15 replies to this topic

[MellonLake]

So I was pondering all things collimation this morning and got to thinking about autocollimators and lasers and secondary collimation.   Then it kind of occurred to me that a laser could be used with an autocollimator.  Basically create an eyepiece with a mirrored flat interior surface with an aperture that the laser passes through.  The beam would go to the primary come back and reflect of the mirrored surface and back to the primary (and back to the collimator.....)  You could then collimate the secondary to make the laser spots converge.

 

My thought is that there are a few problems with this:

1) The reflectivity of the autocollimator would need to be limited such that we do not get too many reflections (like 50% reflective).  

2) The visual autocollimator may be better as the brightness of the laser is a factor in how easy it is to see the reflections.  Also, the bright primary beam may wash out the reflections. 

3) Using a visual autocollimator you are actually making the centre spot reflections converge i.e. looking at a big object to make it converge which may be visually easier and more accurate than making a couple laser points converge.

 

Has anyone tried to make a laser autocollimator for a newtonian? 

 

  

[MellonLake]

So in the last hour or so, I played a little more with secondary collimation...  

 

I was actually doing a secondary collimation using my Farpoint laser with my Glatter Parallizer and 1/2 of a 2" variable polarizing filter attached to the Parallizer to reduce the laser intensity. 

 

FYI...The laser is a polarized light source and if you rotate the laser with the polarizer in you can get from essentially 0% to 90% transmission.  This works to dim the laser making it easier to collimate the secondary (due to the laser being too bright).    

 

Doing this I also noted that the reflected laser beam will return back to the source, reflect back off the back side of the polarizer, return to the primary, and create a second dot on on the primary.  This is just like an autocollimator. 

 

So I then made the reflection of the laser converge rather than just centring the beam.  This resulted in the beam being very slightly to the one side of dead centre of the mirror maker (which suggest my mirror marker is off centre by about 0.5mm).

 

So... Any thoughts on how accurate this method is?    

[Jason D]

There are fundamental problems with the hypothetical tool you have described:

1- When all ongoing and incoming laser beam reflections converge, it means the converged laser beam is aligned with the center of curvature (COC) of the primary mirror. It does not mean the scope is collimated. The converged beam has to strike the primary mirror at the exact center for axial alignment to be achieved. 

2- When collimation is close to be met, the laser beam will mostly overlap with the mirror aperture. That is, when you are close to collimation, the mirror will have little effect.

Jason

[MellonLake]

Ok so I am a bit confused.

 

1) Are not the centre of curvature of the mirror and the mirror centre point supposed to be coincident? If not, why not?

2) Is not the primary mirror axis defined by the parabola and hence pass through the centre of curvature rather than the mirror centre?

3) The primary mirror axis should be aligned with the focuser axis by collimation process. Unless I am missing something, would that not mean that the incoming beam (and reflections) needs to hit the centre of curvature rather than the mirror centre?

4) Is this not what an autocollimator does? It ensures the focuser axis (via the secondary) is pointed adequately at the centre of curvature rather than the mirror centre. In the autocollimation process the light does not know where the "centre" of the primary is, it only knows where there reflections are going, you are at that point directing the light to reflect back by pointing the secondary at the centre of curvature to make the reflections converge. Is this not why the autocollimator is more accurate, it points at the centre of curvature rather than the mirror centre which may be slighly in accurate. Again unless I am missing something.

 

Please direct me where my thought process has gone wrong. 

Edited by MellonLake, 29 November 2020 - 11:43 AM.

[Vic Menard]

Doing this I also noted that the reflected laser beam will return back to the source, reflect back off the back side of the polarizer, return to the primary, and create a second dot on on the primary.  This is just like an autocollimator. 

No, it's not just like an autocollimator. An autocollimator uses images to align the axes. A laser uses a non imaging light beam to show the relative tilt of the optical components.

 

So I then made the reflection of the laser converge rather than just centring the beam.  This resulted in the beam being very slightly to the one side of dead centre of the mirror maker (which suggest my mirror marker is off centre by about 0.5mm). 

Unfortunately, you can make your method work no matter where the outgoing laser beam contacts the primary mirror. This is why it's important to ensure that the outgoing beam should be centered relative to the primary mirror center marker/donut.

[MellonLake]

Vic;

  That makes sense.  The laser is not an image but a beam and will be reflected back to the focuser regardless of where it hits the mirror provided the mirror surfaces are all orthogonal (not sure that is the correct term).  The laser does not require an image to be formed by the parabola of the mirror and hence is not truly collimating the telescope.

 

I am still a bit perplexed though.  If we do a thorough collimation of both the primary and secondary (prior to this laser autocollimation technique) is not the angle of the primary then set so it is pointing at the focuser.  At this point the only spot on the primary that should reflect the beam back to the focuser axis is the centre of curvature?  Then if the dot of the reflected beam is not coincident to the dot of the primary beam (on the primary) does it not mean that the secondary is not pointed at the centre of curvature of the mirror and not correctly collimated?  

[Vic Menard]

...If we do a thorough collimation of both the primary and secondary (prior to this laser autocollimation technique) is not the angle of the primary then set so it is pointing at the focuser.  At this point the only spot on the primary that should reflect the beam back to the focuser axis is the centre of curvature?

With a simple thin beam laser, for all practical purposes, when the axial alignment is correct, it doesn't matter if the primary is a parabola, a sphere, or flat. If there's a residual axial error with the outgoing beam, then the curvature of the primary mirror sees the laser beam originating from the focal point, and reflects it back parallel to the optical axis--unless the primary mirror tilt is adjusted to align the outgoing beam to the laser emitter, which then would align the primary mirror axis to the center of curvature. Remember, when you adjust he primary mirror tilt, you are also moving the COC. That said, the outgoing beam originates from the laser aperture, and to get the return beam to align to that aperture, the outgoing and return beams have to overlap/coincide.

 

...if the dot of the reflected beam is not coincident to the dot of the primary beam (on the primary) does it not mean that the secondary is not pointed at the centre of curvature of the mirror and not correctly collimated?  

I'm not sure I understand what you're asking here. I'm not sure how the dot of the reflected beam cannot be coincident with the dot on the primary mirror, since one originates from the other.

[MellonLake]

Vic;

I am definitely wrong in how I initially describe this and making it confusing as such.  I probably started off with the wrong hypothesis of how the system was working and mucked up the description.    

 

 

Here is the process I am suggesting;

 

1) Collimate the secondary so the laser is centred in the primary mirror marker (no variable polarizer).

2) Collimate the primary via a Cheshrie or Barlowed laser.

 

At this point the system is set up with as I would normally collimate.  

 

Then I put in the laser and the single variable polarizer.  When I put the laser with the single Variable polarizer in, I see three laser dots on the primary:

 

a) The incoming beam dot from the laser.

b) A second slightly less bright dot to the side of the primary beam dot (This is Reflection #1 off of the back of the variable polarizer)

c) A third dim dot diametrically opposite Reflection #1 (This is the second reflection, Reflection #2, back off of the variable polarizer)

 

The picture below shows an exaggeration of what I see (but you can't see the second reflection because the camera does not pick it up).  The second dot to the right of the brighter dot is "reflection #1  

 

3) Then, I very slightly change the tilt/rotation of the secondary and the dots converge into a single dot. 

4)  Repeat 2 and 3 until complete the dots converge and the primary is fully collimated.

 

This sounds an a lot like what is happening with an autocollimator.

 

To quote Catseye..."An AUTOCOLLIMATOR - This is typically a 1" long, 1.25" diameter tube open at one end with a peephole at the other end.  There is a round, flat mirror on the backside of the peephole inside the tube with the mirrored surface facing toward the open end of the tool.  A small amount of the mirrored surface has been removed at the center of the mirror to allow sight through the peephole.  This tool operates on the principal that 2 mirrors (the primary and the autocollimator) faced opposite each other will reflect an image between one another, and if the angle of incidence of both mirrors is "exactly" 90 degrees (i.e. collimated), the reflected "ghost" images will merge together."

 

Unless I am missing something, I am pretty sure this is what I am doing with the laser and the Variable polarizer. 

 

If I twist a spider vane slightly I see the dots separate and when I let go they come back together (the same thing would happen with an autocollimator).  This really sounds like the laser and polarizer are working together as an Autocollimator. 

 

I would bet an actual autocollimator is still more accurate because the laser is so brigth and spot so large that it is hard to get the dots perfectly on top of one another... but I am pretty confident this system is working as an autocollimator.  

 

Edited by MellonLake, 29 November 2020 - 02:55 PM.

[Starman1]

You assume:

--the two surfaces on the polarizer are parallel

--that there are not multiple reflections from deeper layers in the VPF

--that the laser is perfectly collimated

--that the laser beam is not diverged when it passes through the VPF on the way out.

--that the reflection from the polarizer is coincident with the focuser axis as defined by the laser

--that the center of the center marker is or isn't in the center of the mirror (which can be determined with a ruler or a template of some sort)

If it is, and the coincident dots are not in the center, then the laser is not perpendicular to the filter or the filter isn't perpendicular to the focuser, or both.

If it isn't, that adjusting the beams to be coincident means you are adjusting the beam to the center of the mirror (may not be the case).

 

I can see a few ways to test this:

First, make sure the filter, when installed after the laser has already indicated the focuser axis is collimated, does not cause the beam to move on the primary.  If it does, your test won't tell you much of anything.

1) rotate the VPF, but not the laser, to see if the reflected laser dot makes a circle around the main beam if it is not coincident.

2) rotate the laser, but not the VPF to see what happens to both the main beam and the reflected dot.

3) rotate both to see if the pair rotates around a common center or makes a larger circle.

 

Autocollimators that fail (and I've had several that did) have a reflective surface that is not perpendicular to the axial focuser line.

One was collimatable, but had such a large center hole that parallax made the reading poor.

 

I will suspect that your filter is not perpendicular to the focuser axis defined by the laser.

[MellonLake]

Don;

 The image above is when I specifically de-collimate the secondary.   

 

I already checked all of your suggestions.  After I have fully collimated and when I put the laser in the dot does not move and if the laser is in the ideal spot there is no second dot.  

 

1) If I rotate the VPF the dots don't move (if the secondary is de-collimated or correctly collimated).  .

2) If I rotate the laser the dots don't move (either collimated or de-collimated).  They stay put.  I have checked the laser accuracy repeatedly in the Glatter with and without the VPF.

3) If I I rotate both the dots do not diverge significantly. 

 

If I loosen off the VPF the dots spread apart indicating that the VPF is no longer square to the focuser axis.  As such, I am very sure the VPF is square to the focuser axis. (It sits very very rigidly when fully tightened in)    

 

In general I am quite sure this is working as an autocollimator.  I think this is more interesting than practical because I don't think I can see the reflections accurately enough to collimate better than what an autocollimator can do.  

 

This video shows me twisting the spider vane and effect of the dots moving apart due to de-collimation. 

 

Rob  

[MellonLake]

Also, I am now finding the beam is dead centre and not the 0.5mm I noted at the start of the the thread, my primary collimation was very slightly off and when I went through the process a couple of times converged on the centre of the primary marker.  

 

Rob

[Vic Menard]

...I am pretty confident this system is working as an autocollimator.  

 

OK--put an autocollimator in an axially aligned Newtonian focuser.

Now, carefully decollimate ONLY the primary mirror.

You will see the primary mirror center marker and three reflections. The alignment of the primary mirror center marker relative to the reflection behind it magnifies any residual focuser axial error 2X. The distance between the flanking reflections magnifies the primary mirror axial error 8X.

 

...This tool operates on the principal that 2 mirrors (the primary and the autocollimator) faced opposite each other will reflect an image between one another, and if the angle of incidence of both mirrors is "exactly" 90 degrees (i.e. collimated), the reflected "ghost" images will merge together."

 While this is true with a corrected axial alignment (no axial error), the (single pupil*) autocollimator's precision is realized when there IS an axial error.

 

The precision of all tool axial alignments is dependent on the axial reference--in this case, the primary mirror center marker (the tool alignment relative to the focuser is assumed to be coaxial). The autocollimator uses the alignments of the primary mirror center marker, its reflections, and its image. Laser alignment is dependent on its outgoing alignment relative to the primary mirror center marker (for the focuser axis), and its alignment using the return beam (the reflection of the focuser axis) to the laser emitter (for the primary mirror axis), with one half of any residual focuser axis error added to the primary mirror axial read. The outgoing beam alignment read is magnified 1X, and the return beam read is 2X. Because the primary mirror is not flat, the multiple reflection errors you see are not easily interpreted (2X, 4X, 8X for which axis).

 

*The 2-pupil Infinity XLK shows a precision alignment signature for axial parallelism (4X) when the axial alignment is already correct.

[Vic Menard]

Don;

 The image above is when I specifically de-collimate the secondary. 

Just to be clear, when you de-collimate the secondary, the primary will also be decollimated. The question becomes, relative to the primary mirror center marker, what do the reflections show? Since your decollimation occurred at the secondary mirror, it must be some blend of both axes. But the more critical alignment is usually the primary mirror axial alignment, which happens at the laser emitter (and is easily improved with a Barlowed laser procedure or another Cheshire derivative with the read error magnified 2X). 

[MellonLake]

The precision of all tool axial alignments is dependent on the axial reference--in this case, the primary mirror center marker (the tool alignment relative to the focuser is assumed to be coaxial). The autocollimator uses the alignments of the primary mirror center marker, its reflections, and its image. Laser alignment is dependent on its outgoing alignment relative to the primary mirror center marker (for the focuser axis), and its alignment using the return beam (the reflection of the focuser axis) to the laser emitter (for the primary mirror axis), with one half of any residual focuser axis error added to the primary mirror axial read. The outgoing beam alignment read is magnified 1X, and the return beam read is 2X. Because the primary mirror is not flat, the multiple reflection errors you see are not easily interpreted (2X, 4X, 8X for which axis).

 

Got it... This makes sense, as when I move the beam of axis the #1 reflection dot moves much further off axis than the incoming beam spot indicating a factor to the error (i.e. probably 3X, i.e. number of beam passes to produce the reflection dot on the primary,  it is probably an odd factor and not even).    

 

Just to be clear, when you de-collimate the secondary, the primary will also be decollimated. The question becomes, relative to the primary mirror center marker, what do the reflections show? Since your decollimation occurred at the secondary mirror, it must be some blend of both axes. But the more critical alignment is usually the primary mirror axial alignment, which happens at the laser emitter (and is easily improved with a Barlowed laser procedure or another Cheshire derivative with the read error magnified 2X). 

I realize that both will be de-collimated.   Yes the error will be a sum of the de-collimation of both axis but as you iterate between collimating the primary (Barlowed laser/Cheshire) and secondary (my laser autocollimator method) the residual error will go to zero just like you would get with a visual autocollimator (but not nearly as accurate). 

 

Again, I acknowledge this is really just an experiment at this point because I really can't see the beam with the same accuracy as you would see the centre marker in the visual autocollimator.     

[Vic Menard]

...Again, I acknowledge this is really just an experiment at this point because I really can't see the beam with the same accuracy as you would see the centre marker in the visual autocollimator.     

It's not just "seeing" the beam--the accuracy is linear. If the outgoing beam misses the center of the primary mirror center marker, the error is some linear distance (perhaps a millimeter or two). If the return beam misses the emitter by some distance, say 2mm, the actual primary mirror error is 1/2 this amount plus any residual error at the primary mirror center marker.

 

If the primary mirror was flat, a given error would be magnified 2X for each reflection pass (2X, 4X, 8X, etc.), but then you would have to divide the error by the passes to determine the actual error. But the primary isn't flat, which changes the way the error propagates. Worse yet, since the error is a combination of both axes (and these contributing axial errors have both distance and direction), determining the critical primary mirror error is unlikely to benefit from multiple passes.

 

FWIW, I suspect you'll get all the precision you need for your XT10 and 120ST using the outgoing laser beam/Parallizer for the focuser axis (secondary mirror tilt) and a Barlowed laser or other Cheshire derivative for the primary mirror axis.

[MellonLake]

Thanks Vic, I got it.  

 

I know the laser/parallizer is fine (even with the Coma Corrector) for the secondary and the Cheshire is fine for the primary (or Barlowed laser).  It was just an interesting observation I made while I was playing with using half the variable polarizer to reduce the brightness of the laser beam (which really helps to get it centered).  I am an engineer and like thinking about such things.  Collimation is one of those things that the more I think about it the more I understand it.  

 

Rob  

Edited by MellonLake, 29 November 2020 - 07:04 PM.