Status & Plans of Segmentation
Y.P.Viyogi(V.E.C.C)Arun Prakash(B.H.U)
Geometry
Compact MuCh:5 abs.+15 detect layers : Monolithic typePipe_much.geoShield_standard.geoCave.geoTarget_au_250mu.geoMagent_standard.geoSts_same_z.geo
Transport• 10,000 UrQMD + 5 µ+ and 5µ- (Box Generator)• With momentum 2-5 GeV/c• Θ :5-25• Φ: 0-360
Pad size : 2.5 mm to 1 cmDigitization, calculate hits loss = 1-Nhits/NMCPointsResult always ~50%, independent of station/layer and pad size : not quite understandableSome study into the codes, found that even points having good coordinates were many times not getting assigned any pad/sector during segmentation (or digitization), leading to large loss.
Manual Segmentation
Next step…
• Tried with modular geometry : GEM and with different pad sizes
• Small number of events, just for checking the numbers
• Found hit loss negligible, also multiple hits very little
Decide to move ahead with Modular geometry, as this is the next and
practical step.
Geometry : do we need extra radius
• Presently the code adds 20cm to the nominal R_max at each station
• Results in ~25% more number of channels (this is only fictitious) in sectors sitting outside the nominal acceptance
• Points/hits in the regions beyond nominal R_max do not contribute to tracks
Decide to remain within nominal outer radius at each station.
Comparison of two scenario
R_max+20cm R_max+0.6cm
Station-1Station-1
R=70cm
Selection of GEM module and pad size• So far excellent work with various ideal and modular
geometry by the GSI-PNPI-Dubna group in simulation• We slowly move towards realistic detectors• GEM foils : routinely made in 30cm X 30 cm size, sector
shaped GEM foil made at CERN for RD51 collaboration which is ~50cm long. Even 1m long sectors being tried.
• FEE board size and mounting on the modules (too early to decide) : horizontal (parallel to detector plane) preferred. Consequence : pad size to be large enough for reasonable real estate of a 2-nXYter Board, approx. 8 cm x 8 cm
• Optimum pad size : a balance between simulation and hardware efforts
Selection of pad size : particle density
Selection of GEM module and pad sizes
• Nominal size of GEM assumed : 32 cm X 32 cm• Pad sizes 0.5 cm, 1 cm, 2 cm to respect the
condition that number of channels must be 2^n.
Quick check on segmentation with 50 events
Segmentation : First attempt, guided by particle density
Item Pads
No of regions
2 2 2 1 1
No of Channels
256 256 256 256 256 ~636k
V1 R int 32 50 70 - -
Lx 0.5 0.5 1.0 - -
Ly 0.5 0.5 1.0 - -
Rout 70 96 120 150 182.5
Lx 1.0 1.0 2.0 2.0 2.0
Ly 1.0 1.0 2.0 2.0 2.0
Av. Hit loss ~ 1.5%
V1 : Station1 as seen after segmentation
R=70 cm
Item Pads
No of regions
1 1 1 1 1
No of Channels
256 256 256 256 256 ~430k
V2 R int - - - - -
Lx - - - - -
Ly - - - - -
Rout 70 96 120 150 182.5
Lx 1.0 1.0 2.0 2.0 2.0
Ly 1.0 1.0 2.0 2.0 2.0
How bad can it be ?
Av hit loss ~4.8%
Item Pads
No of regions
1 1 1 1 1
No of Channels
256 256 256 256 256 ~120k
V3 R int - - - - -
Lx - - - - -
Ly - - - - -
Rout 70 96 120 150 182.5
Lx 2.0 2.0 4.0 4.0 4.0
Ly 2.0 2.0 4.0 4.0 4.0
Can we worsen it further ?
Av hit loss ~ 14.8%
Item Pads
No of regions
2 2 1 1 1
No of Channels
256 256 256 256 256 ~528k
V4 R int 25 40 - - -
Lx 0.5 0.5 - - -
Ly 0.5 0.5 - - -
Rout 70 96 120 150 182.5
Lx 1.0 1.0 2.0 2.0 2.0
Ly 1.0 1.0 2.0 2.0 2.0
Next stage : fine tuning
Av. Hit loss ~2.20%
V4 : station 1 after segmentation
R=70 cm
Next Steps
Transport : UrQMD + PLUTO events in reasonable mix for modular geometrySegmentation: try V4 firstReconstruction: Do the full reconstruction &calculate the efficiency for signal muons, rho, J/Psi etc.Fine tuning of pad sizes/GEM Geometry and then study again…. And again….