3 ms·

A DIY Astro Tracker in Two Nights

47 points by skazka16 12y ago

codezero 12y ago
Can't wait to see the results of this, good job!
Florin_Andrei 12y ago
There are many projects like this, but a lot of them are done wrong: they use a straight main bolt and they don't compensate for the error in the driver motor. This one is done right - the main bolt is curved, so no error (or very little error) is introduced while tracking. Good job.
- StavrosK 12y ago
  Where is the error from? I don't understand why connecting the motor directly would be different.
  - Florin_Andrei 12y ago
    In most cases, the motor has constant rotation speed, so the speed at which it advances on the bolt is also constant. If the bolt is curved, the radius of the motion, relative to the hinge, is constant through the whole range of motion. Constant linear speed + constant radius = constant angular speed. Some folks prefer to "simplify" the design and just make a straight bolt instead. Again, there's constant linear speed. But this time the radius, relative to the hinge, is not constant (you have a triangle, instead of a circular slice). So the angular speed is not constant anymore. You could vary the motor speed to compensate for the straight bolt, but why bother. Curving a threaded piece of thin metal rod is not that hard.
- lutorm 12y ago
  There is one thing missing though: a means of accurately aligning it to the pole.
- cgearhart 12y ago
  "Wrong" is a strong word. There are compromises in every design. As pointed out by another poster, accurate pole alignment will be the dominant source of error for any reasonably accurate barn door. The short focal lengths, wide field of view, and relatively short "long" exposures of DSLRs (5 minute exposures will induce thermal distortion on my D60 vs >15 minute exposures for film) common in DSLRs don't require extreme accuracy, especially if stacking the final images. Barn door trackers can have very low error without curving the rod. Modern microcontrollers also allow digital compensation compared to older barn doors that often used analog synchronous motors, and variations on this design (e.g., the double barn door [1]) can provide tracking accuracy at the limit of hobby applications. [1] http://davetrott.com/inventions/double-arm-barn-door-drive/ http://davetrott.com/inventions/double-arm-barn-door-drive/
  - Florin_Andrei 12y ago
    Well, curving the rod is a pretty simple way of eliminating one source of errors - why not do it when it's so easy. Of course you still have to do polar alignment and so on.
jhallenworld 12y ago
The overly high tech methods used in the project amuse me: design it with a CAD system, use a laser cutter to cut the wood, design your own gears, write C program for microcontroller to driver stepper motor- (advantage: easily adjust for astrophotography from other planets). It's too much meta-work. So here is my new challenge: make it using only parts and tools available from Home Depot. (even so, I have to get one of these laser cutters..)
- lutorm 12y ago
  I made one a long time ago: two planks, a hinge, two nuts, a threaded rod, and a camera mount. Operation was by turning the threaded rod 1/4 turn every time my HP48SX beeped. :) It's actually advantageous to be manual when you're doing this at -25C on a cold winter night in Sweden. Batteries tend to not work well at all at those temperatures. (And no, you don't need a curved threaded rod, you can just calculate what the correct rate is as a function of position.)
theophrastus 12y ago
Astro-know-nothing question: why wouldn't it be more directly 'tracking' to forget the constant clock drive (which we know will always be off by some amount) and set up a suitably long sight tube on some bright star with a photo sensor at the bottom of it. When the star's light moves out of the tube, a fine-stepped stepper tracks west until it re-appears ..?
- lutorm 12y ago
  That's how automatic guiders work, but they only provide extra signal above and beyond the normal tracking rate. Rather than pulsing on and off completely, you want to start with something you know is almost right and then just provide small perturbations on top of it.

47 points by skazka16 12y ago

Can't wait to see the results of this, good job!

There are many projects like this, but a lot of them are done wrong: they use a straight main bolt and they don't compensate for the error in the driver motor. This one is done right - the main bolt is curved, so no error (or very little error) is introduced while tracking. Good job.

StavrosK 12y ago
Where is the error from? I don't understand why connecting the motor directly would be different.
- Florin_Andrei 12y ago
  In most cases, the motor has constant rotation speed, so the speed at which it advances on the bolt is also constant. If the bolt is curved, the radius of the motion, relative to the hinge, is constant through the whole range of motion. Constant linear speed + constant radius = constant angular speed. Some folks prefer to "simplify" the design and just make a straight bolt instead. Again, there's constant linear speed. But this time the radius, relative to the hinge, is not constant (you have a triangle, instead of a circular slice). So the angular speed is not constant anymore. You could vary the motor speed to compensate for the straight bolt, but why bother. Curving a threaded piece of thin metal rod is not that hard.
lutorm 12y ago
There is one thing missing though: a means of accurately aligning it to the pole.
cgearhart 12y ago
"Wrong" is a strong word. There are compromises in every design. As pointed out by another poster, accurate pole alignment will be the dominant source of error for any reasonably accurate barn door. The short focal lengths, wide field of view, and relatively short "long" exposures of DSLRs (5 minute exposures will induce thermal distortion on my D60 vs >15 minute exposures for film) common in DSLRs don't require extreme accuracy, especially if stacking the final images. Barn door trackers can have very low error without curving the rod. Modern microcontrollers also allow digital compensation compared to older barn doors that often used analog synchronous motors, and variations on this design (e.g., the double barn door [1]) can provide tracking accuracy at the limit of hobby applications. [1] http://davetrott.com/inventions/double-arm-barn-door-drive/ http://davetrott.com/inventions/double-arm-barn-door-drive/
- Florin_Andrei 12y ago
  Well, curving the rod is a pretty simple way of eliminating one source of errors - why not do it when it's so easy. Of course you still have to do polar alignment and so on.

jhallenworld 12y ago

The overly high tech methods used in the project amuse me: design it with a CAD system, use a laser cutter to cut the wood, design your own gears, write C program for microcontroller to driver stepper motor- (advantage: easily adjust for astrophotography from other planets). It's too much meta-work. So here is my new challenge: make it using only parts and tools available from Home Depot. (even so, I have to get one of these laser cutters..)

lutorm 12y ago
I made one a long time ago: two planks, a hinge, two nuts, a threaded rod, and a camera mount. Operation was by turning the threaded rod 1/4 turn every time my HP48SX beeped. :) It's actually advantageous to be manual when you're doing this at -25C on a cold winter night in Sweden. Batteries tend to not work well at all at those temperatures. (And no, you don't need a curved threaded rod, you can just calculate what the correct rate is as a function of position.)

theophrastus 12y ago

Astro-know-nothing question: why wouldn't it be more directly 'tracking' to forget the constant clock drive (which we know will always be off by some amount) and set up a suitably long sight tube on some bright star with a photo sensor at the bottom of it. When the star's light moves out of the tube, a fine-stepped stepper tracks west until it re-appears ..?

lutorm 12y ago
That's how automatic guiders work, but they only provide extra signal above and beyond the normal tracking rate. Rather than pulsing on and off completely, you want to start with something you know is almost right and then just provide small perturbations on top of it.