Optical star ball projector based on a truncated icosahedron

Posted by: Ron Walker Apr 5 2021, 12:04 PM
QUOTE(Scott T @ Apr 5 2021, 12:13 AM) *
Thanks Ron for the condenser information on the other thread - really good stuff.

You may have noticed a pause in the posts - I have hit a simple but time consuming problem.

I have settled on 19000 stars at the moment -about 600 stars per plate.
The spreadsheet sorts the stars to the correct plates in an instant. However, there are around 5000 stars duplicated on more than one star plate (in the overlapping zones of the pates).

I do not want to keep the duplicate stars because if they don't align perfectly (very likely) there will be a higher density of stars in the overlapping regions and if they do align perfectly the stars will be double brightness.

While Excel has a 'remove duplicate' function, it works by retaining the first occurrence and deleting the rest. This would create some large star plates which retain all the stars and some clipped plates where the duplicates are all missing. I want a fairly even spread so that the projections of two star plates merge seamlessly.

The only way I can think of to do this is to simply go down the list and randomly delete duplicate stars, sometimes from one plate, sometimes from another. This is pretty easy it just takes a long time. There are 19000 rows to pick through.

I am pretty sure I will think of a better way to do it when I finally get through them.

I can see an accuracy dilemma raising its ugly head right now. I assume we are talking about 19K stars in the general individual star population. As I said before, one can run a small amount of ambient light to wash out the dimmer stars but that destroys the jet black background that makes an optical/mechanical machine so good. This might not be much of a problem as the reflected light from the Milky Way might just do that job by itself. Thinking this through, one might wonder is it worth projecting more then 6K stars or even less. The Morrison projector only projected 3,800 stars and many consider it was one of the best star fields ever projected. What this all boils down to (in this fevered mind anyway) is how to project 19K stars but only allow approximately 6K to be seen by the average viewer without optical aid. That is, if a viewer comes to a show with birding or opera glasses (ones that can focus down to the dome screen) and with that aid be able to see those very faint stars.

I have read on more then one occasion that people love the bright skies of the Zeiss IX and how well it competes with the video projected at the same time, but then a lot of those same people say the Zeiss VI projected the most accurate skies. It would seem that to project and be seen with a video projection, the stars must be overly bright so the representation of a clear night sky is actually inaccurate. To be really accurate, faint stars should take as many as ten minutes to show up as the viewer becomes accustom to the dark night. This would never work in a now typical rush rush program that only runs twenty minutes or so, but would still work, and work well, in a more expanded hour show.

The problem comes, it would seem, when we try and project all of these various magnitudes of stars where the faint ones don't get brighter because of refraction from such a small star hole or project such a large star that it would be confused with other Suns or Moons in our sky. Since the laws of optics doesn't change, it appears that an optical projector can handle about five stops of magnitude before problems arise at either end of the spectrum. The typical answer has been to add optics for the brightest of stars since there are fewer of them and we could easily add them for the first two levels of magnitude and then let the star projectors take over to magnitude eight. I would imagine the Korsakov projector built for Boston had a spectacular sky as it projected three sets of stars from as many starballs. 1st and 2nd from individual projectors mounted on one ball, 3rd and 4th on another, and finally 5th and 6th on a third. They used the same sized holes but changed the wattage of the lamps. One wonders why they didn't project 1st and 2nd from individual projectors mounted on one ball, 3rd, 4th, 5th, and 6th, from the second, and then 7th, 8th, 9th, and 10th from the third. What a sky...or perhaps the reasoned that no one would be able to see anything projected from the third starball without optical aid, and how many people would bring opera glasses or for that matter, ever care enough to bring them.

So then how do we project these fainter stars and should they be seen even thought they can not be seen in real life. How many stars is enough? Is a Spitz Jr. with only 400 projected stars enough. It certainly can teach locations of the major constellations in their proper positions even in a small room with ceiling and walls. Actually more accurately then many of the newer projectors reproducing many more stars then any of us will ever see. These projectors project a flat "star chart" onto the ceiling of a room. Useful as any but as inaccurate as most. No, to be accurate these very faint stars should not be seen by the average viewer without optical aid. The only way without multiple projectors is to use some kind of neutral density filtering on those dimmer stars which is a whole new thread. Perhaps thr Morrison was right all along...3800 stars are enough.