23rd January, 2003Using a butchered serial mouse to interface to shaft encoders
The LS7266R1: an alternative to using serial mice chips
Controlling stepper motors 23rd January, 2003
On the microcontroller 13th February, 2003
In order to be able to control a telescope to point at a particular star, nebula, galaxy or other object, you need to be able to work out where it is in the sky and then track it as the earth rotates. Without going into any details at this stage, I've found a great book called Astronomical Algorithms by Jean Meeus which provides descriptions of methods and formulae for calculating positions, angle, times for everything astronomical. The book also describes how to account for things such as atmospheric refraction - something you don't generally stop to think about but which has a major impact on the apparent place of objects near the horizon. I've coded parts of this book in the C language and you can read more and get my translation here.
Although I have a preference for designing and building things from scratch, there's been a lot of work already performed on the control of a telescope using a computer. Check out Mel Bartels' site which includes a wealth of links relating to computer control of a telescope and to other nuts' (like myself) pages.
I've chosen to use two stepper motors to control the telescope in the Altitude (up-down) and Azimuth (around and around) directions. I'm hoping that coupling the shafts of the steppers to the roller-skate / roller-blade wheel bearings will yield an acceptable result although I've not seen too many others claiming to have used the actual roller-skate / roller-blade wheels themselves. Many other constructors have used just the ball bearings from roller-skate wheels, not the wheels themselves. Read about this design here.
Stepper motors are good because you can precisely control the amount of turn and the rate of turning using a computer. The motors I'll be using are 200 step per revolution and when half stepped, they'll be 400 step. The shaft diameter is around 6.5mm and the diameter of the mirror box bearing is around 200mm. You can calculate for yourselves that each half step will be a very small fraction of a degree of eyepiece angle indeed.
However the computer should not rely on the step count to track the position of the telescope. You could accidentally knock the tube and cause it to slide, for example. Or you could be leaning against the eyepiece when the computer tries to move the telescope and it could slide (stall) on its bearings. Either of these misfortunes would normally knock your telescope out of alignment with where the computer thinks it should be pointing. There is a rather straightforward solution (I hope).
By coupling a shaft encoder to the telescope bearings through a second roller skate wheel, you now have a method for receiving direct feedback about telescope movement. The idea is that the rotational drag of the shaft encoder should be so low that there is virtually no chance that it will slip, even if the telescope is bumped or held during a slew.
I've obtained 512 count per revolution encoders which, when quadrature encoded, come out at 2k counts per revolution and thus I should be able to read the movements of the telescope to a fraction of step accuracy of the stepper motors.