Ok, so I bit the bullet and decided to get rid of what I believed was tilt in my RC10 focuser...by adding the recommended tip/tilt plate. Here is the process I followed:
1. Installed the plate and adjusted the laser beam to hit the secondary dead center. Rotate laser to check for concenticity, all good.
2. Insert Cheshire - tweak the secondary to centre the reflection of the bright annulus and the secondary centre mark, all good.
Re-check the tilt, adjust and re-adjust secondary as required.
3. Check "hall of mirrors" for misalignment of primary (should be the only error left by now). Get everything concentric.
Re-check the tilt, adjust and re-check and adjust secondary as required.
Now I recon that was pretty good (see attached images), tighten up all and recheck....so now I just have to star test right? But this is where I struggle, visually a defocused star looks good but if I use CCDINspector on the image I get large and highly variable collimation errors reported. Is this bad seeing, tube currents or what? So far any adjustments I have made "under the stars" have made things worse not better.
So do I have to wait for a perfect night for the star test, (well good seeing and a full moon anyway)? I must admit I am uncertain how to differentiate between the various sorts of error by eye (and therefore which element to tweak - 1, 2 or tip?). Again, CCDINspector does the analysis but seems very prone to systematic error if the star field and tracking is not just so, it gives me a shot at separating collimation vs tilt error and curvature at least but again, to date, I have only made things worse and ended up back on the bench to get it back to not bad....grr.
So my question is - for that final tweak - what is the process you use?
I use a Planewave CDK, and for the final tweek, I use a test exposure of a star field, anywhere from 30 sec and up (so long as tracking is good enough). I look in the corners of the exposed image on a full frame chip and make sure the star shapes are round. I do this at camera angles 0, 90, 180 and 270 deg. The star shapes tell me which way the collimation must go and so I know what turning one of the collimation bolts will do to the star field...that's just trial and error to begin with.
HTH...
Hrm... Looks like the secondary could be pointed slightly too low. The takahashi scope can really help here but you might be able to try this:
Get a digital camera with a long lens mount it on a stable tripod and point it through your cheshire view port. You appear to have already done this judging by your photo... But the longer the lens the better. The aim is to better visualise the tiny secondary spot in the middle of the primary reflection. Should be able to see any slight misalignment easily. As you probably know, slight tweaks to the secondary are all that is needed. With my poor eyes it can seem to be aligned but under higher magnification is quite out.
The other thing to try would be to defocus a starfield or globular both intra and extra focal to see if the spherical abberation is balanced. Tweak the secondary only to get the out of shape donughts at the corners to point towards the very centre of the image.
Another thing to ask would be how close do you want it? If seeing is only 2 arc sec or your camera cant resolve the miscollimation then I would be happy.
Ccd inspector can give spurious results if your guiding is slightly off especially important on these long fl scopes.
OK so we had a clear, calm night so I could test a bit more. The results were a bit disappointing - I have had rounder stars...M41 was the test field and short exposures looked very nice but 5 mins starts to show up anomalies. Image scale is 0.75" per pixel. I definitely need to tweak something here - even though CCDInspector seems to be saying I am close...2.4" collimation error and 4% tilt. Not sure if the latter is regarded as significant or not given the field curvature on these scopes.....and I am not sure I believe those numbers in view of the image. The image is not calibrated - colour converted (ST4000) and just stretched a little.
Perhaps the Tak scope should be on my xmas shopping list....
CCD inspector has lied to be before. But if the star test and software agree, then both probably not wrong.
What is with the extra diffraction spikes on your stars? How did you focus your image? Could be slightly out of focus. Are you guiding?
Did you check for triangular stars and pinched optics? Not sure if of is a problem with this type of scope and mirror support but could be worth testing for.
Yup CCDInspector seems to be flaky. Not sure where those extra spikes are from - it may be the power lead to my 2ndry heater has moved of the spider vane - that has happened before I will check it.
Focus - I have a moonlite motor driven focuser and remote focus using CCDSoft, and yes guiding was active for this shot (using the AO8 at about 4Hz). The stars look ok on shorter exposures - and on older shots except in one corner (before I added the tip/tilt plate) so I do not think it is pinched optics, how would I test for that as a meeter of interest?
Regards,
John
Quote:
Originally Posted by tim.stephens
CCD inspector has lied to be before. But if the star test and software agree, then both probably not wrong.
What is with the extra diffraction spikes on your stars? How did you focus your image? Could be slightly out of focus. Are you guiding?
Did you check for triangular stars and pinched optics? Not sure if of is a problem with this type of scope and mirror support but could be worth testing for.
Yep, odd spikes could be a multitude of things but also could be diagnostic of something more... I personally have never had to deal with pinched optics but it is a known issue on some models of these scopes according to my reading.
Plenty of people on these forums would have more experience with this issue or of how to test for it but the best way I would think would be to check the defocus star shape across your FOV. Misshapen or irregular doughnuts (triangular in this case given the mirror support structure or mirror clamps) could be suggestive of this.
Some have reported this is caused by the edge mirror retaining clamps being over tightened at the factory. The problem is solved by simply adjusting or loosening them. Worthwhile reading up on the issue. This can really confound collimation.
Just had a idea. If you mistrust CCDInspect, see what it says with an artificial star. Take all the guiding and seeing out of the equation and give it a perfectly ideal case.
Hi - CCD Inspector is fairly reliable as long as you don't trust it implicitly - it likes data - the more the better.
I'm in the process of finalising the collimation of an iDK 12.5 at the moment and use it to analyse the results. Firstly cut out the tracking/guiding error of your mount - Even though your ST4000 is OSC it can be binned and bin it 2 x 2. Pick a star rich section of the sky - I use the LMC or Milky Way whichever is highest. Use the shortest exposure you can - I use 10 to 15 seconds - focus as well as you can. Then take lots of exposures - I use 50. Put all them in CCD Inspector and Measure all - reject those that are obviously defective then do your 3D curve and plot calculations - you will find the averaged results a lot more consistent and believable.
Pretty sure the image does not show pinched optics - I've had a GSO with that problem. It could be a rolled edge on the primary mirror but more likely microlens flare from the ST4000.
Well that was a long time between drinks.... I have been o/s and then dealing with other priorities. I got back to this collimation issue last week and I think I am now close...it is not perfect but attached are a couple of examples of what I get now. These shots are half res to stay under the 200Kb limit. In the end I abandoned CCD inspector and went for the Howie Glatter collimator with the holographic lens, that made bench adjustment pretty close and quite easy to do. Well a lot easier than a star test anyway. Happy.
Well that was a long time between drinks.... I have been o/s and then dealing with other priorities. I got back to this collimation issue last week and I think I am now close...it is not perfect but attached are a couple of examples of what I get now. These shots are half res to stay under the 200Kb limit. In the end I abandoned CCD inspector and went for the Howie Glatter collimator with the holographic lens, that made bench adjustment pretty close and quite easy to do. Well a lot easier than a star test anyway. Happy.
...and we are very happy to enjoy these beaut shots!
I know there are already many posts on this subject - I found most of them somewhat helpful but none perfect for my purposes. I have been doing AP for a few years with moderate results and modest gear and have used a few bits of kit - a meade ETX90, a Vixen VMC200l and Sphinx, an ED127 and my current scope a GSO RC250cf with a G11. I have never had to dive into collimation before but my current scope got quite dirty (pollen etc on the promary after 3 yrs of regular use) and was in need of a clean - but before I did that I wanted to be comfortable with collimation hence this thread.
I have a cheshire I got with the scope and could tweak the secondary to get the dot in the middle but I always had a corner where the stars had always been little comets (not really bad and could be cropped out but still they should not really be there with my ST2000xm which gives a small fov even when using a FR).
So I read up and came to the conclusion I must have tilt, I bought the collimation ring and a laser collimator to help fix that.
I had a go and though I did not fully understand what was occurring got close to collimation on the bench but could not get a really good result into the corners and my final "star test" tweak attempts would make it worse not better and send me back to the bench. I got some ok results but I was intending to clean it anyway so, having had some success with collimation I got up my courage, took the scope apart and washed the primary and seconday. I got it back to rough alignment as before but my star shapes were not good.
After many iterations of adjusting and testing approaches here is what works for me, I hope it helps someone else "get it".
1. Align the focuser (you do have that collimation ring right? If not stop here now and go get one) to get your laser to hit the secondary centre spot. Do this with the draw tube in the "camera at focus" position (not all the way in or out).
2. Remove the internal baffle tube (you have to remove the secondary ring assembly, then the baffle tube and then reinstall the ring, this is easy to do because its all one piece. Use some tape to mark ring orientation before removal so you can put is back the same way).
Note : If you have the secondary a long way off from previous "adjustment" you will be able to see it now. Please the ring on the bench, 2ndry holder side down on a level surface and put a bubble level on the top, adjust to get the bubble in the centre to get it back to a neutral position.
3. With the secondary re-mounted use the cheshire to line up the vanes by adjusting the primary. This is not possible with the baffle tube in place but is an easy adjustment with it removed. I check alignment with the Howie Glatter laser and concentric ring attachment projected onto a wall I use an A4 sheet and mark the centre and each ring position and make sure the ring spacings are symmetric.
4. Next use the cheshire to adjust the secondary and put the dot in the circle. The laser can confirm this as the beam should now return to the lasers aperture - if it does not you will see a double bright spot.
5. Reinstall the baffle tube.
6. Check everything still looks ok with the Cheshire and laser. You will now be close but not perfect (well you might be if you are lucky). This is where I got stuck. I figured at this point only a tweaks to the secondary would be required to get a good result but I was wrong. I believe the reason there may still be an issue is that, up to this point, I have used mechanical references (centre spot, vanes, etc) to align on and the optical axes of the mirrors do not have to be coincident with the physical, in fact they certainly are not with my scope. Also I really want the ccd centred with and orthogonal to the optical axis not the cheshire or laser and there may be differences in the adapters spacers etc.
7. Next install the camera and perform a star test (remove all filters/reducers first). Use an open cluster and as short an exposure as you can (to eliminate tracking problems but still give a useable image - binning helps), defocus (inside) to get small donuts on your screen and look at the shapes, adjust the secondary (use very small tweaks) to get a good central star shape (concentric rings, Al's collimation tool helps here!). Now look at the rest of the field. If there is still distortion in the shapes it should be even, that is good in the centre with some Radial or Tangential distortion increasing as you get to the corners, other patterns or uneven amounts mean tilt. Tilt is nasty as it can mean a very confusing field.
Unfortunately the most likely outcome seems to be compound after a mechanical collimation, that is we are Ok in the middle and one, two or even 3 corners are ok but one is bad. That was the situation I kept finding myself in. And I thought I had everything in good order and just needed to tweak the seconday a bit. This is NOT true. In my case the residual TILT had to be fixed first. To do that I tweaked the collimation ring to move the best star shapes towards the centre of the field. After that I then check the centre star again, tweak the secondary slightly if required. Keep doing this until you have good shapes all over the field. Then focus and do a 5 min test exposure. If all is ok then reinstall all the kit, reducers, AO, FW, OAG etc and perform final testing, try a few scope orientations, if the results vary you may have flex in the optical train that will need fixing not good news but at least you now know it in NOT a collimation issue.
My current result is attached, a full res shot of NGC3324 using my ST4k at 0.74" per pixel. I think this is ok or even good (the collimation - not the image which has nasty compression to get under 200k!), there may be a little improvement possible but I am happy with this.
