The entrance slit to my spectrometer is only 2.5 arcseconds wide. The equatorial mount that drives my telescope does not have sufficient precision to hold the tracking to that kind of accuracy, so I have been manually guiding the telescope using an additional secondary mirror/eyepiece on the opposite side of the telescope. This worked well enough to get decent spectra, but my tracking efficiency was probably well below 50%, and aligning the crosshair on one side with the slit on the other was a difficult task that had to be done once or twice a night. Hence the motivation to come up with an automatic stabilization system that was right in front of the focal plane of the telescope. The system that was developed may find use on other telescopes since it is compact, inexpensive, and will fit into a 2" eyepiece holder. Without the mirror motors and drive electronics for the motors, it could be easily adapted for use as an autoguider.
The image stabilization system consists of two mirrors rotating around two axes perpendicular to the optical axis of the telescope, about 2" in front of the focus of the telescope. The mirrors are rotated by two stepping motors acting as galvanonometers and not actually stepping (one winding of the stepper provides a restoring force while the other one gets the error signal). The position of the star is sensed by a CCD camera (a Connectix B&W Quickcam , about $100 at your local computer store). A small fraction of the light headed for the focal plane of the telescope is reflected into the CCD camera with a partially reflecting mirror (Edmunds sells them). The camera is controlled by a QuickBasic program which contols exposure time, gain, automatically adjusts the offset, number of lines read out, number of pixels read out and location of pixels to be read out within the full 320 X 240 array if reading out a partial frame. It displays the image on the screen indicating where the target star was located and the correct tracking point and the mirror D/A counts.Block Diagram
The program also centroids the star image and calculates the change in signal required on the two mirror motors to bring the star to the specified location on the focal plane, corrects for stiction in the motors, and shrinks the read-out area down to increase the frame rate after the target star has been successfully acquired. Frame rates in excess of 50 frames/second can be supported on bright targets with a 133 MHz Pentium. The pixel size on the CCD is 10 microns, and the control system seems to be holding the centroid to about one pixel.
The stabilizer consists of three parts - a head that fits into a 2" eyepiece tube, an electronics box containing the D/A converters and drive electronics for the motors, and a control computer. The control computer can be a very modest PC (I use a 25 MHz 386), and get a frame rate of about 10 frames/second.
A schematic of the D/A converter and motor drive electronics is shown in Mirror Drive Electronics
In order to integrate the stabilizer with my spectrometer, I had to move the primary mirror up a bit to accommodate the 2" thickness of the unit. (2" thick by 4" X 4")Outside view (with eyepiece holder for testing) Inside view
Tests of the system have been successful. It holds the star in
position quite well (almost always recording a centroid within
one 10 micron pixel of the target), and the spectra I obtain with
it have much less blurring along the slit. I lost a bit in sensitivity
by diverting some of the light to the CCD tracker, but it has
been made up the ability of the system to hold the star on the
slit better than I could manually. The system is a lot easier
to use as now I don't have to peer through an eyepiece and continuously
guide with hand pressure on the side of the telescope tube. I
can usually trust the system to work automatically for a few minutes
before I need to adjust the RA tracking to keep within the correction
range of the mirrors - enough time to have a sip of wine and make
notes in my log book! Also fun to watch the mirror position readout
bouncing all over the place as it takes out the wind buffeting
of the telescope.
