CCD Pixel Saturation in Light Curves

American Association of Variable Star Observers (AAVSO)
Fri, 07/05/2019 - 21:14



I am trying to analyze a variable star candidate (likely an EA) but am having difficulties with a set of observations I obtained from a telescope. Since I did not personally conduct the observation of the star and do not (think I) have access to the original images, I do not have a full grasp of the specifics of the telescope operation; however, I do know that the telescope in question is typically used for very faint objects (~magnitude 13-17) and is not accustomed to observing stars as bright as my candidate star. I am concerned that this data set contains a number of poor observations caused by CCD pixel saturation and/or CCD overexposure, specifically, large amounts of scattered observations around the maximum magnitude (around 10.8). The reason why I suggest this possibility at all is beacuse I have three other data sets (NSVS, SuperWASP, and ASAS-SN) which all look noticably cleaner and clearer than my fourth data set. Is there any way I can detect CCD pixel saturation by looking directly at the light curves of the observations and comparing them to "more accurate" light curves? As in, is CCD saturation apparent in a light curve when the suspect light curve is compared to a reference light curve?


Many thanks,



pixel saturation

Hi Alex,

As you mention, if you have access to the images, saturation is pretty easy to see.  Normally you do a radial profile plot, which typically shows a Gaussian shape on unsaturated stars.  When they get saturated, the profile goes flat-topped.

Getting saturation just from light curves is a bit harder.  In general, saturation will make the star appear fainter, so if you see other observers continue to rise and your data points trend flat, that is a good indicator.  The other indicator is a very low reported uncertainty, when several data points in a row show much larger variation.  The amount of saturation depends on seeing, which tends to be variable, and so the measurements on subsequent images may be bright or faint depending on the saturation.  They will all report to have ~0.0001mag uncertainties, though!

If you give the specifics of the telescope system: telescope aperture, focal length, camera model, filter, exposure, etc., there are people on the forum who will have similar instruments and can tell you the approximate saturation magnitude for the image.

You might try getting the original images from the observer.  I've found very few that aren't willing to share data!



American Association of Variable Star Observers (AAVSO)
Thank you very much for this

Thank you very much for this advice.


My first instinct was to check the images directly, but I must first find them in my professor's archives (the images are his and have already been processed into raw data). In any case, the maximum brightness of my candidate star from my fourth data set seems to be slightly dimmer than what is recorded in the other three data sets, which would make sense if some pixels are saturated.


I will find the images and examine them for saturation. I could easily obtain the specifications of the telescope, but I do not think many amateur observers would have such a large telescope. Besides, using the actual images would be much more direct.


Once again, thank you for your advice,