Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

RC collimation with no laser or stars

  • Please log in to reply
263 replies to this topic

#101 dg401

dg401

    Vostok 1

  • -----
  • Posts: 101
  • Joined: 11 Jul 2021

Posted 03 August 2021 - 06:00 PM

I wanted to do some real imaging tonight, but the cloud forecast is giving me the middle finger through at least Wednesday evening.

 

So... I just gave my RC6 primary and secondary screws some quarter to half turn abuse and spent 15 minutes putting it back into collimation.  Will re-test tonight, this time with my 9 micron artificial star (0.03 arcsec @ 60 meters).  That's on the dim side, but I've got room to play with ISO and exposure.  It will be interesting to see if I can get the same or better results versus last night's results or whether maybe I just got lucky.


Edited by dg401, 03 August 2021 - 10:18 PM.

  • nebulachadnezzer likes this

#102 dg401

dg401

    Vostok 1

  • -----
  • Posts: 101
  • Joined: 11 Jul 2021

Posted 04 August 2021 - 12:32 AM

Per my last reply, I screwed up the collimation via quarter to half turns of all primary/secondary screws.  I then re-collimated per the OPs procedure.  Star test via 9 micron artificial star at 60 meters.  Focused on-axis and corners at 0.5 sec.  Defocused on-axis and composite at 15 sec.  On-axis tri-bahtinov mask at... I forget.

 

Here is the tri-bahtinov mask on-axis.  It is not perfect, but it is better than my previous collimation.  Not sure that making it perfect would improve real-world photos and might just be an exercise in statistical showing off.  Bright artifacts due to stray neighborhood lighting.  The yellowish horizontal line coming in from left center of frame is the fiber optic cable for the single mode 9 micron optical fiber used for the artificial star.

 

On Axis Tri Bahtinov
 
Here is on-axis defocused.  Again, not perfect, but perhaps close enough to not matter:
 
On Axis Defocus

 

Next, we have 101x101 pixel grabs of (in order from top to bottom) on-axis, upper left, lower left, upper right, lower right.  I might have caught a bit of camera vibration for the on-axis shot:

 
On axis 101x101
upper left 101x101
lower left 101x101
upper right 101x101
lower right 101x101
 
I wanted to show defocused shots with the artificial star moved to different locations all over the frame.  That's a lot of shots, so I abused DSS to make a nice composite.  I find the previously recommended field flattener to be quite over-priced considering I've been told it's just a two element flattener.  I have no reason to believe that using the flattener wouldn't help considerably with field curvature.
 
defocus
 
Top to bottom (APSC) represents around 37 arcsec.  The circle shown below in my image circle versus sensor diagram then represents a sweet spot that's 37 arcsec in diameter on the RC6 at 1370mm focal length... which I daresay covers most any deep sky object I might want to image with this scope.  If I want larger, I'll just shoot with my 81mm/f6.9 apo (and I've got a flattener for that) or my little WO Redcat 51 which is sharp as a tack out of the box.
 
sensors

 

Again, and I can't stress this enough.  These results are using the OPs procedure outlined here using only a Celestron Cheshire eyepiece, daylight for illumination, the hall of mirrors effect, an old eyeball, and a pair of reading glasses to enhance the aforementioned old eyeball.  I collimated the RC6, slapped it on the mount and took the above images.  Once collimation was completed NO SCREWS WERE TOUCHED.

 

I am now even more impressed with TinySpeck's procedure than I was already.  I've not only duplicated my previous star test results, but I believe I've surpassed them.  This means there's more than dumb luck to my results.  This procedure works.

 


  • TinySpeck and nebulachadnezzer like this

#103 nebulachadnezzer

nebulachadnezzer

    Ranger 4

  • -----
  • Posts: 305
  • Joined: 04 Jan 2021

Posted 04 August 2021 - 12:51 AM

Again, and I can't stress this enough.  These results are using the OPs procedure outlined here using only a Celestron Cheshire eyepiece, daylight for illumination, the hall of mirrors effect, an old eyeball, and a pair of reading glasses to enhance the aforementioned old eyeball.  I collimated the RC6, slapped it on the mount and took the above images.  Once collimation was completed NO SCREWS WERE TOUCHED.

 

I am now even more impressed with TinySpeck's procedure than I was already.  I've not only duplicated my previous star test results, but I believe I've surpassed them.  This means there's more than dumb luck to my results.  This procedure works.

Excellent! Based upon your out-of-focus centered star and your artificially created star field, you nailed it. That's certainly as close as I ever bother to get with the DSI method.

dg401 I truly appreciate your dedication to testing and exercising this procedure. bow.gif 

Congratulations, TinySpeck. You've developed a bench RC collimation method requiring one inexpensive and common tool (two if you count the reading glasses lol.gif ).

I'd say the Holy Grail has been found!


  • TinySpeck and dg401 like this

#104 dg401

dg401

    Vostok 1

  • -----
  • Posts: 101
  • Joined: 11 Jul 2021

Posted 04 August 2021 - 01:10 AM

I'm starting to suspect that this will make a good closet scope.

That was me 96.5 hours ago.

 

I am surprised and impressed by what the eye and brain are capable of if challenged to their full potential.  If you're struggling with your GSO Ritchey–Chrétien scope and you just can't get it collimated, then you're in the same collimation Hell I was in from mid June until just recently when I ran across this crazy old retired guy from Seattle spouting collimation heresy.  7 weeks with a perfectly adjusted Farpoint laser and nothing to show for it.  Many sources insisting that the way out was to spend even more money on even more expensive gadgets.

 

Nope, just as TinySpeck claimed upon starting this topic, a Cheshire and an eyeball... then some repetitions to get the bugs worked out of the brain and the eyeball.

 

Also, let's dispose of the notion of the 1/20th turn.  That's 18 degrees on a circle and absolutely HUGE versus the finesse it takes to get the scope zeroed in.  We're into the realm of the "nudge".  I dunno, 1/360th of a turn, 1/200th of a turn?  However little or much it takes to get that dot centered.


Edited by dg401, 04 August 2021 - 01:52 AM.

  • TinySpeck and nebulachadnezzer like this

#105 quilty

quilty

    Aurora

  • -----
  • Posts: 4,806
  • Joined: 07 Oct 2019

Posted 04 August 2021 - 04:32 AM

"RC collimation with no laser or stars" Stargazing would be great if there weren't the flaming stars :-)

Honestly, I'm trying hard to do so on the CC6 which is pretty much the same to this regard without any optical or data processing device even without a chesire eyepiece. For I haven't got any of them.

 

Hall of mirrors. That is great, I think I notice every small tweak in there. For me this serves to align the primary mirror in the first place.

 

Then I see that the collimation mark on the secondary mirror is concentric through the focusser without eyepiece. Adjusting by the three Allen screws at the secondary. Again any tweak is noticeable. 

 

Then an iteration, first hall of mirrors then secondary (hoping and assuming the collimation ringmark marks the optical center of the mirror)

 

This daylight procedure seems to me very reproducable. And seems to lead me very close

 

The last tiny tweak, sorry, will be done at a true star like polaris, aware of the problems involved.

 

Maybe I skipped something of matter but I don't see.

 

I own an artificial star as well and would prefer it by far compared to a true star. My problem is that placing it in the backyard at 40 m distance would need all backfocus extensions and I'm afraid they could change the optical path laterally. If I had a spotlight in 200-300 m distance that would be it.


Edited by quilty, 04 August 2021 - 07:19 AM.


#106 TinySpeck

TinySpeck

    Messenger

  • ****-
  • topic starter
  • Posts: 474
  • Joined: 08 Oct 2017
  • Loc: Seattle area

Posted 04 August 2021 - 10:40 AM

Wow, superb results, @dg401!  I'm with @nebulachadnezzar in my gratefulness for your dedication to this and your careful data reporting.  I think you're right on with your observation about training the eye-brain system too.  You have to really keep raising the bar and looking as closely as you can, but when you get there you really can achieve top-notch collimation before you even head outside.

 

Other stuff has kept me from developing my Cheshire-cam, but I think when that's going it will make this process even easier.  I'll post here when that appears.


  • nebulachadnezzer and dg401 like this

#107 TinySpeck

TinySpeck

    Messenger

  • ****-
  • topic starter
  • Posts: 474
  • Joined: 08 Oct 2017
  • Loc: Seattle area

Posted 04 August 2021 - 10:43 AM

Congratulations, TinySpeck. You've developed a bench RC collimation method requiring one inexpensive and common tool (two if you count the reading glasses lol.gif ).

I'd say the Holy Grail has been found!

Many thanks, @neb!


  • nebulachadnezzer likes this

#108 TinySpeck

TinySpeck

    Messenger

  • ****-
  • topic starter
  • Posts: 474
  • Joined: 08 Oct 2017
  • Loc: Seattle area

Posted 04 August 2021 - 10:46 AM

"RC collimation with no laser or stars" Stargazing would be great if there weren't the flaming stars :-)

Honestly, I'm trying hard to do so on the CC6 which is pretty much the same to this regard without any optical or data processing device even without a chesire eyepiece. For I haven't got any of them.

 

Hall of mirrors. That is great, I think I notice every small tweak in there. For me this serves to align the primary mirror in the first place.

 

Then I see that the collimation mark on the secondary mirror is concentric through the focusser without eyepiece. Adjusting by the three Allen screws at the secondary. Again any tweak is noticeable. 

 

Then an iteration, first hall of mirrors then secondary (hoping and assuming the collimation ringmark marks the optical center of the mirror)

 

This daylight procedure seems to me very reproducable. And seems to lead me very close

 

The last tiny tweak, sorry, will be done at a true star like polaris, aware of the problems involved.

 

Maybe I skipped something of matter but I don't see.

 

I own an artificial star as well and would prefer it by far compared to a true star. My problem is that placing it in the backyard at 40 m distance would need all backfocus extensions and I'm afraid they could change the optical path laterally. If I had a spotlight in 200-300 m distance that would be it.

Thanks for the feedback, @quilty.  It's pretty impressive that you've gotten as far as you have without a Cheshire or apparently a focuser tilt adjustment!



#109 TinySpeck

TinySpeck

    Messenger

  • ****-
  • topic starter
  • Posts: 474
  • Joined: 08 Oct 2017
  • Loc: Seattle area

Posted 04 August 2021 - 11:54 AM

OK, now that it seems that this procedure works, I need help understanding exactly why.  What are the three steps aligning, exactly?  Is there some secret sauce here that aligns the optics even if the secondary donut is not on the optical axis or if the optics aren't perfectly aligned with the OTA?  I may be wrong in any or all of the following musings, but just to get the ball rolling:

 

Step one is centering the Cheshire dot in the secondary donut.  That's arranging secondary tilt so that the center of the donut is perpendicular to the axis between it and the Cheshire center.  This is not necessarily the secondary optical axis.

 

Step two: I think the Hall of Mirrors adjustment ensures that your mirrors are coaxial and parallel.  You're only adjusting primary tilt though, so for this to work I think your secondary optical axis must have been aligned with the Cheshire axis in step 1.  That would rely on the donut being the accurate optical center of the secondary.

 

Step three (focuser tilt) must bring the focuser (camera) axis in line with the coaxial mirror and Cheshire axis from above.

 

If I've got this right, then the procedure does depend on the secondary donut being on the optical axis of the secondary.  That's a big if!  What am I missing?



#110 nebulachadnezzer

nebulachadnezzer

    Ranger 4

  • -----
  • Posts: 305
  • Joined: 04 Jan 2021

Posted 04 August 2021 - 02:46 PM

If I've got this right, then the procedure does depend on the secondary donut being on the optical axis of the secondary.  That's a big if!  What am I missing?

I believe you are correct. 

To me this says a couple of things: 1) GSO manages to get the optical and physical center of the secondary the same in at least some cases such as dg401's scope, and/or 2) the procedure assumes the mirrors are centered in the OTA and optically coaxial, in which case centering the secondary means effectively zeroing the collimation of it to the OTA.

The Hall of Mirrors effect helps verify/adjust coaxial collimation of the two mirrors according to their optical centers. To me that's the "secret sauce" because you've found a tool-free way to do this step whereas most other procedures require either an expensive tool or a star (real or artificial). The question I've had was whether this was accurate enough. It appears that the answer is yes, if you are careful, methodical, and taking due consideration of the limitations of eyesight and the  intricacies of the eye/brain complex.

I think the key thing here is that the the mirrors must be optically coaxial in order to be fully collimated. Wouldn't the HoM effect appear unbalanced with two collimated mirrors that were *not* coaxial? Given the construction of these scopes it is a little hard to see how you could achieve good collimation with either mirror substantially off-center, because it's hard to get the other mirror to shift to compensate. Maybe I'm wrong. dg401 has had his in smaller pieces than I've had mine

If one of your mirrors is off-center, good luck to you. However, if they are centered within the OTA then I believe collimation should be possible in a way that requires no focuser tilt. (You might have to tilt your focuser if your camera's sensor is tilted, but then you have bigger problems.)

@dg401 are you finding you need to tilt your focuser with this procedure? Or did you zero it and call it done?


Edited by nebulachadnezzer, 04 August 2021 - 02:49 PM.

  • dg401 likes this

#111 dg401

dg401

    Vostok 1

  • -----
  • Posts: 101
  • Joined: 11 Jul 2021

Posted 04 August 2021 - 02:52 PM

OK, now that it seems that this procedure works, I need help understanding exactly why.  What are the three steps aligning, exactly?  Is there some secret sauce here that aligns the optics even if the secondary donut is not on the optical axis or if the optics aren't perfectly aligned with the OTA?  I may be wrong in any or all of the following musings, but just to get the ball rolling:

 

Step one is centering the Cheshire dot in the secondary donut.  That's arranging secondary tilt so that the center of the donut is perpendicular to the axis between it and the Cheshire center.  This is not necessarily the secondary optical axis.

 

Step two: I think the Hall of Mirrors adjustment ensures that your mirrors are coaxial and parallel.  You're only adjusting primary tilt though, so for this to work I think your secondary optical axis must have been aligned with the Cheshire axis in step 1.  That would rely on the donut being the accurate optical center of the secondary.

 

Step three (focuser tilt) must bring the focuser (camera) axis in line with the coaxial mirror and Cheshire axis from above.

 

If I've got this right, then the procedure does depend on the secondary donut being on the optical axis of the secondary.  That's a big if!  What am I missing?

The OTA is a red herring.  As long as the mirrors and focuser are all coaxial, it doesn't matter optically if the metal tube is canted, or a truss system, or not there at all and instead the mirrors are held in place by some futuristic force-field.  It might be awkward aiming a scope like this, but the optics don't care about the tube as long as it doesn't shadow the primary.

 

axis

  • AndresEsteban and TinySpeck like this

#112 dg401

dg401

    Vostok 1

  • -----
  • Posts: 101
  • Joined: 11 Jul 2021

Posted 04 August 2021 - 02:59 PM

I believe you are correct. 

To me this says a couple of things: 1) GSO manages to get the optical and physical center of the secondary the same in at least some cases such as dg401's scope, and/or 2) the procedure assumes the mirrors are centered in the OTA and optically coaxial, in which case centering the secondary means effectively zeroing the collimation of it to the OTA.

The Hall of Mirrors effect helps verify/adjust coaxial collimation of the two mirrors according to their optical centers. To me that's the "secret sauce" because you've found a tool-free way to do this step whereas most other procedures require either an expensive tool or a star (real or artificial). The question I've had was whether this was accurate enough. It appears that the answer is yes, if you are careful, methodical, and taking due consideration of the limitations of eyesight and the  intricacies of the eye/brain complex.

I think the key thing here is that the the mirrors must be optically coaxial in order to be fully collimated. Wouldn't the HoM effect appear unbalanced with two collimated mirrors that were *not* coaxial? Given the construction of these scopes it is a little hard to see how you could achieve good collimation with either mirror substantially off-center, because it's hard to get the other mirror to shift to compensate. Maybe I'm wrong. dg401 has had his in smaller pieces than I've had mine

If one of your mirrors is off-center, good luck to you. However, if they are centered within the OTA then I believe collimation should be possible in a way that requires no focuser tilt. (You might have to tilt your focuser if your camera's sensor is tilted, but then you have bigger problems.)

@dg401 are you finding you need to tilt your focuser with this procedure? Or did you zero it and call it done?

There is an implication here that really gets my mind racing considering the physical tube isn't a factor.  I suspect that as long as you have the mechanical motion range on the primary mirror to achieve a precise hall of mirrors with the secondary, the Cheshire will probably be nearly dead center after setting hall of mirrors.  You can't ignore the cheshire entirely as I suspect it has a bit better resolution than HOM.  Convergence is the end all be all here.  You can have the Cheshire aimed at the center of the secondary and have the primary off in whatever odd direction.  But with the following preconditions/assumptions to grant, 1. The focuser/primary are co-axial, and 2. there is such a thing as a "perfect" hall of mirrors, then achieving a perfect hall of mirrors irrespective doing anything with the secondary would achieve a perfect collimation.
 

Edit:  On the focuser question, I treat this as a procedure all to it's own.  I remove the primary/focuser assembly from the OTA, insert my Farpoint laser with it's beam reducer REMOVED for maximum brightness, the shine the beam onto a white card in a dark room.  I'm not trying to focus the dimmer reflection of the beam reflected off the primary onto the beam dot, I try to get it a little inside or outside of focus to create a round target around the laser dot.  Just imagine a tiny "x" drawn on the card with a 3 inch circle around it.  I just pick a distance to fill the circle with the out of focus reflection and look for the laser spot to be directly over the "x".  Done. Lock it down.

 

After that, it's Cheshire & hall of mirrors.  Not another thought is given to the primary/focuser alignment.


Edited by dg401, 04 August 2021 - 03:15 PM.


#113 dg401

dg401

    Vostok 1

  • -----
  • Posts: 101
  • Joined: 11 Jul 2021

Posted 04 August 2021 - 03:29 PM

The Hall of Mirrors effect helps verify/adjust coaxial collimation of the two mirrors according to their optical centers. To me that's the "secret sauce" because you've found a tool-free way to do this step whereas most other procedures require either an expensive tool or a star (real or artificial).

FULL STOP.  AGREE.  SECRET SAUCE HERE.

 

As I mentioned before, one can get a precise Cheshire centering on the secondary and the mirrors can still be far from coaxial.

 

When I was a little kid, my local barber shop had mirrors on both opposing walls with 3 barber chairs in front of each wall/mirror line.  You could look forever into that hall of mirrors... or until you lost it after a couple dozen reflections into the blur of optical imperfection or the ever increasing curve of not quite parallel walls... 

 

This procedure takes the hall of mirrors and pushes it from a parlor trick or a neat but only approximate alignment check into the realm of precision.  You can key in on whatever works best for you.  For me, it is that narrow "smile" shaped sliver of the 2nd reflection of the open end of the 2 inch focuser tube.  I walk all around the tube, rock my head forward and backward, and gauge the size of the sliver in all orientations.  It takes maddeningly fine adjustments, but the sliver can be made identical in all orientations.

 

So on the focuser end, its the Cheshire.  On the hall or mirrors, it's also the Cheshire... the grinning Cheshire cat 2nd reflection of the open end of the focuser tube.


Edited by dg401, 04 August 2021 - 03:30 PM.


#114 TinySpeck

TinySpeck

    Messenger

  • ****-
  • topic starter
  • Posts: 474
  • Joined: 08 Oct 2017
  • Loc: Seattle area

Posted 04 August 2021 - 03:50 PM

I believe you are correct. 

To me this says a couple of things: 1) GSO manages to get the optical and physical center of the secondary the same in at least some cases such as dg401's scope, and/or 2) the procedure assumes the mirrors are centered in the OTA and optically coaxial, in which case centering the secondary means effectively zeroing the collimation of it to the OTA.

The Hall of Mirrors effect helps verify/adjust coaxial collimation of the two mirrors according to their optical centers. To me that's the "secret sauce" because you've found a tool-free way to do this step whereas most other procedures require either an expensive tool or a star (real or artificial). The question I've had was whether this was accurate enough. It appears that the answer is yes, if you are careful, methodical, and taking due consideration of the limitations of eyesight and the  intricacies of the eye/brain complex.

I think the key thing here is that the the mirrors must be optically coaxial in order to be fully collimated. Wouldn't the HoM effect appear unbalanced with two collimated mirrors that were *not* coaxial? Given the construction of these scopes it is a little hard to see how you could achieve good collimation with either mirror substantially off-center, because it's hard to get the other mirror to shift to compensate. Maybe I'm wrong. dg401 has had his in smaller pieces than I've had mine

If one of your mirrors is off-center, good luck to you. However, if they are centered within the OTA then I believe collimation should be possible in a way that requires no focuser tilt. (You might have to tilt your focuser if your camera's sensor is tilted, but then you have bigger problems.)

@dg401 are you finding you need to tilt your focuser with this procedure? Or did you zero it and call it done?

@dg401 and I both would seem to have our secondary donuts exactly on the optical axis if that's a requirement here.  This seems a lot to ask of a cheap scope, but maybe GSO really does this somehow.  I wish they had a responsive tech support team who could weigh in here.

 

I agree on the HOM failing if the mirrors aren't coaxial; not from any optical principles, but intuitively this seems correct.  @dg401 is right that the OTA shouldn't have anything to do with collimation, but the mirrors do have to eventually reflect straight down the focuser tube.  That should be their only tether to the OTA. 

 

You don't need to have the mirrors on the same physical OTA center either, as long as you can adjust tilt on both of them.

 

But look what that implies: in order to make the mirrors coaxial you must adjust tilt on BOTH of them.  Here we're tilting the secondary to center the Cheshire pinhole reflection in the donut, and then the primary to pass the HOM.  This must be making the assumption that the donut is on the optical axis.

 

If we assume that the donut is NOT on the optical axis, I think this procedure fails.  You adjust the secondary for Cheshire alignment which throws off the HOM alignment, and you don't get convergence.

 

I hope I'm wrong about this, but I keep being pushed this way.  It really doesn't seem reasonable to require the donut to be so precisely on the optical axis.  I hope there's some other explanation which doesn't require that.


  • nebulachadnezzer likes this

#115 dg401

dg401

    Vostok 1

  • -----
  • Posts: 101
  • Joined: 11 Jul 2021

Posted 04 August 2021 - 04:21 PM

I just did a test where I loosened one of the secondary screws to push the Cheshire center out of the ring.

 

Try as I might with the primary, I just couldn't get the hall of mirrors right.  I could get it to be close, but there seemed to be some invisible "axis of impossibility" around which I was flailing.

 

So then I messed up both mirrors intentionally and went back to collimation:

 

1.  Approximately centered Cheshire dot

2.  Approximately centered hall of mirrors

 

Noted Cheshire dot fairly off center

 

3.  Precisely centered Cheshire dot.

 

Noted hall of mirrors fairly off center

 

4.  Precisely centered hall of mirrors

 

Noted Cheshire dot a bit off center

 

5.  Precisely centered Cheshire dot

 

Noted hall of mirrors a bit off center

 

6.  Precisely centered hall of mirrors

 

Noted Cheshire dot a hair off center

 

7.  Precisely centered Cheshire dot

 

Noted hall of mirrors a hair off center

 

8. Precisely centered hall of mirrors

 

Re-inspected Cheshire dot and re-inspected hall of mirrors.  Both precisely centered.

 

9. End



#116 TinySpeck

TinySpeck

    Messenger

  • ****-
  • topic starter
  • Posts: 474
  • Joined: 08 Oct 2017
  • Loc: Seattle area

Posted 04 August 2021 - 04:40 PM

I just did a test where I loosened one of the secondary screws to push the Cheshire center out of the ring.

 

Try as I might with the primary, I just couldn't get the hall of mirrors right.  I could get it to be close, but there seemed to be some invisible "axis of impossibility" around which I was flailing.

 

So then I messed up both mirrors intentionally and went back to collimation:

...

You're getting good at this!  Nice to be able to just whip out a collimation, isn't it?

 

I think this shows that the mirrors must be coaxial to pass HOM, which requires both of them to be tilted properly.  With one deliberately mis-tilted you can't get there.  Centering the dot in the donut seems to go along with getting the mirrors coaxial.  Does that mean the donut must be on the optical axis?  It seems like it, but I hope not.

 

EDIT: maybe your testing already proves this, but what would happen if you did the collimation but put the dot in a chosen location NOT centered in the donut?  That would be the same as if the donut were painted on a different spot.  I bet the procedure would not converge (or HOM would fail), and this confirms that the donut really does need to be on the optical axis.  If that's the case, then I do hope that GSO does donut-ize the optical axis so people can generally use this method.  Otherwise @dg401 and I are just very lucky.


Edited by TinySpeck, 04 August 2021 - 04:49 PM.

  • dg401 likes this

#117 nebulachadnezzer

nebulachadnezzer

    Ranger 4

  • -----
  • Posts: 305
  • Joined: 04 Jan 2021

Posted 04 August 2021 - 05:11 PM

You don't need to have the mirrors on the same physical OTA center either, as long as you can adjust tilt on both of them.

 

But look what that implies: in order to make the mirrors coaxial you must adjust tilt on BOTH of them.  Here we're tilting the secondary to center the Cheshire pinhole reflection in the donut, and then the primary to pass the HOM.  This must be making the assumption that the donut is on the optical axis.

 

If we assume that the donut is NOT on the optical axis, I think this procedure fails.  You adjust the secondary for Cheshire alignment which throws off the HOM alignment, and you don't get convergence.

 

I hope I'm wrong about this, but I keep being pushed this way.  It really doesn't seem reasonable to require the donut to be so precisely on the optical axis.  I hope there's some other explanation which doesn't require that.

This is my point. I think this method -- and the GSO RC design -- necessarily assumes that the optical center is the geometric/physical center of both mirrors. Further, it appears to be the case that GSO does a decent job of placing the secondary donut in the geometric -- and thus optical -- center.

It's a bad assumption to make, but at least on a couple of examples here it seems to be the case.

Yes, it's technically possible to have collimated mirrors that are completely independent of the physical axis of the OTA, as indicated in @dg401's diagram. I'm just not sure that GSO's RC OTA design allows for it. The mirror collimation mechanisms only allow for tilt relative to the physical axis (plus inter-mirror distance, of course). They don't allow for pan (perpendicular to the OTA physical axis). If the optical and physical axes differed then collimation with coaxial optics would seem to be impossible in the GSO design. You could collimate the mirrors, certainly, but they would not be coaxial.

I'm not sure if it's possible to induce error on pan in the GSO design. Perhaps one can jam the primary mirror over to one side while screwing the mount back together. (This is one good reason not to take it apart unless absolutely necessary!) Perhaps one can force the secondary to slide one way or the other on the collimation screws. This would introduce significant error. (NOTE: If this happens, a gravity assist is probably the best way to correct it.) The spider isn't adjustable, which differs from GSO's Newtonian designs, but I think that's a good thing. Imagine what accidentally twisting one of those thumbscrews would do to your collimation!

Your process doesn't include correcting these errors if they exist in the OTA. Nor, necessarily, should it. It appears to assume these issues don't exist. The evidence with at least two scopes seems to indicate this is a reasonable assumption.

What are your focuser tilt settings? If they are zero (as I would expect) then that seems to support what I'm saying. If they're non-zero I'm not sure what it tells us, because there could be a lot of reasons for that.

I can tell you that my DSI-collimated RC8 and RC10 both have the focuser collimation zeroed. I'm using Moonlite focusers and an imaging train that I've gone to great pains to ensure is devoid of tilt. My star tests seem to indicate success.



#118 TinySpeck

TinySpeck

    Messenger

  • ****-
  • topic starter
  • Posts: 474
  • Joined: 08 Oct 2017
  • Loc: Seattle area

Posted 04 August 2021 - 05:41 PM

...

Yes, it's technically possible to have collimated mirrors that are completely independent of the physical axis of the OTA, as indicated in @dg401's diagram. I'm just not sure that GSO's RC OTA design allows for it. The mirror collimation mechanisms only allow for tilt relative to the physical axis (plus inter-mirror distance, of course). They don't allow for pan (perpendicular to the OTA physical axis). If the optical and physical axes differed then collimation with coaxial optics would seem to be impossible in the GSO design. You could collimate the mirrors, certainly, but they would not be coaxial.

...

But correction for pan isn't necessary if you have tilt adjustment for both mirrors.  You can point them directly at one another regardless of their shift in a plane perpendicular to the OTA axis.  You do need to adjust the tilt of both mirrors though, in general.

 

...

What are your focuser tilt settings? If they are zero (as I would expect) then that seems to support what I'm saying. If they're non-zero I'm not sure what it tells us, because there could be a lot of reasons for that.

...

I did have to adjust my focuser.  The Cheshire margin and the outermost ring of light were both definitely off-center until I did.


  • nebulachadnezzer likes this

#119 nebulachadnezzer

nebulachadnezzer

    Ranger 4

  • -----
  • Posts: 305
  • Joined: 04 Jan 2021

Posted 04 August 2021 - 06:06 PM

But correction for pan isn't necessary if you have tilt adjustment for both mirrors.  You can point them directly at one another regardless of their shift in a plane perpendicular to the OTA axis.  You do need to adjust the tilt of both mirrors though, in general.

 

I did have to adjust my focuser.  The Cheshire margin and the outermost ring of light were both definitely off-center until I did.

Ok, thinking about this... You can certainly reach collimation while the mirrors aren't aligned with the physical axis of the scope, but since the mechanism only allows tilt relative to the physical axis (by design, anyway) this results in some lateral misalignment of the mirrors. They aren't physically coaxial even if they are optically coaxial, and presumably the baffle and focuser are sitting in the physical center of the primary.

Tilting the focuser can compensate for this misalignment to the OTA, but that locks down the rotation of the focuser which is an undesirable situation (unless you have a model with separate focuser flange collimation adjustments like the GSO truss models). This also results in some slight pan error of the camera vs the optical center of the OTA, but that's not necessarily going to cause visual artifacts (except for putting one edge of the sensor farther away from the sweet spot of the optical system).

However, reaching collimation this way seems to conflict with the necessary (for this method to work) assumption that the secondary spot corresponds to the optical axis of the mirror. Doesn't it?

I can see why you would need to iterate through steps 2-4 of your process with all three of these adjustments in play. However it seems undesirable to have to do so because steps 2 and 4 are somewhat in conflict with each other.

Is the secondary spot correct? Is the Cheshire centered? If either of these is uncertain, there's a problem.

I haven't run through this process, and obviously you have, but I'm not clear how you differentiate between an off-center Cheshire and a tilted set of mirrors. Perhaps you can do so by rotating the focuser/Cheshire, but now we're back to the issue of focuser collimation and why it's undesirable to need to do this (if it can be avoided).


  • TinySpeck likes this

#120 MikeECha

MikeECha

    Viking 1

  • *****
  • Posts: 786
  • Joined: 15 Sep 2018
  • Loc: Charlotte, NC

Posted 04 August 2021 - 06:18 PM

"..If we assume that the donut is NOT on the optical axis, I think this procedure fails.  You adjust the secondary for Cheshire alignment which throws off the HOM alignment, and you don't get convergence..."

That is the reference that is always missing. And that is the reference the laser beam falling back on it self can give you to at least start the collimation process.

 

I can not count how many evenings I spent chasing my tail before I discovered I was not installing the focuser correctly on the scope.

 

But I would not be worried about the donut not being centered. That donut is most likely edged with a machine. That is one of the very few good things about mass produced products is that any decent machine today is light yeas ahead in precision and repeatability to one the same in 1970. I have a 10" dob I bough a  year or so ago at Cosco that has a very nicely edged donut on the primary.

 

On a bigger picture, this successful collimation shows how many inaccuracies are out there about the design of these scopes.

 

If anybody would like to evaluate my collimation results, I have posted them on the link below 

 

https://www.cloudyni...elling-me-here/

 

Frankly, I am skeptical about them. The symmetry of the CCDI graph looks too good to be true. Please take a look.



#121 dg401

dg401

    Vostok 1

  • -----
  • Posts: 101
  • Joined: 11 Jul 2021

Posted 04 August 2021 - 06:36 PM

That is the reference that is always missing. And that is the reference the laser beam falling back on it self can give you to at least start the collimation process.

 

I can not count how many evenings I spent chasing my tail before I discovered I was not installing the focuser correctly on the scope.

 

But I would not be worried about the donut not being centered. That donut is most likely edged with a machine. That is one of the very few good things about mass produced products is that any decent machine today is light yeas ahead in precision and repeatability to one the same in 1970. I have a 10" dob I bough a  year or so ago at Cosco that has a very nicely edged donut on the primary.

 

On a bigger picture, this successful collimation shows how many inaccuracies are out there about the design of these scopes.

 

If anybody would like to evaluate my collimation results, I have posted them on the link below 

 

https://www.cloudyni...elling-me-here/

 

Frankly, I am skeptical about them. The symmetry of the CCDI graph looks too good to be true. Please take a look.

I'm starting toward the opinion that all this statistical analysis is about on par with running rose petal extract through a GC/mass spectrometer to find out why it smells nice.  It perhaps misses the point.

 

Are your defocused stars round?  Do they focus to pinpoints?  Is the corner frame astigmatism balanced?

 

I think I'm a little crazy right now because with all the analysis, I have to ask myself "When is the last time I actually imaged something to produce a nice photo?"  That's partly because of all the clouds lately, and the smoke... but a good bit of it is due to picking apart results.  Not a bad thing, but there comes a time to smell the roses, or image the heavens.
 


  • TinySpeck likes this

#122 MikeECha

MikeECha

    Viking 1

  • *****
  • Posts: 786
  • Joined: 15 Sep 2018
  • Loc: Charlotte, NC

Posted 04 August 2021 - 07:19 PM

I'm starting toward the opinion that all this statistical analysis is about on par with running rose petal extract through a GC/mass spectrometer to find out why it smells nice.  It perhaps misses the point.

 

Are your defocused stars round?  Do they focus to pinpoints?  Is the corner frame astigmatism balanced?

 

I think I'm a little crazy right now because with all the analysis, I have to ask myself "When is the last time I actually imaged something to produce a nice photo?"  That's partly because of all the clouds lately, and the smoke... but a good bit of it is due to picking apart results.  Not a bad thing, but there comes a time to smell the roses, or image the heavens.
 

I know what you mean.

 

NINA consistently focuses with R^2= >.95 and the stars look "pinpointy" within reason as I am doing this @ .36 arcsec/pix. The defocused stars look very round on the corners and center and I have to go up to 100% zoom to see a very slight out of concentricity/flare on the outer ring on corner stars. So visually my "rose petal extract smells good" to me.

 

However, eliminating the visual (accuracy or inaccuracy) form the process is my "holy grail". I am going the route of "redesign" the secondary holder. The defocused picture I posted there was taken with the primary completely retracted (i.e. reset) and zero adjustment to it. Only secondary adjustments with cheshire to center the dot were made. Not even laser this time. So I think I am close. I just need to spend some time designing an "attachment" to precisely control the installation and removal of the secondary.

 

I am glad you got it with the HoM but that never worked for me. Probably my eyes are to blame.



#123 dg401

dg401

    Vostok 1

  • -----
  • Posts: 101
  • Joined: 11 Jul 2021

Posted 04 August 2021 - 08:47 PM

Experimental determination of the optical center of a convex elliptical/spherical, parabolic,or hyperbolic mirror (I think that covers all the conic sections).

 

1.  We don't know where the optical center of the RC secondary mirror is, but we know that GSO does a good job figuring the mirrors, otherwise we wouldn't bother with these scopes at all.

 

2.  A laser reflecting off of the secondary mirror at a fixed circular radius away from the optical center will be reflected at an identical angle.  If we can control the parameters, we can use this fact to determine the optical center experimentally.

 

Here is the setup:

 

setup

 

We have my RC6 scope with the secondary cell removed.  It is mounted and stationary with a well-collimated Farpoint laser mounted in the focuser.  We have the RC6 secondary cell mounted on a photo tripod ball head with a hand grip to allow adjustment of position.  We have the ability to rotate the secondary mirror via it's distance adjustment ring (as long as we are careful not to rotate it too far out and have the whole thing fall to the floor).

 

I have the laser striking the secondary a distance from the center mark (around 18mm away from the center mark).  I have the secondary adjusted so that the beam is reflected back directly on top of the laser source.  I mark the spot where the beam is hitting the secondary with a Sharpie.  I rotate the secondary 60°. If I simply rotate and mark every 60°, I'll just be making marks equidistant from the MECHANICAL center.  That doesn't do me any good.  But here's the trick:  I make a fine adjustment with the hand grip to re-center the reflection back on the laser source with each 60° rotation.  It only takes a very small adjustment. This small centering adjustment creates the offset between the mechanical center and the center of the circle you've already figured out that I'm putting together that rings the optical center.  I mark the spot again where the beam hits the secondary.  I repeat this process until I've made 6 roughly equally spaced marks on the secondary, re-centering the reflection on the laser each time with a small adjustment.  I photograph the secondary.

 

I wipe away the marks with some 91% isopropanol because Sharpie on my secondary?  Yuck.

 

So now we have the photo of the secondary:

 

center
 
Making a fine mark on a mirror with a Sharpie isn't as easy as you would think.  Sometimes I nailed it, other times I boogered it up.  Where the mark is more a blob, I tried to make the center of the blob correspond to the laser dot.  I overlaid a visual "best fit" circle over the marks on the secondary.  I bisected the circle with 2 perpendicular lines across the radius, and there we have a sloppily experimentally determined optical center.
 
Yes, there is a LOT of room for improvement and I've not convinced myself that all assumptions are fundamentally sound.  But the experiment did a pretty good job of approximating a circle, and the center did have an offset from the mechanical center mark of about 1mm.  Not an authoritative test and pretty sloppily constructed.  But perhaps not entirely useless either.

 

 

 



#124 MikeECha

MikeECha

    Viking 1

  • *****
  • Posts: 786
  • Joined: 15 Sep 2018
  • Loc: Charlotte, NC

Posted 04 August 2021 - 09:48 PM

 

Experimental determination of the optical center of a convex elliptical/spherical, parabolic,or hyperbolic mirror (I think that covers all the conic sections).

 

1.  We don't know where the optical center of the RC secondary mirror is, but we know that GSO does a good job figuring the mirrors, otherwise we wouldn't bother with these scopes at all.

 

2.  A laser reflecting off of the secondary mirror at a fixed circular radius away from the optical center will be reflected at an identical angle.  If we can control the parameters, we can use this fact to determine the optical center experimentally.

 

Here is the setup:

 

 

 

We have my RC6 scope with the secondary cell removed.  It is mounted and stationary with a well-collimated Farpoint laser mounted in the focuser.  We have the RC6 secondary cell mounted on a photo tripod ball head with a hand grip to allow adjustment of position.  We have the ability to rotate the secondary mirror via it's distance adjustment ring (as long as we are careful not to rotate it too far out and have the whole thing fall to the floor).

 

I have the laser striking the secondary a distance from the center mark (around 18mm away from the center mark).  I have the secondary adjusted so that the beam is reflected back directly on top of the laser source.  I mark the spot where the beam is hitting the secondary with a Sharpie.  I rotate the secondary 60°. If I simply rotate and mark every 60°, I'll just be making marks equidistant from the MECHANICAL center.  That doesn't do me any good.  But here's the trick:  I make a fine adjustment with the hand grip to re-center the reflection back on the laser source with each 60° rotation.  It only takes a very small adjustment. This small centering adjustment creates the offset between the mechanical center and the center of the circle you've already figured out that I'm putting together that rings the optical center.  I mark the spot again where the beam hits the secondary.  I repeat this process until I've made 6 roughly equally spaced marks on the secondary, re-centering the reflection on the laser each time with a small adjustment.  I photograph the secondary.

 

I wipe away the marks with some 91% isopropanol because Sharpie on my secondary?  Yuck.

 

So now we have the photo of the secondary:

 

 
 
Making a fine mark on a mirror with a Sharpie isn't as easy as you would think.  Sometimes I nailed it, other times I boogered it up.  Where the mark is more a blob, I tried to make the center of the blob correspond to the laser dot.  I overlaid a visual "best fit" circle over the marks on the secondary.  I bisected the circle with 2 perpendicular lines across the radius, and there we have a sloppily experimentally determined optical center.
 
Yes, there is a LOT of room for improvement and I've not convinced myself that all assumptions are fundamentally sound.  But the experiment did a pretty good job of approximating a circle, and the center did have an offset from the mechanical center mark of about 1mm.  Not an authoritative test and pretty sloppily constructed.  But perhaps not entirely useless either.

 

Even with all the inaccuracy in making, you can see a trend to be at the center. Again, I am pretty sure that mass produced secondary mirrors are not center spotted by hand, not in the 21 century. 



#125 dg401

dg401

    Vostok 1

  • -----
  • Posts: 101
  • Joined: 11 Jul 2021

Posted 04 August 2021 - 10:48 PM

Even with all the inaccuracy in making, you can see a trend to be at the center. Again, I am pretty sure that mass produced secondary mirrors are not center spotted by hand, not in the 21 century. 

The center spot on my RC6 is at mechanical dead center (referenced against the mirror locking ring).




CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics