The heavy ionising particles in CMS tracker

1
The heavy ionising particles in CMS tracker

• Hip effect description.
• The PSI beam test.
• Measurements of hip rate and induced dead time.
• Impact on Cms tracker performances.

Benjamin TROCME (IPN Lyon) on behalf of CMS
collaboration
2
CMS tracker and hip introduction

• The Cms all silicon tracker
• 16k modules with 75k read out chips(Apv).
• 128 strips/Apv, 4 or 6 Apvs per module.
• Inelastic hadronic interactions generate hips
• Up to 1000 mips signal in a 500mm Si detector.
• Well known effect, whose rate is predicted by
  theory of energy deposition in silicon.
• Impact on Apv first observed in 120GeV p test
  beam at Cern in october 01.

Lhc
Test beam
3
The hips and their effects on Apvs

• Inverter supply common to 128 strips
• In case of hip sharing of the current
  ''consumed'' by few hit strips.
• Hip topology
• High signal shared between few strips.
• Downward saturated baseline in the remaining
  strips of the apv.
• Apv may remain saturated during 100-700 ns

0 ns
25 ns
50 ns
75 ns
100 ns
125 ns
150 ns
175 ns
200 ns
225 ns
250 ns
275 ns
300 ns
325 ns
350 ns
375 ns
400 ns
425 ns
450 ns
475 ns
500 ns
525 ns
550 ns
575 ns
600 ns
625 ns
650 ns
Raw data(ADC)
Pedestal
300
Digital zero
100
1 128 Strips
4
The PSI test beam

• Goalestimate hip rate and induced Apv deadtime.
• 12 modules of different thicknesses and with
  different Rinv.
• Two trigger modes
• (PM1PM2)?1 data read outhip rate estimate.
• (PM1PM2!PM3) (enriched hips sample)?30 data
  read out spaced by 25ns Apv recovery following.

300MeV pions(/-) 20ns bunched beam
PM1 / PM2
PM3
5
The PSI test beam (contd)

• For the first time, use of Xdaq framework
• Stable with high performances.
• Efficient online monitoring.
• Quasi online analysis with Orca (C
  reconstruction framework).

Beam profile
S/N20
Outer barrel module
Endcap module
6
Hip caracterisation
Pedestal
CMN

• Basic method
• Common Mode Noiseevent by event Apv noise.
• Because of limited digitizer dynamic
• CMN gt -CMN0.
• An Apv is said ''hipped'' if CMN/CMN0 lt -0.8
• A 2nd method makes use of baseline flatness
• An Apv is said ''hipped'' ifRms(Raw data)lt1
• Not considered here but leads to similar results.
• Warningall numbers are preliminary.

7
Hip rate measurement

• Distribution normalised to number of good
  trigger,incident pions.
• Hip rate increases with
• Thickness.
• Rinv (100W by default).

Hip rate
Thickness500mm
Thickness320mm
Cut on CMN/CMN0
8
Induced dead time estimate

• Principle
• Use of data sample with 30 consecutive readouts.
  Only 6 detectors considered (500mm thick).
• Looks for a hipped Apv in a detector.
• In following samples,looks for tracks
  intercepting hipped Apv.
• Compute tracking efficiency
• Normalise it by regular efficiency (i.e non
  hipped Apv).

9
Induced dead time estimate(cont'd)
Corrected
Not corrected

• Average Apv dead time
• t S(1-etrack.) x t 165ns

10
Impact on physic

• Hip effect included in Cms full simulation
• 2 input parametership rate r and induced dead
  time t.
• Simple simulationwhen a hip occured,all apv data
  are lost during t.
• At high luminosity1.6 (0.2) equivalent hadron
  per bunch crossings in first (last) layer of
  silicon.
• Track angle with detector is taken into account.
• Pessimistic parameters choice

11
Impact on physic(cont'd)

• High P tracks (gt1Gev) from B jets
• Impact on HLT performances

12
Conclusion

• Due to the presence of an inverter supply common
  to 128 strips,the Apv chip can be disabled by
  heavy ionising particle.
• The occurrence of such events and the induced
  deadtime has been studied during a dedicated
  testbeam.
• Preliminary studies showed a very limited impact
  on physic performance.