The draft SCOC recommendation are for 1x29.2s exposures for if single snaps are implemented. How precise of control is there are LSSTCam exposures? Does LSSTCam have subsecond exposure time precision?
Hi Meg!
Sorry for the delay in a response! It looks like your message flew under the radar for a bit. Sorry!
I’ve hunted around the documentation and I don’t yet have a definitive answer, but here are some things I have found:
- In the LSST System Science Requirements Document (LPM-17), Section 3.3.2.4, the shortest exposure time will be no greater than:
- 5 sec (design spec)
- 10 sec (minimum spec)
- 1 sec (stretch goal)
-
In the Observatory System Specifications (OSS) Document (LSE-030), specification OSS-REQ-0335 states that “[s]hutter timing errors shall result in an RSS contribution of not more than [0.4 millimags] to the repeatability of magnitudes for bright isolated point sources,” and specification OSS-REQ-0291 states that "[t]camera shall be able to obtain a single exposure with an effective minimum exposure time of no more than [1 sec], with a goal of an effective minimum exposure time of [0.1 sec]. (This latter spec seems to contradict the values from the LSST System Science Requirements Document above (last revised January 30, 2018), but LSE-030 is a more recent document (last revised on May 8, 2024).
-
From a presentation I found describing the requirements on the shutter for the LSST Commissioning Camera (“ComCam”), I found the following:
- Minimum exposure time: < 1 ms
- Exposure time error: <300 ÎĽs
- Exposure inhomogeneity: ±200 μs over the full field view.
So, I have yet to find an explicit shutter timing requirements for the LSSTcam as measured in seconds, I suspect it would be no worse than, say, 10x that of the smaller ComCom, or < 10*300 μs (< 3 ms), which, if I remember correctly, is comparable to the shutter timing requirements for the Blanco 4m’s DECam imager.
I will continue searching, but wanted to let you know what I found so far.
Thanks!
Hi Meg,
Thanks for bringing this up on slack at Lsst@Europe. I asked @reil directly — I was pretty sure the answer is yes, but figured I would ask him.
He said he could do 0.1s exposures no problem with the camera. So yes we can do 29.2: exposures … whether Zeljko has a good point on “that’s a funny number to use”, but there are reasons and it is technically possible for the camera.
Now, will it be exactly 0.1s instead of 0.15s?
That aspect of “control” I didn’t ask, but there are even tighter constraints on measuring what the actual shutter speed and exposure time is (I’ll see if I can look up the requirement specification, it’s in the camera requirements).
I would like to know what is the maximum exposure time with the full moon at the center of the frame, like the example image for the LSST field of view, just not to overexpose the sensors. Would 3ms be ok? And with a thin crescent or a lunar eclipse? And with Sirius at the center of the frame?
Thanks!
Hi Meg,
I just got additional info from @reil and @arasmus. This is what I learned:
From @reil, the LSST Shutter Verification Matrix doument (LCA-16648) states:
-
Requirement C-SHUT-002 Exposure duration accuracy The accuracy of the shutter exposure duration shall be < 50 milli-seconds Applies for all shutter orientations per C-SHUT-036 Test Exposure Trajectory Test
-
Requirement C-SHUT-066 Exposure duration knowledge The exposure duration shall be known to within 300 msec with a goal of knowing the duration to 5 msec, rms. This provides the required knowledge of 2% rms for the standard exposure of 15 seconds Test Exposure Trajectory Test
-
Requirement C-SHUT-006 Min exposure time capability The shutter shall be able to obtain a single exposure with an effective minimum exposure time of no more than 1 second, with a goal of an effective minimum exposure time of 0.1 seconds Test Exposure Trajectory Test
From @arasmus, “Until the shutter belts stretch significantly, we can use a linear combination of the measured start & stop times and it’s within about 1ms of a more careful analysis.”
1 Like
Hi Renato!
Interesting question!
Very roughly, the LSSTcam is expected to saturate for a point source at around mag=16 in a 15sec exposure.
The Full Moon is about mag=-12, but it covers an area of sky of about pi x 0.25^2 square degrees, which is about 0.2 square degrees or about 2.5e6 square arcsec. The LSSTcam pixel scale is 0.2 arcsec/pixel, so the full moon would cover about 2.5e6 square arcsec/(0.2arcsec x 0.2arcsec) = 6.4e7 LSSTcam pixels.
So, the average surface brightness of the Full Moon is about -12 + 2.5*log10(6.4e7) = +7.5 mag per LSSTcam pixel. (Admittedly, the Full Moon will be brighter in some pixels compared with others, but this is a very back-of-the-envelope calculation. We are not even considering the size of the PSF or other effects here!).
A +7.5 mag object is 10^(-0.4(7.5-16.0)) = 2500 times as bright as at 16th mag object. Thus, a +7.5 mag object should saturate the LSSTcam detectors in 1/2500th the time of a 16th mag object, or 15sec/2500. = 0.006 sec or about 6 milliseconds.
So, we probably do NOT want to try to image the Full Moon with LSSTcam! 
This would probably be roughly the same for a thin crescent Moon, since it is the brightness per LSSTcam pixel of the illuminated portion of the Moon that is important (and should be roughly the same as the Full Moon’s brightness per LSSTcam pixel), not the crescent Moon’s total brightness.
Hard to say for a lunar total eclipse – they are so variable in brightness.
Sirius is a point source with a brightness of roughly mag=-1. A mag=-1 object is 10^(-0.4(-1-16)) = 6.3e6 times as bright as at 16th mag object. Thus, a mag=-1 point source should saturate the LSSTcam detectors in 1/6.3e6 the time of a 16th mag object, or 15sec/6.3e6 = 2.4 micro-seconds.