Mindboards

Hi,

You should not try to control the speed of the motor. Instead you should try to control the position of the motor. This is far more direct and has the possibility to recover from previous errors. I know NXT-G will allow you to control the position of a NXT motor, but I do not know how good this is. If it is not good enough you should write a PID-controller to control the position of the motor even better.

But even then the task will be difficult. The position of the NXT motors is known to the degree. This resolution won't be enough to keep a camera exactly focused to a specific star. So you will need to increase the resolution by gearing down the motors. However, gearing down the motors with lego gears introduces slack, this can be quite a lot, too much for your purpose. But you could try to use levers (made from liftarms) instead. This will limit the range of your device, hopefully not too much.

Aswin

Slack is not too much of a problem if you use worm drives and avoid the Lego gearbox, then most of the slack is in the Lego motor gear train.

The weight of the camera/ scope usually means that the tilting gears are always in contact even after changes in direction (unless your tilting axle is very stiff!).

When panning, you can build in an extra motor movement at each change in direction to take up the slack.

You can get a good estimate of the amount to take up using the MK 1 eyeball by writing a short program which just reverses the motor on command and displays the rotation count.

Run in one direction, stop and note the count, reverse until the rig starts moving again, stop and note the new count.