The BRITE Instrument

1
The BRITE Instrument Optics, Filters, Detector
Stefan Mochnacki Dept. of Astronomy
Astrophysics University of Toronto (Canada Day,
2008)?
2
Performance RequirementsFilters and Passbands

• Two Austrian and two Canadian BRITES two with
  Blue filter, two with Red filter.
• Passbands chosen to allow for mode separation and
  equal average weighting of signal (hence B filter
  narrower than R).
• Interference filters chosen for well-defined
  passbands.
• Direct measurement of delivered filters being
  done with DDO spectrograph (Pribulla).

3
(relative efficiency)?
4
Performance RequirementsPhotometric Accuracy

• Stars with V3.5 will be observed in
  differential mode.
• Each measurement (defined as up to 15 minutes of
  integration within a single orbit) shall be
  better than 0.001 mag.
• The error amplitude spectrum in photometric
  measurements for periods longer than a month
  shall be less than 20 ppm.

5
S/N, Exposure time Calculations

• Simple spreadsheet used to compute S/N given
  basic parameters and assumptions, at 5500 A for
  1500 A bandpass.
• Only noise sources source photon noise (vN),
  sky background 18 mag/sq), dark current (5 e- _at_
  20 C), readout noise (13 e-).
• Gaussian PSF (6 pixels), integrating aperture
  dia. 2 x FWHM. Approx. 110 pixels in
  integrating aperture.
• Plate scale 26.56/pixel. PSF 3', ap. dia.
  6'.
• Bright star limit set by saturation of central
  pixel.
• 400 exps. co-added (0.5 1.4 )s x 400 13
  min.
• No real penalty on brighter stars due to
  multi-exp.

6
(More accurate estimates have been computed by
A.Kaiser)?
7
Telescope

• We chose a 30mm aperture, 70mm focal length,
  which with a 35mm-format sensor 9µ pixels has an
  image scale of 26.6 arcsec/pix.
• The field of view is between 22 and 25 degrees.
• The lens design was driven by the need for
  adequately sampled images and good baffling,
  which also yielded an image-space telecentric
  telescope.

8
Optical Design (Blue)?
9
Red Brite spot matrix
10
Blue BRITE Point Spread Function
11
Blue BRITE design PSF simulation by Rainer
Kuschnig- approximate simulation - Zemax PSF
data resampled/binned to the CCD frame
BRITE Blue PSF _at_ 8.5 deg
Zemax PSF
Desired Gaussian PSF
12
BRITE Telescope Side View
13
BRITE Optical Cell
14
Instrument Baffle and Pupil Stop
15
Angled View (Materials)?
Aluminum
Magnesium
Aluminum
Stainless Steel
16
Telescope and Startracker Assembly
CCD Header Tray and Enclosure
17
BRITE Instrument

• Photometry
• Differential photometry with 0.1 precision.
• Error amplitude spectrum lt20 ppm, gt 1 month.
• Timing
• Exposure times 0.1-100s, known to 0.01
• Absolute time accuracy better than 0.1s.
• Optics
• Point Spread Function Gaussian, 5-7p FWHM.
• No vignetting, telecentric, minimum ghosting.
• Filter either 390-460nm or 550-700nm.
• 3 cm aperture telescope, 24 degrees FOV.
• Detector
• Detector temperature low, measured to 0.1ºC.
• SNR 1000 per 100s exposure at V3.5
• Design out sun stare risk, no shutter or door.
• Stray Light Baffle and light-tight instrument.

18
Front and Side Views of Instrument in the BRITE
Satellite
19
The BRITE Satellite
20
Choice of Detector

• After a CMOS effort, we chose the Kodak KAI-11002
  CCD Interline transfer, microlenses.
• Very low dark current at room temperature (no
  active cooling on BRITE), avg. 5e-/s/pix.
• Low readout noise 13e-/pix possible, BRITE
  prototype yields 25e-/pix.
• Good quantum efficiency 50
• Fast readout 12 MHz

21
CCD Sensor
Kodak KAI-11002
22
Quantum Efficiency Curve for BRITE CCD
23

• Issues affecting sampling of KAI-11002M CCD
• It is an INTERLINE TRANSFER device, which is
  essential for a CCD without a mechanical shutter.
• It uses microlenses to compensate for the 70
  dead space. The peak quantum efficiency goes
  from 16 to 50 when microlenses are fitted,
  suggesting there should be little effective dead
  space.
• This means that microlenses probably eliminate
  most of the dead space as far as its
  contribution to undersampling is concerned. Even
  back-illuminated CCDs have intra-pixel
  sensitivity variation.
• The angular response is quite good lt 12 degrees.

24
Variation of Quantum Efficiency with Angle of
Incidence in KAI-11002M (Kodak data).
25

• Prototype camera using KAI-11002 chip has been
  built and is used extensively for testing.
• Most important are tests on real star fields and
  on simulated star fields to measure PSFs and
  sampling.
• The dark current distribution is very promising
  if longer exposures are desired.
• We have adjustable bias level and gain. At
  present bias 100 ADU, gain 3e-/ADU. The gain
  is shown in photon transfer curves

26
Saturation effect
Slope 0.33 gt gain3 e/ADU
27
Another transfer curve (unsaturated)?
28
Typical light distribution of target imaged for
transfer curve measurements
29
60 second dark, bias subtracted. Note spikes,
typical of Kodak CCDs.
30
Dark current from 10 second exposure.
31
Dark current for 60 sec. integration
32
A blow-up of lower dark current levels (60 second
integration).
33
Same thing, linear count scale. 10 ADU bins.
34
Breakdown of imaging for T120 s.
35
Conclusions re Dark Current

• In a 60 second exposure, we see less than 5
  e-/p/s in 99 of pixels. The hot pixels appear
  always to be the same ones. This will be studied
  in greater detail.
• The device begins to break down at exposure times
  gt 60 seconds at 20C.
• Temperature dependence of dark current and
  breakdown could constrain temperature regulation
  by heating.

36
Real and Simulated Star Fields

• Real star fields observed without tracking, 1
  second exposures. These were made to focus the
  prototype and examine PSFs.
• Some star images at different positions are shown
  on the next slide.
• Artificial star fields in the lab used to compare
  photometry when small movements of the camera are
  made.
• All testing is now being done with the BRITE
  prototype lens (vibration test model) and
  electronics.

37
Prototype Blue BRITE and electronics
38
(No Transcript)
39
Real star images
40
Central image, -0.125 mm out of focus.
41
Artificial star field, in focus
42
Artificial field, 0.5mm out of focus
43
Test results

• Simulated star field tests show some
  undersampling errors when the BRITE lens is
  sharply focused, requiring remedial action (Alex
  Kaiser will discuss). Fully sampled images have
  shown no aliasing in previous testing.
• A wide range of focus positions is being tested.
• Real star field tests show the variation of the
  PSF over the whole field, and test the instrument
  sensitivity.