Observations are taken from the James Karle Observatory, on the Lewis & Clark campus, located about 5 miles south of downtown Portland.

Prior to 2000, observations were made by utilizing a 10-inch Newtonian reflector to focus light into a dry ice cooled 1P21 photomultiplier tube. Light output, indicated by the magnitude of the electrical current from the photomultiplier tube, was converted to a voltage and fed through a General Radio DC amplifier into a LabView program on a Macintosh computer. (Light levels were originally recorded on a Brown chart recorder, but later data were guided into a Labview program Colby Jurgenson tailored to the purpose.) Relative photometric techniques were employed by the observers, by comparisons of the intensity of the variable star (44i Bootis) to that of a nearby comparison star of constant brightness (47k Bootis), using HD135558 as the check star. To account for long term changes in the atmosphere, the variable star is compared to the average of comparison measurements, taken both before and after the variable reading. Background light is accounted for by taking regular measurements of the sky, between variable and comparison observations. Constant monitoring of atmospheric conditions is not necessary for relative photometry, as they apply equally to both variable and comparison stars; also, since the stars are about a degree apart, the data was left uncorrected for differential extinction.

During 2000, we began observations with an 11-inch Schmidt-Cassegrain Celestron to focus light into a thermoelectrically cooled CCD camera. Light output, indicated by the pixel count of the CCD camera, is recorded directly to a Macintosh computer. Comparisons of the light intensity of variable star (44i Bootis) to that of a nearby comparison star of constant brightness (47k Bootis) are still made. Extraneous noise is accounted for both by cooling the CCD camera to -15 degrees Celcius and subtracting a dark frame from the means of about every 10 star images.

Photomultiplier data analysis began sending it through a computer program created by Jerritt Collard, which output the average deflection of each star measurement . The measured brightness of the sky was then subtracted from each data point, and average voltage values are attained. Next a ratio between each variable deflection and the averaged comparison deflection was computed and converted into a corresponding magnitude difference between the variable and comparison stars at that point in time. The magnitude differences were then graphed against the Julian date of the observation, to generate a light curve. Note that, for close binary systems, time is usually expressed as phase, and one phase corresponds with the orbital period. The times of minima were determined by locating the intersection of the best fit lines for the changing magnitudes of both sides of the eclipse. Below is a light curve graph generated from data taken in the summer of 1997.

Using Karle's data, Jurgenson and Prices's data (summers 1996 and 1997), Pereira, McInnes, and Price's data (summer 1998), and the results of previous observations by other researchers, we obtained a long term record of the system's behavior. Below is a graph which was created to compare the calculated measurements for the expected times of minima with the actual observed times; this particular graph is based on Oprescu's (1991) ephemeris: JD 2443604.5880 + 0.26781753E.

This o-c comparison resulted in a clearly non-linear graph. The curve of the graph indicates that over the last 30 years the eclipse period of the star system has been constantly changing. The trend of eclipses appears consistently later than the calculated times, indicates the stars' increasing period.

With the addition of the equation of the least-fit line to the O-C curve, the original ephemeris becomes a quadratic ephemeris, which accounts for the consistently changing period, and indicates that the stars' period increases linearly over time.

CCD data analysis begins with the program MIRA. Dark frames are subtracted and flat fieldscorrections are made and statistics are run. Essentially, the background mean is subtracted from binary and comparison star means and multiplied by the corresponding pixels, resulting in a light count (photon intensity). Then these intensity differences are compared and graphed against the Julian date of the observation, to generate a light curve.

By analyzing the light curve's general shape and measuring how long the eclipses diminish the light and by how much, astronomers can derive a model of the system. If they also observe the eclipsing binary as a spectroscopic binary, they can combine data from the light and velocity curves to determine such properties as radii, masses, densities, temperatures of the stars, and the true orbital sizes.

Program Introduction

Created by: MelissaPereira 4/2000

For more information, contact Olsen@lclark.edu