I'd like to suggest folks interested in the details of photometry to have a look at the recent paper by David Rubin et al:
This deals with spectrophotometeric calibration of mostly hot white dwarfs and subdwarfs. Some of this is pretty heavy going, but you needn't grasp all of it. Many of the stars are high-weight Landolt standards, and I observe several of them intermittently as part of the mix of general all-sky calibration. Some simply tie down the (very) blue end of the transformations. Several that lie between about -25 and -30 Dec I use to help with measuring atmospheric extinction in combination with northern fields . Among the intentions is to get calibration for other stars in these fields, so one can have reliable photometry for a diverse range of star-colors in a single CCD image. Many of the fields have one or two mid-K giants as well as stars of intermediate-color to go with the (usually) blue Landolt star, but lack full-up standardization. The GD 71 field already has calibration on "many" nights from Henden and from Stetson, all tied to Landolt's few added stars here.
Looking at Table 1 in the paper, I have data on:
NGC 7293 (PN central star)
There is an extended discussion in the Rubin et al paper appendix B about the possible variability of the F-subdwarf BD+17 4708, which is the 'original' SDSS standard star. I'm pretty sure it is not variable in any meaningful way for mere mortals doing broadband photometry. My 57 nights of data over the last two observing seasons show rms scatter of only 0.0031 mag, about as good as you can do from the ground. Published V magnitudes since the early 1950s show a range of only ~0.02 mag, mostly attributable to calibration scatter. But Arlo Landolt claimed the star brightened by about 0.04 mag between 1986 and 1991. However, the Hipparcos data, which overlap two seasons of Arlo's series, show no variation; it is dead flat over seven successive seasons in ASAS-3. Arlo's mean V matches Nancy Roman's determination in the early 1950s, other data from the 1960s, the adjusted Hipparcos mean V, and the ASAS-3 mean. This is an old-old high-velocity metal-poor halo star, about the most stable stars out there. I think Arlo's data are just 'funny' for this particular star (his mean values are correct, however).
An excellent plot worth examining is Figure 7. This shows the run of extinction (for Mauna Kea) as a function of wavelength across the visible and far-red. Among other things, it shows why there is _no_ second-order extinction in the V passband: the Chappuis bands of ozone flatten out the downward trend of the Rayleigh curve (from gaseous air). By contrast one pretty much must include the second-order extinction in B. Johnson himself ignored the U-band (actually U-B) extinction, and typically one compensates by having the U transformation include linear and quadratic color terms in B-V.
I saw this sentence, "Among the intentions is to get calibration for other stars in these fields, so one can have reliable photometry for a diverse range of star-colors in a single CCD image," and saw a means of bumping up differential techniques to derive actual, standards-anchored magnitudes a la all-sky techniques. Am I reading this correctly? If so, what have people been doing up to this point with this concept? If you have a ready list of references/links that save time for your response, I would gladly welcome those!
At least for fields with full-up calibration, one can get zero-point and color terms from a single set of images. This assumes there is a significant range in color in the calibrated field stars. The AAVSO way uses just two stars as comps for variables, but there is no reason not to use several, and if they span a good range in color, then they can be used to beef up an all-sky solution that includes bona-fide standard fields (Landolt or other). Fields away from the Milky Way tend not to have red/blue stars in them without poking around to find them, as I've been trying to do with the isolated Landolt stars like the spectrophotometric standards and others. Then those fields must be calibrated on 'many' nights.
You can get calibration in zero-point and color even with a single picture by switching the order of the terms in the usual transformation equation, like so:
v = V + zp + a(B-V) + b(B-V)^2 + kv(X)
where v is the instrumental magnitude (in this case), V and B-V are magnitudes of the standards, and kv(X) is the extinction. Notice that you don't need a B (blue) image to go with this, but a wide color-range of standards in the image.
My reading on this subject is from Peter Stetson, and includes these papers:
...specifically read only Appendix B and sections 2 and 3 of the text;
This mainly extolls the virtues of point-spread-function fitting of stars (skip that for now!), but more importantly (section III.B near the end) tells why to reverse the order of the terms from the traditional "Henden & Kaitchuck" transformation method.
...outlines the general idea of building up a library of high-weight standard stars;
...describes standard-star creation in more detail ,and how one can go wrong. Besides the early sections relating to reductions, be sure to read the Appendix here relating to statistics, and why one reduces CCD data as separate magnitudes rather than V mags and color-indices.
This conference talk gives some additional details about calibrating CCD frames and an example of use (plus a good picture of Peter Stetson).
I am grateful for these resource Brian. -Tom