434 replies to this topic

[Don W]

Slick. Me likey.

[turtle86]

Thanks, Jim. Looks great!

[Jason D]

Here's what I use and recommend:

Coast Cutlery #TT75331CP LED clip light retrofitted with a bright 3000 MCD Red LED (Gilway E184 found here) and clipped to the spider; I get the best illumination with the light pulled in close to the Secondary as shown in the pics.

Ummm, I wonder if these points of light on the side have enough brightness to illuminate the CAM?

[Spaced]

Jason, my compliments on another superbly written and illustrated how-to manual. This is just excellent! Thanks so much for sharing your knowledge!

I'm currently using only a Glatter laser + Blug to collimate my f/4.5. My secondary is properly positioned with a Catseye sight tube. Assuming I'm using these tools properly and with care, am I able to achieve sufficiently accurate collimation, such that improvement with an AC wouldn't provide noticeably better views? To make your answer of more general interest, what do you think is the fastest focal ratio for which just a good barlowed laser + sight tube provides sufficient collimating accuracy for visual use?

Thanks much.

[Jason D]

First, thank you for the kind words, Mike.

You asked good questions. Here is my take:

I believe the combination of:
- Quality dual-pupil autocollimator and calibrated cheshire
OR
- Dual-pupil autocollimator+CAM
will give the most accurate collimation available today. We are talking sub 0.2mm accuracy for both the focuser and primary axial alignment.

The above is factual and measureable.

What is not as factual and measurable is how accurate collimation should be? Will the use of a sub 0.2mm collimation accuracy for FAE/PAE produce noticeable difference at the EP compared to sub 0.5mm collimation accuracy? Answering this question is somewhat subjective – after all, we are not talking about the difference between 1mm and 0.2mm accuracy but between 0.5mm and 0.2mm. The scope specs, average seeing conditions, observing experience, and whether the setup is meant for imaging or visual can sway the answer from one end of the spectrum to the other. Therefore, I can’t say that the additional accuracy provided by the dual-pupil autocollimator + (cheshire or CAM) will make a noticeable difference for everyone. But what I can tell you is that you have quality collimation tools that are capable of delivering superb collimation.

Jason

[Starman1]

One way to look at it is that collimation is an envelope of accuracy.
If you are inside of the envelope, your collimation is probably good enough. You don't have to be dead center in that envelope to see good images.
However, telescopes are not static devices--they flex, sag, change shape with temperature changes.
You want to make certain your telescope stays within the envelope of good collimation--the greater accuracy you achieve, the more likely will your telescope stay within that envelope as it sags, twists, and contracts.

Plus, achieving a superior accuracy in collimation when you first set up may eliminate any necessity to check collimation during the night.

Last is the benefit of being able to see tiny changes in collimation as the scope moves up and down. Tracking down sources of sag and flexure and fixing them means the scope is better able to hold its collimation after you've spent the time to make it good.

So do you NEED to collimate to such a high precision? It depends on the scope, its targets, and its conditions of use. But, there are never any deleterious effects from collimating better than is required. After all, developing the skills of collimating using accurate tools will pay off when you get that 20" f/3 scope that uses a Paracorr.

[Vic Menard]

I believe the combination of:
- Quality dual-pupil autocollimator and calibrated cheshire
OR
- Dual-pupil autocollimator+CAM
will give the most accurate collimation available today. We are talking sub 0.2mm accuracy for both the focuser and primary axial alignment.

Sub 0.2mm? I can read a calibrated Barlowed laser (PAE) to about 0.01-inch (about 0.005-inch PAE) pretty consistently--that's a direct read of PAE with no parallax error. Taking registration into consideration I would add +/-0.005 to 0.01-inch (about 0.01- to 0.015-inch PAE). Assuming you're interpreting PAE from CAE and LAE, what do you consider the read accuracy of the individual alignments--or are you assuming both can be consistently read to better than 0.1mm accuracy (without detailed examination of a closeup image)? I know that both errors (CAE and LAE) contain PAE and FAE components, but balancing one against the other doesn't necessarily improve the resolution of the interpreted errors. What do you feel are the best, and worst case scenarios?

If Mike is using a Glatter laser and Blug to align the axes of his 14.5-inch f/4.5 optic, he should have no problem keeping the FAE in tolerance (+/-0.4-inch), even if he's using a Paracorr (+/-0.07-inch). The PAE high magnification tolerance (+/-0.018-inch) should also be manageable with a Barlowed laser solution (although I would recommend the 1mm Glatter aperture stop over the Blug for the best precision).

I apologize for coming into the discussion this late (three pages already). It was a busy holiday with lots of family get-togethers.

Twenty-plus years ago, when I first persuaded Tom Clark to include the "complicated" autocollimator in the 3-tool Tectron arsenal, I never imagined it would be studied in such detail and modified to utilize the new CAE and LAE alignments! If you continue on this path, the autocollimator will eventually have to either auto collimate or make coffee--but I'm not sure which the end user would demand first!

Great post!

[Spaced]

The scope specs, average seeing conditions, observing experience, and whether the setup is meant for imaging or visual can sway the answer from one end of the spectrum to the other. Therefore, I can’t say that the additional accuracy provided by the dual-pupil autocollimator + (cheshire or CAM) will make a noticeable difference for everyone. But what I can tell you is that you have quality collimation tools that are capable of delivering superb collimation.

I understand that too many variables prevent a concise answer. I infer that your answer to my first, specific question is, "You're probably doing just fine without the AC."

One way to look at it is that collimation is an envelope of accuracy.
If you are inside of the envelope, your collimation is probably good enough. You don't have to be dead center in that envelope to see good images.

That description resonates.

So do you NEED to collimate to such a high precision? It depends on the scope, its targets, and its conditions of use. But, there are never any deleterious effects from collimating better than is required.

You just won't make a decision easy, will you?

After all, developing the skills of collimating using accurate tools will pay off when you get that 20" f/3 scope that uses a Paracorr.

Oh yeah, I'd better start practicing right away!

If Mike is using a Glatter laser and Blug to align the axes of his 14.5-inch f/4.5 optic, he should have no problem keeping the FAE in tolerance (+/-0.4-inch), even if he's using a Paracorr (+/-0.07-inch). The PAE high magnification tolerance (+/-0.018-inch) should also be manageable with a Barlowed laser solution (although I would recommend the 1mm Glatter aperture stop over the Blug for the best precision).

That's reassuring.

[nsldvd]

Vic, Jason (and all)
As I go through the steps (using the Glatter laser [w/1mm aperture stop] and Blug) collimation cannot be complete without the AC. But using the offset pupil seems to have eliminated two previously used procedures, the first being the center AC pupil. Apart from showing the disappearing act, cannot all AC collimating procedures be accomplish without the center AC pupil and just using the offset AC pupil? Secondly, doesn't the AC offset pupil eliminate CDP?

dave

[Vic Menard]

Vic, Jason (and all)
As I go through the steps (using the Glatter laser [w/1mm aperture stop] and Blug) collimation cannot be complete without the AC.

Why not? What is the aperture and focal ratio of your Newtonian? And are you using a coma corrector? I also suggest that you verify the Blug alignment against the self-Barlow attachment to be sure that the axial alignment with normal focuser registration is the same as the self centering alignment achieved at the bottom of the focuser drawtube.

But using the offset pupil seems to have eliminated two previously used procedures, the first being the center AC pupil. Apart from showing the disappearing act, cannot all AC collimating procedures be accomplish without the center AC pupil and just using the offset AC pupil? Secondly, doesn't the AC offset pupil eliminate CDP?

Stacking P-3 via CDP is only useful from the center pupil, allowing a direct read of FAE (magnified 2X), which may be critical in some applications. As Jason has already noted, you can iterate between CAE and LAE and achieve precise axial alignment only limited by the read precision of the signature alignments (CAM alignment and P-2 alignment from the offset pupil). But the iterative procedure may require several repetitions if the necessary precision adjustment isn't easily effected.

I think, perhaps, the more significant question is, if one finds a discrepancy between the calibrated Barlowed laser and the autocollimator using the offset pupil to assess CAE and LAE, should one trust the calibrated Barlowed laser or the autocollimator when correcting PAE? (Or should one anticipate that the residual error is exclusively FAE, since FAE is a component of both CAE and LAE?)

[Jason D]

Vic,

With XLK+CAM, I consistently achieved as good of PAE reduction as a cheshire – I would say even better. However, I am more experienced and knowledgeable about autocollimators and collimation than an average user. It won’t be fair to extrapolate my success to others. The 0.5mm and 0.2mm quotes are maximum errors for average users. The true average errors will have to be dedicated by others.

Here is a thought. If CDP can be performed with great accuracy which puts our FAE at ~0 then the CAM will magnify PAE by 4X as opposed to only 2X by the cheshire. Not only that but the CAM readability where a bright ring gets eclipsed by a dark ring is easier to discern than centering a triangle within a larger ring – of course, I am referring to the cheshire. In other words, starting with FAE=0 as our initial point, I do not see how the cheshire can outperform the CAM. Actually, I do not know of any tool that magnifies PAE by 4X other than the scenario I described. Of course, the catch here is that FAE needs to be eliminated. But the objective of the XLK+CAM is to eliminate all residual errors. Those who do not mind living with few millimeters of FAE should not use any kind of autocollimators, PERIOD. Actually, the very same users can be happy with a cheaper mass produced laser with small miscollimation error. They can use it to achieve an FAE with few millimeters error then use it again with a barlow to eliminate PAE. I just happened to belong to the camp who strives to eliminate all collimation errors for the reasons Don mentioned in his last post. I will not tolerate any FAE. That is why I am interested and a strong supporter of the autocollimator tool. That is my belief which others might not agree with. But that is OK. After all, this is only a hobby where people are free to do whatever they want and spent their money whichever way they like. There is no right or wrong.

By the way, here is an old animation that shows readability difference between a cheshire and a CAM. The CAM used in the animation is an old makeshift CAM that I build at home.

click here to view the animation

In general, let us compare the conventional steps of the single-pupil AC + cheshire to dual-pupil AC + cheshire. How can the first even be at the same level as the second? Not only everything in the first tool combination is covered by the second combination, but here is what you get with the second tool combination:
1- You get the CAM alignment which is more accurate than P+1 stack because it involves reflections located at the same plane which means both are sharp and parallax-free – not to mention the CAM has better readability than P+1
2- You get P+2 alignment which again involves sharp and parallax-free reflections.
For those who own the XLK and a chechire, if you need to iterate at the end then you should iterate between the cheshire and P+2 stack via the offset pupil as opposed to iterating between the cheshire and P+1 via the central pupil. For future XLK+CAM owners, assuming such product makes it to market, then the cheshire becomes optional. However, for those who are not interested in eliminating FAE then the cheshire would be a requirement even for the XLK+CAM. As I mentioned earlier, those who are not interested in eliminating FAE then they should not use any kind of autocollimator because it will not work.

Jason

[Jason D]

Vic, Jason (and all)
As I go through the steps (using the Glatter laser [w/1mm aperture stop] and Blug) collimation cannot be complete without the AC. But using the offset pupil seems to have eliminated two previously used procedures, the first being the center AC pupil. Apart from showing the disappearing act, cannot all AC collimating procedures be accomplish without the center AC pupil and just using the offset AC pupil? Secondly, doesn't the AC offset pupil eliminate CDP?

dave

Dave, if you have an autocollimator with only an offset pupil (no central pupil) equiped with the CAM, then you can achieve highly accurate collimation without a cheshire -- or the need for a central pupil. That is how I collimate.
However, without the CAM you will need either a cheshire or the central pupil. As Vic mentioned, if you start off with P+3 stack (CDP) via the central pupil then you will cut down on the number of iterations. That is why I recommend to start off with CDP provided reflection 3 can be clearly discerned.
Jason

[Jason D]

I understand that too many variables prevent a concise answer. I infer that your answer to my first, specific question is, "You're probably doing just fine without the AC."

My point is that an autocollimator will provide more accuracy; however, I can't tell you that the additional accuracy will always translate to a discernable improvement at the EP. This is akin to premium optics. In many nights, you will not easily discern the difference between good mass produced optics and guaranteed premium optics. But in those few nights when everything is perfect to get the most out of your premium optics then the additional collimation accuracy will make a difference.
Sorry, I just can’t make a general statement that the autocollimator will always produce discernable improvement at the EP all the time especially when you are using a quality laser collimator.
However, as the trend towards sub F3.0 reflectors continues, I can see more importance to the autocollimator in the future.
Jason

[Moggi1964]

..................If you continue on this path, the autocollimator will eventually have to either auto collimate or make coffee--but I'm not sure which the end user would demand first!

Great post!

Vic, there you go, limiting your expectations: let's have one that 'auto' collimates AND makes coffee!.

This is a great thread

[Jason D]

I am working on it. The central pupil will dispense caffeinated coffee and the offset pupil will dispense decaffeinated. This feature will help keeping you awake in case you spend the whole evening collimating.

[erick]

This feature will help keeping you awake in case you spend the whole evening collimating.

Right, that's the collimation done. Phew! What's that to the east? Drat! Dawn!!

Thanks for the thread. Time I learnt how to use my autocollimator on my coming scope.

[Vic Menard]

[quote name="Jason D"]The 0.5mm and 0.2mm quotes are maximum errors for average users. The true average errors will have to be dedicated by others.[/quote]
If 0.2mm is the maximum error, then you are expecting accuracies better than 0.1mm for CAE and LAE, is that correct?

[quote]If CDP can be performed with great accuracy which puts our FAE at ~0 then the CAM will magnify PAE by 4X as opposed to only 2X by the cheshire.[/quote]
I understood CAE to be the same as P-1, which is 4X for FAE and 2X for PAE (assuming the axes intersect at the COC). But P-2 magnifies both FAE and PAE 4X, so if FAE is zero, then P-2 would magnify PAE 4X. But getting FAE perfect, there's the rub. I've tried using the Barlowed laser to minimize PAE and then followed with FAE correction at the secondary mirror to minimize P-2, but the FAE correction creates a PAE error, so reiteration is necessary to achieve maximum precision.

[quote]Not only that but the CAM readability where a bright ring gets eclipsed by a dark ring is easier to discern than centering a triangle within a larger ring – of course, I am referring to the cheshire.[/quote]
I'm typically not using the Cheshire--I use a variation of the Barlowed laser (1mm aperture stop) with a concentric diffraction pattern that reads to very high precision (like the CAM, only up close where I can use reading glasses!)

[quote]In other words, starting with FAE=0 as our initial point, I do not see how the cheshire can perform better than the CAM.[/quote]
But you have to perfect the FAE first. Even with a CDP, P-3 only magnifies FAE 2X, and there are focus and parallax issues. No doubt, P-3 is better than a sight tube cross hair alignment, and possibly better than laser alignment with a 1mm aperture stop (curiously, also performed with a CDP). But I'm not ready to rely on P-3 as the reference alignment PAE will be collimated to.

[quote]Actually, I do not know if any tool that magnifies PAE by 4X other than the scenario I just described.[/quote]
Me either.

[quote]Of course, the catch here is that FAE needs to be eliminated.[/quote]
Yup.

[quote]But the objective of the XLK+CAM is to eliminate all residual errors.[/quote]
That's a worthy goal, but PAE and FAE are the only ones that matter. I don't know about eliminating all residual errors (especially considering registration tolerances), but reducing errors, if possible to the resolution of the tool, may be necessary for some applications. Using CAE and LAE to reduce PAE and FAE is the goal I see.

[quote]Those who do not mind living with few millimeters of FAE should not use any kind of autocollimators PERIOD.[/quote]
But what about those f/3 Dobs? A 20-inch f/3 can tolerate 2.5mm of FAE, with a Paracorr, but requires a PAE correction of 0.13mm (0.005-inch)! That's assuming, of course, that an f/3 scope is capable of high magnification performance.

[quote]...I just happened to belong to the camp who strives to eliminate all collimation errors for the reasons Don mentioned in his last post. I will not tolerate any FAE. That is why I am interested and a strong supporter of the autocollimator tool.[/quote]
I'm sure you know you're preaching to the choir. But there are tolerances, and there is also useful readability. Of course, you can carefully examine an image of the alignment and extrapolate a fairly reliable PAE (usually the more critical axial alignment). I know you know what I mean and the implications.

[quote]...let us compare the conventional steps of the single-pupil AC + cheshire to dual-pupil AC + cheshire.[/quote]
Instead of a Cheshire, let's compare the two autocollimators combined with a calibrated Barlowed laser.

[quote]How can the first even be at the same level as the second?[/quote]
If precise PAE is the ultimate goal, unless you can find a way to zero FAE first, the calibrated Barlowed laser has some real advantages for many users. If precise FAE is the goal, P-2 in the offset pupil should eventually (after several iterations) deliver an FAE nearly as precise as the PAE delivered by the Barlowed laser.

[quote]...here is what you get with the second tool combination:
1- You get the CAM alignment which is more accurate than P+1 stack because it involves reflections located at the same plane which means both are sharp and parallax-free – not to mention the CAM has better readability than P+1[/quote]
That's true, but a perfect CAM alignment does not in and of itself indicate perfect PAE alignment (neither does a perfect P-2 alignment in the offset pupil). Because both errors (PAE and FAE) can be present in both alignments (CAE and LAE), PAE and FAE must be extrapolated from the CAE and LAE alignments. Of course, if you can align CAE and LAE perfectly (whatever the read accuracy actually is), then PAE and FAE must be very good, if not perfect.

[quote]For future XLK+CAM owners, assuming such product makes it to market, then the cheshire becomes optional.[/quote]
Like you, I want the best alignment possible (for my f/4 primary). I tend to rely on redundant verification when possible. So while the XLK+CAM may indeed allow me to derive the PAE correction I need, I'll keep my Barlowed laser nearby so I can actually look at the PAE correction.

[quote]As I mentioned earlier, those who are not interested in eliminating FAE then no autocollimator derivative tool should be used.[/quote]
I disagree, because I prefer redundant verification. My autocollimator is still the quality control inspector for my axial alignment (even though my FAE tolerance, with Paracorr, is 2.8mm).

[Jason D]

If 0.2mm is the maximum error, then you are expecting accuracies better than 0.1mm for CAE and LAE, is that correct?

The 0.2mm is an estimate for the maximum error an average user might run into. It is my estimate. Real life users will determine what this number should be. I can do much better but I am not an average user.

I've tried using the Barlowed laser to minimize PAE and then followed with FAE correction at the secondary mirror to minimize P-2, but the FAE correction creates a PAE error, so reiteration is necessary to achieve maximum precision.

For the benefit of others, Vic just described the strength of the XLK. Using XLK+Cheshire might require additional iterations but the extra effort should not be viewed as negative. See, those who use XL+cheshire will get to a point where all visual cues will indicate perfect collimation and that would be the end of collimation. However, those who use XLK+cheshire might flag additional unforeseen residual errors which would require additional iterations. Those who decide not to perform the additional iterations will end up with collimation as good as the XL+cheshire. It is a choice.

That's a worthy goal, but PAE and FAE are the only ones that matter. I don't know about eliminating all residual errors (especially considering registration tolerances), but reducing errors, if possible to the resolution of the tool, may be necessary for some applications. Using CAE and LAE to reduce PAE and FAE is the goal I see.

I prefer to treat all FAE, PAE, LAE, and CAE with the same importance. After all, my goal is to eliminate all four. Mathematically, the goodness of the axial alignment is determined by the smallest two residual errors out of the mentioned four. Since LAE and CAE can be reduced with the most accuracy using the XLK+CAM, they should provide the smallest PAE. It is just a different paradigm of the way I view axial alignment.
Having said all the above, owning a cheshire with XLK+CAM is not a bad idea. Actually I would recommend it. Not only it would server as a redundent check but in case the owner knows he/she has an residual FAE but do want want to take the time to correct it then a cheshire will be handy to eliminate PAE and live with the residual FAE.

But what about those f/3 Dobs? A 20-inch f/3 can tolerate 2.5mm of FAE, with a Paracorr, but requires a PAE correction of 0.13mm (0.005-inch)! That's assuming, of course, that an f/3 scope is capable of high magnification performance.

Vic, you know how I feel about the 0.034D FAE tolerance derivation. In my opinion, the premise behind the formula is too loose. I am still of the opinion to use the edge of the field stop as the proper reference to calculate the FAE tolerance – not the edge of the coma “free” zone. According to my premise, FAE tolerance will even be tighter than the current tolerance calculation for a paracorr.

However, let us proceed with the 2.5mm FAE and 0.13mm PAE scenario you have described. Getting below 2.5mm FAE is not a problem. Even a tuned mass produced laser collimator can bring FAE below 2.5mm. A quality calibrated cheshire (or a quality barlowed laser) used by an experienced user can bring PAE at 0.13mm. Then my question is why should an autocollimator be used? After all, we can bring such a setup within tolerance using a stock laser collimator and a quality cheshire!!

As I mentioned earlier, those who are not interested in eliminating FAE then no autocollimator derivative tool should be used.

I disagree, because I prefer redundant verification. My autocollimator is still the quality control inspector for my axial alignment (even though my FAE tolerance, with Paracorr, is 2.8mm).

I guess what I am saying is that those who accept the 0.034D FAE tolerance formula will not need an autocollimator. On the other hand, those who strive to eliminate as much as of their FAE as their PAE then an autocollimator would be of great benefit.

Consider this: Someone who accepts the 0.034D formula stops when FAE=2mm and PAE=0mm. How would that someone use the autocollimator? If that someone inserts a quality single-pupil autocollimator, unstacked reflections will be seen. How would that be helpful? I guess I am having hard time reconciling the 0.034D formula with the use of an autocollimator. I do not know how an autocollimator can be used to eliminate PAE yet tolerate few millimeters of FAE -- it can't be done. What is your take, Vic?

Jason

[CatseyeMan]

... Consider this: Someone who accepts the 0.034D formula stops when FAE=2mm and PAE=0mm.

I think this is an unrealistic scenario. What tool out there reads PAE independently of FAE permitting the user to obtain a PAE of "zero" in the presence of 2mm FAE ? In the presence of a 2mm FAE, a "perfect" Chesire or Barlowed laser read will leave undetected PAE. The beauty of the P-3 stack of the CDP is that it DOES allow an independent read and subsequent correction of FAE prior to addressing PAE thus maximizing the effectiveness of the PAE tools to optimally reduce PAE within the tool(s) capability.

[Vic Menard]

The 0.2mm is an estimate for the maximum error an average user might run into. It is my estimate. Real life users will determine what this number should be. I can do much better but I am not an average user.

I don't consider myself an average user, but I have real problems keeping CAE and LAE less than 0.1mm (4 thousandths of an inch read at 88-inches) with my XLK+CAM autocollimator. I suspect most average users will have a hard enough time achieving a 0.2mm (0.008-inch) read for both signatures at 50-inches, which, worst case scenario, could deliver a PAE about twice that amount.

...my goal is to eliminate all four...Since LAE and CAE can be reduced with the most accuracy using the XLK+CAM, they should provide the smallest PAE.

If you're limited to using just the XLK+CAM. The question I posed earlier, "...if the "perfect" XLK+CAM alignment differs from the calibrated Barlowed laser alignment, which one should the end user choose?", is the reason for the discussion. My concern is that a "better than average" user with a moderately large aperture, short focus optic will have difficulty achieving the correction resolution you're expecting. While I find that errors in the LAE and CAE signatures are useful indicators that there may be a problem with PAE and/or FAE, I'm not willing to use what I consider to be good LAE and CAE signatures as the final arbiter of best PAE alignment. But I do feel it's important to have a tool that is sensitive enough to verify the axial alignment I can achieve with my PAE alignment tool, and even cause a PAE reevaluation when necessary.

...a tuned mass produced laser collimator can bring FAE below 2.5mm. A quality calibrated cheshire (or a quality barlowed laser) used by an experienced user can bring PAE at 0.13mm. Then my question is why should an autocollimator be used? After all, we can bring such a setup within tolerance using a stock laser collimator and a quality cheshire!!

While your closing comment is essentially correct, I think I've answered your question already. The autocollimator is simply too good at what it does not to be used! And for a scope with such demanding PAE tolerances, it's worth the effort to reduce FAE far below the tolerance threshold to be able to see the other alignment signatures and prove the alignment accuracy of the other tools!

[sixela]

... Consider this: Someone who accepts the 0.034D formula stops when FAE=2mm and PAE=0mm.

I think this is an unrealistic scenario. What tool out there reads PAE independently of FAE permitting the user to obtain a PAE of "zero" in the presence of 2mm FAE

You're selling it - a BlackCat placed precisely so that the focal plane is between the pupil and the ring.

? In the presence of a 2mm FAE, a "perfect" Chesire or Barlowed laser read will leave undetected PAE.

Nope. If placed properly, it directly detects PAE (and nothing else).

Of course, many people don't know where to place the tool, but I do .

Yeah, I'm only precise in placement to about 3mm, so I guess 2mm FAE will translate into one micron extra PAE...I can read the BlackCat well, but not that well (assuming "zero" with apologetical quoting means as little as you can detect with the tools at hand).

[Vic Menard]

... Consider this: Someone who accepts the 0.034D formula stops when FAE=2mm and PAE=0mm.

I think this is an unrealistic scenario. What tool out there reads PAE independently of FAE permitting the user to obtain a PAE of "zero" in the presence of 2mm FAE ? In the presence of a 2mm FAE, a "perfect" Chesire or Barlowed laser read will leave undetected PAE. The beauty of the P-3 stack of the CDP is that it DOES allow an independent read and subsequent correction of FAE prior to addressing PAE thus maximizing the effectiveness of the PAE tools to optimally reduce PAE within the tool(s) capability.

I get what you're saying, but I don't agree with it entirely. I think claiming a PAE of "zero" is unrealistic--it's certainly possible but the read accuracy is going to place a limit on what you can actually see. And while Cheshire derivatives and Barlowed lasers are relatively insensitive to residual FAE, a 2mm defect at 60-inches (about 0.001-inch lateral displacement +/-0.75-inch from the focal plane--assuming collimation takes place at or near the focal plane and the focuser travel is linear) is unlikely to impact the corrected PAE enough to be cause for concern at or near the focal plane.

But I do agree that it's best to address FAE first (when possible), and always make the final tweaks to correct PAE last.

[Jason D]

I don't consider myself an average user, but I have real problems keeping CAE and LAE less than 0.1mm (4 thousandths of an inch read at 88-inches) with my XLK+CAM autocollimator. I suspect most average users will have a hard enough time achieving a 0.2mm (0.008-inch) read for both signatures at 50-inches, which, worst case scenario, could deliver a PAE about twice that amount.

My XT10 has a focal length of 47.6-inches and I do not have an issue seeing 0.2mm LAE and CAE errors. Attached is a representation of how the 0.2mm LAE and CAE would look like. But I’ll grant you that going below 0.2mm to around 0.1mm then experience, illumination, and vision acuity play a large factor. That is why I quoted the 0.2mm maximum error. Of course typical errors are less than that.
But there is always advantage for owning and double checking with a cheshire. Not only as a redundancy check but also as an alternative method to quickly collimate a scope without worrying about FAE elimination and the additional work needed to eliminate FAE.

"...if the "perfect" XLK+CAM alignment differs from the calibrated Barlowed laser alignment, which one should the end user choose?",

There is no simple answer. If that happens to me, I would analyze both tools using my camera to determine the root cause of the discrepancy. Is it the quality of the tools? Is it readability? Is it consistency? Or is it something else? Of course, this kind of analysis will be done only once – not with each collimation session.

While your closing comment is essentially correct, I think I've answered your question already. The autocollimator is simply too good at what it does not to be used! And for a scope with such demanding PAE tolerances, it's worth the effort to reduce FAE far below the tolerance threshold to be able to see the other alignment signatures and prove the alignment accuracy of the other tools!

Then for those who intend to use the autocollimator should discard the 0.034D FAE tolerance formula and should tackle the FAE reduction with the same vigor as PAE reduction. That is, FAE becomes as important as PAE when an autocollimator is used. Of course, elevating FAE importance to the same level as PAE also implies elevating both CAE and LAE to the same high standards. I agree with this approach. Those who intend to get the most out of their scopes should not compromise any part of collimation including FAE elimination.

Jason

Attached Thumbnails

• 

[Jason D]

And a 0.2mm PAE representation using a cheshire

Attached Thumbnails

• 

[hudson_yak]

Here's where I wound up with this, based on the particular scope I have.

After adding tension to the spider vanes and making sure the truss clamping was holding, the scope maintains good FAE through a normal range of elevations, by good I mean holding it within 1-2mm drift. On the assumption this was good enough, I set FAE once after scope setup and don't mess with it the rest of the evening, using a decent laser that I have verified for repeatable results. Since FAE isn't going to be exactly 0 much of the time, and I don't care that it isn't, I don't see any need for better accuracy than that.

However, that minor FAE drift does cause a similar PAE drift, and, given the tighter tolerance for PAE, I do care about making sure it is close to correct several times during an evening, particularly when I want to look at a planet. Another issue I see sometimes is that the mirror seems to shift a bit in the cell during an evening.

It helps that, unlike the secondary's closely-spaced adjustment screws and flexy spider vanes, the primary's widely-spaced screws with large thumb knobs lend themselves to making precise adjustments easily and quickly.

For a few months I used a cheshire for this, and decided the best tool would be one that is self-illuminated, remotely readable so that primary adjustment can be made while observing in real time, similar in weight to an eyepiece so that the scope wouldn't move up or down when switching between eyepiece and tool, accurate, didn't require FAE to be zeroed out first, and convenient. Only one tool fit all these requirements, the 1.25" Glatter tuBlug.

Most of the time, only a small tweak, if any, is necessary, which also tells me nothing bad has happened to the focuser axis. I can also tell that by racking the focuser in and out to make sure the reading doesn't change too much. If something does seem to be seriously amiss, it probably means a truss shifted and it is time to revisit the FAE.

This is why I haven't found the AC to be worth getting. That could change if I ever got a bigger coma-corrector-equipped scope capable of holding FAE more precisely.

Mike