CHAPTER 4
THE AUTOMATIC PHOTOELECTRIC TELESCOPE

The automatic photometric monitoring of cool stars is the major activity of the Automatic Photoelectric Telescope (APT) Service, which is a joint undertaking by the Fairborn Observatory and the Smithsonian Institute (Genet et al. 1988). APT's are a specialized telescope and instrument combination designed to perform nightly differential photoelectric photometry without the need of human intervention. The benefits of such an automated system are its relatively low cost (there is no need for observers quarters for example) and its high observing efficiency. The disadvantage is that unique situations are impossible to handle and usually result in lost or bad data. The need for long term, routine photoelectric observations, however, far outweighs the disadvantages of the systems.

One of the scientific justifications for the first generation APT, from which the data for this dissertation was taken, is that it can provide a time-serial photometric data base of known quality, for a large number of objects over an extended time period. These long term observations can then be used by other scientists in order to develop better models to describe the long term photometric behavior of the systems observed. Unfortunately, the accessibility of the data has been largely limited to the I.A.U. Commission 27 Archives of Unpublished Observations of Variable Stars. It takes several months to obtain a hard-copy of the data. The large amount of data obtained by the APT on any one object (see Appendix B) makes the reentering of the data an awesome task. Fortunately, I was able to acquire the first two years of APT data on a set of MS-DOS formatted diskettes. The Fairborn Observatory has since changed their computer control and acquisition system and they have been unable to supply additional data on magnetic media. This chapter briefly describes the first two years of observing by the APT.

4.1 The First Generation APT

The Fairborn observatory's first generation APT came into regular operation in 1983. The system is made up of a 10-inch f/6 Newtonian with a standard side-on 1P21 photomultiplier and a Johnson UBV filter set (Genet and Baliunas 1986). The photometer is uncooled and the filters are not temperature regulated.

The UBV photometric system was defined and established by Johnson and Morgan (1951, 1953) who wanted to establish a three color photoelectric standard that would yield magnitudes comparable to the yellow and blue colors of the International System and be closely tied to the Morgan-Keenan spectral classification system. The system was developed around the RCA 1P21 photomultiplier and three broad band filters. The visual magnitude filter (V-band) is yellow with a peak transmission near 5500Å. The long wavelength cutoff at this band is determined by the response of the photomultiplier tube and not the filter. The blue filter (B-band) is centered around 4300Å and relates well to earlier blue photographic magnitudes. This filter has a strong cutoff at the ultraviolet to eliminate the effects of the Balmer discontinuity. The ultraviolet (U-band) filter is centered around 3500Å, but there is a "red leak" which must be accounted for by performing near infrared observations or be blocked by a separate filter. The short wavelength cutoff for this band is determined by the earth's atmospheric ultraviolet transmission. A good introduction to the elements of broad band photometry has been written by Henden and Kaitchuck (1982).

The observing strategy of the APT is to observe a source "group" with a series of ten second integrations in each of three colors (U, B and V). A group consists of the check star, comparison star, variable star and the sky for a total of thirty-three ten second observations; approximately ten minutes are required for a complete observation (including telescope moves). Table 4-1 shows the observing sequence for a single group. After all the groups for an evening have been observed, the observing pattern begins again. Thus the telescope normally acquires one datum per star per evening, but sometimes two observations can be made.

The first two years of raw APT data had observations for sixty-nine groups, but I found that the data set contained thirty-six RS CVn binaries. These stars, along with their comparison and check stars, are listed in Table 4-2. The raw data, in both graphical and tabular format, can be found in Appendix B. The complete data sets from the APT can be found in the I.A.U. Commission 27 Archives of Unpublished Observations of Variable Stars in various files, beginning with file Number 131.

The data from the first quarter of 1983 to the last quarter of 1985 had some known problems (Hall 1986) such as sticky filter wheels and loose filters that resulted in the corruption of the data. These problems are listed in Table 4-3. These bad data are easily identified in the light curve and were simply edited out. One electronics problem, however, had the effect of reducing the observed brightness of very bright stars which was equivalent to not applying a dead time correction to the photometer. This problem had the potential for affecting all the data for this work. Fortunately, only DK Dra was bright enough to be noticeably affected and the first order corrections used on the data are described in Chapter 6.

The resultant observations of the APT show a statistical error of about 0.01^m, which is typical for photometric systems. The semi-regular spacing of the observations, on the order of one day, results in a pseudo-Nyquist harmonic near 2p for all the systems (see Chapter 6) and makes period variability determinations for the short period groups impractical. Unfortunately, the data set is also over too short a time scale to investigate period variability for most of the long period systems, however, this problem will be remedied with future observations. Overall, the APT data is not of the quality I would have liked for a reliable analysis. For the RS CVn systems, the distortions in the light curves are often on the order of only a few thousands of a magnitude, which is well into the statistical noise and precision limits of the APT data. Fortunately, the modified Scargle periodogram, described in Chapter 2, is well suited for the analysis of data with a signal to noise ratio less than unity.

The data for this work is from a first generation APT system and it is adequate for the analysis to be performed for this dissertation. It is expected, however, that as the systems improve, the quality of the data will improve also.

Table 4-1
APT Observing Sequence

Table 4-1 shows the observing sequence for the APT. At each position, all three colors are observed in ten second integrations, so the entire sequence results in thirty-three observations. The data released by the Fairborn observatory are the differential magnitudes for the nine variable star measurements (three in each color) corrected for background, differential extinction and transformed into the Johnson system.

Table 4-2
APT Comparison and Check Stars

Table 4-2 is a list of the 36 RS CVn systems observed by the APT from the last quarter of 1983 to the first quarter of 1985: Column one is the star name. Columns two, three and four are the comparison star, its visual magnitude from the Henry Draper (HD) catalog and its spectrum. The last three columns are the name, HD visual magnitude and the spectrum of the check star.

Table 4-3
APT Observing Problems

Table 4-3 is a list of the known observational problems from the first two years of APT observations. The first column lists the range of dates over which the problem occurred and the second column is a brief description of the difficulty. In general, all bad data identified within these ranges has been deleted.

05/22/02 ern

Copyright (c) 1988-1997, Eric R. Nelson, Ph.D.