View
221
Download
0
Category
Tags:
Preview:
Citation preview
Plan of the talk
● Work involved in MTCC2● Data analysis..i.e● LED data- Local runs for HCAL● Global data runs..● Things learnt from this analysis..
Pre-requisites for the analysis..
● Location of the data.● Account on various machines to
access the data and the analysis codes and the HO map file.
● How to use the CMSSW and ROOT.
Initial Work involved....
● Understanding the procedure of the data taking..that is using the xdaq .
● Configuration of the test setup involved..● The various aspects of the CMSSW● The understanding of the existing codes for
data analysis ..and how they need to be modified..
Recalling Some useful facts..
● The Readout box components :• Readout modules(RMs), Calibration Module,
CCM, LEDs● RM :HPD(18 pixels); each pixel: each tile● LED data Run: on pixel data and off pixel data● LED local data : reflection of the electronics● Global data : reflects quality of the subdetector
Actual HO RBXs installed for MTCC
● Ring0 sector 10
● Ring1 sector 3,4 ,9 and 10
● Ring2 sector 3,4,9 and 10
Chasing the Noise in the HCAL
● The global HCAL data showed noise that increased with the rampup of the Magnetic field..
● Investigation of the different sections of the hcal was therefore done for local LED data too..
● Strong corelation of noise(of local and global data) seen .
Noise in global data (timing of HCAL events above 10fC wrt
CSC trigger)
Nu
mb
er o
f ch
ann
els
wit
h E
>10
fC
How we realized, we have a problem:,Over 6k entries for 20k triggers,Timing of signals not synchronized DT trigger
timing of HCAL signals>2GeV,
r3533, 2T field (10kA)
Good timing,(HO is ok at 10kA),Only 74 evts per 20k triggers
Local LED/ped R1302,HB, w15/, Bfield=2T
Event number
HT
R c
han
nel
nu
mb
er
Single eventMultiple hits
New observations of HO Noise
● eta-phi distribution is not flat ● Distributions of noise are mirror symmetric (for
HO, YB+1, s9 and s10)● They map out into HPD pixels in specific way● Noise seems to originate from a single location on
the HPD surface (edge)● We see multiple pixels seeing noise in a single
event● Noise data consistent with HV discharges..
Noise count vs HPD pixel position
Each eta-phi tower on detectorCorresponds to a single pixel on HPD
Maps are same for:YB1/S9 (phi=47, 48, 49)YB1/S10 (phi=58, 57, 56)
20 26 28
222 286 357
72 149 190 272
32 126 71 74
57 58 x 133 201
RM2, phi=47-49
Noise count vs HPD pixel position
Each eta-phi tower on detectorCorresponds to a single pixel on HPD
Maps are same for:YB1/S9 (phi=47, 48, 49)YB1/S10 (phi=58, 57, 56)
43 44 76
216 254 359
94 156 189 344
53 146 91 145
83 64 x 182 314
RM1, phi=56-58
New observations
● eta-phi distribution is not flat ● Distributions of noise are mirror symmetric (for
HO, YB+1, s9 and s10)● They map out into HPD pixels in specific way● Noise seems to originate from a single location on
the HPD surface (edge)● We see multiple pixels seeing noise in a single
event● Noise data consistent with HV discharges..
RBX setup in b157 (meyrin))
● In B157, we have magnet going up to 1T● Data being taken with HCAL RBXes to study pixel
cross talk vs Bfield● This setup was used to understand HCAL noise• Noise effect was reproduced in HO RBX at
B=0.3T (no noise at B=0T)
Conclusion
Effect depends on Bfield Effect depends on HV Effect does not depend on Bias voltage Effect stays with RBX (or RM). It is not caused by
faulty Bulgarian power supply Effect is seen in both pedestal (local trigger) and
muon-triggered global data Effect is present in both LED on and off Effect is present during ramp-up, flat-top of the
magnet and during ramp-down, even when all magnet power supplies are disconnected
Recommended