434 replies to this topic

[Jason D]

Towards the end of this lengthy thread, I promised to start a new concise thread, and hopefully less intimidating, about conventional autocollimators and improvements (2-pupil and CAM).

The autocollimator is by far the most misunderstood collimation tool. Not only misunderstood by many users but also misunderstood by some of the vendors who manufacture them. One vendor explains that achieving a darkened background implies perfect collimation – very wrong. Another vendor explains how you can achieve perfect collimation by a final tweak to the secondary mirror to “stack” all reflections – very wrong. Many believe miscollimation errors are magnified which each additional round trip reflections between the autocollimator and primary mirrors – very wrong.

Reflections formed when light bounces back and forth between a flat mirror and a concaved mirror are very different from reflections formed when light bounces back and forth between two flat mirrors which seems to be one source of the confusion. Understanding how autocollimator reflections are formed and figuring out how to get the most out of an autocollimator is not intuitive. Fortunately, most of the autocollimator secrets have already been revealed by the work of Nils Olof Carlin, Vic Menard, and Jim Fly. Nils Olof Carlin was the first to publish a thorough mathematical analysis of the autocollimator reflections (link). Vic Menard was the first to publish the optimal procedure for using autocollimators (link). Jim Fly was the first to simulate and publish autocollimator images and animations to improve understanding of the tool.

The objective of this thread is to explain the autocollimator and improvements with many supporting photos, illustrations, and animations. It took quite a bit of time and effort in my part to complete this thread.

[Jason D]

Let us start off with reviewing the fundamentals of collimation:

Collimation consists of:
1- Secondary mirror alignment: Optimizes illumination distribution within the FOV – less critical
2- Axial alignment : Minimizes “coma” and brings the whole FOV to focus – more critical

Secondary mirror alignment involves centering and rounding the secondary mirror under the focuser. Doing so will optimized the 100% illumination field placement within the eyepiece FOV. Any star located within the 100% illumination area will have its entire reflected light cone intercepted by the secondary mirror. Star located outside the 100% illumination area will have their light cone partially eclipsed by the secondary mirror edge; subsequently, these star images will underutilize the primary mirror aperture. The following animation shows how a centered star and an adjacent one have their entire reflected light cone intercepted by the secondary mirror. However, a third star which is way off-axis has its reflected cone eclipsed by the secondary mirror edge.



Referring to the next animation, centering the secondary mirror under the focuser will center the 100% illumination field within the FOV – an optimal setup.



The following illustration explains the “good”, the “bad”, and the “ugly” for secondary mirror alignment.
1- The “good” diagram shows a centered 100% illumination in the FOV – ideal.
2- The “bad” diagram shows an off-centered 100% illumination in the FOV. Visual observers can live with it but not astrophotographers.
3- The “ugly” diagram shows an off -FOV 100% illumination. The reflected light-cone for the on-axis star will be partially eclipsed by the secondary mirror edge. This setup will not utilize the primary aperture to its fullest.



The autocollimator, which is the main subject of this thread, is NOT the proper tool to align the secondary mirror. That is the job of a quality sight-tube or a quality holographic laser collimator. Refer to the following thread for more insight about secondary mirror alignment – just the first few pages of the thread.

[Jason D]

Secondary mirror misalignment as was described in the previous post (even the “ugly” case) will not introduce the infamous on-axis view killer known as coma. Axial misalignment is the main culptrit for on-axis coma. Attaining axial alignment using an autocollimator will be the main subject for the remainder of this thread.

Axial alignment involves:
1- Coinciding the focal points of the primary mirror and the eyepiece
2- Coinciding the focal planes of the primary mirror and the eyepiece

If we could see the focal points and focal planes of the primary mirror and the eyepiece then axial alignment would be a breeze. Unfortunately, we can’t. We use indirect means to achieve axial alignment. Coinciding the primary mirror and the focuser axes when using a quality focuser implies the eyepiece’s focal plane/point will be well-aligned with the primary’s mirror focal plane/point.

Refer to the following:
The “Good” diagram is ideal. Both the eyepiece and primary mirror focal points and planes coincide.
The “Bad” diagram might be acceptable for visual observers but might not be for astrophotographers. Both eyepiece/primary focal points coincide but the focal planes are at an angle (or tilted)
The “Ugly” diagram shows a bad case where the focal points of the eyepiece and primary are separated laterally. Even a 1mm separation is enough to ruin the view of a fast reflector.



Different collimation tools are available to guide us to coincide both the focuser and primary axes. There are four fundamental collimation axial errors targeted by different collimation tools:

CAE (COC Axial Error) corresponds to the distance between the focuser/primary axes at the COC plane. COC stands for Center Of Curvature. COC point is located on the primary mirror axis at twice the focal length distance. It is the virtual center of the sphere which the primary mirror surface curvature follows.

PAE (Primary Axial Error) corresponds to the distance between the focuser/primary axes at the focal plane.

FAE (Focuser Axial Error) corresponds to the distance between the focuser/primary axes at the primary mirror surface.

LAE (paralleL Axial Error) corresponds to the distance between a parallel projection to the focuser axis starting from the COC point and the center of the primary mirror. It is a measure of parallelism between the focuser and primary axes. If LAE=0 then both the focuser and primary axes are parallel and possibly coincident.



Every known collimation tools is designed to provide visual cue to eliminate at least one of mentioned four errors. For example, the cheshire is designed to eliminate PAE. Unbarlowed laser is designed eliminate FAE and CAE. Barlowed laser is designed to eliminate PAE. The conventional autocollimator is capable of eliminating CAE, FAE, and potentially PAE. The enhanced autocollimator (to be discussed later) is capable of eliminating CAE, LAE, FAE, and potentially PAE.

Eliminating any two of the 4 errors is enough to achieve axial alignment. That is, eliminating any two if the 4 errors ensures that both the focuser and primary axes coincide. Furthermore, eliminating any two of the 4 errors also implies the remaining two errors will also be eliminated. For example, if CAE and FAE are zeros then LAE and PAE must be zeros.

[Jason D]

(courtesy of Catseye)

A conventional autocollimator (AC) with a single/centered pupil consists of a perforated mirror mounted inside an eyepiece. Looking through the autocollimator pupil hole after the scope has gone through rough collimation will reveal several reflections of the primary mirror center spot. Typically, four reflections are seen. Identifying each reflection and interpreting its relative location is important.

Reflection P: Sharp and bright. It is the only reflection that does not move when you wiggle the AC in the focuser.

Reflection 1: Has the same orientation as P but could look somewhat fuzzy and bloated.

Reflection 2: Sharp and rotated 180 degrees with respect to P.

Reflection 3: Rotated 180 degrees with respect to P but could look somewhat fuzzy and bloated. It is the dimmest reflection and the hardest to see.

In addition to the above 4 reflections, there is the reflection of the AC pupil which will always be located at the center of the AC foreground reflection.

[Jason D]

Stacking/aligning/centering reflection “P” against each of the other reflections mentioned in the previous post will tell us something about how the focuser axis is positioned in relation to the primary axis:

Aligning P+pupil means the focuser axis intersects the primary axis at the focal point.
Aligning P+1 means the focuser axis intersects the primary axis at the COC point
Aligning P+2 means the focuser and the primary axes are parallel
Aligning P+3 means the focuser axis intersects the primary axis at the primary mirror center



The apparent distance between P and the pupil reflection equals two times PAE
The apparent distance between P and reflection 1 equals two times CAE
The apparent distance between P and reflection 2 equals two times LAE
The apparent distance between P and reflection 3 equals two times FAE

But how can we make use of the above information?

Mathematically, any two lines that share two distinct points are coincident. Using the same principle, if we can ensure the focuser and primary axes share at least two distinct points then both axes have to be coincident -- which is the definition of axial alignment.

Each of the four shown alignments can be accomplished by either adjusting the secondary mirror or the primary mirror with only one exception which is P+3 alignment. Adjusting the primary mirror will not alter reflection 3 relative position to reflection P. Only secondary mirror adjustment is capable of stacking P+3. We will take advantage of this exception later.

[Jason D]

In theory, if we align “P” with any two of the four reflections (pupil, 1, 2, 3) then we would achieve axial alignment. But which two should we select?

As it turns out, the best two alignments to use are P+1 and P+3.

P+pupil alignment is not a good choice because it is hard to discern the black pupil refection against the darkened background.

P+2 alignment is not a good choice either because reflection 2 disappears right before axial alignment is reached. Each frame of the following animation represents a case where the focuser and primary axes are parallel (P+2 is aligned in each frame). The only difference between the frames is the distance between both axes. As the parallel focuser axis gets closer to the primary axis, reflection 2 will start to fade away gradually until it disappears when both axes are within 0.5mm from each other. Also note how it is hard to discern the pupil reflection against the darkened background



P+1 is a good choice because reflection 1 does not disappear when the final axial alignment is reached

P+3 is another good choice because it is the only alignment that is insensitive to the primary mirror adjustment. Only adjusting the secondary mirror will move reflection 3 in reference to reflection P. The remaining (P+pupil, P+1, P+2) alignments are sensitive to both the secondary and the primary mirrors adjustments. P+3 insensitivity to the primary mirror adjustment allows us to isolate FAE elimination from others – very handy!!!!

[Jason D]

How to can we apply the knowledge presented thus far to achieve axial alignment?

The following are the recommended steps for using the autocollimator as described by Vic Menard.

Step 1: Use other collimation tools to achieve rough collimation. The objective of this step is to bring collimation close enough to be able to see most of the center spot reflections from the autocollimator pupil. Without this step, it would be almost impossible to make use of the autocollimator.

Step 2: Start off with stacking P+3 by ONLY adjusting the secondary mirror. Stacking P+3 means the focuser and the primary axes will intersect at the primary mirror center (FAE=0). You will need to adjust (decollimate) the primary mirror a little to move reflections 1 and 2 out of the way to be able to accurately stack P+3. This is our first intersection point between the two axes. Note: Make sure you are stacking P+3 reflections in this step – not P+2 reflections.



Step 3: Follow the above step with stacking P+1 by ONLY adjusting the primary mirror to intersect the focuser axis with the primary axis at the COC point (CAE~0). This is our second intersection point between the two axes. Since adjusting the primary mirror will NOT impact P+3 stack, step 3 is independent of step 2. That is, stacking P+1 by adjusting the primary mirror will not undo the P+3 alignment achieved in step 2.

After the third step, both the focuser and primary axes will have two intersection points which mathematically implies both axes are coincident. Since CAE and FAE are ~zeros, PAE and LAE should also be ~zeros.

Again, the following two images are key to using the conventional single-pupil autocollimator. The left photo represents P+3 (FAE=0) case and the right photo represents P+1 (CAE=0) case.



NOTE: Step#2 (P+3) is called CDP (Carefully Decollimated Primary) collimation protocol introduced by Vic Minard based on the mathematical analysis published by Nils Olof Carlin. For more details, refer to the following link

[Jason D]

There are two pitfalls which autocollimator users need to watch out for. Here is the first pitfall:



Some would stop when the final view looks similar to the left photo thinking a “centered” hexagram represents stacked reflections. After all, reflection 2 is 180 degrees rotated which should form a hexagram when stacked on the top of reflection P.

Seeing a hexagram on the final autocollimator view via the central pupil is incorrect.

The final view should look similar to the right photo. Only a single reflection against darkened background. The reason we do not see a hexagram is because when proper collimation is achieved, both reflections 2 and 3 (both are 180 degree rotated reflections) will disappear. The technical reason is beyond the scope of this thread but it has to do with the central pupil hole.

I will cover the second pitfall in a later post.

[Jason D]

So far, all the information presented is based on the conventional autocollimator with a single/centered pupil.

But, can the conventional autocollimator be further improved? If yes, what kind of shortcomings the conventional autocollimator has and how can it be improved?

If you followed the steps outlined earlier, you should end up with an outstanding axial alignment. However, there is a potential of persisting residual errors introduced by:
1- Reflections 1 and 3 fuzziness gets worse with shorter focal length scopes. It is harder to accurately stack reflections with fuzzy edges; therefore, the accuracy of stacking P+1 and P+3 might degrade for scopes with shorter focal lengths. The technical reason for why reflections 1 and 3 get fuzzier with shorter focal length primary mirrors is beyond the scope of this thread.
2- Stacking P+1 and P+3 are susceptible to parallax though this issue can be minimized by reducing the size of the AC pupil and pulling the eye slightly away from the pupil.
3- The final autocollimator view will not reveal residual PAE and FAE but rather only residual CAE. The remaining of this post and the next few ones are dedicated to explain this point.

Consider the following scenario:

Let us say you’ve been invited to a star party with your friends, Adam and Brian. Both collimate their identical scopes using their own collimation tools and methods. Both believe to have achieved perfect collimation. They ask you to assess their collimation; however, they emphasize that you are not allowed to touch any adjustment screw because they want to preserve their so-called “perfect” collimation. You insert your quality and trusted conventional autocollimator to assess both scopes. You get the views shown below.



Do both scopes have “perfect” collimation?

Both views look quite similar, don’t they? A single center spot reflection against darkened background. But is that enough to declare both scopes have “perfect” collimation?

[Jason D]

The final view of a conventional single-pupil autocollimator could be somewhat deceiving. The single center spot reflection against darkened background does not necessarily mean axial alignment has been achieved. All it means is that both focuser/primary axes intersect at the COC point – the same as P+1 alignment. In other words, the final view tells us that CAE is zero but it does not tell us if we have residual PAE or FAE.

Back to Adam’s and Brian’s scenario: Since both views include a single center spot reflection against darkened background, the only fact we can deduce is that the focuser axis intersects the primary axis at the COC point (CAE~0). But that is only one intersection point. In order to verify axial alignment, we need to assess whether both axes intersect at a different point. Unfortunately, examining both autocollimator views does not provide enough accurate information to assess a second intersection point:
P+pupil can’t be easily seen (PAE??)
P+2 can’t be evaluated due to the disappearance of reflection 2 (LAE??)
P+3 can’t be evaluated due to the disappearance of reflection 3 (FAE??). It is possible to decollimate the primary mirror until reflection 3 appears but this an intrusive method. We need a good method to assess final collimation without the need to touch any collimation adjustment screw.

One solution is to use another quality collimation tool which evaluates a different intersection point – other than CAE. Using a quality cheshire is a good choice. Using a quality barlowed laser is another good choice – both evaluate PAE.

Continuing with Brian’s and Adam’s scenario, to assess a second intersection point you remove the autocollimator and replace it with a quality cheshire. Now it is clear who has an excellent collimation and who does not. In fact, Adam’s scope is WAY OFF. It has a PAE of 1mm which is enough to show coma for F5 scopes. If CAE~0 and PAE=1mm then FAE=2mm.



Brian’s scope has an excellent axial alignment because:
1- The autocollimator tells us both focuser/primary axes intersect at the COC point (CAE~0)
2- The cheshire tells us both focuser/primary axes intersect at the focal point (PAE=0)
Since both axes share two distinct points then both axes must be coincident.

Adam’s scope is axially misaligned because:
1- Though the autocollimator tells us both focuser/primary axes intersect at the COC point (CAE~0)
2- Unfortunately, the cheshire tells us that both axes are 1mm apart at the focal plane (PAE=1mm)
Since both axes share only one point then both axes are not coincident. In other words, as long as we can show at least one of the four errors (PAE, FAE, LAE, or CAE) is non-zero then our scope is axially misaligned.

In the photo loaded in the previous post, I purposely obscured the pupil reflection of the autocollimator because the camera can see it but the unaided eye can’t. In the photo of this post, I did not obscure the pupil reflection. You can see how the cheshire and pupil reflection agree. If we could only see the pupil reflection with the clarity depicted in the photo, then we would not need to use the cheshire.

IMPORTANT: As was shown in Brian’s and Adam’s scenario, the autocollimator view of Adam’s scope did not catch the miscollimation issue but the cheshire view easily caught it. HOWEVER, NO one should draw the conclusion that cheshires are more accurate than autocollimators. Both tools are NOT competing tools but rather complementary tools. In Adam’s setup, CAE~0 and PAE>0 which favors the cheshire. I could easily concoct an opposite scenario where the cheshire view will not catch a miscollimation issue but the autocollimator view would. That scenario will include an axial relationship where PAE=0 and CAE > 0.

Bottom line: It is highly recommended to use a complementary tool with the autocollimators. The word “complementary” refers to a quality collimation tool that aligns/assess an axial intersection point other than the COC point (CAE). Quality cheshires and barlowed lasers are good choices which evaluate (PAE).

[Jason D]

In the previous scenario it was shown how the final view of an autocollimator consisting of a single center spot reflection against darkened background is NOT enough to constitute axial alignment. All it means is that the focuser/autocollimator axes and the primary axis intercept at the COC point (CAE~0) – same as P+1. Refer to the photo below. Both photos show a single center spot reflection against darkened background but the left photo has a whopping 4mm PAE and 8mm FAE – that is a HUGE error.



THIS IS THE SECOND PITFALL OF USING AN AUTOCOLLIMATOR. Many are under the impression that getting darkened background means axial alignment has been met -- this is far from the truth.

In fact, consider the following hypothetical experiment. Suspend a conventional autocollimator from a string attached to the COC point then lightly swing the autocollimator. The autocollimator axis will always run through the COC point (CAE=0), hence, it will always have a darkened background with a single center spot reflection showing.

Note: Hold off on the view depicted in the lower left corner of the animation for now.

[Jason D]

Can the autocollimator be improved to flag collimation problems – similar to the one described in Adam’s scope -- without the use of a cheshire?

The answer is YES.

First, here is a quick review: Adam’s final autocollimator view only told us that the focuser and primary axes intersect at the COC point (CAE~0). That is all we can deduce from the autocollimator final view. We need to be able to evaluate the proximity of the focuser and primary axes at a different point to determine whether Adam’s scope has reached axial alignment (PAE, FAE, or LAE). Unfortunately, that final view of the autocollimator does not provide the ability to assess other points since the autocollimator other alignments (P+2 LAE, and P+3 FAE) are not visible when focuser/primary intersect at the COC point and since (P+pupil PAE) in not easily discernable. That is why we had to resort to a different collimation tool such as the cheshire.

As it turns out, it is possible to solve the visibility of P+2 alignment (LAE) even when the focuser and primary axes intersect at the COC point (CAE~0).


(Courtesy of Catseye)

By utilizing the simple solution of adding an offset pupil to the conventional autocollimator, the P+2 disappearance problem is solved!!!! Solving P+2 disappearance will improve the ability to assess the final axial alignment.

We can use the central pupil to evaluate the proximity of the focuser/primary axes at the COC point (CAE) then use the offset pupil to evaluate parallelism of the two axes (LAE). If the central pupil tells us that both the focuser/primary axes intersect at the COC point (CAE~0) and the offset pupil tells us that both axes are parallel (LAE=0), then mathematically both axes are coincident and we have attained axial alignment.

Refer to the following photos:



Top row: Brian’s scope. Left photo via the central pupil indicates the focuser axis intersecting the primary axis at the COC point (CAE~0). Right photo via the offset pupil indicates parallel focuser/primary axes (LAE=0). Conclusion: Brian’s scope attained axial alignment. Two different alignments are zeros (CAE and LAE)

Middle row: Adam’s scope. Left photo via the central pupil indicates the focuser axis intersecting the primary axis at the COC point (CAE~0). Right photo via the offset pupil indicates unparallel focuser/primary axes. Conclusion: Adam’s scope has NOT attained axial alignment since one of the 4 alignments is non-zero, namely, LAE > 0.

Bottom row is for a scope with gross miscollimation: PAE=4mm and FAE=8mm despite CAE~0. The left photo is via the central. It shows a single center spot reflection against darkened background because the focuser axis runs through the COC point (CAE~0). If we could only see reflection 2 via the central pupil, then it will be apparent how grossly reflections are unstacked but we can’t. If you look carefully you can actually see a very dim reflection 2 via the central pupil detected by a long exposure camera shot – not visible via the unaided eye. However, the offset pupil shows reflection 2 very clearly. The relative positions of reflection 2 to reflection P is the same from both pupils; however, it is only visible from the offset pupil when CAE~0 – see the following animation. The animation also shows how both reflections P and 2 relative positions are parallax-free. Even when you move your eye by as much as 0.5” across pupils, the relative positions remain intact -- the AC mirror has to be located within few millimeters from the focal plane.



Below are the same photos but re-arranged for clarity to show how the second pupil manifests Adam's scope miscollimation.

[Jason D]

Can the autocollimator even be improved further?

The answer is YES.

When the focuser axis intersects the primary axis at the COC point (CAE~0), then we end up with the familiar view of a single center spot reflection against darkened background. Well, both axes do not really need to precisely intersect at the COC point to end up with the familiar view. In reality, when the focuser axis is approximately within 0.5mm from the COC point, we will see the familiar view (CAE <= 0.5mm). The ramification is that the 2-pupil autocollimator has a maximum residual error of 0.5mm. FAE, PAE, and CAE are <= 0.5mm for the 2-pupil autocollimator though LAE will be zero.

Refer to the following:



The top row is for a perfectly collimated scope (CAE=0,PAE=0,FAE=0,LAE=0).

The middle row is for a scope with the focuser axis parallel to the primary axis (LAE=0); however, a shift of 0.5mm was purposely introduced between both axes (PAE=0.5mm, CAE=0.5mm, FAE=0.5mm). Note how the offset pupil view indicates axial parallelism (P+2 are stacked) but the central pupil view shows only a single center spot reflection against darkened background even though CAE=0.5mm. The views from both pupils look quite similar for both cases. The 0.5mm is the maximum error that can be introduced – typical errors will be less than 0.5mm for the 2-pupil autocollimator.

The bottom row is for a scope with parallel axes just like the top/mid setups; however, the separation between the focuser and primary axes is “slightly” greater then 0.5mm. Note how reflection 2 started to appear via the central pupil and how P+1 unstacking is getting clearer. Therefore, any errors above 0.5mm for either CAE, PAE or FAE will be detected by the 2-pupil autocollimator.

It is possible to reduce that maximum 0.5mm error by adding a CAM to the 2-pupil autocollimator. The CAM which stands for (COC Alignment Mask) consists of two small rings positioned exactly opposite to each other. One ring is dark and the other is reflective. The backside of both rings is darkened to avoid internal light scatter.



When collimation is close, two reflections of the CAM will be seen: A foreground reflection on the top of a background reflection. The background reflection is always 180 degree rotated. Both reflections will stack only when the focuser axis precisely intersects the primary axis at the COC point (CAE=0). There is no longer a 0.5mm error to deal with.



Stacking the CAM is precise because it involves stacking two reflections that reside on the same visual plane which means both reflections will look sharp and parallax-free. In comparison, stacking P+1 reflections involves the somewhat fuzzy reflection 1 which is also susceptible to parallax because reflections P and 1 do not reside on the same visual plane.

When the CAM foreground reflection stacks precisely over the background reflection, the foreground dark ring will completely eclipse the reflective ring background reflection. Eclipsing the reflective ring with the dark ring is visually clearer, hence, more precise.

IMPORTANT: The central pupil can’t be used to stack the CAM for the same reasons the central pupil can’t be used to stack P+2. The background reflection will virtually disappear when CAE <= 0.5mm. If has to be done via the offset pupil.

Refer to the following photos:



Top row: Focuser axis is parallel to the primary axis but both axes are 0.5mm apart (CAE=0.5mm, PAE=0.5mm, FAE=0.5mm, LAE=0). The view from the central and offset pupils will not flag the 0.5mm residual error – using the 2 pupil autocollimator

Middle row: Uses the same 2-pupil autocollimator on above setup but with an added CAM. By illuminating the bright ring, it is clear that the CAM reflections are not perfectly stacked. The 0.5mm error is easily flagged.

Bottom row: When the CAM reflections (CAE=0) are perfectly stacked and P+2 reflections are also perfectly stacked (LAE=0), then we know we have achieved perfect collimation.

[Jason D]

SUMMARY:

1- Axial alignment is important to optimize the quality of your view. Axial alignment occurs when both the focuser and primary axes coincide. More accurately, when the eyepiece and primary axes coincide.

2- To align/assess axial alignment, at least two different points of intersection between the focuser/primary axes need to be evaluated.

3- There are 4 different focuser/primary alignments (or errors) that can be evaluated by typical collimator tools:
a. Distance between both axes at the COC point (CAE)
b. Distance between both axes at the focal point (PAE)
c. Distance between both axes at the primary center (FAE)
d. Parallelism between both axes (LAE)

4- If you can prove that at least two of the above errors are eliminated, then your scope is axially alignment. On the other hand, if you could show at least one of the above four errors is non-zero then your scope is axially misalignment.

5- The proper use of the conventional single pupil autocollimator (AC) involves first eliminating FAE by stacking P+3 then eliminating CAE by stacking P+1.

6- The main shortcomings of the conventional single pupil AC which could introduce residual errors are:
a. The final conventional autocollimator view evaluates only CAE. That is only one evaluation point which is not enough to ensure both focuser/primary axes are coincident. FAE can’t be evaluated based on the AC final view.
b. P+3 and P+1 stacking is susceptible to parallax and could involve fuzzy reflections which in turn could impact the accuracy of stacking.

7- The above shortcomings can be solved by adding a second pupil to the AC which will enable an additional error evaluation, namely, parallelism (LAE) by stacking P+2. P+2 stacking involves two parallax-free and sharp reflections. That makes P+2 less susceptible to errors. In addition, the offset pupil along with the central pupil will allow more precise evaluation of the final collimation setup.

8- Adding a CAM will further enhance the autocollimator by replacing P+1 stacking with CAM stacking. The latter, just like P+2, involves sharp and parallax-free reflections. In addition, CAM stacking is more readable compared to P+1.

9- In terms of residual errors -- the bottom-line:
a. If you do not follow the single-pupil autocollimator protocol as outlined by Vic Menard including the use of CDP and a cheshire, then you might end up with a relatively large error -- as much as 1mm to 2mm PAE and FAE. This is enough to ruin your view.
b. With the dual-pupil autocollimator, the maximum residual error you might end up with is 0.5mm PAE and FAE without the use of CDP or a cheshire.
c. With the dual-pupil autocollimator and CAM, residual error is virtually eliminated without the use of CDP and a cheshire. Actually, you do not even need to use the central pupil. Having said that, the use of CDP will reduce the number of iterations; therefore, it is highly recommended to use CDP if it is feasible.

[Jason D]

How to evaluate the quality of an autocollimator

First, you need to have a good quality focuser. That is, if your drawtube has a noticeable slop then using an expensive quality autocollimator might not help much because the additional collimation precision achieved by the autocollimator will be overshadowed by the focuser slop error.

Assuming you have a quality focuser, roughly collimate your scope then insert the autocollimator. Adjust the secondary mirror (or the primary mirror) until reflections P and 2 overlap – not stacked. Rotate the auto collimator and monitor the following via the CENTRAL pupil:

1- If you can discern the pupil reflection, make sure it remains at a fixed position relative to reflection P. If the pupil reflection rotates, then the autocollimator perforation is not centered. This would be a concern.
2- Note the shape of reflection 2. If it is deformed and changes shape as you rotate/tilt the autocollimator, then the autocollimator mirror is either non-flat or mounted under stress. This would be a concern.
3- Note the relative positions of reflections P and 2. If both reflections remain stationary as you rotate the autocollimator then you have a good autocollimator. If reflection 2 jitters a little, that is OK. This is a very sensitive test. However, if reflection 2 noticeably shifts back and forth then this would be a concern.
4- Look for reflection 3. If you can’t see it, then you will be unable to perform CDP; however, if this is the only issue, the autocollimator can still be used to produce accurate collimation. In this case, you will need to re-iterate between the autocollimator and a quality cheshire or a quality barlowed laser -- unless you have the 2-pupil autocollimator. Not having high reflectivity autocollimator mirror will contribute to dimmer reflection 3. Furthermore, scopes with short focal length will increase the fuzziness of reflection 3 – sometimes beyond recognition.

[Jason D]

Understanding which axial error each collimation tool eliminates is important. For example, someone with a quality scope and quality collimation tools might find that the cheshire agrees with the barlowed laser but both disagree with the autocollimator. This could happen because both the cheshire and barlowed laser check for PAE but the final view through a conventional autocollimator checks for CAE.

Understand which axial errors are checked for by your collimation tools.
Ensure that at least two axial errors are covered; otherwise, you will not be able to achieve good axial alignment.

The following table is a good reference. It tells you what each collimation tool practically checks for.

[Jason D]

Links for further reading…

The Autocollimator and its reflections
Addendum to the Fifth Edition of New Perspectives on Newtonian Collimation
Passive Tool Collimation and the Newtonian
Catseye vs. Howie Glatter and Blug??
New idea to improve the autocollimator (AC) tool
Questions RE the use of the Infinity XLK A/C?
Collimate with 1-1/4" or 2" laser???
Autocollimation question
The Catseye XLK 2-pupil collimation tool
Collimaton Tools Don't Agree
The INFINITY XLKTM Collimation Procedure
Vic Menard's Carefully Decollimated Primary (CDP) Collimation Protocol

[CatseyeMan]

Jason,

This series is no doubt a "Masterpiece" of elegance, detail, clarity and tutorial excellence! Bravo, my friend! With the additional insight you bring to the table completing the final chapter in the quest for AC reflection understanding, no doubt you have have earned your place amoung the ranks of the "elite few" who comprehend this special niche of optical physics. Your special talents of novel spacial/mathematical analysis and graphic presentation artistry have been beautifully blended synergistically here in your best work to date.

[rockethead26]

Hear! Hear!

[EdTheEdge]

Wow that's a lot of reading. Unfortunately for me it will take reading this thread quite a few more times before I can begin comprehend it all.

But I applaud you for taking the time to help newbies like myself!

THANKS!

[turtle86]

Great work! I've saved this as a favorite thread. Would love to see this in article/pdf format.

[Jason D]

Thank you everyone for the kind words.

This thread is a labor of

I do realize that there is so much material in this thread and it is inevitable that some of the presented information is unclear and would need more clarification. Please feel free to ask questions and to challenge some of the concepts presented. If you have questions but you are hesitant to post them because you think they are too basic to ask, be assured that many others will have the same questions in mind but are also hesitant to post them for the same reason. Just post your questions no matter how basic they are.

Jason

[Tim L]

Stickey! Stickey!

[Don W]

I will put this in with the Best of Reflectors list.

[backwoody]

Jason, I've got some work ahead of me to understand all the detail of your explanation. Thanks for the lessons from a guru...

c/s,