It is possible to get perfect stars right across the diagonal of an APS-size chip such as the Starlight Xpress M25C or M26C – but only if your Hyperstar imaging system is in very good collimation.  Getting good collimation is not a trivial exercise, but once you have it set up then it only becomes a matter of very small tweaks if any mirror movement requires re-collimation.

 

The very first thing you must do is make sure that your imaging chip is perfectly normal to the optical axis.  For the Starlight Xpress cameras mentioned above there are adjustment screws to move the camera chip with respect to the front plate which connects to your OTA.  Starlight give a method here http://www.sxccd.com/maintenance_info/Aligning_CCD.pdf for getting the chip flat using a laser pen.  I have found that this method works extremely well and you can get precise chip flatness very quickly and easily once you have gone through the process a couple of times.  The only change I made to the Starlight setup was that I work in a vertical plane with the laser pen pointing upwards from the ground as using this geometry you don’t have the camera mounted sideways against gravity.  Being vertically mounted makes it a lot easier to maintain the camera flat against whatever rotation holder you use.

 

With the camera chip perfectly flat to the front plate, fix the camera to the Hyperstar and get a good focus.  I recommend the electric MicroFocuser and MicroTouch software for all your focusing.  You can do it manually, but it takes much longer using that route.  Get a good focus and take a 20-second image.

 

Unless you are extremely lucky you will find that stars towards the centre of the image look pretty good, but stars out towards the edge are “tailed” like comets – and this is the collimation needing adjusting.  You will need the program CCDInspector to get really good collimation, and to get a numerical feedback that your collimation is good – but the initial “rough” collimation is done by eye by referring to the image on the monitor.

 

Pick any one of the Hyperstar collimation adjusters and move it either in or out a touch, where a touch is no more than a quarter of a turn.  Retake the image and note the result on the outermost stars.  It will soon become obvious which adjustments “push the tail” of the stars up into the main star until you end up with a decent looking round star.  With the collimation adjustments made so that you can see you have decent round stars in all 4 corners, you are good to go, and you will get extremely good results from your Hyperstar – but you can fine tweak things a little better than your eye can pick up just looking at the monitor – and this is where CCDInspector comes in.

 

Take a 20-second image of a star field preferably without galaxies or nebulae in the field as these can “throw” the CCDInspector values out.  Put CCDInspector into arc seconds mode rather than giving “pixels” output, as this gives you more accuracy in the next part of the setup.  You will see values for the tilt in the X and Y directions, and the aim now is to get the smallest value you can for these.  The next adjustments of the collimation screws on the Hyperstar are TINY, no big movements or you will throw away all your hard work so far.  You now need to make tiny adjustments of the collimation screws on the Hyperstar and note their effect on the tilt X and Y values as you adjust them.  You will need to write these figures down as relying on memory alone, especially if you are working out in the freezing cold, simply doesn’t work.  You will also need to re-focus after each adjustment and this is why it is a very good idea to have computer-controlled focusing, otherwise this step will simply take far too long.  Make adjustments to the collimation screws to get the smallest values you can for the tilt in X and Y in CCDInspector, and don’t forget to refocus after each adjustment.  You will see that a natural by-product of small X and Y tilt values is a small value for the collimation and for the field curvature.  Proceed iteratively until you get the smallest X and Y values possible.

 

Finally, when you have managed to get the Hyperstar collimated as best you can, it is a good idea to then re-run all your V-curves for FocusMax so that they apply to your new, nicely collimated system.

 

That’s all there is to it – Happy Imaging J

Be Sociable, Share!
17 Responses to “Collimating the Hyperstar III”
  1. Michael Sullivan says:

    Thank you Greg,
    I last wrote to you about five years ago when I finished reading your book “Making Beautiful Deep Sky Images”. Dean at Starizona had reccomended it when I purchased my Hypetsar lens. Since then I went through years of fighting with my neighbor and his flood lights and my interest in astrophotography was disrupted. However, he finally lowered the lights and I bought a new starlight camera. Now, my frustration is trying to get perfect collimation. I just read this post and have renewed confidence that there is hope. My question for you is, when in perfect collimation, what curvature of field does the CCDinspector might I expect. I have a C 14 and it’s showing 17% which seems high. I was expecting womething flatter like 9%. Can you possibly e-mail a screenshot from the CCDinspector so that I might have a reference goal? Thank you in advance if you can help. Michael Sullivan

  2. Greg Parker says:

    Hi Michael,
    If you look further up the page you’ll see CCDInspector readings. I would say that a 17% curvature for a C14 is absolutely fine, at best you would only get a couple of percent better. Very sorry to hear about the neighbour and lights problem – it’s all too common 🙁

  3. Greg Parker says:

    Another thing I should point out. In your CCDInspector image make sure you only have stars – no nebulae and no galaxies, or they will throw the results out. Also – don’t have any very bright stars in the frame that will give ghost flares or this too will throw out the results. Good luck with the collimation 🙂

  4. Greg Parker says:

    And finally – make absolutely sure the chip on your SX camera is perfectly flat BEFORE you bolt it onto the Hyperstar or you will never get the magic 0, 0 for the X and Y flatness in CCDInspector.

  5. Michael Sullivan says:

    Hi Greg,
    Thank you for the response. I showed my camera to Dean at Starizona and he used a dial caliper to measure the gap between the plates holding the camera chip. Right away there was a discrepancy that was easily corrected with the collimation screws. However this was before I saw to the link that you provided above with the laser collimating gig. It makes more sense knowing that a laser beam would deflect in different directions as you rotated the camera if the chip was not flat. With that knowledge, using the dial caliper probably only works if the chip starts off being perfectly flat. I’ll find out in a few hours from now as I begin testing my images. Thank you again for you knowledge and inspiration!
    Michael

  6. Greg Parker says:

    Ahh – well there’s your problem then 🙂 The chip is NEVER flat in the first place so it’s no good having the front plate parallel to the camera front face. If you use the SX laser method to flatten the chip I think you’ll find it all suddenly falls into place 🙂 Good luck.

  7. Michael Sullivan says:

    Thank you Greg,
    As you predicted the images were actually worse last night with the front plate parallel to the camera front plate. Apparently, adjustments had already been done, perhaps at the factory, but they were not quite perfect. I’ll use the SX laser method and make the appropriate adjustments. I like your suggestion of using a vertical setup and plan to cut a hole in a table and place a two inch adpter in it to hold the camera as I turn it 360 degrees. Below I will support a laser which I will point up at the camera window. It seems to me that the further the distance to the floor the greater the deviation of the beam as I rotate the camera. With this in mind I should be able to make a significant enhancements towards flattening the chip. Thank you again. I will keep yopu posted.
    Michael

  8. Greg Parker says:

    Yep – that is exactly right, and I have the laser about 3 feet below the camera. I think this is plenty for getting a pretty accurate flattening of the chip. If you manage to get the reflection completely stationary then it will be fine to go onto the Hyperstar. Took me over an hour to do this the first time, but you get a lot quicker at it when you’ve done it a few times 🙂

  9. Michael Sullivan says:

    Hi Greg,
    I made a jig and used a red laser until the battery wore out and still could not get the dot to stay stationary. As I rotated the camera it carved out a smaller and smaller circle but I could never get it to perfection since the beam was too dim. So, I tried it with the hyper star hoping for the best but things were actually much worse. I have a friend with a green laser that I am going to borrow and give it another try. I will keep you posted.

  10. Greg Parker says:

    Hi Michael – I suppose I should have said I use a high-power green laser to do this job. It does take a LONG time to get the chip flattened the first time you do it. The main thing to be very careful about is making sure you are looking at the correct bright dot! Take a good look at the image on the SX site to make sure you are looking at the correct dot. Now before using your friend’s green laser – extend all three screws on the front plate to give you maximum movement of the front plate – and then try again. You might need to do this a couple of times until you get it right. You can get yourself into a position where it doesn’t look like you can flatten the chip – this is because you are too far out for the method to work. Extend the screws out to maximum and then try again – eventually you will get there. If the spot is stationary and the result on the Hyperstar looks terrible – then it’s the Hyperstar that is well out of collimation and that will then need adjusting.

  11. Michael Sullivan says:

    Thank you Greg,
    I’m heading over to my friends house today to pick up the green laser. One thing that I noticed when I started with the plate perfectly flat was that I could see the smaller and dimmer diffraction pattern dots. At some point the diffraction pattern disappeared and I was working with just one reflected laser dot that was impossible to center. I think as you alluded to that I “too Far” out and needed to start over with the plate once again perfectly flat.

  12. Michael Sullivan says:

    Good news Greg!
    Using the green laser was absolutely essential to my success. You need that kind of power to produce a bright enough diffraction grid that allows you to pick out the correct point of light reflecting off the chip. If the light beam falls off center just slightly and the diffraction grid goes away an interesting flaw enters the equation as the brightest light beam remaining is not the one you want to be looking at…it’s the reflection off the window above the chip. I demonstrated this to my wife over and over as i moved the laser pen off axis while drawing attention to both the correct dot and the one reflected off the window. When the diffraction grid disappears so does the correct dot. Using the red laser may have wasted some time but I gained the experience I needed to do it correctly. I used a dial caliper and set the gap at all three set screws to 0.l05 in and when I was done they ended up at 0.096, 0.100 and 0.103 inches when I was done. My wife stood over me and verified that the correct dot was stationary as I rotated the camera 360 degrees. I got lucky this time, it only took thirty minutes! When I attached it to the Hyperstar I only had to adjust one of three collimation screws from a clock position from 6:30 to 6:35. The CCD inspector showed a 3% tilt and 10 % curvature using over 200 stars. The X and y both showed 0.4 but the star images were very nice. Thank you for your guidance and experience, it is exceptionally rare to find help for something this specialized and I am very appreciative.

  13. Greg Parker says:

    Success 🙂 🙂 Glad you had a positive outcome. Now do some very fine tweaking on the Hyperstar itself to get the magic 0, 0 for the X and Y values and 0% chip flatness. Curvature won’t be better than about 10%. Collimation will be 0 or very small. Happy Hyperstar imaging 🙂

  14. Olav Norrud says:

    Hello Greg,
    Came across your blog post – well written and exactly what I needed. Will try as soon as the night is back up here in Norway, using my C11HD + HS3 + QHY8L.
    Do you think your “premium collimation” will be kept even if removing/reinstalling the Hyperstar?
    What about transporting the OTA/HS/CCD collimated and fully assembled in a big padded case, going to a dark site with a bit bumpy ride?

  15. Greg Parker says:

    Hi Olav – no, you won’t keep collimation under those conditions – I wouldn’t even breathe too hard on the scope once you’ve got it spot on 🙂 🙂

  16. Olav Norrud says:

    That’s what I suspected, just needed to have it spelled out!
    Thanks a lot, Greg.

    Btw, there’s very little literature on Hyperstar imaging, and I see you have the “Making Beautiful Deep Sky Images” with a dedicated chapter. Dean has of course some good articles, but stuff like complete process walkthroughs would be great, to pick up anything that has been overlooked.
    An updated and Hyperstar-dedicated eBooklet perhaps? 🙂

  17. Greg Parker says:

    Funny you should mention that 🙂 Springer have just asked me to update “Making Beautiful Deep-Sky Images” and I will be doing that.

  18.  
Leave a Reply