Digitization in EMC simulation

1
Digitization in EMC simulation

• Dmytro Melnychuk,Soltan Institute for Nuclear
  Studies,Warsaw, Poland

2
General scheme of digitization
APD
PWO
Preamplifier/shaper
FADC
Digi

• Conversion from deposited energy to photons
• Signal conversion in APD
• Signal transformation in pre-amplifier/shaper
• FADC
• Extraction of signal amplitude from FADC output
  (digital filter or simple parabolic fit)

3
Design parameters which affect digitization

• Noise (preamplifier APD)
• Photon statistics (photoelectrons per MeV)
• Shaping time (integration, differentiation)
• Scintillation time constant
• FADC frequency
• Foreseen but not used at the moment
• Noise dependency on the channel (measured values
  which will be taken from Condition Data Base
  (CDB))
• Non-uniformity of light production

4
Values of parameters

• 2 options for read-out electronics
• - Quad-preamplifier (developed by Basel group),
  noise value 1600 e- at -25 C (0.12 MeV),
  external shaping 250 ns
• - ASIC read-out, noise 4400 e- at -25 C (0.34
  MeV), built-in shaping 100 ns (?)
• Other parameters
• APD contribution to noise excess noise factor
  F1.38 at gain M50, dark current (?)
• 400 photoelectrons per MeV at -25 C, QE70
• 80 MHz FADC
• crystal scintillation time 6 ns

5
Energy resolution dependence on digitization

• 1) Single gammas with different energies were
  generated with uniform angular distribution (10,
  50, 100, 200, 500, 1000 MeV)
• 2) Dependence of energy deposited in 1 cluster on
  parameter of electronics
• (cluster group of neighboring crystals) Energy
  threshold for single crystal 1MeV

6
Photon statistic dependence
E10 MeV
E50 MeV
RMS (MeV)
7
Noise dependence
RMS (MeV)
8
Shaping time dependence
RMS (MeV)
9
Summary

• 1) Energy resolution is sensitive to photon
  statistic only at lowest energies
• 2) Noise contribution to energy resolution is on
  the level of 3 and decrease with growth of
  energy
• 3) Energy resolution does not depend
  significantly on the shaping time but it a little
  bit worse for smaller values
• To do
• Event mixing
• Realistic digi format from DAQ