RVS Calibration

1
RVS Calibration

• RVS Calibration First Look Workshop, Paris
• Mark Cropper

2
Contents

1. Which parameters will be calibrated?
2. How will they be calibrated?
3. Type/volume of calibration data?
4. Auxiliary data required for the calibrations?

3
1. Which Parameters? Guidelines

• The aim of calibration is to transform
  measurements from the internal instrument system
  to a universal system (cgs/SI).
• Simply all those parameters that are necessary
  to do this transformation should be included.
• Obvious ones easily come to mind examples
• wavelength reference point
• throughput as a function of wavelength
• etc.
• However, a complete list is more difficult to
  compile and depends on a number of other
  parameters, such as temperature, and others we
  may not realise yet ? some flexibility needs to
  be retained.

4
1. Which Parameters? Guidelines (ctd)

• Also, calibration has to be practical and driven
  by scientific needs (no point in measuring
  something for the sake of it) ? implications for
  the policy on sub-system level testing, eg.
• information per optical surface
• information per CCD etc.
• In-orbit calibration uses as a basis the ground
  calibration/performance verification, together
  with
• calibration file structures
• infrastructure (conceptual, harware/software
  systems etc.)
• simulation and analysis software
• which it is sensible to re-use to some extent if
  this is helpful (but issues of information and
  knowledge transfer).

5
1. Which Parameters? Guidelines (ctd)

• Calibration has to take note of the particular
  circumstances for Gaia
• no dedicated observations will be made
• all data will be taken in TDI mode
• fixed instrumental configuration (very helpful)
• the dataset will be very large ? automation is
  required
• Calibration data must be sufficient to
• provide scientific data products
• allow data to be combined on board prior to
  telemetry
• provide long-term monitoring of instrument health
• All of these considerations need to be combined
  in making decisions on which parameters to
  include.

6
1. Which Parameters? Suggestions

• Calibration items required all of these are
  f(tlong-term)
• photometric throughput as f(CCD,y,?)
• AC line spread function as f(CCD,y,?,scan_law)
• AL line spread function as f(CCD,y,?)
• distortion map as f(CCD,y,?)
• wavelength scale and zero reference as f(y,?,
  tshort-term)
• CCD bias as f(CCD)
• CCD readout and dark noise as f(CCD,y)
• CCD TDI flat field as f(CCD,y,?)
• CCD blemishes as f(CCD,y)
• signal linearity as f(CCD)
• saturation level as f(CCD)
• scattered light/ghosts as f(CCD,y)

we come back to these later
7
Which Parameters? Suggestions (ctd)

• Many of these need to be known for each
  individual CCD (CCD) this is particularly true
  for the on-board processing software.
• ? is this also true for the ground processing?
• Data at individual CCD level is available only
  via diagnostic mode
• Could adopt the policy for ground processing that
  the science calibrations operate only on the
  co-added data.
• In this case have two calibration streams
• diagnostic mode data ? calibration for on-board
  software
• science data ? science calibration (mostly via
  SGIS)
• Else - use individual diagnostic mode
  calibrations to construct end-to-end calibration

? end-to-end ? detailed calibration ?
8
1. Which Parameters? Suggestions (ctd)

• On consideration, probably best to separate
  requirements and use two-stream approach
• ensures best compatibility between measurements
  for where they are required
• may be difficult to combine the detailed
  information via the diagnostic stream to use in
  the science calibration
• NOTE however,
• ground calibration will be responsible for both
  streams (on-board and science processing)
• long term trends and monitoring will need to be
  done on both streams
• cross checks will need to be made to ensure
  compatibility between the two calibration streams

9
1. Which Parameters? Suggestions (ctd)

• Splitting the two streams, then, on-board
  processing requires the following parameters
• photometric throughput as f(CCD,y,?)
• AC line spread function as f(CCD,y,?,scan_law)
• AL line spread function as f(CCD,y,?)
• distortion map as f(CCD,y,?)
• wavelength scale and zero reference as f(y,?,
  tshort-term)
• CCD bias as f(CCD)
• CCD readout and dark noise as f(CCD,y)
• CCD TDI flat field as f(CCD,y,?)
• CCD blemishes as f(CCD,y)
• signal linearity as f(CCD)
• saturation level as f(CCD)
• scattered light/ghosts as f(CCD,y)

10
1. Which Parameters? Suggestions (ctd)

• Science processing requires the following
  parameters
• photometric throughput as f(CCD,y,?)
• AC line spread function as f(CCD,y,?,scan_law)
• AL line spread function as f(CCD,y,?)
• distortion map as f(CCD,y,?)
• wavelength scale and zero reference as f(y,?,
  tshort-term)
• CCD bias as f(CCD)
• CCD readout and dark noise as f(CCD,y)
• CCD TDI flat field as f(CCD,y,?)
• CCD blemishes as f(CCD,y)
• signal linearity as f(CCD)
• saturation level as f(CCD)
• scattered light/ghosts as f(CCD,y)

note CCD
11
2. How will they be calibrated? overview

• In-orbit calibration needs to proceed in as
  automated manner as possible, especially as
  mission proceeds
• Calibration objects will be observed as part of
  the routine observations
• Ground processing needs a database (or database
  partition) identifying the calibration objects
  and what they are used for
• Standard ground processing pipelines will process
  calibration objects, after which they will be
  extracted and piped to the calibration pipeline
  for further processing
• Calibration pipeline will derive instrument
  calibration parameters
• Calibration files will then be updated
  automatically
• Instrument monitoring and alerts will be produced
  automatically

12
2. How will they be calibrated? science data
stream

• Calibration objects will be
• objects identified a priori by humans as suitable
  calibrators
• objects identified by the Gaia-RVS processing
  itself (SGIS)
• A mixture of both is required, using the SGIS
  approach for the major (self) calibration, and
  the external calibrators used to provide zero
  points/references to the SGIS calibrations
• In any case, suitable calibrators will need to be
  identified before launch, and this may require
  specific observations from other telescopes
• In the SGIS approach, suitable robust techniques
  will also need to be identified this will take
  place when the data centres are established post
  the ESA-AO.

13
2. How will they be calibrated?

• Standards will be required for
• spectrophotometric throughput and out-of-band
  rejection
• linearity of photometric response
• radial velocity zero points
• rotational velocity determinations
• astrophysical parameter determinations (Teff,
  log(g) etc.)
• Suitable stable stars selected by SGIS will be
  sufficient to monitor
• LSF in AL and AC directions
• distortion map
• dispersion law
• radial velocity stability
• spectrophotometric throughput stability

14
2. How will they be calibrated? science data
stream

• SGIS works by monitoring the output spectra from
  the single-transit pipelines
• bright stars, but not too bright to induce
  non-linearity/saturation
• range of appropriate types (F, G, K) - may be
  useful to have input from photometry to select
• SGIS then selects a subset of those which produce
  a consistent wavelength, throughput, position
  etc. parameters
• Note these are not constant parameters, as there
  will be drifts at some level, particularly on the
  wavelength scale.
• SGIS will need to allow for and track parameter
  changes in order for consistent sample to be
  selected ? some learning/iterative process?

15
2. How will they be calibrated? science data
stream
16
2. How will they be calibrated? science data
stream (ctd)
17
2. How will they be calibrated? science data
stream (ctd)
18
2. How will they be calibrated? diagnostic mode
data stream

• Diagnostic mode data has to be compared with
  co-added data from all 10 CCDs to
• update the on-board software parameters/look-up
  tables
• determine the additional calibration to operate
  on the co-added data telemetered from the RVS
• Sequence
• determine the calibration parameters
• update the calibration files automatically
  (calibration database version control and
  management)
• visualise the calibration parameters and provide
  tracking as a function of time
• provide automated alerts for deviations
• All of this requires specific software tools
  (both manual and automatic) to be developed, and
  for the process to be overseen in some way by
  humans

19
2. How will they be calibrated? diagnostic mode
data stream
20
3. Type/volume of calibration data

• Calibration data for the science stream will not
  add any particular telemetry overheads - these
  are normal observations
• Calibration data for on-board processing will
  incur overhead telemetry with the diagnostic mode
  data stream
• data rate can be adjusted according to timescale
  that variations occur in the calibration
• limited by overall telemetry allocation
• priority when data is to be lost before
  telemetry?
• Additional volume will be required to host the
  calibration database and the calibration
  pipelines
• SGIS
• external standards
• calibration for on-board processing
• tools
• status monitoring etc. TOTAL volume
  TBD

21
4. Auxiliary data

• Auxiliary data is required for the processing of
  the calibration (and other) pipelines. This
  includes
• instrument/spacecraft parameters (temperatures,
  solar angles, time, attitude, operating voltages)
  these provide the correction for long-term and
  short-term calibration drifts
• MBP/BBP photometry
• astrometry from Astro
• catalogue information (especially early in
  mission astrometric positions etc.)
• external standards (radial velocity,
  spectrophotometric etc.)
• Auxilliary data should include time history of
  the processing applied on the (calibration
  source) data up to that point.
• List is non-exhaustive further thought required