170
12 High-Precision Differential Photometry
All four of these factors contribute to the overall uncertainties in your photometric
observations. Additionally, there are other uncertainties such as the variability in the
check stars and atmospheric conditions such as high cirrus.
12.2 Calibration Noise
One of the most controllable forms of noise is in the calibration of the images. The
most obvious way to start is to ensure that any bias, dark, and flat-field frames you
are using are as up to date as possible. Bias frames are the most stable over time.
Dark frames are relatively stable, but they need to be taken regularly and whenever
there is an equipment change in the observatory that might cause additional noise,
for example any cable change or a power supply change. Flat-fields should be done
every day if possible. Historically, I have found that super dome flats produced using
a flat-field luminescent panel tend to be more stable than others. You should always
use super bias, flat, and dark calibration frames when processing science frames for
high-precision photometry. These are produced by taking the mean of several (at
least ten) frames. Recall that dark and flat field frames are temperature specific and
that flat frames are also filtered variant. Also, be aware that on a warm day, camera
coolers work far harder than in the evening and can produce some additional read
noise, so it is best to take dark and flat fields after sunset.
Dark current is proportional to both sensor temperature and integrated exposure
time. You should try to cool the sensor to the lowest temperature that your imager
cooling system can comfortably cope with. It is common to use scaled darks for
imaging. This means that if you don’t have a 60 s dark for your exposure but have
a 30 s dark, you subtract the 30 s dark from your image twice. This works, and for
imaging it is a perfectly acceptable practice. It is also acceptable for photometry, but
when you are trying to reduce your uncertainties to the limit, it is best to have the
correct dark frame.
The total noise contribution from the calibration frames is the square root of the
sum of the squares of the standard deviations of the bias, dark, and flat-field frames.
12.3 Bias and Dark Noise
Dark and bias frames are used to remove noise from the read process and thermal
electrons from a science frame. However, the dark noise is effectively random, as
is to some extent the bias noise. Neither calibration frame can remove all the noise,
and in fact, in a low noise calibrated frame, calibration frames might introduce noise.
Hence, the noise contribution by dark and bias frames is not the standard deviation
with the calibration frame but the standard deviation pixel to pixel between multiple
calibration frames.
12 High-Precision Differential Photometry
All four of these factors contribute to the overall uncertainties in your photometric
observations. Additionally, there are other uncertainties such as the variability in the
check stars and atmospheric conditions such as high cirrus.
12.2 Calibration Noise
One of the most controllable forms of noise is in the calibration of the images. The
most obvious way to start is to ensure that any bias, dark, and flat-field frames you
are using are as up to date as possible. Bias frames are the most stable over time.
Dark frames are relatively stable, but they need to be taken regularly and whenever
there is an equipment change in the observatory that might cause additional noise,
for example any cable change or a power supply change. Flat-fields should be done
every day if possible. Historically, I have found that super dome flats produced using
a flat-field luminescent panel tend to be more stable than others. You should always
use super bias, flat, and dark calibration frames when processing science frames for
high-precision photometry. These are produced by taking the mean of several (at
least ten) frames. Recall that dark and flat field frames are temperature specific and
that flat frames are also filtered variant. Also, be aware that on a warm day, camera
coolers work far harder than in the evening and can produce some additional read
noise, so it is best to take dark and flat fields after sunset.
Dark current is proportional to both sensor temperature and integrated exposure
time. You should try to cool the sensor to the lowest temperature that your imager
cooling system can comfortably cope with. It is common to use scaled darks for
imaging. This means that if you don’t have a 60 s dark for your exposure but have
a 30 s dark, you subtract the 30 s dark from your image twice. This works, and for
imaging it is a perfectly acceptable practice. It is also acceptable for photometry, but
when you are trying to reduce your uncertainties to the limit, it is best to have the
correct dark frame.
The total noise contribution from the calibration frames is the square root of the
sum of the squares of the standard deviations of the bias, dark, and flat-field frames.
12.3 Bias and Dark Noise
Dark and bias frames are used to remove noise from the read process and thermal
electrons from a science frame. However, the dark noise is effectively random, as
is to some extent the bias noise. Neither calibration frame can remove all the noise,
and in fact, in a low noise calibrated frame, calibration frames might introduce noise.
Hence, the noise contribution by dark and bias frames is not the standard deviation
with the calibration frame but the standard deviation pixel to pixel between multiple
calibration frames.