerator in 2016.The Dewar could be opened inside the experimental cavern.The hooking zones have since been re-defined and protections have been put on all the valves of the Dewar to prevent collisions.
3) Consolidations and Upgrades: Additional consolidations and upgrades of the cryoplant have been performed as well [20].A full redundancy of the two compressor stages was installed during LS1, allowing to reduce the downtime and to recover helium gas in case of a compressor equipment failure.The powering supplies of the compressors and their water-cooling pumps were changed for more consistency, to have them all on the same AC power distribution network.New connection flanges were also implemented on the compressor cooling circuit to be able to install a backup chiller, in case of a major unavailability of the water-cooling network.The LN 2 transfer line insulation vacuum ports had to be modified due to interferences with surrounding metallic structures, preventing to pump properly the insulation volume which deteriorated the transfer line performance.All the I/O cables of the CMS Coldbox control system had to be replaced in 2017 due to the degradation of the insulation caused by UV from lights in the service cavern, and protection were put in place to avoid this issue again.

E. Cryostat Pumping
1) System Description: Two stainless steel diffusion pumps are connected to the magnet cryostat and the phase separator cryostat.Both cryostat volumes are linked by a chimney through which pass all the helium pipes connecting the phase separator to the coil circuits.Two redundant fore vacuum pumping groups, with roots and rotary pumps, are located in the service cavern.All pumps are commercial off-the-shelf models.A 200-mm diameter vacuum line connects the diffusion pumps to the fore pumping groups.The fore pumps are operated alternately to extend their lifetime, with a swap each week.Maintenance is performed according to the recommendations of the supplier.
In operation at 4 K, the high vacuum valves are closed, and the static vacuum volume pressure is stable and stays below 10 −6 mbar.
2) Issues With the Diffusion Pumps: A water leak inside the 400-mm diameter diffusion pump of the magnet cryostat was discovered in October 2019, after a tightness check of the high vacuum valve.The magnet was at 25 K, and the static vacuum was stable at 4.10 −7 mbar.Luckily, the two high vacuum valves were closed and tight, therefore no water cryo-pumping effect occurred.About 20 liters of water were drained out of the diffusion pump.The primary groups were polluted as well with water.
Authorized licensed use limited to the terms of the applicable license agreement with IEEE.Restrictions apply.Due to a difficult access to the diffusion pump in a narrow location, and the associated risks of breaking the vacuum during the pump removal, it was decided to warm up the coil to room temperature.The warm-up took place during the Covid-19 lockdown in spring 2020.The pumping capacity was reduced, with only the 250-mm diameter diffusion pump on the phase separator in operation, connected to the magnet cryostat through the low conductance chimney.The warm-up took about 40 days, limiting the pressure rise in the cryostat due to outgassing humidity by reducing the magnet current to slow down the warm-up.
The origin of the leak was the brazed copper-to-steel tube connections on the cold cap cooling circuit of the diffusion pump.X-ray Computed Tomography (CT) scans were performed at CERN [21] and revealed a significant lack of filling of the brazed connections, with less than 10 % at some spots (Fig. 5).Other flaws were also discovered in the stainless-steel welds (cracks, lack of filling, root porosity), but did not result in a performance degradation.The small diffusion pump had a water leak on the external cooling circuit which generated corrosion in proximity of a weld.
The non-destructive inspections carried out at CERN revealed welding and brazing qualities below standards to be expected for a diffusion pump on a cryostat with a cold mass inside, although it has been in operation for 13 years.The detailed welding and brazing quality levels should have been specified during the tender process to avoid such a mishap.
A full consolidation of the vacuum pumping circuit was done in 2021, with the replacement of the two diffusion pumps and the overhaul of the two primary pumping groups.The pumping circuit was cleaned, and the layout was modified to increase the conductance on the primary pumping.New flanges and valves were added to allow the connection of back-up pumps if necessary.

A. Magnet Availability
The magnet has been in operation at 3.8 T and 4 K for 45202 hours from 2008 to 2022.This is a cumulative total equivalent to 5.15 years.Such time neither includes the annual commissioning carried out at the magnet restart, nor the time for ramping up or down to the requested magnetic field.The magnet availability, defined as the ratio of the actual over the planned running time at 3.8 T, is given in Fig. 6.It has been excellent over the LHC run 1 and run 2, respectively 97.1% and 99.1%  in average.In 2015, the magnet availability was reduced due to the cryoplant pollution and operation plans had to be revised.In 2021, the repair of the pumping system took place during the LS2 and did not limit the short operation period, with 2 weeks at 3.8 T (100% availability).During run 3 for the first year (2022) the availability was 98%.

B. Causes of Magnet Systems Disruptions
The total number of faults per sub-system and per year from 2008 to August 2023 are given in Fig. 7.Not all of them have limited the magnet operation time, as, in many cases, the issue appeared either with magnet off or could be solved without disrupting the magnet operation.
Due to the clogging issue, the cryogenics largely dominates with 76.1% of the faults encountered and 105 events registered.It is followed by the powering systems (10.1%, 11 power network glitches, plus one fault on the power converter, on a UPS, and on a 3.3-kV breaker safety), then the human interventions (7.3%, 4 on powering systems, 5 on cryogenics and one on MCS), the water-cooling systems (3.6%, 5 faults), and finally the magnet systems (2.9%, 4 faults).

A. Mechanical Cycles
Ramping up and down the magnet current results in a cycle of the magnetic forces applied to the coil.As these forces are proportional to the square of the magnetic field, the number of equivalent mechanical cycles is considered and defined by: Where B, B init and B nom are respectively the field value at the end of the ramp, the initial field value before ramping, and the nominal field value (3.8 T), all in Tesla; j is the current ramp (up or down) number.
Since 2006, including the first commissioning of the magnet before its transfer to the experimental cavern, up to August 2023, there has been 103 equivalent magnetic cycles.A total of 56 cycles were done on request, and 47 were due to faulty conditions.Faults on cryogenics account for 54.9% of the equivalent mechanical cycles, and 17.2% are with powering systems, 14.1% are with human interventions, 9.3% are with cooling systems and 4.5% are with magnet systems.Details over the years are given in Figs. 8 and 9.

B. Fast Discharges
A total of 12 fast discharges (FD) were triggered over the entire operation period since 2006 up to August 2023, but only 6 occurred unintentionally during operation, with 5 FDs at 3.8 T. The causes of these 6 FDs are listed here below: 1) A ground fault on the main switch breakers command circuit, during a power glitch on the CERN supply line (2009).2) A faulty position controller, or its power supply (both were later changed), of a control valve used on one of the two magnet current leads (2011).3) A valve left in manual mode on the cryoplant control system, during a cryoplant stop due to a power glitch (2011).4) An unsuccessful attempt to reconnect the coldbox with the magnet at 3.8 T (2012).5) A valve on the low-pressure helium circuit with a wrong setting causing a reduced flow on the current leads that triggered a quench during unintentional cryoplant stop (2023).6) A stop of the cryoplant due to a faulty crimped wire on the fast stop protection of the 3.3-kV power distribution to the cryoplant, later followed by a faulty human action on the cryoplant control system that stopped the liquid Helium transfer from the Dewar while trying to restart the compressor to recover helium gas (2023, FD triggered at 3.2 T).

C. Temperature Cycles
Four temperature cycles from room temperature to 4 K were performed.Fig. 10 gives the temperature of the coldmass over the years.Short temperature cycles between 4 K and about 100 K take place each year-end during the cryoplant stop for Authorized licensed use limited to the terms of the applicable license agreement with IEEE.Restrictions apply.maintenance.Only in 2016, the temperature reached 162 K when the repairs on the cryogenics took place causing a longer stop of the cryoplant.
The thermal shields display similar temperature variations, but in the range of 40 K to 300 K.The current leads follow the same trend, but the cold feet reach quickly about 220 K in about 3 days when the cryoplant is stopped.

D. Superconductor Stabilizer RRR
The CMS magnet is built with a compound reinforced superconductor including a high purity aluminium (HPA) stabilizer.The initial Residual Resistivity Ratio (RRR) of the HPA in the conductor was measured above 3000 [10].The RRR is defined at zero magnetic field by: Where ρ is the aluminium electrical resistivity.The coil electrical resistance was measured with the use of an external DC power supply, providing 45 A in the coil.The measurements were performed with a magnet temperature stabilized at about 15 K. From the voltage measurement, the HPA resistivity was calculated, taking into account the contribution to the total resistance of the other conductor components.As a first approximation it was assumed that the HPA resistivity is uniform over the stabilizer cross section.This is a fair assumption as the stress gradient is small and mostly localized around the Rutherford cable [22].It takes several days to have a stabilized temperature with a gradient below 1 K, and this is not always compatible with the operating plans of the CMS experiment.Some results were excluded either because the temperature of the coil was too high, above 30 K, or because the temperature gradient in the coil was too high (during the cooldown of the coil), or because the current leads were not in the correct temperature range.
The RRR measurements are given in Figs. 10 and 11.At the end of 2022, the RRR was measured equal to 2374.The typical accuracy of the RRR measurement is a few percent.
A degradation of the HPA RRR is expected with the mechanical cycles of the CMS conductor, but there should be a recovery during the warm-ups at room temperature [23] which has been observed only once (Fig. 10).Recovery takes place at 300 K as expected, but much less pronounced after the LS2 warm-up (2020-2022), possibly because fewer mechanical cycles were performed during run 2 compared to run 1.The year-end short warm-ups during run 2 may have also influenced the results, having the coil for about 40 days with a floating temperature increasing up to 100 K, and up to 160 K in 2016, causing possible limited recoveries not observed with these measurements.
Calculations [2] indicated that a RRR above 800 provides acceptable stability margin for a safe operation of the CMS magnet.Experimental results had shown that the RRR for HPA at zero field and after 1000 strain cycles at 0.15% remains above 950 [24].It was also demonstrated that the mechanical cycles performed with the CMS magnet do not damage the bonding between the superconductor and the stabilizer [25].However, the RRR may be lower locally in the aluminium volume surrounding the Rutherford superconducting cable, due to the slightly higher thermo-mechanical stresses in this region, where the current diffusion would start in case of local loss of superconductivity.Therefore, the magnet stability could still be slightly reduced, but with the comfortable margin available, this should not affect the operation.

VI. CONCLUSION
After 15 years of operation at 3.8 T and 4 warm-ups at room temperature, all the CMS magnet parameters have remained at their nominal values.All the issues with the cryogenics and the diffusion pumps have been understood and resolved.Over the past years, the CMS magnet availability averaged 91.1% (97.6% not including the year 2015) thanks to the efforts of the teams involved in the operation and maintenance, reducing the mechanical and thermal cycles to a minimum.The CMS magnet performance is well within expectations thanks to the continuous and meticulous efforts of the CERN specialized teams.In the event of an issue, dedicated analyzes are systematically carried out to trigger the corrective and preventive actions, and more generally to draw lessons from the problems.Quality and efficiency are maintained through carefully prepared consolidations and upgrades campaigns, and planned replacement of the many components before their expected end of life.
The CMS compound reinforced aluminium-stabilized Nb-Ti/Cu superconductor has proven highly reliable over the 15 years of operation at CERN-LHC, and it gives confidence for successful operation during the HL-LHC phase in the next decades.
Finally, the lifetime of this experimental facility will encompass several generations.The knowledge transfer of its operation and maintenance, as well as the underlying conceptual aspects, is another challenge for the next decades to maintain the CMS magnet performance at the highest level and avoid human errors that could severely limit data taking.

Fifteen
Years of Operation of the Compact Muon Solenoid Detector Superconducting Magnet Benoit Curé , Gilles Le Godec, Maciej Ostrega , and Udo Wagner (Invited Paper) Abstract-The Compact Muon Solenoid (CMS) detector magnet has been in operation since 2008 at CERN's Large Hadron Collider (LHC).It will have to operate until the end of the High-Luminosity LHC run, beyond 2040.The CMS magnet comprises a large superconducting solenoid coil providing a magnetic field of 3.8 T with a free bore of 6 m in diameter and a length of 12.5 m.The coil is constructed with an aluminium stabilized Rutherford Nb-Ti/Cu cable and operates at 4 K with indirect conduction cooling in thermosiphon mode with boiling helium.The magnet reached 4 T and a record stored energy of 2.6 GJ when it was commissioned in 2006 in the surface hall at CERN Point 5.It was then transferred in 2007 to the underground experimental area, where it was recommissioned and successfully operated at a nominal field of 3.8 T since then.A summary of the magnet operating data is presented in this paper along with the observed progressive change of the Residual Resistivity Ratio (RRR) of the pure aluminium conductor stabilizer as a function of operating cycles and magnet warm-ups.The technical problems encountered, and the solutions implemented with the cryogenics and the vacuum pumping of the cryostat are described, as well as the upgrades carried out during the LHC shutdown periods on the control system, the cryogenics and the powering circuit where a freewheel thyristor system has been implemented.Index Terms-Aluminium-stabilized superconductors, detector magnet, superconducting magnets.

Fig. 3 .
Fig. 3. Quench detector voltage signals during magnet ramp to 3.8 T. The maximum spike amplitude during a ramp is 120 mV for less than a second.

Fig. 10 .
Fig. 10.High purity aluminium stabilizer RRR variation with coil temperature profile over time.

Fig. 11 .
Fig. 11.High purity aluminium stabilizer RRR variation with cumulated equivalent mechanical cycles over time.
Manuscript received 20 September 2023; revised 14 November 2023; accepted 21 November 2023.Date of publication 29 November 2023; date of current version 18 December 2023.This work was supported by CERN for the CMS collaboration.

TABLE I CMS
MAGNET MAIN CHARACTERISTICS