Get rid of RBI phenomena

ddemartin

For many type of observations (if not always), it is important to get an image as "clean" and without artifacts as possible.
I've often been cheated by RBI phenomena, i.e. ghost images left by previous observation, that persist on the sensor as faint blobby "nebular" features.
A way to get rid of these is chosing a readout with RBI flood. The sensor is "preflashed" before each image, resetting the base level of each pixel. The downside of this method is that each preflash takes time. If you do many short exposures, it is not a viable way to do, as the preflashes take more time than the actual observation. In addition, this makes the reduction more tricky.
A few weeks ago I and Lukas tested a method that add at the beginning of the observing run a single short exposure (1 or 2 secs) with RBI flood, followed by the actual exposures taken with the normal readout (with no RBI flood). Lukas made the trick software-side and I tested the results. Which are very interesting. With this method, I never got RBI phenomena in my images. In addition, for some reasons that I can't fully explain, I noticed that the quality of the frames is very stable, with a limiting magnitude very consistent along the session.
Thanks to Lukas for his kindness in supporting this test. Lukas has implemented this feature as an advanced setting.
Davide

lukas_demetz

Just to give an idea of a result (thanks David):

SkG

Is this a problem just for the Namibian cameras or does it affect the New Mexican and Spanish imagery as well?

lukas_demetz

@SkG Well, it is a "problem" of the 16803 CCD's, basically. Now, with wideband imaging one would not really notice it. RBI can become disturbing when one has little signal (narrowband) and is imaging something dim right after the sensor took bright images. This solution is helpful for photometry as it starts with a known level right before the first exposure. We did not test it for pretty picture imaging, though.

SkG

Does the appearance or level of RBI depend at all on the readout frequency (speed)?

lukas_demetz

Nope. The residual image depends mostly on the object which has been imaged prior to the current frame, and the temperature of the CCD (lower temperature = longer persistance of the ghost image)