Seeing in radio frequencies
  • chironexchironex March 20
    So it took me and my friend 2 months, but we managed to build our first radio telescope. LINK
    Easily my favorite part of the whole project was at the end when we were left with an image of something, in a frequency that you will never be able to see. We were able to see satellites over 35,000km away, glowing as bright spots in the sky. To really drive the point home, we took a 360 visual image and then overlaid the radio image on top. In this way we could literally see the ring that surround the earth. 

    I feel like, and other variations at other frequencies would be awesome to have in something like the hololens. Being able to look around and feed in data from telescopes like this, would make staring at the sky super interesting

  • ThomasEgiThomasEgi March 21
    That's pretty darn cool. If you want to build such a system again in a DIY way, I'd have a whole set of ideas which would increase stability/accuracy of the mechanics while cutting back on 3d printed parts pretty much entirely. 
  • I want one. I'm half tempted to build one of these with a friend. Do you think you could use enough of these to create a near real-time image of the sky, and then pass that to one of those VR headsets?
  • chironexchironex March 21
    Thanks :) If you've got a 3d printer, i'd actually advise using MORE pieces, not less. Increasing the torque at least double would be ideal. And using a worm gear in the bottom set. Or if you've got a machine shop laying around, im sure there's better ways to do it. 

    Do it! this was super fun, though if you do, maybe consider getting at leasta  90cm dish. That's the minimum requirement to see the milk way. Or use a different style antenna. If you just want to see the satellites though, this is fine. For the live image thing, if you had 7, all scanning a different bit of the sky over and over, it'd take about 3-5 minutes per image. So not quiet real time, but pretty close. Again, you'd want to use a bigger dish or you won't see anything fun. But it would be really cool. I mean that's sort of what they do at the really big observatories. Though they're looking for more specific stuff. I think it'd be really cool to try.  We're planning a few big community projects to collaborate with people all over the world to do large scale stuff.
  • ThomasEgiThomasEgi March 21
    A few mechanical suggestions from my end. 
    Use proper bearings for the vertical axis (azimuth?). I'd suggest UCF type bearings. They cost about 10 per pcs bucks depending on where you get them. Easily found on Ebay. Pillow bearing blocks might also be an interesting alternative. Proper bearings are vital to get a more rigid setup, which again is vital to operate the entire thing under non-ideal environments, such as wind loads etc.
    I recommend two of those bearings or the vertical axis to lock down all undesired motions. If you can use a metal rod instead of a broomstick, i recommend that too.

    As for the elevation bearing I have no handy recommendation. Ballbearings work but would be a bit too much effort to get in place. Bronze bushing and a steel rod could work out okay. There are also a number of dry running polymer based bearings for such a task. Again, 2 bearings would be the way to go, preferably spaced apart as far as reasonably possible (my recommendation would be > 5cm )

    Now for driving the thing. Using specially printed gears with low backlash for the elevation is a nice idea, but in this particular situation it has no benefit over cheap off the shelf parts. The gears are under permanently pre-loaded by the weight of the dish. So you never deal with backlash so no need to reduce it.
    I'd also recommend to mount the small thread further away from the rotation axis to give it a bigger lever effect. This will reduce load on your gearing and motor, giving you more accuracy and  a ted more stiffness. Best option would be to run it along a circular surface (similar to oil pumps), that way you get a linear relationship between steps and elevation.

    Controlling the azimuth is a bit trickier and there are more options. The wormgear idea is one option. Another would be to avoid gears as those suffer from backlash (and in this case it does affect the operation). One alternative would be to construct a disk (maybe 30 to 40cm diameter) and run a timing belt around it and the motor, with a small tensioning mechanism. The disk doesn't need groves like other timing pulleys. If you are good with a limited amount of rotation you can even screw it to the disk in one place.
    You could also try to work with friction gearing. Use a disk on the azimuth pole again (printed/wood/metal, all works) and simply press the stepper motor axis against it (preferably with something rubbery on one side). Rubber-paint on the disk , or shrinking tube on the motor might already do the trick. I'd try the timing belt first tho but if you'r on a budget, worth a shot.

    Once you get your mechanics to be rigid and non-wiggely you can, and should, go big. Reason being simple: bigger the dish, the more narrow your beam. Not only will the signals you pick up be stronger, you also get a higher "resolution" when scanning the sky.
    So yes size does matter, I'm not kidding: 
    I've had the honor to climb that thing halfway up during a guided tour. It's darn amazing. Not quite within the DIY budget but crazy cool.

    What else to say. Signal measurement: keep the noise out and away. At this stage the receiver amp is probably the part that benefits most from improvement. Not only would a proper circuit itself help, but also properly designed encasing to block noise, proper power supply (no switched stepup, at least not without very good linear regulators and filtering). Moving radio emitters out of the way (away from behind the dish too).
    Just the many little things there are. But yeah, keep up the work, continue hacking! Next time you have such a cool thing going on be sure to drop some note on IRC so i can throw in my 2 cents.