OSIRIS SNR Calculator Help

Updated: September 2011

  1. Introduction
  2. Telescope and CCD parameters
  3. Throughput
  4. Atmosphere
  5. Sky background
  6. Calculator modes
  7. Broad-band imaging
  8. Spectroscopy
  9. Tunable filter imaging
  10. Final comments

Introduction

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:

GTC+OSIRIS throughput


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:

Atmospheric transmission




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.


Sky emission

 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:

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,


formula



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:

 


Broad-band imaging

      The default readout mode is Imaging with a readout speed of  200 kHz. There are three options for this observing mode:

  1. Broad-band imaging with the standard filters u'g'r'i'z'. 
  2. Intermediate-band imaging with the order sorter filters. Note that the blue order sorter filters are not available (September 2009).
  3. User-provided filter. In this mode the user should specify filter characteristcs: central wavelength, width, and transmission.
    In case (1) the input target magnitude corresponds to the selected filter. In cases (2) and (3) AB magnitudes are used. In all cases, target magnitude is assumed to be constant trough the filter bandwidth.

    Output:  

Spectroscopy

    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:

     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:

  1. 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.
  2. 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:




Tunable filter imaging

      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:

  1. 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.
  2. 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:


   In this mode, signal-to-noise ratios are separated in continuum, line and total.

  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.