I would like to make some observations of T Mon (6.62-5.58V) and RT Aur (5.82-5.0V) and capture them on a CCD.
For various reasons (including availability of filters) I have opted to use iTelescope's T5, which is located in Utah, and uses a Takahashi Epsilon 250 (presumably 10-inches of aperture). The attached camera is an SBIG ST-10XME.
The chosen telescope would definitely be overkill, were it not for the fact that the comparison stars on offer from the AAVSO range from 8.7 - 10.4 (for T Mon) and 10.6 - 13.4 (for RT Aur).
Capturing both the target stars and the comparison stars in a single image, without blooming of the target star and still having comp stars with an SNI > 100 seems to be quite a balancing act.
Indeed, both target stars show blooming at exposures of 60 seconds. Reducing exposure times to 15 seconds seems to improve things (the streaks disappear) and a reduction in exposure to 5 seconds appears to be perhaps OK. However, a few questions:
- Given the ideal that comp stars ought to be at roughly the same magnitude as the target, how is the quality / accuracy of measurements of the target affected (if at all) by such a disparity in magnitudes?
- Is a short exposure of 5 seconds too short? In other words, should I be looking at some minimium exposure (of say 30 seconds)?
- How can I tell if blooming is still occurring in the target star? Is it enough that the vertical streak has disappeared from the target image?
- Given that the target image is quite a bit larger on the plate than the comparison stars, I assume that I have to make the inner aperture large enough to contain the whole of the target image. Is this correct? And, if so, what effect does that have on the measurement of the comp stars when they appear as a relatively small dot in the middle of a largely-empty aperture?
I don't think that I have a great solution, if you are constrained to a small field of view. Trying to compare a bright target star to faint comparison stars is tough. All I can advise is sticking as much as possible with the guidance on exposure times and number of exposures in chapter 4 of the AAVSO CCD photometry guide. An exposure time of at least 10 seconds is generally recommended, with 3 seconds something of a minimum. You'll be trying to get a number of shorter exposures instead of one longer exposure. The long periods of the Cepheids will help, since they won't be changing brightness rapidly. I might suggest starting by checking the duration of exposure times when you start to run into problems with saturation or non-linearity. That will help you know what range of exposure times are feasible. Slight defocusing can spread the light of the bright target, but won't get rid of the faint comparison star problem. Some other reader of this forum might come up with better suggestions.
These are good questions, and this sort of observing problem inevitably shows up in various parts of the sky or for various types of projects. Obviously you'll have to determine the exposures necessary to avoid saturation or getting into the non-linear regime for the bright targets. If that exposure is indeed short (few seconds), and assuming the camera shutter is reasonably good, then simply take bunches of frames at each visit to the field. Five of 'em is not too many, and one might compare two batches of five (say), offsetting the telescope by the equivalent of 10 pixels (say) to see how the averages of the two batches compare. The idea is to beat down the errors due to scintillation noise as well as beefing up the statistics for the faint comps (use multiple comps if possible).
I use the above procedure for a number of fields containing T Tauri stars. Because of the general obscuration by the dust clouds, often there are only a couple of rather faint comparison stars that aren't themselves other T Tauris, even after offsetting the telescope to pick them up (i.e. not centering on the variable itself). It has worked out well enough to just take batches of frames and average the results. Similarly, I follow the carbon Cepheid RU Cam, where the four suitable comps are quite a bit fainter than the variable, and the star is rather bright for the telescope. I use exposures in V of only 1 or 2 seconds, but take at least five at each visit. The nightly scatter is ~0.007 mag rms, so these are entirely satisfactory. Exposures at B and U are long enough that this is not so much of an issue.
In re the apparent size difference of the bright/faint stars: in general the actual size as measured by the full-width-at-half-maximum will not be much different; it is only that you are not seeing the wings of the profiles of the faint stars that makes them look smaller. (Yes, that is very difficult to get accustomed to!). It is true, however, that the optimum measuring aperture is larger for the bright stars. Choose an aperture that is about 3x or 4x larger _diameter_ than the fwhm; you'll have to live with whatever the errors are on the fainter comps (there's no free lunch). You can experiment with the results by also measuring in smaller/larger apertures to see what is most consistent for the particular instrumental set-up, seeing variations, bright Moonlight, etc. By having many frames and multiple comp stars (three to five is plenty), the mean values should come out pretty consistent. You can check this sort of thing by carrying along in your measurements some other convenient star in the field (check that it is not a known variable at least!). That will give an independent (and realistic) estimate of the night-to-night scatter produced by the ensemble of comp stars. Reduced in a consistent manner, you'll find these sorts of experiments on your data to be very instructive. They'll tell you what you _must_ do (like not saturating any targets/comps), but also limits of the data, and what you can get away with (some thin clouds, relatively noisy data but large-amplitude variations, etc).
I have used a similar telescope, and also in Utah, for objects of this type. Currently I use a 6" telescope to try for the brighter stars, and still have issues brighter than about 7th magnitude. So, what Horace said is very accurate. It is often best to defocus a bit, or more, to allow for an increased exposure time. If you are seeing spikes on your stars you are very likely in the non-linear range if not fully saturated. The best is to check peak counts, or if you can plot a radial profile you can see if it starts to flatten out.
As for measuring the stars on your frames. The shape of all the stars will be the same in terms of the FWHM in pixels. The highest single-to-noise point is at about 1 FWHM, but I normally use about 1.5xFWHM. Steve Howell's book on photometry gives a very good description of how errors work when measuring counts off of a CCD. Using the small aperture also allows you to get the most out of your faint stars. I know is sounds counter-intuitive to give up some of the counts that are coming from the star, but you are reducing the noise at the same time. When I do photometry I generally use VPHOT at AAVSO or AstroImageJ. Both of these allow you to adjust your aperture frame by frame to maximize the photometry. If you can find a couple of comparison stars in the frame you can use the statistical advantage of averaging over comparisons to get a better final value.