A tripod mounted camera tracker for DSLR astrophotography - by P.C. Kuhn - May 2011
My interest in astronomy began when I built my first telescope in 2007. Upon discovering the multitude of fascinating objects in the night sky, my longstanding interest in photography surfaced and I was soon dabbling in astrophotography. Finding that wide-angle “starscapes” held the most appeal for me, I began using a tripod-mounted DSLR camera with standard lenses. I quickly learnt that star-trails caused by the earth’s rotation severely limited the range of shots I could take. A well-known rule-of-thumb stipulates that the maximum usable exposure in seconds without trailing is given by 600 (for a full-frame sensor), or 400 (for a crop sensor) divided by the focal length of the lens (35mm equivalent). In practice this meant that I was confined to focal lengths of around 24mm which allowed a maximum exposure of 15s with quite high ISO settings. I soon longed to explore longer focal lengths and exposure times.
This requires the camera to follow or track the apparent motion of the stars so that they appear stationary in the photograph, even at long exposures and when using telephoto lenses. This is often done by “piggybacking” the camera on a telescope which is equipped with an equatorial drive. I preferred a compact tracker which would mount on my camera tripod and could easily be included with the photographic kit which accompanies me on all my trips. A standard tripod head with its tilt and pan adjustments would allow the required polar alignment and it seemed to me that the trackers popularly known as “Barn door”, “Scotch- or Haig mounts would fit the bill nicely. Researching these, I discovered that there were several distinct types which seem to have evolved over time, to overcome the effects of tangential tracking errors which arise from the simple geometry of the barn door concept. These included the Single arm, Single arm isosceles, Dual arm and Curved rod types. I chose the curved rod type which has, in theory at least, a zero tangential error. Despite the curved geometry, it seemed fairly straightforward to construct. The ideas embodied in my design are not unique but were collected from currently available information, gleaned mostly from the internet and combined in a form that best suited my needs.
As can be seen in the accompanying photographs, my tracking platform consists of two flat plates 265 x 120mm, joined at one end by a two solid brass hinges. It is important that the hinges are really solid with a truly snug action, as any sloppiness here will cause havoc with tracking accuracy. I used a length of 10mm laminated flooring board for the plates as it is stable, easily machined and looks good. The 6mm threaded rod (1mm pitch) needs to be curved to a radius equal to the distance from the hinge pin to where the threaded rod passes through the end of the plate. A former of the correct radius was cut from a wood scrap and used to carefully shape the rod to form a smooth curve. For this size platform, the rod travel is about 65mm per hour’s tracking, so the rod was cut to an overall length of 130mm (to provide at least an hour’s worth of tracking, plus some safety margin) and fixed with two nuts to the end of a plate opposite the hinge. The other end of the rod passes through a hole in the other plate with enough clearance so that it does not touch the plate at any point of its travel. This end can have a stop nut fitted to prevent the arms flopping open too far. The rod curvature should be as smooth and true as possible in order to ensure good tracking accuracy. It makes life easier if a longer length of rod is curved and then cut to length.
A set of two matching plastic gears with a 3:1 ratio was procured from the local hobby shop and the larger one tapped for a 6mm thread to match the rod. The smaller gear is fitted to the shaft of the drive motor, which is mounted firmly on the second plate such that the smaller gear will mesh with the larger when this is threaded over the 6mm rod. A glance at the photos will clarify the arrangement. For the motor I used a small unipolar stepper with an integrated 80:1 reduction gearbox. With the motor’s 48 steps per rev, this gives effectively 48*80*3=12000 steps per 1 revolution of the larger gear. The speed required to move the camera at the sidereal rate of 15 degrees per hour, is dependant on the length of the hinged arm and the pitch of the threaded rod. For this example, the required rate is close to 1rpm. This requires a stepping rate of 200 pulses per second, which is well above the mechanical resonance of the tracker/camera package, thus minimizing vibration problems. The threaded hole in the large gear needs to be well centered and perfectly vertical, as any wobble will cause periodic errors. Another important aspect is that the rod’s radius of curvature should be long enough to prevent any hint of the rotating gear binding on the rod. This in turn sets a minimum length for the hinged arm. On the other hand, making the hinged arm longer can reduce the effects of periodic error and vibration as long as stiffness is preserved. There is of course a practical limit to the overall length if the unit is to remain compact.
Motor control is implemented with a microcontroller. I used a Microchip PIC16F684 type, programmed in assembler code. A digitally sampled potentiometer provides fine speed control and a tricolor LED battery monitor and power switch complete the package which is built into a small plastic enclosure mounted under the tracking unit. A 9V PM3 size rechargeable NMH battery provides up to 5 hours of tracking on a charge. A small mains adaptor can be plugged in to the control unit to charge the battery. The electronics can be somewhat simplified if a stepper motor control kit is purchased from an electronic store. Alternatively a d.c. motor with gearbox and a simple variable voltage speed control offers an even simpler solution. There are even trackers which are manually operated via a hand crank!
The bottom plate of the tracker is fitted with a quick release tripod mounting plate. The camera mounts to the tracker on a simple alt-az mounting as seen in the photos. A bottle cap containing a T-nut potted in auto body filler acts as a tightening knob for the altitude adjustment and the azimuth relies on a friction adjustment. A terry clip carrier holds a small finder which greatly eases aiming the camera, especially when fitted with a telephoto lens!
The tripod mounted tracker needs to be polar aligned such that the hinge pin points parallel to the earth’s axis. This is achieved by tilting the tripod head at the latitude angle of the location and orientating it N/S. I use an inclinometer for the former operation and a compass for the latter (remembering to allow for the magnetic declination of the specific location). Final alignment is carried out by means of the star drift method.
For my first trial I used a 30mm focal length and a 150s exposure. Results were encouraging with nice round stars. I then tested “longer” lenses up to 200mm (equivalent to 320mm with my crop sensor). As expected, results were a function of the polar alignment. I found that with a little care I could reliably achieve exposure times up to 30 times longer than were possible without tracking (e.g. 60s with a 200mm or 240s or more with a 50mm lens). This gives the flexibility to use lower ISO settings to control noise levels and allows me to use my 24-105 F4 lens to compose the majority of my shots and employ the 200mm F2.8 lens for “close-ups”.