This is a very particular question, I realize, but I am still curious...

The Vera Rubin telescope is supposed to take a 15-second exposure, then take 5 seconds to move to its next spot...

But I also read that it will take two consecutive 15-second exposures at each stopping point, then spend 5 seconds moving to its next position...

On Wikipedia, in the middle of the article on the Vera C. Rubin LSST Telescope, two consecutive paragraphs contradict each other, the first saying it will spend only 15 seconds at each spot before moving on, and the second saying it will take two consecutive 15- second exposures at each point.....


1 Answer 1


How the Vera C. Rubin telescope will physically carry out the actual survey of the sky (the Legacy Survey of Space and Time; LSST) is still subject to evolution and refinement. The broad survey strategy is set out in the LSST System Science Requirements Document which is (intentionally) relaxed on how it is actually carried out. The main requirements are:

  • The main survey, of at least 18,000 square degrees must be uniformly covered do that each field of the camera's 9.6 square degree field of view gets an average of 825 30s visits over the 10 years of the survey. (The 825 visits are totalled over all 6 filters)
  • Rapid revisits (anywhere from 40 seconds to 30 minutes) must be acquired over at least 2000 square degrees to discover fast transients. (This can mostly be done by overlapping the fields of view of adjacent fields slightly

Other science requirements such as needing to detect moving objects (one of the four main science goals) and being able to reach the stated parallax and proper motion goals also impose some constraints on how often and when you need to revisit a field.

These constraints apply to the main "wide-fast-deep" (WFD) survey which is what will cover the 18,000 $\deg^2$ area above. Within the main WFD survey area, the baseline plan is make 2 visits per camera field (9.6 $\deg^2$) in either the same or different filters each night (This allows the moving objects and rapid transients to be found and alerts sent out). This pair of visits is repeated every 3-4 nights during the time the field is visible. Each visit consists of 2 back-to-back 15s exposures in the same filter, separated by the readout time of 2 seconds. These two 15s exposures are then combined in the pipeline to make the final 30s image.

Exactly how the WFD field visits are strung together, how far the telescope moves between fields and in what filters it observes and when, and what the observatory spends the other ~10% of time on, is a very complicated problem and depends a huge amount on what objects you want to be sensitive to and what type of science you want to do. These various tradeoffs are still ongoing as part of the Survey Cadence and Optimization Committee. The report of the large number of survey cadence simulations has recently come out but a final survey strategy will not be decided on until April 2023 (timeline summary) after more meetings, reports and simulations. Also, while 1x30s exposure at each visit (rather than 2x15s which has higher overhead) and a shorter time spent on each field is preferred (it increases the survey throughput, letting you cover more fields/area per night), this will depend on the final LSST Camera performance. This is because the susceptibility of the cameras' CCDs to cosmic rays, and whether you need the 2x15s exposures to remove them, is still partly unknown until it arrives at the telescope and has been operating for some time.

The simulations are based on a telescope model, whose code is available on github. The time taken to go from field to field depends on a quite a few things such as how far the telescope has to move (as there is both acceleration and deceleration limits as well as optics settling time depending on how far it moved), how much the dome has to move and whether you are changing filters (a 120s operation). In general, the total time spent on a field is going to be: $$ t_{slew}+t_{settle}+t_{exp}\times num. exp. + t_{shutter}\times num. exp. + t_{readout }\times (num. exp.-1) $$ where $t_{slew}$ is the slew time (depends how far you move), $t_{settle}$ is settling time (also depends how far you move but minimum of 3s), $t_{exp}$ is exposure time, $t_{shutter}$ is shutter open/close time (1s; it's really big...) and $t_{readout}$ is the readout time (2s; the $(num. exp.-1)$ comes about because you can slew to the next field while reading out the last exposure).

So for 2x15s exposures, we have minimum of $3s+(15s\times2)+(1s\times2)+(2s\times1)=37s$ between fields. For the 1x30s, we have minimum of $3s+(30s\times1)+(1s\times1)+(2s\times0)=34s$ between fields. Exactly how often the telescope will be moving about to a new field will depend on the survey cadence decided on as outlined above, but there will be a strong preference to keep the slew times short and the number of filter changes small to maximize the amount of fields and sky covered per night.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .