Most professional photometrists recommend you get new flats each night.
Something may have changed since you made your last set such as some new dust donuts, a shift in the equipment that affects vignetting, and so on. If you have
the routine down, these don’t take too long and can be done before dark settles
it and it’s time to start observing. Many choose to build libraries of flats. For the most part, this is acceptable but keep in mind that things can and do
change. If you’re looking for the most accuracy, try to get a fresh set of
flats; if not every night then frequently.
There seems to be a difference of opinion about darks in this regard. There are many excellent CCD imagers who use cloudy nights to build libraries of darks. This is where they get a number of raw darks at various combinations of binning, temperature, and exposure length so that they can create master darks to be used later on. If you’re using scaled darks, meaning you need to get bias frames as well, you may want to get fresh darks with every run. There’s less time involved and so why not? I’ve found using scaled darks a little trickier than I’d like. I don’t use too many different combinations of temperature, exposure, or binning; for me, building a library is not quite as involved. There is something to be said for consistency and the easier path.
As for when you get the flats, if you’re doing twilight flats, you should definitely get them in the evening twilight, as the sky is getting darker, not lighter. That way you have a bit more control during the short window when the sky is neither too bright nor too dark. If you’re doing light box or dome flats, I think either before or after the run is fine. The same can be said for darks. One of the considerations that might control the situation is where your target is at the onset of dark and just before twilight. For example, consider the situation where you’re working a target this is well up when darkness sets in and it sets long before twilight starts. To shoot the flats and darks at the start of the run would take precious time away from the short observing window. In this case, shoot the darks and flats after the target dips below 30°.
Some observers get a dark and use that one dark for all their images for a run. It’s that or they’ll use a library dark. Sometimes, they’ll also merge in a flat from a library of flats as the images are taken and saved to disc. Don’t do this. You have far less control over what happens and, with today’s software, you’re not really saving any time. Get your flats and darks for the run and merge them with the raw images after the run but before measuring the images.
If the night is starting off clear and your window for shooting the target is
not overly restricted, get the images of the standard fields for the Hardie method right off. That way they are done and you’re more confident of the
conditions under which they were shot. The same consideration as mentioned above
regarding when to get darks and flats applies here, too. Namely, if the field is
already high up, you have a limited time to get images, it looks like conditions
will hold, and there is plenty of dark time after the run, then you can wait
until after the run to get the transformation images.
This has been covered previously in several places in other contexts, so I’ll give a brief overview here. Your goal is to get images where the comparisons and target have a high enough signal-to-noise ratio (SNR) to get the precision you want or need. In addition, you want to avoid excessive trailing, to get as many images as conditions allow, and to take images at a rate such that you work the target effectively. Many of these conditions are interactive: changing one affects one or more others.
If you’re looking for 0.01m precision, you need at least a 100 SNR on all objects, not just the target. For an average scope of 0.25m and a main belt asteroid, that means exposures of about 2 minutes. In general, you should keep exposures at 1%-2% of the period. However, that’s not a hard-fast rule. A 5-minute exposure on a target with a two-hour period is about 4%, but if that’s what it takes, that’s what you should do. Eventually, there is a limit; if you’re working an asteroid with a suspected period of 12 minutes and using two-minute exposures, or 16% of the period, you’ll have a hard time getting a usable lightcurve.
There is another general guideline that says you should sample the lightcurve at a rate of no less than 1/20 the period, e.g., if the period is 20 hours, you should get a measurement at least every hour. My general recommendation is to use an interval between exposures that is 2/3 to 2 times the exposure time. So, if my exposures are 60 seconds, I’ll wait anytime between 40 seconds to 120 seconds.
Again, this is not a hard-fast rule and assumes a target with a period on the order of 2 hours or more. This can result in significant oversampling of the curve for a target with long period but it’s better than undersampling. There is one more point regarding “how often.” There is nothing gained by having your scope sitting idle and so waiting a minute or two between exposures is letting photons go unused. If the target is on the faint side, go ahead and shoot almost as fast as the system allows an image to be taken and downloaded.
If your camera can take the extra workload, the additional images let you “beat down the noise”. As noted above, in this instance, oversampling is good. Usually, I don’t change the exposure or intervals after the first session. This is to maintain consistency whenever possible. If you’re making lots of changes, it’s hard to keep things straight. Of course, if circumstances dictate, adjust the exposure time and interval as needed. Rules are meant to be broken.
Experience is your best guide.
For those following asteroids, you have to weigh several
factors. How fast is the asteroid moving and what is your field of view (FOV)? The
answers dictate, in part, how you’re going to shoot through the night. If the answers say that you can keep the asteroid and the same comparisons in the field
for the entire run, then you can aim the scope at the asteroid, or maybe a position where it will be halfway through the run, and go about your business. I
often run scripts where I update the scope’s position to the asteroid every ten
images or so. For a main belt asteroid, this means I move about a third of the field of view over an eight-hour run. In this case, I pick comparisons that
favor the side of the image to which the asteroid is moving. As time goes by, they move towards the center of the image and not away from it and so I can use
the same set of comparisons for the entire run.
If the asteroid is going to cross the field or if by keeping up with it you lose
the original comparison stars, you have to decide what to do.
Staying on Target
If you decide to periodically update the scope’s position, e.g., every ten images or so, then at some point those comparison stars might get too close to
the edge or go off the frame entirely. As long as there are stars to replace them, you can continue on. You’ll just split the data into subsets with each
subset using a different set of comparison stars. In this case, it helps to have
at least one common comparison star in any two successive fields. Another reason for the scripts is to make frequent calls to a GotoAsteroid-type
command. This is required for those using German Equatorial Mounts (GEM) so that
the controlling software can tell the scope to perform a “meridian flip”. This
is when the target moves far enough west that if the scope continues in its current position, the tube or part of the mount would hit the pier. At that
point, the scope is sent to the other side of the pier so that it can continue tracking.
Keeping a Point of View
You may prefer to keep a specific field and, when the asteroid moves too close to the edge, switch to an entirely different field on which you can sit for
another period of time. This isn’t as easy to put into a script since you have
to know how often you need to change the field. You still have the issue of having subsets of data that use different comparison stars.