OSIRIS
is a versatile instrument and
provides many observing modes (see http://www.iac.es/project/OSIRIS/), some of which are new or at least non
standard. The OSIRIS SNR Calculator is a tool that covers the
main observing modes of the instrument. These are: broad-band imaging,
spectroscopy and tunable filter (TF) imaging. This tool estimates
signal-to-noise ratios for a given exposure time.
The
approach to do that is to compute the
expected signal at the detector by taking into account all the elements
involved in the detection process (telescope, lenses, filters,
detectors, etc.)
as well as the external factors contributing mainly to photon losses
or noise (atmosphere, sky brightness, Moon brightness). Many of the
factors are well
known or may be measured accurately, for instance the instrument
efficiencies,
while others must be estimated or assumed (e.g. sky brightness).
The
OSIRIS Signal-to-noise Calculator has
been developed by the author as a Java program to be executed through a
web
navigator. At present it can be accessed at http://www.ifca.unican.es/users/gserrano/OsirisETC/Calculators.html.
To run the application Java JRE 1.6 or later should be installed on the
client
machine. Also, the JNLP MIME type should be set to application/x-java-jnlp-file to be
opened by
javaws on the client navigator.
As the GTC/OSIRIS commissioning is progressing, real behaviour and
efficiencies are incorporated into the calculator. Also, not all the
subsystems are available yet: blue TF and blue order-sorter filters.
All (except the blue TF) are implemented in the calculator
but this
DOES NOT
IMPLY THAT THEY ARE WORKING AT THE TELESCOPE. Visit the GTC
OSIRIS page to check what is available and what is
not.
Telescope
and CCD parameters
The collecting area of the GTC is taken to be
741,400 cm2.
The
CCD pixel size has been measured during commissioning and it is 0.12718
arcsec.
Although it is possible to define CCD binning of 1x1, 2x2,
1x2,
and 2x1,
only binning
2x2 is operative in all modes
(September 2009). There are three readout modes: Spectroscopy, Imaging
and Acquisition, but the last one it is not suitable for scientific
applications, in general.
The CCD characteristics used in the calculators are listed in the next
table and correspond to commissioning values:
CCD characteristics
| Readout
mode |
Imaging |
Spectroscopy |
Acquisition |
| Readout velocity (kHz) |
200 |
100 |
500 |
| Gain (e-/ADU) |
0.95 |
1.18 |
1.46 |
| Binning |
2x2 |
2x2 |
2x2 |
| Readout
time (s) |
24.0 |
42 |
7.8 |
| Readout
noise (e-) |
4.5 |
3.5 |
8 |
Preliminary measurements of dark current after changing to the new
cryostat (February 2010; Antonio Cabrera private communication)
indicate a mean dark noise of 1.8 e-/hour/pixel.
Throughput
Before OSIRIS was mounted at the
telescope, most
of the transmissions and efficiencies of the optical subsystems and
CCDs were measured in laboratory conditions. Broad-band and order
sorter filters transmissions have been obtained during commissioning,
as well as for the available grisms. TF transmission has been
measured at some spectral ranges and has been updated (~80% lower that
theoretical).
The throughput of the
instrument (without filters or grisms) is plotted in the following
figure:
Atmosphere
The
simulator computes atmospheric extinction and absorption. Atmospheric
extinction at each wavelength is obtained from
standard curves measured at the observatory, which are given in
magnitudes per
airmass. Atmospheric
transmission is also taken into account using the following
curve:

Sky
background
Night
sky emission is also taken into account both in TF imaging and in
spectroscopy.
The spectrum used is from (Hanuschik, R.W. , 2003, A&A, 407, 1157) which is a high-resolution
flux-calibrated
spectrum from Paranal.
The
spectrum used in the calculator is shown in the next
figure and has been scaled to La
Palma
night-sky spectrum and degraded
to a resolution of 1Å.

For TF imaging, spectroscopy, and imaging using order sorter
or
user-defined filters, the sky background is computed from that
spectrum. In broad-band mode using the standard ugriz filters,
sky background is assumed to be the mean at the observatory.
To take into account moonlight we have classified
nights into three types: dark, grey, and bright. For
all modes, the surface brightness of the sky background is increased by
an
amount depending on the type of the night and on the spectral range.
These are
listed in the next table:
Moon contribution
to sky background
(in mag/arcsec2)
|
|
U
|
B
|
V
|
R
|
I
|
Z
|
|
Range(Å)
|
3400-4000
|
4000-4950
|
4950-5950
|
5950-7150
|
7150-8700
|
8700-10000
|
|
Grey
|
2.1
|
1.1
|
0.4
|
0.3
|
0.2
|
0.2
|
|
Bright
|
5.0
|
3.2
|
1.8
|
1.0
|
0.7
|
0.7
|
This approach is the same as that used in the ESO ETC calculators.
Take into account that sky brightness vary from night to night
and
even during one night. Also, Moon contribution is strongly
depending on the angular distance from target to Moon and on airmass. Therefore,
these
are the main factors affecting estimations of signal-to-noise.
Calculator
modes
We
describe here and in the
following sections the operation of the OSIRIS Calculator.
For
all modes, user must choose:
- Focal
station (Nasmyth or Cassegrain)
- CCD
readout mode
- Binning
- Type of
source (point-like or extended)
- Airmass
- Seeing
FWHM (arcsec)
- Moon
(bright, grey or dark)
- Fraction
of flux (if source type is point-like)
in %
- Number of
exposures
- Exposure
time for a single exposure (s)
Fraction of flux refers to the percentage of the total flux of the
source to be considered for SNR calculations. Here a Gaussian point
spread function is assumed for point sources. It is related to the
radius of the aperture as,
where FWHM is the seeing and EF is the fraction of flux.
For extended sources this parameter is irrelevant.
Time
to saturation is also
computed assuming that the maximum counts per pixel is 50,000. For
point-like
sources it is assumed that source is centred at one pixel, therefore,
the saturation time computed
is an upper limit.
When
the tool starts, a window with three independent tabs is open. Each tab
corresponds to an observing mode. All parameters are initialized at a
default value, so check all parameters to meet your requirements. By
clicking on 'Compute'
the
calculations are done and the output is shown at the right panel. If
some error or warning exists, a red message will appear below that
button. If all is OK, the message will be '...done'. You can
change from panel to panel as many times as you need, but take into
account that parameters ARE NOT TRANSFERRED between them.
The tool can be stopped by clicking 'Quit' at any of
the panels.
It is possible to send the output to a printer by clicking button 'Print'. For the spectroscopic
mode there are two queries to the printer: one for the text
output panel and the other for the SNR/pixel plot.
The output is different from mode to mode, but there is some
information that is common to all modes:
- Focal station
- Type of source
- Readout mode and ADU
- Moon
- Binning
- Encircled energy or fraction of flux
- Band information
- Total exposure time and total time (i.e. including
readouts ONLY). .
The
default readout mode is Spectroscopy with a readout speed of
100kHz. User should select a grims from the list.
Characteristics of grims are listed in the following table:
|
Grism
|
lc
(Å)
|
Spectral
range
(Å)
|
Dispersion
(Å/mm)
|
Dispersion
(Å/pixel)
|
Resolution
(dl) with slit =0.6”
|
Peak
eff.
|
|
R300B
|
4420
|
3500
– 7000
|
188.7
|
2.83
|
13.6
|
0.64
|
|
R300R
|
6650
|
5000
– 10000
|
282.7
|
4.24
|
20.3
|
0.42
|
|
R500B
|
4730
|
3500
– 7000
|
130.7
|
1.96
|
9.44
|
0.53
|
|
R500R
|
7170
|
5000
– 10000
|
171.3
|
2.57
|
12.3
|
0.36
|
|
R1000B
|
5450
|
3700
– 7580
|
76.7
|
1.15
|
5.51
|
0.54
|
|
R1000R
|
7470
|
5300
– 10000
|
92.7
|
1.39
|
6.65
|
0.41
|
|
R2000B
|
4750
|
4000
– 5580
|
31.4
|
0.47
|
2.26
|
0.66
|
|
R2500U
|
3990
|
3500
– 4560
|
22.1
|
0.33
|
1.58
|
0.81
|
|
R2500V
|
5180
|
4480
– 5960
|
28.7
|
0.43
|
2.06
|
0.76
|
|
R2500R
|
6560
|
5640
– 7550
|
36.4
|
0.55
|
2.64
|
0.53
|
|
R2500I
|
8650
|
7420
– 9980
|
47.9
|
0.72
|
3.46
|
0.35
|
The efficiency curves used for
all the grisms are those measured with real calibration stars and
updated - July 2011.
The input includes the slit width (in
arcsec), the
binning, and the size along slit. For the binning there are four
options:
- 1x1 standard pixel size
- 1x2 binning along dispersion
- 2x1 binning along spatial direction
(slit)
- 2x2 binning along both directions
The
size along the slit, given in units of the seeing, determines the
number of pixels
along the slit to compute signal from source, sky background, readout
noise and
dark current. For point sources, seeing profile is assumed to be
Gaussian. For extended sources this parameter does not apply and
computations are made per pixel along slit. In this mode there are two
options:
- Target is a pure continuum source. In this case, the
spectrum is
assumed to be of constant magnitude trough the spectral range of the
grism and magnitude is in the AB system.
- Target is a continuum + emission line source. For the
continuum,
assumptions are the same as in case (1). For the emission line, user
can define the central wavelength, FWHM and total intensity of the line
(erg/s/cm**2). A Gaussian profile is assumed for the emission line.
Output:
- There are two output panels.
- Central panel is a plot of the signal-to-noise ratio per
pixel along spectral direction.
- The
right panel includes information about grism, pixel size,
effective resolution, fraction of total flux that enters the slit (if
the source is point-like), and total exposure and observing
times.
- For extended sources the plot corresponds to one pixel
along slit.
- Note that noise depends on the sky background, therefore in
the SNR
plot sky emission and absortion features will appear.
The default readout mode is Imaging with a readout speed of
200 kHz.
First, the TF should be selected. At
this moment
(February 2010) blue TF is not available so that a message will show
up if this is selected. Desired TF wavelength and FWHM should be
selected. The tool checks the possible FWHMs for the given wavelength
and chooses the nearest one. Then, the order sorter filter is selected
automatically from the list of available filters. If no filter exists,
a message will appear below the 'Calculate' button. Use the
TF Setup Tool to obtain the
possible values of the FWHM for a given wavelength.
As in spectroscopy, user has two
options:
- Target is a pure continuum source. In this case, the
spectrum
is assumed to be of constant magnitude trough the spectral range of the
TF+order sorter filter and magnitude is in the AB system.
- Target is a continuum
+ emission line source. For the continuum, assumptions are the same as
in case (1). For the emission line, user can define the central
wavelength, FWHM and total intensity of the line (erg/s/cm**2). A
Gaussian profile is assumed for the emission line.
Output:
- User should look carefully at
the output window to check that central wavelength and FWHM of the TF
meet his/her requirements. Some information related to the TF is shown
in the output panel. The Free Spectral Range (FSR) is the size of the
spectral range (centred at the tunned wavelength) where only one
interference order is transmitted through the etalon. Note that no order sorter filter will be
selected if theFSR is smaller
than the width of any of the filters available, and then, no
calculations will be made.
- Total efficiency of the system at the selected wavelength
is shown.
- The combined effect of atmosphere extincion and
absorption is computed and shown (in %), i.e. a high value means a high
extinction and absorption. Check this number and use the TF Setup Tool to select a
wavelength free of strong telluric lines.
- If a strong sky emission line is +/- 5 Å from the
central wavelength, a message will appear, showing the wavelength of
this line. Check that and use the TF Setup Tool to select
a wavelength free of strong sky emission lines.
In this mode, signal-to-noise ratios are separated
in
continuum,
line and
total.
- Continuum
SNR
takes into account the
continuum of the source as the signal only, and the noise is computed
from the source, sky background and detector noise.
- Line
SNR
takes into account the emission line as the signal only, and
the
noise is computed from the source emission line, source continuum, sky
background and detector noise.
- Total SNR
takes into account both continuum and emission line as the signal.
The remaining of the output is similar to that of the
Broad-Band mode.
Final
comments
The author (J. Ignacio
González-Serrano) is
the only responsible for this tool. An effort has been made to make a
realistic simulation of the OSIRIS instrument although some effects are
very difficult (or even impossible) to predict accurately.
Please contact the
author
for comments, suggestions, complaints, and
errors or inconsistencies found; they are all welcome.