Photoelectric Photometry Newsletter

Photoelectric Observations of Eclipsing Variables

With the permission of the author and editor David B. Williams, I am pleased to reprint the following article from the AAVSO Eclipsing Binary Update, Number 5, October 1997.

Some Thoughts on Photoelectric Photometry of Eclipsing Binaries

At some point, every visual observer of eclipsing binaries looks at his ragged, asymmetric light curves and imagines how much better his minima timings would be if he used a photoelectric photometer or CCD cmnera. Photoelectric techniques, whether based on the photomultiplier tube (PMT) or CCD chip, eliminate the systematic errors involved in visual estimates and offer much higher precision in the observed magnitudes and, therefore, in the resulting time of minimum.

Of course, there are trade-offs. The experienced visual observer can follow several minima at the same time, moving his telescope from star to star and making estimates of each one every 10 or 15 minutes. A determined visual observer can time five or ten minima per night. This mass production of relatively low-precision visual timings is what allows the AAVSO eclipsing binary program to monitor the periods of hundreds of eclipsing binaries and fill in the O-C diagrwns that are featured in our recent publications. The photoelectric observer can usually observe only one star at a time. But while the PEP observer may obtain only one minimum timing per night instead of ten, that one timing is ten times as precise as a visual timing (the uncertainty is usually +0.001 day for PEP versus +O. 0 1 day for visual).

With these considerations in mind, visual and PEP observers need to adopt appropriate observing programs. The visual observer should focus on eclipsing binaries with relatively deep and rapid minima. For such stars, visual precision is perfectly adequate to monitor period changes. The PEP observer should concentrate on stars whose minima are too slow or too shallow for visual observation and stars for which there is a specific need for high-precision timings, such as apsidal motion in eccentric binary systems.

Some of my own observations of eclipsing binaries illustrate these points. See, for example, my paper on visual and PEP observations of DM Per in AAVSO Journal Vol. 20, No. 1, p. 14. DM Per varies from 7.9-8.6 V, but the eclipse is 11 hours long, so the rate of variation is only about 0.15 magnitude per hour. The O-C diagram (right, open circles = vis, solid circles PEP) shows that visual timings of DM Per Minima are wildly scattered and,worse, systematically earlier than the PEP timings. This suggests that. during such a slow eclipse, visual observers anticipate the minimum and imagine that the variable is dimming faster than it realiv is. At 8th magnitude with an eclipse depth of 0.7 magnitude. DM Per is attractive, but it is not a suitable star for good visual timings.

A year ago, I observed PEP minima of Algol and RZ Cas. From a strictly scientific point of view. Algol is a star that visual observers do not need to observe. Algol has been investigated many times in great detail, and the remaining questions about this binary system can only be answered with data of photoelectric precision. Tne star is bright and can be observed photoelectrically with the smallest instruments.

My observations were made with a 125-mm refractor and Optec SSP-3 photometer. The night was marginal for photoelectric observations, with patchy clouds until half an hour before the observations began, and more clouds soon after the observing run was concluded. The light curve (below) shows some scatter in the individual points, but reduction of this data by the Kwee-Van Woerden method yielded an uncertainty in the time of minimum of +0.0005 day (43 seconds).

I observed the minimum of RZ Cas with a 28-cm Schriiidt-Cassegrain reflector and Optec SSP-5 photometer. Although the seeing was poor that night, the photometric conditions were superior. The light curve (above) is extraordinarily smooth, and the Kwee-Van Woerden reduction shows an uncertainty of only +/- 0.0001 day (9 seconds!). I have to admit, however, that this was not the best use of my observing time. RZ Cas has deep, rapid minima and is observed visually at least 20 or 30 times per vear. The sum of these visual timings is more than adequate to follow this star's period changes in detail.

The booming popularity of CCD cameras has expanded the opportunities for eclipsing binary observers to perform high-precision photometry. Perhaps the most wonderful characteristic of the CCD chip is its high sensitivity. A CCD camera on a 25-cm telescope can reach about the same limiting magnitude as a PMT photometer on a 1-m telescope. For CCD observers, a particular advantage of observing eclipsing binaries is that minima can be timed without purchasing and calibrating expensive photometric filters.

Photoelectric photometers based on photomultiplier tubes or photodiodes have to be pointed at each object to be measured: the comparison star, the sky next to the comparison star, the variable, the sky next to the variable, then the comparison star again, etc. This means that measuring a differential magnitude takes time - four to five minutes for each cycle of measurements, and all this manipulation of the telescope and photometer keeps the observer quite busy. A great advantage of the CCD is that the variable, the comparison star, and the surrounding sky are all recorded in the same exposure. The telescope doesn't move, and all the pointing and centering on the individual star images is performed later by software.

AAVSO observer Dan Kaiser of Columbus, Indiana, is experienced in both classical PMT and CCD photometry. His light curve of AD And (right) was obtained with a 35-cm Schmidt-Cassegrain and SBIG ST-6 CCD camera. Each exposure was 40 seconds, and the chip read-out took about 50 seconds, so he wasable to obtain a differential magnitude every 1.6 minutes. This produced a densely sampled light curve. The data were obtained while AD And varied by only 0.2 magnitude during an interval of 1.7 hours. The scatter appears large due to the vertical scale of the light curve but is only about +0.015 magnitude. The uncertainty in the time of minimum is +/- 0.0003 day.

When Dan observed with an Optec SSP-5 photometer, he had to spend every moment of an observing run operating the telescope, centering stars in the photometer diaphragm, turning the flip-mirror knob on the photometer, etc. Now, with the CCD cameras he only visits his telescope to acquire the field and position the variable and comparison star on the CCD chip. Then he retires to the shelter of his enclosed porch, where he can relax in a comfortable chair and munch popcorn while watching a computer log one CCD frame after another, hour after hour. In photometry, the Age of Heroes is over!

But CCDs also have their limitations. CCD chips, especially those in the cameras that most amateurs can afford, are very small and can only record a patch of sky a few arcminutes across. This means that it is usually impossible to observe stars brighter than 8th or 9th magnitude with telescopes in the common 20-cm to 35-cm range. because it is impossible to get the target variable and a comparison star of similar brightness on the chip at the same time.

High-precision CCD photometry also requires a very accurate clock drive. Dan has upgraded the 7.5 -inch R.A drive gear on his Losmandv EQ 150 mount with an 11-inch gear and added a guidescope and CCD autoguider in a continuous battle to make star images sit still on the CCD chip. A PMT photometer can tolerate some image wandering during an integration, as long as the star stays near the center of the photometer's diaphragm.

My Algol-light curve and Dan's AD And light curve illustrate one welcome advantage of photoelectric photometry, whether PMT or CCD - the observing runs of 2.5 hours and 1.7 hours were substantially less than a visual observer would need for these stars, because a good timing could be obtained while the stars dimmed and brightened by less than 0.3 magnitude. There are many interesting eclipsing binaries with total amplitudes of less than 0.3 magnitude, and these are the stars that PEP observers should concentrate on. -DBW