The MIRI Sensor Chip Assembly (SCA) simulator

1
The MIRI Sensor Chip Assembly (SCA) simulator

• Steven Beard
• UK Astronomy Technology Centre
• Royal Observatory, Edinburgh, UK

ROE Workshop Following the Photon, 10-12
October 2011
2
JWST Instruments and Detectors
MIRI
NIRCAM
NIRSPEC
FGS-TFI
3 SiAs detectors
2 HgCdTe detectors
2 HgCdTe detectors
10 HgCdTe detectors
3
JWST Detector Readout
Each integration is made from several
non-destructive reads, divided into groups.
4
MIRI Summary
Imager Coronagraph Low Res. Spectrograph

• Medium Res. Spectrograph
• Short wavelength
• Long wavelength

5
MIRI SCA Subarrays
Subarray name Rows Cols First row First col Light sensitive cols
FULL 1024 1032 1 1 1024
MASK1065 256 256 1 1 224
MASK1140 256 256 229 1 224
MASK1550 256 256 452 1 224
MASKLYOT 320 320 705 1 275
BRIGHTSKY 512 864 1 1 512
SUB256 256 608 1 1 256
SUB128 128 132 897 1 128
SUB64 64 68 897 1 64
SLITLESSPRISM 512 68 348 1 64
The detector can be read out full frame in a
variety of pre-defined subarrays (windows).
6
MIRI Modes to be Simulated

• MIRI has a number of different modes to be
  simulated, each of which include several common
  requirements

7
MIRI Modes to be Simulated

• MIRI has a number of different modes to be
  simulated, each of which include several common
  requirements
• MRS mode is simulated by Specsim.

8
MIRI Modes to be Simulated

• MIRI has a number of different modes to be
  simulated, each of which include several common
  requirements
• MRS mode is simulated by Specsim.
• All the imager modes are simulated by Mirim Sim
  (Rene Gastauds presentation?).

9
MIRI Modes to be Simulated

• MIRI has a number of different modes to be
  simulated, each of which include several common
  requirements
• MRS mode is simulated by Specsim.
• All the imager modes are simulated by Mirim Sim
  (Rene Gastauds presentation).
• MTS Sim simulates the telescope simulator used
  for Flight Model (FM) testing.

10
MIRI Modes to be Simulated

• MIRI has a number of different modes to be
  simulated, each of which include several common
  requirements
• MRS mode is simulated by Specsim.
• All the imager modes are simulated by Mirim Sim
  (Rene Gastauds presentation).
• MTS Sim simulates the telescope simulator used
  for Flight Model (FM) testing.
• All the simulators share the same detector
  simulation provided by SCA Sim.

11
The MIRI Simulator Suite

• Having a suite of simulators ensures that every
  problem is solved only once.
• But we did miss the opportunity to share
  simulate targets and simulate background.
• SCASim provides a common detector simulation
  service for the other simulators.
• It converts detector illumination information
  from any MIRI simulator and generates simulated
  MIRI data in a choice of formats accepted by MIRI
  pipeline and analysis software.

12
Required SCA Simulator Steps

• The SCA Simulator simulates
• Quantum efficiency
• Reference pixels and outputs
• Bad pixels
• Dark current and hot pixels
• Persistence
• Readout modes
• Poisson noise and read noise
• Bias, gain and non-linearity
• Cosmic ray effects
• Subarray (window) modes
• The simulation is controlled by
• Input parameters
• e.g. Readout mode
• Configuration information
• e.g. Detector properties
• Configuration measurements
• e.g. Dark current vs temperature

X
X
X
X
13
Design Decisions

• I used an Object-Oriented design that would make
  SCASim more flexible and reusable.
• I also wrote the simulator in Python an
  object-oriented language with several useful
  scientific and array processing add-ons (numpy,
  scipy, matplotlib, pyfits, etc)
• Since the JWST detector readout modes are very
  similar, I chose to make the SCASim workable with
  any JWST detector not just the MIRI detectors.
• So ngroups ? nframes. Useful for NIRCAM and
  NIRSPEC as well?
• I also chose to encapsulate as much of the
  detector information in parameter files, rather
  than in software constants.
• By modifying these parameters and/or the class
  methods, SCASim could be adapted to similar kinds
  of detector.
• The simulator modules were developed along with
  unit tests.
• This ensured that changes didnt generate
  unwanted side effects.

14
The Python language

• Good things

• Bad things

• An OO language.
• An interpreted language.
• Ease of experimentation.
• Dynamic typing
• Functions and objects can be treated like
  variables.
• Dictionaries and lists.
• A good way of storing parameters.
• Built in doc strings.
• Built in unit tests.
• Can be extended with many useful utilities.
• numpy, scipy, pyfits, matplotlib, etc

• Weak OO protection.
• Naming convention used to denote __private
  attributes.
• An interpreted language.
• Performance issues.
• C-based extensions greatly improve the
  performance.
• Dynamic typing!
• Variables can change type without your knowledge!
• numpy supports fixed types.
• Installation issues.
• It sometimes takes several attempts to install
  properly.

15
Python Utilities Used

• numpy
• Array processing utilities.
• Supports sparse addressing and slicing.
• Written in C, so improves performance.
• scipy
• A collection of science-oriented data processing
  functions (rather like the Fortran NAG library).
• pyfits
• Reads and writes FITS files.
• PIL (Python Imaging Library)
• Reads and writes image files (JPEG etc).
• matplotlib
• Plotting utilities (mathcad).
• Sphinx
• Document generation

16
SCASim Design
Name of class

• The SCA simulator has an object-oriented design.
• The core of the simulator is a Detector Array
  class.
• The operational interface is generic All
  detectors are illuminated, reset, integrated and
  read out, or can be hit by a cosmic ray.
• The detailed implementation of the methods
  simulates the effects of the MIRI detector.
• This makes the design adaptable and reusable.

Attributes
Operations
17
SCASim Design

• The detector uses a helper class, the Poisson
  Integrator, which encapsulates the Poisson
  statistics.
• Other associated classes are used to describe
  detector characteristics, such as bad pixels, hot
  pixels, dark current and quantum efficiency.

18
SCASim Design

• Each detector may be read out by one or more
  amplifiers (4 in the case of MIRI ref. output).

19
SCASim Design

• Each detector may be read out by one or more
  amplifiers (4 in the case of MIRI ref. output).
• Each amplifier is responsible for reading a
  particular slice (or zone) on the detector
  surface.
• Each amplifier has an associated gain, linearity
  and read noise.
• Note that dark current and read noise are derived
  from a generic Measured Variable class, used to
  described laboratory measurements (in this case
  measuring how dark current and read noise vary
  with temperature).

20
SCASim Design

• Each detector may be read out by one or more
  amplifiers (4 in the case of MIRI ref. output).
• Each amplifier is responsible for reading a
  particular slice (or zone) on the detector
  surface.
• Each amplifier has an associated gain, linearity
  and read noise.
• Note that dark current and read noise are derived
  from a generic Measured Variable class, used to
  described laboratory measurements (in this case
  measuring how dark current and read noise vary
  with temperature).

21
JWST Detector Readout

• MIRI detectors have 4 dark columns at the left
  and right edges.
• Detectors are read out by 4 amplifiers.
• Every 5th readout is interlaced with reference
  output columns.
• These are processed as shown before the data are
  transmitted to the ground as Level 0 FITS Data.
• Instrument metadata is added to make Level 1
  FITS Data.

22
SCASim Design

• The detector and the amplifiers can both be hit
  by cosmic rays during an integration, depending
  on the integration time and their target area.
• A Cosmic Ray Environment class describes the
  cosmic ray environment (solar condition etc),
  and can generate Cosmic Ray events.
• Cosmic ray events are selected from a library of
  simulated events created by Massimo Roberto at
  STScI (for all JWST detectors).

23
SCASim Design

• The SCA simulator also defines classes describing
  how an exposure is constructed from a series of
  integrations.
• The Integration class sequences the
  reset/integrate/readout operations in the
  Detector Array.
• An illumination map describes the intensity and
  wavelength of the illumination across the
  detector surface.

24
SCASim Design

• Finally, these additional classes show how the
  contents of the input file are distributed, and
  how the data associated with each exposure is
  written to an output file.
• The Sensor Chip Assembly class manages the
  simulation and provides a selection of interfaces
  to the outside world (not shown here).

25
SCASim External Interfaces
26
SCASim Design Process 4Class Diagram for MIRI
Exposure
Detector pixels represented by Detector Array
class with properties in Measured Variable
class. Illumination Map describes detector
illumination. Poisson Integrator is at the heart
of the simulation. It represents the potential
wells. Amplifier objects read out the
detector. Cosmic Ray objects dump energy onto
the detector. Properties of the Exposure and
Integration objects determine the structure of
the output file.
27
MIRI SCASim Class Summary

• DataMap
• Created from input files
• DarkMap
• BadPixelMap
• IlluminationMap
• QuantumEfficiency
• MeasuredVariable
• Dark current vs T
• Read noise vs T
• DetectorArray
• Amplifier

• PoissonIntegrator
• Detector pixels.
• CosmicRayEnvironment
• Generates cosmic rays
• CosmicRay
• Hits detector or amplifier
• ExposureData
• Manages FITS header
• Generates output file
• Exposure info
• Integration info

28
detector_properties.py

• def _config_file(name, detector_id)
• filename "ss.fits" (name, detector_id)
• abspath os.path.join(_datapath, 'detector',
  filename)
• if os.path.exists(abspath)
• return abspath
• _sca493
• _sca493'SCA_ID' 493 Numerical
  SCA ID
• _sca493'FPM_ID' "FPMSN106" Unique FPM
  ID
• _sca493'NAME' "Sensor Chip Assembly 493 with
  Focal Plane Module 106"
• _sca493'DETECTOR' "MIRIMAGE" ASCII SCA
  ID
• _sca493'CHIP' 'SiAs' Type of
  detector chip
• _sca493'ILLUMINATED_ROWS' 1024
• _sca493'ILLUMINATED_COLUMNS' 1024
• _sca493'LEFT_COLUMNS' 4 Reference
  columns on detector
• _sca493'RIGHT_COLUMNS' 4 Reference
  columns on detector
• _sca493'BOTTOM_ROWS' 0 There are no
  extra rows at the bottom
• _sca493'TOP_ROWS' 256 Reference rows
  in level 1 FITS image

• DETECTORS_DICT '493'_sca493,
• '494'_sca494,
• '495'_sca495
• READOUT_MODE
• READOUT_MODE'SLOW' (10, 2, 1, 0, 10,
  1, 1, 1)
• READOUT_MODE'SLOWINTAVG' (10, 2, 1, 0, 1,
  4, 1, 4)
• READOUT_MODE'SLOWGRPAVG' (10, 2, 1, 0, 4,
  1, 4, 1)
• READOUT_MODE'SLOWGRPGAP' (10, 2, 4, 8, 4,
  1, 1, 1)
• READOUT_MODE'FAST' (1, 0, 1, 0, 1,
  10, 1, 1)
• READOUT_MODE'FASTINTAVG' (1, 0, 1, 0, 1,
  4, 1, 4)
• READOUT_MODE'FASTGRPAVG' (1, 0, 1, 0, 4,
  1, 4, 1)
• READOUT_MODE'FASTGRPGAP' (1, 0, 4, 8, 4,
  1, 1, 1)
• DEFAULT_READOUT_MODE 'FAST
• SUBARRAY
• SUBARRAY'FULL' None

29
amplifier_properties.py

• _amp493_1
• _amp493_1'ID' 'SCA493_1'
• _amp493_1'NAME' "Amplifier 1"
• _amp493_1'COMMENTS' "Reads every 4th detector
  column starting at column 0"
• _amp493_1'TYPE' "illuminated"
• _amp493_1'REGION' "0" Starting column
• _amp493_1'AREA' 5.0e5 Area of amplifier
  surface in square microns
• _amp493_1'BIAS' 2 Bias in electrons
• _amp493_1'GAINTYPE' 'POLYNOMIAL'
  Type of gain function
• _amp493_1'GAIN' (-6.6e-8, 0.1683, 0.0)
  Coefficients for gain function
• _amp493_1'MAX_DN' 65535.0 Maximum DN output
  by this amplifier
• _amp493_1'READ_NOISE_FILE' _config_file('read_
  noise', '493', '1')
• _amp493_1'READ_NOISE_COL' 1 Column number
  where to find read noise

• _amplist494 (_amp494_1, _amp494_2, _amp494_3,
  _amp494_4, _amp494_5)
• _amplist495 (_amp495_1, _amp495_2, _amp495_3,
  _amp495_4, _amp495_5)
• _amplist493 (_amp493_1, _amp493_2, _amp493_3,
  _amp493_4, _amp493_5)
• AMPLIFIERS_DICT '494'_amplist494,
• '495'_amplist495,
• '493'_amplist493

30
cosmic_ray_properties.py

• CR_FLUX
• CR_FLUX'NONE' 0.0
• CR_FLUX'SOLAR_MIN' 4.8983E-8
• CR_FLUX'SOLAR_MAX' 1.7783E-8
• CR_FLUX'SOLAR_FLARE' 3.04683E-5
• CR_LIBRARY_FILES
• CR_LIBRARY_FILES'NONE' None
• CR_LIBRARY_FILES'SOLAR_MIN'
  _config_file('CRs_SiAs', 'SUNMIN', 0, 9)
• CR_LIBRARY_FILES'SOLAR_MAX'
  _config_file('CRs_SiAs', 'SUNMAX', 0, 9)
• CR_LIBRARY_FILES'SOLAR_FLARE'
  _config_file('CRs_SiAs', 'FLARES', 0, 9)
• CR_LIBRARY_FILES'MIN' 0
• CR_LIBRARY_FILES'MAX' 9
• CR_NUCLEONS ('H', 'He', 'C', 'N', 'O', 'Fe',
  '??')
• CR_COUPLING ((0.0,0.015,0.0),(0.015,0.94,0.015),
  (0.0,0.015,0.0))

• CR_ELECTRONS'SOLAR_MIN' \
• (
• (1.0e1, 0.0),
• (1.8e1, 0.0),
• (3.2e1, 0.0),
• (5.6e1, 0.02),
• (1.0e2, 0.05),
• (1.8e2, 0.05),
• (3.2e2, 0.05),
• (5.6e2, 0.07),
• (1.0e3, 0.08),
• (1.8e3, 0.10),
• (3.2e3, 0.20),
• (5.6e3, 0.3),
• (1.0e4, 0.5),
• (1.8e4, 1.0),
• (3.2e4, 0.6),
• (5.6e4, 0.15),
• (1.0e5, 0.08),

31
SCASim Demonstration.Start with some test
illumination data.
Intensity data
Wavelength data
32
Add Reference Pixels and Outputs
33
Apply Quantum Efficiency
34
Add Bad Pixels
35
Add Dark Current Hot Pixels
36
Apply Gain and Non-linearity
37
Add Poisson Noise (frame 2/18 shown)
38
Add Read Noise (frame 2/18 shown)
39
Add Cosmic Ray Events(frame 8/18 shown)
40
Build up of signal with group
41
Build up of signal with group
42
Build up of signal with group
43
Build up of signal with group
44
Build up of signal with group
45
Build up of signal with group
46
Build up of signal with group
47
Build up of signal with group
48
SCASim Experience

• Good

• Bad

• Too many inputs.
• Nobody has yet edited the configuration files or
  provided their own calibration files.
• But the simulation is only as good as the
  calibration and configuration info. given to it.
• Only known effects can be simulated.
• The underlying causes of the first and last
  integration effect and the pixel lag effect
  are not yet known.
• Perhaps caused by leaving detectors too long
  without flushing?
• Subarray vs full frame.
• But the simulator can help investigate such
  effects by trying out ideas.

• OO design made SCASim highly flexible and
  reusable.
• Rapid development time.
• Additional effects (such as variable dark current
  and persistence) were easy to add.
• Now a useful tool adaptable for other detectors.
• SCASim made successful predictions for MIRI FM
  testing.
• Expected S/N and exposure times.
• Effect of cosmic ray hits.
• It also helped development and testing of
  analysis software.
• DHAS cosmic ray detection

49
The MIRI Sensor Chip Assembly(SCA) Simulator -
Summary

• What does it do?
• It simulates the behaviour of the MIRI detector
  chips and focal plane electronics.
• This simulation is common to all MIRI simulators,
  so it saves duplication of effort.
• What is it for?
• MIRI observation planning.
• MIRI Flight Model test planning.
• JWST Pipeline development and testing.
• Design Features
• OO design, written in Python.
• Highly adaptable and reusable.
• Valid for all JWST detectors could be adapted
  for other instruments.

50
SCA Simulator Documentation

1. Online documentation athttp//miri.ster.kuleuven
   .be/bin/view/Internal/ScaSim
2. Source code at http//svn6.assembla.com/svn/jwst/t
   runk/teams/miri/
3. MIRI Sensor Chip Assembly (SCA) Simulator User
   Manual, Steven Beard.
4. MIRI Software Installation Instructions, Julien
   Morin.
5. MIRI Sensor Chip Assembly (SCA) Simulator
   Software Design Document, Steven Beard.
6. MIRI SCAsim Software Reference Manual, Steven
   Beard.

51
Questions?

• ?