Planning an observation

The exposure-time calculator is the first scientific tool in the normal workflow. It predicts the signal-to-noise ratio in user-selected wavelength bins for a source spectrum and instrument configuration, or calculates the AB magnitude that reaches a requested S/N in each bin for a fixed exposure time.

Inputs

The core calculation accepts a reference or target spectrum whose wavelengths are in the observer frame, together with:

  • exposure time;

  • observer-frame wavelength-bin centers and bin size;

  • sky background (dark, grey, or bright);

  • detector/camera model;

  • square detector binning;

  • grating choice;

  • numerical airmass;

  • fiber length;

  • fiber-coupling efficiency; and

  • either a target AB magnitude and LSST band for a forward S/N calculation, or a target S/N and LSST band for a limiting-magnitude calculation.

The ETC supports scaling a template spectrum to LSST g, r, or i photometry in the AB system. This is useful when the spectral shape is known or assumed but only broadband target photometry is available. For a limiting-magnitude calculation, the template still supplies the spectral shape, but its absolute normalization does not affect the result since the limiting observable magnitudes are the output.

The ETC does not apply a redshift correction. If a template is in the rest frame, transform it to the observer frame before supplying it to the ETC. The requested wavelength-bin centers and bin size must likewise describe the observer frame.

Fiber-coupling efficiency is a fraction from 0 to 1 representing point-source light lost before entering the fiber. It reduces source counts but not sky counts. The default of 1.0 assumes perfect coupling.

Detector binning is selected separately from the camera model. It defaults to 2×2 for Aurora and 1×1 for Kepler and QHY268. The selected factor must be a positive integer that evenly divides both native detector dimensions. Effective detector sampling, fiber pitch, extraction fraction, read noise, telescope collecting area, and the component throughput model are derived from the selected camera, binning, and shared classi-sim instrument model rather than entered as independent GUI parameters. The default instrument configuration is the FLI Aurora AR571 camera with 2×2 binning and the Newport 1294 grating.

Throughput accounting

The calculation keeps the major throughput terms separate before multiplying them into the total response. This makes it possible to inspect detector, grating, fiber, atmosphere, and lens contributions individually and to disable a term for diagnostic comparisons. The fiber term uses the CeramOptec UVNS attenuation profile distributed by classi-shared-data.

The atmosphere term uses the LSST v1.9 standard-atmosphere transmission profile distributed by classi-shared-data. It includes aerosol, telluric, and line-absorption structure; the nominal airmass-1.0 throughput is raised to the selected numerical airmass. ETC predictions near atmospheric features inherit the profile’s sampling and should be treated as planning estimates rather than an observation-specific telluric correction.

The dark, grey, and bright choices select line-resolved DESI sky spectra. Each sky spectrum is integrated on its own finely sampled wavelength grid over the fiber’s circular on-sky area, preserving narrow airglow lines. Because these spectra represent surface brightness at the observatory, atmospheric extinction is not applied to them again.

The spatial extraction box is one fiber pitch wide, extending halfway toward the centerline of each neighboring trace. Source and sky counts are multiplied by the fraction of the assumed Gaussian fiber profile enclosed by that box. Fiber coupling is then applied only to the source. Dark-current and read-noise variance use the same extraction-box pixel count. The detector temperature is not a user input: the ETC uses each camera’s native-pixel dark-current value at -20 °C and records that assumption in the result metadata. For square binning factor b, the simulator readout model scales dark current per output pixel by b**2 and read noise per output pixel by b.

Source and sky spectra are linearly interpolated at the exact requested bin boundaries before integration, so an input spectrum must cover every complete bin. The spectral width of each bin is computed from the nonlinear readout_spectrograph.wavelength_to_x() mapping for the selected camera and binning rather than from binsize / dispersion. Consequently, the spectral and total pixel counts and their read-noise and dark-current contributions are reported separately on each bin result.

The reported result for each wavelength bin also includes source counts, sky counts, S/N, and mean component throughputs.