ASTEROID LIGHTCURVES - PART II
After attending the TA AGM at Alston Hall last October and listening to Petr Pravec’s excellent talk and workshop on asteroid photometry I set myself the objective of obtaining a good lightcurve by the end of the year. By ‘good’ I mean;
- minimal scatter (<0.02 mag, ideally <= 0.01 mag)
- agreement with known period
- ‘thumbs up’ from Petr Pravec who had agreed to analyse my results
Initial trials (and tribulations)
The method used in this project was that of Unfiltered Differential as opposed to All-sky Photometry. My set-up included; a 10” Orion Optics Newtonian reflector on a Vixen Polaris equatorial mount, Skysensor 2000 hand controller and a Starlight Xpress MX516 CCD camera. I had a small amount of experience of CCD photometry and had worked up a reasonable procedure including; set-up, obtaining calibration frames and imaging. For imaging I used ASTROART with the Starlight Xpress plug-in. Although I have recently purchased CANOPUS for analysing the images I used AIP4WIN in this case as I did not yet feel sufficiently practised with the former (CANOPUS will by my preferred analysis package as this allows multiple comparison starts to be selected). The data generated by AIP4WIN was then loaded in to an EXCEL spreadsheet to produce the actual lightcurve.
Photometry demands a better quality image than astrometry (most of my CCD imaging to date had been NEO astrometry related) so I really had to get rid of what looked like numerous short horizontal scratches on my images. My observing notes suggested that ASTROART was the culprit as images obtained with Starlight Xpress’s own software were ‘clean’. This was finally fixed, after a logical approach had failed, by varying all the Astroart plug-in options in turn until I found the one that worked (checking ‘Long cable Delay’ ?!?).
The first asteroid I imaged, in attempting to obtain a lightcurve, was 1103 Sequoia (mag 14.9). This asteroid was imaged on October 26/27 2003 with an exposure time of 60 secs. The lightcurve obtained, shown below, was extremely noisy (+/- 0.1 to 0.2 mags).
A new book ‘A Practical Guide to Photometry
and Lightcurve analysis’ by
I checked the linearity of the MX516 CCD camera and found it to be reasonable down to mag 16. Aperture and annulus sizes in AIP4WIN were set as suggested in the book.
As to matching pixel size to seeing Brian Warner suggests that each pixel should be about ½ the FWHM. My set-up gave me a choice of 1/3 or 2/3 with binning and I chose the former.
Choosing the correct exposure time requires a number of factors to be considered. The maximum exposure time is that which avoids trailing and saturating pixels (50% max for CCD cameras with anti-blooming gates such as the MX516). To avoid trailing I used the Stephen Laurie formula; Exposure time (mins) = FWHM of star image/motion in arc secs. Astroart has useful facilities which enable the maximum pixel value to be measured for the whole image or individual stars. Long exposures require that the telescope tracks well and the Starlight Xpress Simultaneous Track and Record feature (STAR 2000) is excellent for this purpose. Once set-up the telescope will image continuously with little manual attention (very handy on cold nights!).
I did a little better with 196 Philomela (mag 11) and 98 Ianthe (mag 12.3) in that the noise reduced to approximately (+/- 0.02 -0.03). Only the latter showed any variation, see chart below. 98 Ianthe was imaged on the 15/16 December 2003 with an exposure time of 120 secs.
On 17/18 December 2003 I observed asteroid 250 Bettina (mag 12.2), exposure time was 150 secs. The asteroids parameters were well defined (in a list published on the internet by Alan Harris at http://cfa-www.harvard.edu/iau/lists/LightcurveDat.html). The magnitude variation, as the asteroid rotates, will depend on the position of its poles wrt the Earth. While not ideal it did look as if the amplitude would be large enough to show a reasonable light curve. The amplitude, according to the above mentioned list, is between 0.11 and 0.6 mag and the period between 5.05 and 5.10 hrs so an early start would allow a complete rotation to be observed by around . This attempt was successful and the lightcurve is shown below.
For those of you who understand these things Petr Pravec commented ‘the r.m.s. residual of the best-fit-fifth-order Fourier series decreased to 0.010 mag’. (I think that means it’s OK !). The ‘decreased to’ statement was brought about because the first set of results (published in the January 2004 edition of ‘The Astronomer’) showed a residual of 0.02 mag and the overlapping portions of the light curve did not tie up. The improvement came when I selected two different comparison and check stars after finding some variation with the previous pair plus the comparison star image was very close to saturation. The gap in the lightcurve occurred when I had to reverse the telescscope. This necessitated removing the CCD and thus I had to obtain calibration frames before doing so.
I plan to image some fainter asteroids (mag 13 -15) to test the limitations of my set-up and partake in observing campaigns such as the Photometry and Shape Modeling Program being run by the Minor Planet Section of the Association of Lunar and Planetary Observers. Details can be found at http://www.bitnik.com/mp/alpo/.
My thanks to Petr Pravec (Ondrejov Observatory at http://sunkl.asu.cas.cz/~ppravec/neo.html) for his assistance and to the late Andy Hollis and Richard Miles (at the time this article was written, Director and Assistant Director respectively of the Asteroid and Remote Planets Section of the BAA) for their comments.