Performance of the CMS Muon Detectors in early 2016 collision runs

UPDATED 14.06.2016 - add link to CMS-DP-2016-023.

This twiki page contains the plots in CMS-DP-2016-023, which summarizes the performance of the CMS Muon subdetectors (CSC, DT, and RPC) in the early LHC running of 2016. All data are from pp collisions at √s = 13 TeV and with full magnetic field, B = 3.8 T. Clicking on a small image will open a full-size PNG. If a PDF version of a plot is available, click the pdf version link at the top of the image to display or download it (web-browser dependent.)

Abstract

The performance of the three CMS muon subdetectors: DT, RPC and CSC, was evaluated with the first collision data in 2016 and compared to 2015.

Contacts

CMS DPG conveners of the Muon subdetectors:
  • RPC <cms-dpg-conveners-rpc@cern.ch>
  • DT <cms-dpg-conveners-dt@cern.ch>
  • CSC <cms-dpg-conveners-csc@cern.ch>

References

This is the link to CMS-DP-2016-023

The following published CMS papers are useful for understanding the terminology of the Muon system, the techniques used in obtaining these performance results, and the concepts of muon reconstruction:

Efficiency

Figure Caption
EffiPhiLay2016.png

DT single hit efficiency 2016 - Phi Layers

The Drift Tube (DT) efficiency to detect a single hit was defined and measured as the ratio between the number of detected and expected hits. The position of expected hits was determined using sets of well reconstructed track segments: at least 7 or at least 3 hits were required to be associated to a segment, in the Phi and Theta view respectively. Moreover the segment itself was required to cross the chamber with an inclination lower than 45 degrees. The intersection of such a high quality track segment with a DT layer determined the position of the expected hit. The DT was considered efficient if a hit was found within the tube where it was expected to be. Such efficiency can be computed choosing different detector granularities. We present results for Phi layer efficiency, Theta layer efficiency, Phi superlayer efficiency and chamber efficiency. This plot shows the Phi Layer efficiency.

EffiTheLay2016.png

DT single hit efficiency 2016 - Theta Layers

The Drift Tube (DT) efficiency to detect a single hit was defined and measured as the ratio between the number of detected and expected hits. The position of expected hits was determined using sets of well reconstructed track segments: at least 7 or at least 3 hits were required to be associated to a segment, in the Phi and Theta view respectively. Moreover the segment itself was required to cross the chamber with an inclination lower than 45 degrees. The intersection of such a high quality track segment with a DT layer determined the position of the expected hit. The DT was considered efficient if a hit was found within the tube where it was expected to be. Such efficiency can be computed choosing different detector granularities. We present results for Phi layer efficiency, Theta layer efficiency, Phi superlayer efficiency and chamber efficiency. This plot shows the Theta Layer efficiency.

EffiPhiSL2016.png

DT single hit efficiency 2016 - Phi Superlayers

The Drift Tube (DT) efficiency to detect a single hit was defined and measured as the ratio between the number of detected and expected hits. The position of expected hits was determined using sets of well reconstructed track segments: at least 7 or at least 3 hits were required to be associated to a segment, in the Phi and Theta view respectively. Moreover the segment itself was required to cross the chamber with an inclination lower than 45 degrees. The intersection of such a high quality track segment with a DT layer determined the position of the expected hit. The DT was considered efficient if a hit was found within the tube where it was expected to be. Such efficiency can be computed choosing different detector granularities. We present results for Phi layer efficiency, Theta layer efficiency, Phi superlayer efficiency and chamber efficiency. This plot shows the Phi Superlayer efficiency.

EffiChamber2016.png

DT single hit efficiency 2016 - Chambers

The Drift Tube (DT) efficiency to detect a single hit was defined and measured as the ratio between the number of detected and expected hits. The position of expected hits was determined using sets of well reconstructed track segments: at least 7 or at least 3 hits were required to be associated to a segment, in the Phi and Theta view respectively. Moreover the segment itself was required to cross the chamber with an inclination lower than 45 degrees. The intersection of such a high quality track segment with a DT layer determined the position of the expected hit. The DT was considered efficient if a hit was found within the tube where it was expected to be. Such efficiency can be computed choosing different detector granularities. We present results for Phi layer efficiency, Theta layer efficiency, Phi superlayer efficiency and chamber efficiency. This plot shows the Chamber efficiency.

EffiTheLay1516.png

DT single hit efficiency 2015/2016 comparison - Phi Layers

The same 2016 results above, now compared to 2015 results. This plot shows the Phi Layer efficiency.

EffiTheLay1516.png

DT single hit efficiency 2015/2016 comparison - Theta Layers

The same 2016 results above, now compared to 2015 results. This plot shows the Theta Layer efficiency.

EffiPhiSL1516.png

DT single hit efficiency 2015/2016 comparison - Phi Superlayers

The same 2016 results above, now compared to 2015 results This plot shows the Phi Superlayer efficiency.

EffiChamber1516.png

DT single hit efficiency 2015/2016 comparison - Chambers

The same 2016 results above, now compared to 2015 results This plot shows the Chamber efficiency.

barrelEff 8June cmsStyle.png

RPC overall efficiency - Barrel

The overall efficiency distribution for the Barrel part of the RPC (Resistive Plate Chambers) system. The distribution is obtained using 2016 collision data at √s = 13 TeV, B = 3.8 T and an integrated luminosity of about 80 pb-1. Details of efficiency calculation method can be found in (J.Instrum.8(2013)P11002). The mean RPC efficiency was calculated to be 94.6 % - 95.1 %. The few chambers with low efficiency have known hardware problems.

endcapEff 8June cmsStyle.png

RPC overall efficiency - Endcap

The overall efficiency distribution for the Endcap part of the RPC (Resistive Plate Chambers) system. The distribution is obtained using 2016 collision data at √s = 13 TeV, B = 3.8 T and an integrated luminosity of about 80 pb-1. Details of efficiency calculation method can be found in (J.Instrum.8(2013)P11002). The mean RPC efficiency was calculated to be 94.6 % - 95.1 %. The few chambers with low efficiency have known hardware problems.

pdf version
seg eff 2016BMuonJSON May27subMay31.png

Measured efficiency (%, with statistical uncertainty) of each CSC in the CMS endcap muon detector to provide a reconstructed muon track segment.

The efficiency (in %, with statistical uncertainty only) of each Cathode Strip Chamber in the CMS endcap muon detector to provide a reconstructed muon track segment. A segment in a CSC is a straight-line track segment reconstructed from the hits on the 6 layers of the CSC. Segments are used as seeds for the full CMS muon track reconstruction algorithm, in combination with tracks reconstructed in the Silicon Tracker, in both the CMS High-Level Trigger (HLT) and CMS offline muon reconstruction. These efficiencies were obtained using a Tag & Probe technique in which Z→μμ candidates are selected based on the invariant mass of the combination of a reconstructed global muon (tag) with a reconstructed track (probe). The probe track is projected into the CSC system and a matching segment is searched for in each CSC the track traverses. To reduce backgrounds and ensure the probe actually enters the CSC under consideration, compatible hits are also required in a downstream CSC. In rings ME2/1, 3/1, and 4/1 each chamber covers 20° in φ all other chambers cover 10° in φ. There are a few (out of the total 540) chambers with known inefficiency usually due to one or more failed electronics boards which cannot be repaired without major intervention and dismantling of the system. There are also occasional temporary failures of electronics boards, lasting from periods of hours to days, which can be recovered without major intervention. Both contribute to lowered segment efficiency.

pdf version
seg 1D 2016BMuonJSONMay27subMay31.png

Overall Efficiencies of CSCs for providing reconstructed Muon Track Segments.

Measured efficiency of each CSC in the CMS Endcap Muon detector to provide a reconstructed muon track segment. There is one entry per CSC. Note that there are 540 CSCs in the system, but that the ME1/1 chambers are divided into two strip regions, labelled ME1/1A and ME1/1B giving effectively 612 separate detector regions, thus accounting for the total number of entries of 612 in each plot.

Spatial Resolution

Figure Caption
HitReso phiTheta 2016.png

DT single hit resolution.

The DT single hit resolution was computed from the width of the distribution of the observed distance between any reconstructed hit and the fitted segment it belongs to. Relying on the azimuthal symmetry of the detector, hits reconstructed within the same station of the same wheel were added together. However Phi Super Layers (measuring position on the R Phi plane) and Theta Super Layers (measuring position on the R z plane) were kept apart in order to take their geometrical differences into account. The present results show no differences from 2015 results shown in 2015 DT resolution and CMS DP-2015/061.
Features: (1) Within every station, both Theta and Phi Super Layers show a symmetric behaviour w.r.t. to the z=0 plane, as expected from the detector symmetry. (2) In Wheel 0, where tracks from the interaction region are mostly normal to all layers, the resolution is the same for Theta and Phi SL's. (3) Going from z=0 towards the forward regions, tracks from the interaction region have increasing values of pseudorapidity: this feature affects Phi and Theta SL’s in opposite ways. In fact the theta angle lies on the measurement plane of Theta layers, while it is orthogonal to it for Phi layers. The result is that in Theta SL's the increasing inclination angle, by spoiling the cell linearity, also worsen the resolution. Instead, in Phi SL's the inclination angle increases the track path within the tube (along the wire direction), thus increasing the ionization charge and improvingmthe resolution. (4) The poorer resolution of the Phi SL's in MB4, compared to MB1-MB3, is because in MB4 no corrections are applied to the hit position in order to take into account the muon time-of-flight and the signal propagation time along the wire.mIn fact in the MB4's no position information is available in the direction parallel to the wires.

clsREm4 8June cmsStyle.png

RPC cluster size.

The RPC cluster size is defined as the number of adjacent strips fired in response to the passage of a single particle. The example 2-dimensional plot represents the mean value of the cluster size distribution for every particular RPC eta partition (RPC roll) for one of the Endcap stations. The X axis corresponds to the sector numbers. There are 36 sectors per ring in the endcap stations and every sector covers 10° in azimuthal direction. The Y axis corresponds to the ring number and the RPC eta partition’s names. The rolls installed at lower eta are shown on the top of the plot, while the rolls at higher eta (closest to the beam pipe) are shown on the lower part of the plot. The cluster size depends on the strip pitch and because of this it is higher for the innermost eta partitions (Ring 2, Rolls C) and it is smaller for the outermost ones (Ring 3, Rolls A).

barrelCLS 8June cmsStyle.png

RPC cluster size - Barrel.

The cluster size distribution obtained for the RPC Barrel. The mean value of cluster size should be ~2 strips and no larger deviations from this value have been observed. The mean cluster size of 2 strips is in agreement with the previous measurements and also with the TDR specification..

endcapCLS 8June cmsStyle.png

RPC cluster size - Endcap.

The cluster size distribution obtained for the RPC Endcap. The mean value of cluster size should be ~2 strips and no larger deviations from this value have been observed. The mean cluster size of 2 strips is in agreement with the previous measurements and also with the TDR specification..

pdf version
csc resol 2016.png

Spatial Resolution of CSCs in 2016

A CMS Cathode Strip Chamber (CSC) contains 6 layers of gas, each with cathode and anode planes. The cathode planes are divided into radial strips and allow a precise measurement of the azimuthal position of a hit from a muon traversing the layer. This is the direction of bending of a muon in the solenoidal field. A muon typically leaves a hit in each layer, with charge deposited on a few neighboring strips, and a muon track segment is a straight line fit to these hits. To estimate the spatial resolution one hit is removed from a segment and the segment refit from the remaining hits. The residuals between segment and hit are quite Gaussian so a simple fit to a Gaussian is used to measure a ‘resolution’ per layer. There are 10 types of CSC in the CMS Endcap Muon system, labelled MEi/j where i labels the station and j the ring. There are two independent regions of strips in the ME1/1 CSCs, with the split at |η| = 2.1.The inner region, closest to the beam line, is called ME1/1A and the outer ME1/1B. In non-ME1/1 CSCs, alternate layers are offset by half a strip width between each other. Better resolution is obtained if a hit is near the edge of a strip, rather than near the center, because then more charge is shared between strips. Therefore resolutions are measured separately for center and edge regions.The layer measurements are combined to give an overall resolution per CSC (‘station’) by 1/σ2station = 6/σ2L (ME1/1) and 1/σ2station = 3/σ2C + 3/σ2E (non-ME1/1 chambers.) The figure shows typical residuals for some CSCs: in ME1/1A, ME1/1, and in ME2/1 and ME2/2 for hits near the center and near the edge of the central strip in a hit cluster.

pdf version
csc resol 1516.png

Spatial resolutions of CSCs in 2015 and 2016 (μm per station)

The table summarizes the resolutions per station measured for all chamber types in the CMS CSC system in early 2016 data, and values measured at the end of 2015 for comparison. Statistical uncertainties from fits are negligible, and systematic uncertainties dominate. These arise primarily from variation of the resolution with atmospheric pressure (the gas gain has been measured to increase by 7-8% as atmospheric pressure decreases by 1%, and this improves the spatial resolution), with angle of incidence of the muon, and with muon momentum. The apparent improvement in resolution in 2016 is likely due to the difference in average atmospheric pressure in the data collection periods. All values well within the design specifications of the CSC system.

Time Resolution

Figure Caption
dt timing 2016.png

DT time resolution.

DT time information is obtained from 3-parameter fit of segments, where position, direction and time of a crossing track are determined at once. This method was shown to allow high reconstruction and identification efficiency for Out-Of-Time tracks and at the same time to improve the resolution for In-Time-Tracks. See DT 2015 results. The slight asymmetry that was observed in the central peak of the DT time distribution at the first implementation of the algorithm, was traced back to be due to remaining delta rays and muon showers. In fact whatever ionized charge reaching a DT wire before the signal actually produced by a muon track, causes an anticipated hit that, due to the intrinsic electronic dead time, has the effect to hide any hits occurring in the next 150 ns, including the correct one. A new iterative pruning mechanism was implemented in 2016: it discards "outlier" hits when performing the timing calculation for a muon track. As the calculation makes use of all hits associated to a muon track (provided by all the crossed DT stations), the hits from delta rays and showers can be rejected within an individual chamber using the other chambers as reference. The DT time resolution obtained, measured as the width of the central peak of the DT time distribution, is 1.4 ns: significantly improved w.r.t. the 2015 value of 2.0 ns.

pdf version
csc rechit time.png

Time resolution for reconstructed hits in the CSCs.

The distribution of times measured by the cathode strips from reconstructed hits in the Cathode Strip Chambers of the CMS endcap muon detector, for muons originating in pp collisions at √s = 13 TeV in 2015 and in 2016. Only rechits associated with segments contributing to reconstructed global muons with pT > 5 GeV/c and passing the standard CMS Muon ‘Tight Id’ requirement, but without a z-vertex restriction, are used. The hit times are obtained from a template fit to the digitized cathode strip pulse height distributions. Corrections are applied to each CSC to account for the muon time-of-flight from the interaction point and for signal propagation in cables and the electronics, and additional adjustments are made to account for variations in the time response of the system. The average value in each ring of CSCs in the system is adjusted to zero for muons from pp collisions in 2015, but this has not yet been done for the 2016 data. The overall distribution of times in each year shows a mean value close to zero and an r.m.s. width of approximately 7.5 ns. For ease of comparison, the 2016 data have been scaled by a factor of 2.6 so that the areas of both distributions are the same.

pdf version
csc segment time.png

Time resolution for reconstructed muon track segments in the CSCs.

The distribution of times measured for reconstructed track segments in the Cathode Strip Chambers of the CMS endcap muon detector, for muons originating in pp collisions at √s = 13 TeV, in 2015 and 2016. The segment time combines the times measured by the cathode strips and anode wires of all reconstructed hits composing a segment, after time calibration of the strips and wires. The overall distribution of times shows a mean value close to zero and a width of approximately 3.1 ns.

Local Trigger Efficiency

Figure Caption
Approval effic vs eta.png

DT local trigger efficiency a a function of η.

The DT Local Trigger (DTLT) efficiency was defined and measured as the ratio between the number of observed and expected triggers in events which contained a "Global Muon" reconstructed in the barrel. The expected triggers were defined requiring the presence in a chamber of a reconstructed track segment, belonging to the Global Muon and having at least 4 associated hits in the Phi view. The DTLT was then considered efficient if a trigger primitive was delivered at the correct bunch crossing in the same chamber. Note that no efficiency can be computed if for any reason no reconstructed segments are available. The computed efficiency is shown as a function of the Global Muon direction and transverse momentum. This plot shows the dependence of the efficiency measured in 2016 on pseudorapidity, and is compared to measurements from 2015.

Approval effic vs pt.png

DT local trigger efficiency a a function of pT.

The DT Local Trigger (DTLT) efficiency was defined and measured as the ratio between the number of observed and expected triggers in events which contained a "Global Muon" reconstructed in the barrel. The expected triggers were defined requiring the presence in a chamber of a reconstructed track segment, belonging to the Global Muon and having at least 4 associated hits in the Phi view. The DTLT was then considered efficient if a trigger primitive was delivered at the correct bunch crossing in the same chamber. Note that no efficiency can be computed if for any reason no reconstructed segments are available. The computed efficiency is shown as a function of the Global Muon direction and transverse momentum. This plot shows the dependence of the efficiency measured in 2016 on transverse momentum, and is compared to measurements from 2015.

Approval phi vs eta.png

DT local trigger efficiency station by station vs η and φ of DT track segment.

The DT Local Trigger (DTLT) efficiency was defined and measured as the ratio between the number of observed and expected triggers in events which contained a "Global Muon" reconstructed in the barrel. The expected triggers were defined requiring the presence in a chamber of a reconstructed track segment, belonging to the Global Muon and having at least 4 associated hits in the Phi view. The DTLT was then considered efficient if a trigger primitive was delivered at the correct bunch crossing in the same chamber. Note that no efficiency can be computed if for any reason no reconstructed segments are available.
Comments: (1) Significant hardware upgrades of the trigger system, downstream of the DTLT, were performed in 2016, but are not expected to affect DT trigger primitives at the chamber level. However, the electronic chain that the primitives go through before being readout was actually replaced so that in principle we could expect some effects. The MB3 and MB4 DTLT efficiency observed in 2016 is slightly better than in 2015. More accurate studies are in progress to assess the impact of the new hardware at all trigger levels. (2) Temporary problems with the readout electronics may prevent the trigger efficiency calculation in an entire chamber: this was the case for the run shown here, where an MB2 chamber of Wheel +2 was in a faulty readout state and did not provide reconstructed segments to perform the efficiency computation. (Nonetheless the DTLT was working normally)..

pdf version
lct eff 2016BMuonJSON May27subMay31.png

Measured efficiency (%, with statistical uncertainty) of each CSC in the CMS endcap muon detector to provide a trigger primitive for the CMS Level-1 trigger.

The efficiency (in %, with statistical uncertainties only) of each CSC in the CMS Endcap Muon Detector to provide a Trigger Primitive for input to the CMS Level-1 Trigger. A Trigger Primitive in a CSC is a pattern of hits consistent with arising from a muon track crossing the chamber. These efficiencies were obtained using a Tag & Probe technique in which Z→μμ candidates are selected based on the invariant mass of the combination of a reconstructed global muon (tag) with a reconstructed track (probe). The probe track is projected into the CSC system and a nearby trigger primitive is searched for in each CSC the track traverses. To reduce backgrounds and ensure the probe actually enters the CSC under consideration, compatible hits in a CSC downstream are also required. In rings ME2/1, 3/1, and 4/1 each chamber covers 20° in φ; all other chambers cover 10° in φ.

pdf version
lct 1D 2016BMuonJSONMay27subMay31.png

Overall Efficiencies of CSCs for providing Trigger Primitives.

Measured efficiency of each CSC in the CMS Endcap Muon detector to provide a Trigger Primitive for input to the CMS Level-1 Trigger. There is one entry per CSC. Note that there are 540 CSCs in the system, but that the ME1/1 chambers are divided into two strip regions, labelled ME1/1A and ME1/1B giving effectively 612 separate detector regions, thus accounting for the total number of entries of 612 in each plot.


Topic attachments
I Attachment History Action Size Date Who Comment
PNGpng Approval_effic_vs_eta.png r1 manage 38.3 K 2016-06-10 - 18:29 TimCox  
PNGpng Approval_effic_vs_pt.png r1 manage 33.7 K 2016-06-10 - 18:29 TimCox  
PNGpng Approval_phi_vs_eta.png r1 manage 46.9 K 2016-06-10 - 18:29 TimCox  
PNGpng EffiChamber1516.png r1 manage 15.1 K 2016-06-10 - 15:33 TimCox  
PNGpng EffiChamber2016.png r1 manage 12.0 K 2016-06-10 - 15:33 TimCox  
PNGpng EffiPhiLay1516.png r1 manage 16.4 K 2016-06-10 - 15:33 TimCox  
PNGpng EffiPhiLay2016.png r1 manage 13.0 K 2016-06-10 - 15:33 TimCox  
PNGpng EffiPhiSL1516.png r1 manage 14.2 K 2016-06-10 - 15:33 TimCox  
PNGpng EffiPhiSL2016.png r1 manage 13.1 K 2016-06-10 - 15:33 TimCox  
PNGpng EffiTheLay1516.png r1 manage 13.7 K 2016-06-10 - 15:33 TimCox  
PNGpng EffiTheLay2016.png r1 manage 12.1 K 2016-06-10 - 15:33 TimCox  
PNGpng HitReso_phiTheta_2016.png r1 manage 46.5 K 2016-06-10 - 17:24 TimCox  
PNGpng barrelCLS_8June_cmsStyle.png r1 manage 16.6 K 2016-06-10 - 17:24 TimCox  
PNGpng barrelEff_8June_cmsStyle.png r1 manage 17.8 K 2016-06-10 - 16:15 TimCox  
PNGpng clsREm4_8June_cmsStyle.png r1 manage 45.4 K 2016-06-10 - 17:24 TimCox  
PDFpdf csc_rechit_time.pdf r1 manage 195.9 K 2016-06-10 - 18:29 TimCox  
PNGpng csc_rechit_time.png r1 manage 108.1 K 2016-06-10 - 18:29 TimCox  
PDFpdf csc_resol_1516.pdf r1 manage 27.1 K 2016-06-10 - 17:24 TimCox  
PNGpng csc_resol_1516.png r1 manage 163.6 K 2016-06-10 - 17:24 TimCox  
PDFpdf csc_resol_2016.pdf r1 manage 406.5 K 2016-06-10 - 17:24 TimCox  
PNGpng csc_resol_2016.png r1 manage 518.8 K 2016-06-10 - 17:24 TimCox  
PDFpdf csc_segment_time.pdf r1 manage 195.9 K 2016-06-10 - 18:29 TimCox  
PNGpng csc_segment_time.png r1 manage 101.1 K 2016-06-10 - 18:29 TimCox  
PNGpng dt_timing_2016.png r1 manage 18.0 K 2016-06-10 - 18:29 TimCox  
PNGpng endcapCLS_8June_cmsStyle.png r1 manage 16.3 K 2016-06-10 - 17:24 TimCox  
PNGpng endcapEff_8June_cmsStyle.png r1 manage 17.2 K 2016-06-10 - 16:15 TimCox  
PDFpdf lct_1D_2016BMuonJSONMay27subMay31.pdf r1 manage 13.8 K 2016-06-10 - 16:33 TimCox  
PNGpng lct_1D_2016BMuonJSONMay27subMay31.png r1 manage 79.7 K 2016-06-10 - 16:33 TimCox  
PDFpdf lct_eff_2016BMuonJSON_May27subMay31.pdf r1 manage 45.7 K 2016-06-10 - 16:33 TimCox  
PNGpng lct_eff_2016BMuonJSON_May27subMay31.png r1 manage 460.3 K 2016-06-10 - 16:33 TimCox  
PDFpdf seg_1D_2016BMuonJSONMay27subMay31.pdf r1 manage 13.7 K 2016-06-10 - 16:33 TimCox  
PNGpng seg_1D_2016BMuonJSONMay27subMay31.png r1 manage 78.4 K 2016-06-10 - 16:33 TimCox  
PDFpdf seg_eff_2016BMuonJSON_May27subMay31.pdf r1 manage 45.5 K 2016-06-10 - 16:33 TimCox  
PNGpng seg_eff_2016BMuonJSON_May27subMay31.png r1 manage 435.8 K 2016-06-10 - 16:33 TimCox  
Edit | Attach | Watch | Print version | History: r5 < r4 < r3 < r2 < r1 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r5 - 2016-07-28 - TimCox
 
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    CMSPublic All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright &© 2008-2021 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
or Ideas, requests, problems regarding TWiki? use Discourse or Send feedback