76
7 The Astronomical Detector
Bell Labs. However, the fabrication of CCDs was considerably easier than that of
CMOS chips, given the microprocessor manufacturing techniques used at the time.
Consequently, CCDs received a considerable development advantage. Improvements
in the production of very small scale circuitry and the large demand for consumer
CMOS chips (most modern phone cameras are CMOS, as are most DLSR cameras)
has resulted in the advantages between CCD and CMOS for astronomical imaging
becoming increasingly similar. At the time of writing the second edition of this book,
there was a gradual move by amateur astronomers away from CCD images towards
CMOS, mostly because of the lower cost of CMOS compared to CCD cameras of the
same size. Universities tend to spend more on imaging cameras but replace them less
often, so it is likely, if you are using a university-based telescope or even a research
instrument, that it will still be CCD based.
From a practical point of view, there is little difference between CCD and CMOS
cameras. They use identical software and the same basic technology.
Both CMOS and CCD cameras use an array of light-sensitive cells, or pixels. You
can consider each pixel to be a well with a series of balls (or in this case, electrons)
around the rim. When a photon hits the well, an electron is knocked into it. At the end
of the exposure, the number of electrons in each well is proportional to the number of
photons that hit that pixel, with the number of hits each pixel received being known
as its count.
Looking at the literature supplied with an astronomical camera or a manufacturer’s
website, you will see a large number of facts, functions, and figures associated with
a camera. For example the frame transfer rate and whether it is backlit or frontlit. In
general, this information will be at most interesting. However, there are a number of
pieces of information you do need.
The array size is the size of the CCD or CMOS in pixels. The array size is not
the physical size, as pixel sizes vary. In general, bigger is better. However, very large
CCDs or CMOSs might not be fully illuminated by the telescope, so your camera
should match your telescope.
The pixel size is the physical size of each pixel, which in most cases are square.
Pixel sizes are typically in microns. You can use the array size and pixel size with
the telescope’s focal length to find the image scale, as discussed in Sect. 3.1.
The quantum efficiency (QE), as mentioned earlier, is the measurement of how
effective the chip is at capturing photons. The QE is wavelength-specific and hence
will change depending on the object being observed. For chips designed for optical
observing, the peak QE is around 500–600 nm, with a very sharp dropoff in the blue
end and a less sharp but significant dropoff in QE at the red end. Hence, you may
find that some objects need longer exposures than others if their peak emission is
outside the peak QE wavelength. Furthermore, QE is very sensitive to temperature,
and you will not be able to achieve the expected sensitivity if the CCD or CMOS is
running hotter than specified, one of the reasons astronomical detectors are cooled.
Additionally, the published QE for a camera is unlikely to be achieved by every pixel
in the array. This variation in QE can be a significant source of error, and it can be
dealt with for the most part by a process known as flat fielding, which we will discuss
at length later in this chapter.
7 The Astronomical Detector
Bell Labs. However, the fabrication of CCDs was considerably easier than that of
CMOS chips, given the microprocessor manufacturing techniques used at the time.
Consequently, CCDs received a considerable development advantage. Improvements
in the production of very small scale circuitry and the large demand for consumer
CMOS chips (most modern phone cameras are CMOS, as are most DLSR cameras)
has resulted in the advantages between CCD and CMOS for astronomical imaging
becoming increasingly similar. At the time of writing the second edition of this book,
there was a gradual move by amateur astronomers away from CCD images towards
CMOS, mostly because of the lower cost of CMOS compared to CCD cameras of the
same size. Universities tend to spend more on imaging cameras but replace them less
often, so it is likely, if you are using a university-based telescope or even a research
instrument, that it will still be CCD based.
From a practical point of view, there is little difference between CCD and CMOS
cameras. They use identical software and the same basic technology.
Both CMOS and CCD cameras use an array of light-sensitive cells, or pixels. You
can consider each pixel to be a well with a series of balls (or in this case, electrons)
around the rim. When a photon hits the well, an electron is knocked into it. At the end
of the exposure, the number of electrons in each well is proportional to the number of
photons that hit that pixel, with the number of hits each pixel received being known
as its count.
Looking at the literature supplied with an astronomical camera or a manufacturer’s
website, you will see a large number of facts, functions, and figures associated with
a camera. For example the frame transfer rate and whether it is backlit or frontlit. In
general, this information will be at most interesting. However, there are a number of
pieces of information you do need.
The array size is the size of the CCD or CMOS in pixels. The array size is not
the physical size, as pixel sizes vary. In general, bigger is better. However, very large
CCDs or CMOSs might not be fully illuminated by the telescope, so your camera
should match your telescope.
The pixel size is the physical size of each pixel, which in most cases are square.
Pixel sizes are typically in microns. You can use the array size and pixel size with
the telescope’s focal length to find the image scale, as discussed in Sect. 3.1.
The quantum efficiency (QE), as mentioned earlier, is the measurement of how
effective the chip is at capturing photons. The QE is wavelength-specific and hence
will change depending on the object being observed. For chips designed for optical
observing, the peak QE is around 500–600 nm, with a very sharp dropoff in the blue
end and a less sharp but significant dropoff in QE at the red end. Hence, you may
find that some objects need longer exposures than others if their peak emission is
outside the peak QE wavelength. Furthermore, QE is very sensitive to temperature,
and you will not be able to achieve the expected sensitivity if the CCD or CMOS is
running hotter than specified, one of the reasons astronomical detectors are cooled.
Additionally, the published QE for a camera is unlikely to be achieved by every pixel
in the array. This variation in QE can be a significant source of error, and it can be
dealt with for the most part by a process known as flat fielding, which we will discuss
at length later in this chapter.