The throughput curves in this directory should be considered 'baseline' for the current modeled behavior of Rubin. An X=1.2 atmosphere has been used to compile the 'total' throughputs and average losses over time have been included.
Note that these throughput curves are subject to change as our knowledge of the Rubin systems improve.
m1.dat, m2.dat, m3.dat represent the current mirror throughputs lens1.dat, lens2.dat, lens3.dat represent the current lens throughputs detector.dat is the current detector sensitivity
filter_u / g / r / i / z / y. dat represent the current filter (filter only!) throughput curves.
atmos_std.dat is a fiducial atmosphere throughput at Rubin at 1.2 airmasses.
total_*.dat throughput curves represent the combination of all components in the Rubin system - mirrors, lenses, filter, detector, and the zenith atmos_std.dat atmosphere.
(hardware_*.dat curves are the hardware only, without atmospheric component).
All curves are in nanometers, with throughput represented by a number between 0 and 1.
The source of this information is syseng_throughputs
SysEng-approved LSST throughput curves The latest m5 depths are available in the notebooks, such as in notebooks/Overview Paper.ipynb.
This repository provides the ultimate source of the throughput curves in the repository lsst/throughputs.
The components directory contains the response curves
for each individual component of the camera and telescope. In each
directory, there is also a *_Losses directory that contains the
time-averaged ten-year losses due to contamination or condensation on
the surfaces of the component. In some directories, there is also a
*_Coatings directory, which contains information on coatings applied
to the surface, such as the Broad Band Anti-Reflection coatings on the
lenses.
These components curves are maintained and updated by the LSST system engineering team.
Python utilties to read and combine these various curves appropriately are maintained in this repository, in the syseng_throughputs directory. In particular, note the utilities provided in bandpassUtils.py. At this time, we expect most users to use the throughputs repository instead of this repository directly - the curves in the throughputs repository are constructed from these curves, and can be traced through the git SHA1 and release tags.
The python code requires rubin_sim to run.
After installation of rubin_sim, install syseng_throughputs into the same python environment using
pip install -e .
This updates uses the same throughput components are previous versions, but moves from Al-Ag-Al mirror coatings to Ag-Ag-Ag (triple silver, or 3Ag) mirror coatings. This results in increased throughput in redder bands, at the cost of lower throughput in u band. However, since more survey time is spent in r and i bands (and redder bands in general) than u, the overall impact on survey efficiency is positive.
This update includes as-measured filter throughput curves and as-measured glass and coating measurements for the lenses. The mirror reflectivities are also all updated to as-measured curves for Al and Ag, but the mirror coatings are assumed to remain Al-Ag-Al in this tag.
The M2 reflectivity was updated based on witness sample measurements from M2 coating run in July 2019. The PR for this update is lsst-pst/syseng_throughputs#12. The notebooks showing what has changed and the updated m5 calculations are found in the "documentation" subdirectory.
The mirror reflectivity was updated based on measurements from coating samples from June 2019. The PR for this update is lsst-pst/syseng_throughputs#11. The notebooks showing what has changed and the updated m5 calculations are found in the "documentation" subdirectory.
This is a minor update for throughputs (the lens2 glass and BBAR coating curves have been extended in their wavelength information, but the curves themselves are the same as previously). However it is a major update for documentation and process information, as reflected in the "documentation" subdirectory.
The primary update here is in the lens2 response curves. The BBAR coating has been updated.
Other minor updates include bug fixes in the python code in sedUtils.py, updating of the jupyter notebooks, and the addition of notebooks evaluating the effect of the mixed vendor detector focal plane and recreating the inputs for the LSST Overview Paper.
The primary update here is in the detector response curves. The QE response curves here are the result of measurements of multiple chips provided by each vendor, ITL and E2V. The measurements have been averaged across multiple CCDs; the default (single) 'generic' curve remains the minimum QE response at each wavelength between both vendors. These curves were provided by Steve Ritz in December, 2017.
Other minor updates include additional python code to allow scaling of the FWHM at different airmasses and wavelengths (according to details provided in Document-18208 and Document-20160), and a jupyter notebook which can provide latex-formatted content of Table 2 from the overview paper.
This is primarily an update to the python code in the repository, using corrected and updated readnoise values (which results in corresponding changes to m5, particularly in the u band).
- Detector: There are two separate detector response and loss curves, corresponding to the expected response (QE response + AR coatings) of the CCDs provided by each of the two vendors under consideration. For most purposes (including the detector curve reported in the throughputs repository), we use a 'generic' detector response that is generated by combining both of these throughput curves using the minimum QE response at each wavelength. The response curves from each vendor correpond to a response measured in LSST labs, using vendor-provided prototypes. The loss curves provided for each vendor represent a simulated effect of contamination buildup over time; the loss curves are identical for both vendors and are the average expected values over ten years. Note that some values in the 'contamination' loss file for the detectors are > 1; this is because the contamination is primarily a thin film of water, which at some wavelengths can enhance the performance of the AR coating on the detector -- this is only true for the detector.
- Lenses: There are three separate lenses in the camera, each with an
identical base
*_Glass.datcurve that represents the fused silica throughput of the lens itself. This throughput curve must be smoothed using the Savitzy-Golay smoothing function. The fused silica lens transmission curves are based on vendor-provided expected transmission curves. The silica base of the len must also be combined with the BroadBand AntiReflective (BBAR) coatings response in the*_Coatingsdirectory. There are two coatings; one for each side of the lens. The BBAR coating response is based on vendor-provided models, consistent with LSST requested coating requirements. There are small differences between the glass components used for each lens; there are also small differences in the BBARS, including a difference from one side of the lens to the other. In each lens, there are also several files in the*_Lossesdirectory, representing the time-averaged condensation and contamination losses for each surface of each lens. The losses are based on models developed by Andy Rasmussen at SLAC. These vary depending on the direction the lens is facing and the location of the lens in the camera. The final response curves for all lenses are similar in shape, however lens3 has a slightly higher overall throughput due to slightly lower losses (only by 1-2%). - Filters: For each filter, a goal throughput envelope has been
provided. This is the goal throughput envelope provided to the
filter vendors; tolerances on this envelope have also been
provided. Note that this is not the expected performance for an
as-manufactured filter, which would likely include some out-of-band throughput leaks
(within a specified limit), and represents a change compared to
previously provided throughput curves (which represented one simulation of
an expected as-provided filter set). In the
*_Lossesdirectory, there are also ten-year-average simulated contamination and condensation losses for each surface of the filters, based on models developed by Andy Rasmussen.
- Mirrors: Each mirror has a reflectivity curve, which should be
coupled with the respective losses curve found in the relevant
*_Lossesdirectory. The reflectivity of mirror1 (primary mirror) and mirror3 (tertiary) is based on using a protected aluminum surface; the reflectivity of mirror2 (secondary) is based on using a protected silver surface. These mirror reflectivities are based on lab measurements of pristine witness samples. The losses represent the ten-year average, based on performance degradation measurements from historical telescope performance, modified for the expected LSST maintenance schedule. Currently mirror cleanings are scheduled yearly, with resurfacing every two years.
- Atmosphere: The atmosphere throughput is modeled by using MODTRAN to produce a 'standard US Atmosphere', which does not include aerosols. To better represent the expected atmospheric transmission on site, aerosols have been added to the resulting throughput curves, using the python script addAerosols.py. The atmospheric transmission curves are in the siteProperties directory, with an X=1.2 and X=1.0 atmosphere, with and without aerosols. To represent 'typical' throughput, the X=1.2, with aerosols atmosphere curve should be used. To represent zenith, optimum throughputs, the X=1.0, with aerosols atmosphere curve should be used.
- Dark sky: The expected dark sky, zenith, background spectrum can be found in darksky.dat. This is used to calculate expected zenith, dark-sky limiting magnitude values. The dark sky SED is based on data from UVES and Gemini Near-IR, combined with ESO sky data from Ferdinand Patat, modified slightly at the red and blue ends to match observed dark sky broadband skybrightness values reported by SDSS.