- René Descartes
Realigning Newtonian Collimation
Updated July 2003, July 2004, December 2004, April 2006
April 2006
Revised focuser axial tolerances
Updated “the Barlowed laser”
December 2004:
New images and movie links in the autocollimator section.
Removed “THE FOCUSER DIAGRAMS” section--replaced with actual images in the new...
PRACTICAL APPLICATIONS OF THE AUTOCOLLIMATOR section and
ANALYSIS OF THE AUTOCOLLIMATOR REFLECTIONS section.
performed additional minor reformatting and corrections throughout the text.
July 2004:
added section THE PRIMARY MIRROR CENTER SPOT AND “STAR COLLIMATION”
added information on “edge induced astigmatism”
added information on “the Barlowed laser”
performed some minor reformatting in preparation for the next addition
For the less informed, you’ve probably seen someone go through the motions of aligning his or her optics with a laser, Cheshire eyepiece, cross hair sight tube, or perhaps even that complicated tool called the autocollimator. From person to person, ’scope to ’scope, the basic mantra has always been, “align the diagonal to the primary mirror, align the primary mirror to the focuser...” With modern high performance optics, if you miss the mark, the image quality suffers and the observing experience is reduced to an exercise in frustration.
Until recently, I approached the process of fine collimation as a qualitative exercise, where the variability of the result was controlled by the sensitivity of the tools, and the goal was, well, perfect, collimation. Last December, (2002), because of a lengthy online debate, I experimented with changing the process to a quantitative analysis, restraining the collimation accuracy to an acceptable tolerance window. A year later, online again, with a new wide field autocollimator and digital images, the autocollimator reflections were finally analyzed and defined. The implications were at times unexpected and often surprising. The quest for perfect collimation seems almost magically just out of reach, although realistically we seem so close. In the following pages, presented in question and answer format, I’ll reexamine the original protocols, the underlying mathematics, and the refined procedures. Taking the time to understand the collimation tolerances that fit your ’scope and your observing preference is a worthwhile exercise. Until then, perfect collimation always delivers the best possible image, so push for perfection!
The fourth edition of New Perspectives on Newtonian Collimation is now nearly 5 years old. In the past five years the landscape of Newtonian/Dobsonian technology has changed significantly. The illustrations in the fourth edition are typical of a 20-inch f/5, a ’scope I used in the late 1980s. Reference is also made to the 20-inch f/6.2 I had been using from 1992 to 1998. Modern Newtonian owners use affordable laser collimators, precision collimatible focusers, notebook paper reinforcement rings to define the primary mirror center spot (the ubiquitous donut), and of course, very fast f/4.1 to f/4.3 large aperture ’scopes with tracking platforms and GoTo drives—and many more economy 6- to 10-inch ’scopes than I expected! Although the fourth edition is still fundamentally correct, the fine collimation procedures, in particular when using the primary mirror center spot as the reflection generator in the autocollimator, need some updating. With the proliferation of donut (and triangular) center spots, it should be stressed that all tools should demonstrate accurate collimation, to deliver precision collimation of the primary mirror (Cheshire), diagonal mirror (laser or sight tube), and focuser (autocollimator) simultaneously! This is the only method that effectively minimizes all image defects caused by misalignment. Also, after any collimation procedure, it is important that you verify that the primary mirror alignment has not been compromised. Whether you confirm this with the accurately stacked autocollimator reflections, the reflection of the pupil of the autocollimator in the perforation of the primary mirror center spot, or reexamine the Cheshire alignment (or the Barlowed laser...), remember to always check and make sure that the primary collimation is adjusted spot-on!
Incidentally, the fifth edition is in the works!
GETTING STARTED
Q. My viewing conditions are often turbulent, so why should I concern myself with accurate collimation?
A. The same logic could be used to dismiss the need for premium optics. When the viewing conditions permit, the ability to go diffraction limited answers your question.
Q. What effects will less than perfect collimation have on images in the eyepiece?
A. Given excellent optics and good seeing, defects in collimation can ruin the image in the eyepiece and reduce the effective aperture of the ’scope. Check out http://perso.club-internet.fr/legault/collim.html
Q. Why does my 12.5-inch f/5.6 seem to out perform my 20-inch f/4.3?
A. At f/5.6, the combined focuser and primary axes collimation tolerance of the 12.5-inch is almost four times that of the f/4.3 20-inch. And it’s much more likely the seeing will be steady enough for diffraction limited performance with the 12.5-inch than the 20-inch. The result is superb planetary views with the 12.5-inch ’scope. However, on those not too rare nights when the air is steady enough to provide short periods of diffraction limited viewing with the 20-inch, if the optics are of similar quality and the collimation is right, the 20 will outperform the 12.5 every time.
Q. My images look astigmatic. Is my mirror bad?
A. It’s not likely that your mirror is bad. But it could be the problem. When defocused images flip 90 degrees on either side of focus, astigmatism is the culprit. Assuming the problem is not your eye or the eyepiece, the problem is confined to the primary and diagonal mirrors and collimation. If the mirrors have previously delivered good images and the ’scope is accurately collimated, there are only a few possible contributing factors. Optics that are pinched by retaining clips can look astigmatic. You should also make sure your mirror rests freely on the mirror cell flotation pads (and edge supports)—a hot day in the sun at a star party (with the 'scope parked in the horizontal position) can cause surfaces to adhere to each other. Large aperture “thin” mirrors are susceptible to edge support induced astigmatism, particularly as the ’scope is moved from a vertical to horizontal orientation and the mirror weight is shifted from the flotation cell to the edge support. After thermal issues have been addressed (mirror cooling, boundary layer, and other “seeing” effects), edge support induced astigmatism is probably the most prevalent defect encountered in the collimated Newtonian system. There are several solutions for edge support problems with mixed results depending on the size of the mirror and the magnitude of the aberration. From stainless steel wire slings to actually gluing the primary mirror to the flotation cell pads, the science of supporting thin mirrors is a continuously evolving process.
Q. Is the procedure offset or centered?
A. The final collimation will be offset, even though the spider is mechanically centered.
Q. If the optical axis isn’t perfectly centered in the optical tube assembly, will my DSCs still work?
A. Digital setting circles with encoder resolutions of 0.1 degree will continue to perform with little or no discernible accuracy degradation following the collimation procedure.
Q. Does my focuser have to be perfectly square to the front cage/tube assembly?
A. How would you measure that? The preliminary mechanical adjustment of the focuser axis should be as close to perpendicular to the telescope tube axis as possible. With the addition of a collimating focuser base, fine adjustments can be made to the focuser axis to bring it into coincidence with the optical axis. It should be noted that with the collimation procedure that will be discussed, the focuser axis will probably not end up perfectly square to the optical tube assembly, the primary mirror optical axis will not be precisely coincident with the optical tube axis, and the intercept angle will likely not end up exactly 90 degrees. Yet we will be able to achieve collimation as close to perfection as is measurably possible, and easily fall within very strict tolerances.
Q. Is a collimatible focuser really necessary?
A. If you need to adjust the focuser alignment, and you don’t want to use loose shims, the ability to easily collimate the focuser quickly becomes more than just a luxury. With faster focal ratio (f/4.5 and under), larger aperture (12-inch and bigger), Newtonian systems, I consider a collimatible focuser with fine focus capability a worthwhile investment.
Q. Why does the diagonal have four adjustment screws instead of three?
A. Actually, four makes more sense than three, allowing the adjustment to be confined to paired, opposing screws. When these paired adjustment screws coincide with the directional adjustment needed to set the angle of the diagonal/intercept, the collimation procedure is intuitive and predictable. Plus, four adjustment screw diagonals work better with four vain spiders!
Q. Could you clarify single pass and multi-pass collimation terms?
A. Single pass is what you do with a sight tube, a Cheshire, or a laser. The accuracy of these tools is determined by one reflection off the primary mirror and two reflections off the diagonal (one pass—round trip through the system). The autocollimator uses multiple passes to create the reflected images used in the stacking or convergence procedure. Each additional pass increases the angular accuracy of the test while the observed reflected images in the autocollimator magnify and define the axial offsets of the focuser and primary mirror axes.
Q. When collimating a 14.5 Starmaster (f/4.3), using the Tectron Tools, how far in do you place the sight tube? I had heard that you stick it in as far as it will go. I have had trouble judging whether the secondary is exactly centered under the focuser because of the wiggle in the sight tube.
A. I like to use the tools flush to the head when possible—the heads of the tools are fabricated with a CNC lathe that can generate parts that are accurate to about 0.0001 inch. The tubes themselves are necessarily undersized so they will fit in the focuser or 2- to 1.25-inch adapter. They're kept to a tolerance of approximately +0 / -0.03 inches, so yes, they can wiggle a bit in some focusers and adapters.
Q. I had heard that you should have tools that match your ’scope. For instance, you should have an f/4.3 sight tube and Cheshire. Will Tectron make custom tools?
A. Tectron does not make custom tools. The sight tube is good for about f/5 or slower focal ratios. If you're faster than that, the Cheshire can be used as a makeshift sight tube for f/3 to f/5. This is important when evaluating the concentricity of the bottom edge of the collimating tool, the real outer edge of the diagonal, the reflection of the outside edge of the primary mirror, and the reflection of the front opening of the tube assembly. If you have a 2-inch focuser, Jim Fly (catseyecollimation.com) offers a variable focal length sight tube.
Q. How do I calibrate my Cheshire eyepiece for precision primary mirror collimation?
A. If you’re using a square center spot (Tectron), collimating the four corners of the square with the perforation of the Cheshire gives precision results—accurate to 1 or 2 hundredths of an inch. If your ’scope has a notebook paper reinforcement ring to define the primary mirror center spot, (StarMaster, Obsession, etc.) it’s impossible to see the center perforation, so the outside of the ring (donut) is collimated to the outside of the bright ring in the Cheshire. There are problems with this approach. First, the reinforcement ring or the bright annulus in the Cheshire may not be perfectly concentric around the respective axes they define. Notebook paper reinforcement rings are not manufactured to any kind of precision tolerance. The center perforation in the Cheshire is always properly centered, but the window in the Cheshire may be slightly misaligned. This was not an issue when the central perforation defined the calibrated center spot. The other issue is the size differential between the bright outer diameter of the Cheshire and the smaller diameter of the donut. The large window was designed to facilitate location of the center spot when collimating in darkness. Ideally, a calibrated Cheshire should be very close to the size of the center spot to be able to read axial offsets as small as 0.01-inch. Various makeshift remedies include modifying the center spot or changing the fieldstop of the Cheshire. The CatsEye™ triangular center spots (in white or reflective red) meet this calibrated criteria, and work with Tectron tools as well (see images below.) A clear acetate overlay is also available with the CatsEye™to facilitate precise placement of the center spot.
THE PRIMARY MIRROR CENTER SPOT AND “STAR COLLIMATION”
Q. How critical is the placement of the primary mirror center spot?
A. Somewhere around this point in the discussion, shortly after digesting an axial accuracy as small as .01 inch, the question regarding the accuracy of the placement of the primary mirror center spot is bound to come up. Mathematically, the allowable error of the “mechanical” placement should be no greater than twice the allowable error of the optical axis at the focal plane. It’s actually not too difficult to place the center spot within 0.01-inch of the measured center of the primary mirror’s face. It’s often said that we assume the mirror is a figure of revolution with its optical axis at the center of the mirror, but after much discussion, I would change the wording slightly to, “The primary mirror is a figure of revolution with its optical axis defined by the edge of the mirror!” Modern mirror manufacturing procedures seem to consistently deliver a parabolic surface that behaves according to this axiom, so it’s important to measure carefully. Accurately center spotting the primary mirror defines the physical edge of the parabola and also allows the focuser to be accurately aligned to the mirror surface. Mel Bartels has also weighed in on alignment accuracy: “I suggest consistent pointing within 0.1 or 0.2 millimeters of the primary mirror center. That's 1/10 of 1 percent of mirror size. What you are doing is making the focuser axis coincident with the primary's center. That's critical at f/4, because it's extremely challenging to collimate properly otherwise (about the only sane way to do it is with star testing.)” He adds, “...that's why I suggest star testing for final collimation particularly in situations where there is uncertainty as to the focuser/eyepiece's aim at the center of the primary.” (Numerical values for collimation tolerances are discussed further under “Autocollimator Evaluation”.)
Q. Isn’t “star collimation” the ultimate collimation test/procedure?
A. The discussion continues to rage on regarding “star” collimation, or, “collimating to the sweet spot”. Apparently, some observers have noticed the “sweet spot” in the actual eyepiece’s field of view is not always centered. The argument is that the primary mirror collimation should be adjusted to center the “sweet spot”. Concerns have been voiced about the subsequent focal plane being tilted, and some would advocate moving the primary center spot to “approximate” the sweet spot alignment. Nils Olof Carlin commented, “I am reluctant to believe a quick star collimation using a well defocused star is really sensitive enough—I think it takes good seeing and near or at focus imaging to catch the coma near the minimum detectable amount (whatever that is), and perhaps taking the average of several readings.” We both noted that in our combined years of experience we had not encountered anyone, who actually understood the principles of collimation, who had successfully and accurately repositioned his primary mirror center spot “off center” to deliver a consistently better Airy disk than a mechanically centered mirror spot.
In my experience, after observing with a multitude of “precisely” collimated 14.5 to 42-inch fast focus motor driven Newtonians (with good seeing, eyepieces, etc.), the sweet spot has always been in the center of the field. If it isn’t, and I know the collimation is good, the culprit is always something else--thermals, pinched diagonal, jammed primary, dew accumulation, wedged eyepiece, loose truss tube(s)--but not something you want to "collimate out". Star testing can become very subjective when a large aperture is being pushed to its theoretical limit. Averted vision and moments of clarity work together to build a better image in the brain, but using this expanded envelope to evaluate collimation quickly reaches a point of diminishing returns with apertures that are operating well beyond the average seeing. Collimating to the “sweet spot” may be a reasonable solution when the performance of the telescope is in question and the collimation is suspect.
THE VIEW IN THE CHESHIRE
The next two images were captured in a 14.5-inch StarMaster Dobsonian during Cheshire collimation.
The first image is after diagonal collimation and rotation correction with a laser, but before primary mirror collimation. Note the obvious paired screw heads in the upper right quadrant of the diagonal silhouette, and the less obvious unpaired screw heads—one at nine o’clock, the other at five o’clock.
View in the Cheshire—first image (image by Dallas Rogers)
In the next images, the primary mirror has been collimated. Notice in the second image, the bright annulus is not perfectly concentric. The center spot is offset about 0.5mm (approximately .02 inch) towards the eleven o’clock position. This minor offset becomes more problematic during autocollimation. A calibrated Cheshire (third and fourth images) improves the accuracy.
View in the Cheshire—second image (image by Dallas Rogers)
View in the Cheshire—third image (image by VM)
11-inch f/5.4 Newtonian with a white perforated triangular center spot.
Notice the Tectron Cheshire perforation is just visible behind the three sides of the triangle.
View in the 2-inch CatsEye™ calibrated Cheshire—fourth image
The triangle fits precisely within the large perforation.
LASER EVALUATION
Q. What do you use the laser for?
A. I primarily use the laser for three tasks. First, to quickly evaluate collimation, second, to collimate the diagonal angle adjustment, and third, in combination with a visual collimation tool such as the Cheshire to quickly remove accumulated rotation and angle errors in the diagonal collimation.
Q. How do you evaluate collimation with the laser?
A. Because the entire optical system is collimated to the focuser axis, the first step is to make sure the laser spot (or holographic target) falls directly on the center of the primary mirror. If it doesn’t, it must be corrected before attempting to collimate the primary mirror with the laser or the Cheshire eyepiece. The laser provides a quick visual interpretation of the tilt of the diagonal—an elegant solution similar to the Cheshire’s handling of the tilt of the primary mirror. Both steps can be accomplished (with a little more sweat and a little less accuracy) with the sight tube, and that used to be the accepted method. Modern fast focus Dobsonians will quickly test your patience if you limit yourself to sight tube collimation. The process will continue to evolve as new tools become available.
Q. How do you set the diagonal adjustment with a laser?
A. If you have already done the preliminary mechanical and coarse alignment procedures, the laser can improve the accuracy of the diagonal’s angle collimation. This angle is defined by the intersection of the focuser axis and the optical axis. The diagonal may also require adjustment of the lateral, or transverse, adjustment screws on the diagonal. These set the skew of the diagonal and usually only require minor tweaking. However, they can get way out of adjustment with laser only collimation.
Three things have happened in the last few years that change the mechanics of diagonal collimation. First, inexpensive laser collimators, second, no-tool knobs on the back of the diagonal holder, and third, plastic washers on the rotational adjustment to provide complete “no-tool” diagonal collimation. Those little no-tool knobs just beg you to adjust them when the laser beam doesn't hit the center of the primary mirror. They're so easy to adjust, a lot of Newtonian owners have completely forgotten to consider diagonal rotation as the probable cause of the alignment error. The plastic washers that provide no-tool rotational adjustment allow the diagonal to rotate out of collimation when the ’scope is transported to a remote observing site.
When you use a laser to collimate the diagonal, and the laser beam hits the primary mirror above or below the center spot, correction can be applied by either rotating the diagonal or adjusting the transverse adjustment (top/bottom, up/down, forbidden, whatever) screws on the back of the diagonal holder. The question is, "Which way is the right way?" The answer can't be determined with a simple laser, but can be seen with any of the collimating tools. Look at figures 7 and 8 in New Perspectives... You can see positioned around the diagonal holder the silhouette of the four screw heads that secure the diagonal shell to the backplate (where the adjustment screws push or pull.) If the rotation and tilt angle are correct, the two unpaired screw heads will be barely visible. If not, only one will be visible. If the one towards the top is visible, it means the diagonal is skewed and pointing below the center of the primary mirror. Using the laser, adjust the diagonal adjustment screws to point the beam an inch or so above the center spot, then rotate the diagonal to bring the beam back down to the center. Check again visually and continue to adjust the diagonal until you can see the two unpaired screw heads equally. The two paired screw heads may be more or less visible than the two unpaired screw heads. They will only appear identical all the way around (as in figures 7 and 8) if the angle of the intercept is very close to ninety degrees. If the intercept angle is not precisely ninety degrees, the paired heads may look different from the unpaired heads—this does not affect collimation. Failing to correct a minimal skew error will also not affect the final accuracy of the collimation procedure. Because the error can become very pronounced with routine laser collimation procedures, it should not be overlooked. Remember, when it's right, the view in the collimation tools should look pretty much "textbook".
Q. What about the Barlowed laser?
A. Described by some as “the best collimation method for ’scopes with wobbly focusers”, the Barlowed laser (ref: Sky and Telescope, January, 2003, Collimation with a Barlowed Laser by Nils Olof Carlin) does an excellent job impersonating a Cheshire eyepiece. Where the Cheshire eyepiece can be sensitive to small eye position variations, the Barlowed laser is parallax free. In use, a laser is inserted into a Barlow lens that has been fitted with a perforated paper target on the lens side facing the primary mirror. The diverging laser beam exits the perforation and casts a circular spot an inch or two in diameter on the center of the primary mirror, illuminating the donut center spot. The primary mirror reflects the silhouette of the donut center spot back to the target fixed to the bottom of the laser, where the silhouette is aligned with the perforation by adjusting the primary mirror. If you don’t have a Cheshire eyepiece and rely on a laser for your collimation needs, this protocol is a necessity for precise primary mirror collimation. There are a few issues...
The Barlowed laser is another axial alignment tool--it will not resolve diagonal alignment errors (i.e., combined rotation/tilt errors.) You may encounter some difficulty with higher profile focusers where the Barlowed laser target can end up deep inside the focuser draw tube. This usually means the person collimating the ’scope will have to use a separate mirror to peer up into the focuser or will have to observe the reflection in the primary mirror at a distance. Neither solution offers the anticipated precision we’ve come to expect during Cheshire collimation (which also benefits from an axial pupil, and subsequently, an axial read.) It’s also important to keep the target inside the focal plane and the laser source outside the focal plane to keep the tool balanced (insensitive to focuser angular errors.) Recently, a commercially available Barlowed laser attachment was introduced that is inserted in the bottom of the focuser draw tube and provides a 45-degree slanted target area. The accessory should work with almost any simple laser collimator and facilitates the visual read from both the back and front of the optical tube assembly.
The only other concerns I’ve encountered with my Barlowed laser is potential light trespass and an impact on my dark adaptation. If the laser is bright enough to read in early evening twilight, you (and your observing neighbors) may find it uncomfortably bright in fully dark conditions.
AUTOCOLLIMATOR EVALUATION
Q. What exactly does the autocollimator do?
A. The autocollimator is uniquely capable of displaying and correcting axial errors of the primary mirror and focuser simultaneously. The multiple images that will ultimately be lined up behind the actual primary center spot can be collimated while the primary mirror axis is simultaneously observed with Cheshire-like accuracy. Alternately, the focuser axis can be resolved first and the autocollimator’s power can be turned on the primary mirror’s axial error. Laser for diagonal—Cheshire for primary mirror—and autocollimator for final alignment of the focuser axis to the primary mirror axis. If you experience difficulty stacking the multiple reflections on axis, you will need to finish with a precision (calibrated) Cheshire collimation of the primary mirror to minimize coma. You might want to try the new 2-inch Infinity II™ autocollimator (see below)--with its wider field of view and highly reflective first surface mirror it eliminates the guesswork as well as the need for multiple iterations.
The autocollimator is also the ultimate collimation “quality control” tool. Here, the autocollimator is used to turn the telescope optics back on themselves so that the power of the multiple reflections can be used to "see" the residual accumulated errors of the other tools. This is probably the most important part of the collimation procedure--the comparative analysis of the collimation provided by each of the tools. Once the scope is well collimated, it becomes a test bed for the tools themselves, and the user’s collimation skill. Subtle differences in eye positioning (parallax), inconsistencies in the focuser mechanisms, tools that misbehave when rotated... Any residual error(s) left over from hurried or careless single pass collimation procedures become glaringly obvious. I would venture a guess that there are many Newtonian users who have never had their ’scope precisely collimated!
When I "fine" collimate a scope, I carefully align the focuser axis with a laser--and the primary mirror axis with the Cheshire. I take my time, I look for subtle parallax errors, and I push the tolerance of the tools to their respective limits. I think my "proficiency" using the tools is, "above average". Even so, the autocollimator always reveals subtle errors that I can easily ferret out through the iterative process or using the carefully decollimated primary protocol. I'm not sure if visually I can detect the subtle collimation improvement, but then these errors do accumulate, and I like pushing my high Strehl ratio optics to their theoretical limits--which I do when the seeing permits (and it often does in Florida!)
For someone with less collimation "expertise" (and a fast focal ratio scope), the autocollimator may seem to be too critical and difficult to use or interpret. In this situation, the autocollimator plays the role of the "quality control" inspector, forcing the user (through the iterative process), to improve laser and Cheshire procedures--the end result being better collimation.
Finally, the autocollimator is not absolutely necessary for every Newtonian telescope or every Newtonian telescope owner! Some people have a knack for getting the collimation right with minimal tool support, others need help (or assurance), to finish the job. I don't think I can stress enough how much easier it is to use the new 2-inch Infinity II! If you've had problems with the 1.25-inch autocollimator, you will quickly become proficient with the 2-inch (and then fully understand what you were seeing "just off the edge of the field" in the 1.25!)
Q. How significant is obtaining the condition where the view through the autocollimator goes totally black as far as the reflection of the autocollimator mirror is concerned--what does this mean regarding the degree of collimation? I've been able to obtain this dark reflection, but don't think I've seen multiple images of anything (as has been discussed here.)
A. Making the autocollimator go black is significant because if it is not darkened you probably won’t see all of the reflected images (the wider view in the 2-inch autocollimator helps here!) A completely brightened autocollimator means the axial alignments are far enough off that the focuser axis reflects past the diagonal and out through the front of the tube assembly. You can check this yourself by setting the collimation to brighten the autocollimator and then waving your hand in front of the tube assembly—you will see the reflection of your hand inside the reflection of the autocollimator.
*When the combined axial errors are reduced sufficiently to contain the reflections between the primary mirror and the diagonal, the autocollimator will appear darkened (or mostly darkened with a 2-inch autocollimator.)
*When the collimation errors are further reduced to contain the reflections between the primary mirror and the autocollimator, the autocollimator will appear darkened and multiple reflections (three donuts if you’re using a donut center spot—or three triangles if you’re using a CatsEye™) will be observed. A fourth reflection may be observed with the more reflective first surface mirror in the 2-inch autocollimator.
*When all axial errors are eliminated, the autocollimator will appear darkened and at least two of the three (or four) reflections will appear stacked—either on or off the focuser axis—don’t stop here! Interestingly, when the axes intersect at the radius of curvature (the axes are not parallel) one of the reflections will disappear behind the primary mirror center spot and the remaining offset reflections can be observed to fade out completely off axis. There are variations here, but it’s obvious if there are remaining axial errors.
*When both axial errors are eliminated, the autocollimator will appear darkened and the multiple reflections will disappear, stacked on axis behind the first reflection of the primary mirror spot—the pupil of the autocollimator will also be centered in the reflections of the stacked images. The two axes coincide.
***The autocollimator is a very powerful tool when any axial error is present in the final collimation. Even when the optical axis error is perfectly parallel to the focuser axis, the resulting autocollimator evaluation will reveal multiple images (two separate stacks!), magnifying the collimation error. If the final autocollimator iteration leaves a close jumble, and the total offset observed is within the acceptable tolerance window for the focuser axis, you can finish collimating with the Cheshire and disregard the remaining focuser axis error revealed by subsequent autocollimator evaluation. Or, you can systematically remove the remaining error.
Q. What should I look for in the autocollimator reflections?
A. The view in the autocollimator is at first glance very similar to the Cheshire eyepiece—the actual reflection of the primary mirror center spot should appear centered in the reflection of the autocollimator mirror. Under close scrutiny, the autocollimator pupil should also appear centered in the actual reflection of the primary mirror center spot. Like the Cheshire or the Barlowed laser, collimation evaluation takes place very close to the actual focal plane. When the autocollimator pupil is observed in the center of the bright reflected ring (or triangle), the visual appearance is similar to the Barlowed laser with the added benefit that you are able to evaluate the collimation axially (through the pupil.) But wait, there's more! The other reflections that appear behind the mirror center spot reflection magnify any remaining errors, whether the errors are the result of focuser axis misalignment, primary mirror axis alignment, or both! When all of the reflections are stacked behind the actual primary mirror center spot, the autocollimator very accurately defines precise angular and axial collimation. One tool reveals all!
View in the autocollimator —notice the actual primary mirror center spot reflection (bright ring) is slightly offset to the one o’clock position (the pupil of the autocollimator is easily visible slightly decentered in the bright reflection.)
The brighter (focal plane) reflection to the right is the reflection of the center spot as it would appear if it were one focal length behind the primary mirror. The fainter (inverted) reflection to the left appears as if it were three focal lengths behind the primary mirror, and is displaced far enough from behind the primary center spot reflection to reveal part of the ring reflection’s perforation. Look closely, inside the perforation of the actual primary mirror center spot reflection (on the right side) and you will notice the perforation of the focal plane reflection just coming into view. At f/4.3, the angular collimation is very close, but the axial collimation (pupil) still needs improvement.
View in the autocollimator with a white perforated triangle. Autocollimator evaluation immediately following “critical” laser (diagonal) and Cheshire collimation. Note the bright upright triangle reflection slightly offset to the right, and the inverted triangle reflection to the left. (Image by VM—11-inch f/5.4 StarMaster)
Q. Why can’t I see the pencil or knife reflections in my StarMaster?
A. Because StarMaster telescopes use a notebook paper reinforcement ring to mark the primary mirror center, most of the autocollimator mirror reflection will be obscured by the ring. If you illuminate the ring itself (lift the cloth truss cover), you can use the reflections of the ring (instead of a pencil or knife) to evaluate and fine tune your collimation. After dark, the center spot can be illuminated with a red flashlight to evaluate collimation with the autocollimator (make sure that you observe both of the flanking reflections converging behind the actual primary mirror center spot.) If you’re using the new Infinty II™ 2-inch autocollimator--the knife reflections will be visible in the wider field of view.
Convergence with a small screwdriver blade used as the target. Notice the inverted reflection coming in from near 12 o’clock terminating behind the triangular center spot and the upright reflection positioned in front of the center spot. Also notice the brightening of the autocollimator mirror caused by the center of curvature image of the silver screwdriver blade.
Q. The reflections in the autocollimator are stacked—am I done?
A. Not necessarily. If the reflections are “closely” stacked but not centered on the autocollimator pupil, you may have created a condition where the focuser and primary axes are not perfectly coincident but they intersect at the radius of curvature. You may see additional offset reflections that “wink” in and out of view if you gently twist the diagonal or push one side of the focuser. You may also encounter a situation where the reflections are closely stacked, but there’s still some “undefinable” problem (more on this in “Practical Applications...”) If you check your collimation with the laser, the beam may indicate the primary mirror center spot, but the Cheshire eyepiece will indicate the primary mirror axis is offset (or vice versa or combination of the above.) When the reflections are precisely stacked, the actual primary mirror center spot will be all that is visible centered in the darkened field of the autocollimator. At this point you can verify that the Cheshire (or Barlowed laser) and the laser indicate correct alignment.
View in the autocollimator after minor correction to diagonal alignment (first iteration.) The alignment can be improved with additional iterations, ultimately resulting in the background reflections disappearing behind the bright primary mirror center spot. See movie clips... (Image by VM—11-inch f/5.4 StarMaster)
Movie Clip 1 - (244KB) Image sequence demonstrating the effect of diagonal decollimation in the autocollimator.
Movie Clip 2 - (1.1MB) Image sequence demonstrating the effect of primary mirror decollimation in the autocollimator.
Go to catseye.com to view the image sequence in .avi format.
Q. What are the allowable tolerances for precise collimation?
A. The allowable axial error at the focal plane is defined by the diffraction limited field diameter created by the primary mirror. Beyond this diameter, the effects of coma reduce the image performance below 0.8 Strehl. If your goal is to collimate the primary mirror so that at least the edge of the diffraction limited field just contacts the focuser axis, the Cheshire collimation (or equivalent) should not exceed 0.00035 times the focal ratio cubed. At f/4 that’s about .022 inch. Again, that’s the soft tolerance. A strict tolerance for high magnification, high definition observing would probably be more like half, or one third of the soft tolerance—one hundredth of an inch or less. At f/4.5 the soft tolerance window is .032 inch, and at f/5, .044 inch. Remember, the soft tolerance is really the minimum tolerance. You should also note that whatever tolerance you decide to allow, the placement of the primary mirror center spot should be accurate to no more than twice that amount. At f/4.x, it’s all about splitting hairs.
While the allowable tolerance window for the primary mirror axis collimation is defined by the coma “free” field at the focal plane, the tolerance window for the focuser axis collimation is defined as a defocusing tolerance at the focal plane (constrained to the coma free field.)
The depth of focus at f/4 with an allowable +/- 1/10 wavelength defocusing, is only .014mm (0.00056 inch). You could probably push for a tighter focus, but I'm not sure you would be able to achieve that level of focusing accuracy even with a modern 10:1 reduction gear focuser. One full revolution of the fine-focus knob on a FeatherTouch or JMI-DX1 results in a total axial travel of approximately 0.07 inch. To achieve focus within the +/- 1/10 wavelength defocusing tolerance, the fine focus knob position will need to be precisely adjusted to within 1/100th of a full revolution—not much more than a careful tweak!
To hold the defocusing error contribution from a tilted focuser axis to no more than 10-percent of the total wavefront error (1/12 wave) at the edge of the coma free field diameter, the axial tolerance should be no more than 0.03 times the primary mirror diameter, or 0.005 times the primary mirror diameter when a Paracorr is being used. For a fast focal ratio 10-inch aperture (with Paracorr), the tolerance is 0.05-inch--a bit more than 1mm. This tolerance, added to a 1/10th wave focuser tolerance, still falls comfortably inside the 0.8 Strehl at the edge of the coma free field of view. When the focus depth is barely more than the focal plane tilt, snap focus characteristics may be impacted, or the tilted focal plane may present a pseudo off-axis sweet spot.
There’s no penalty for arriving at an axial collimation that’s better than the tolerance threshold. Mechanical tolerances (flexure, lateral shifts, torque, etc.) can also impact axial collimation and force tighter optical tolerances. Digital imaging systems that push the optical resolution beyond Rayleigh (or even Sparrow) seem to demand all the precision you can muster.
PRACTICAL APPLICATIONS OF THE AUTOCOLLIMATOR
Q. How do I get the images to stack and remain centered in the autocollimator?
A. There are currently three protocols:
The first is the simple iterative process. Using a laser or sight tube, carefully set the diagonal alignment, then follow with a precision calibrated Cheshire alignment. Now insert the autocollimator and adjust the focuser or the diagonal (either will work when the primary mirror and focuser axes are already this close) to stack the reflections as closely as possible. Repeat the process until all tools indicate precise axial collimation.
I’ll call the second approach the “educated guess”. If you study the reflections, you may determine that after carefully aligning both axes, that the primary axis is contributing the observed error. (See “Analysis of the Autocollimator Reflections”.) If this seems to be the case, you can try carefully adjusting the primary alignment, and if subsequent Cheshire evaluation reveals a collimated condition--stop here. If the reflections indicate a focuser alignment error, you’re back to at least one more iterative cycle.
The third protocol I refer to as “the carefully decollimated primary” protocol. If you can see the second inverted (dimmest) reflection, you can use this method to minimize or even eliminate the iterative steps. Begin by carefully collimating the diagonal and primary mirror. Insert the autocollimator and carefully decollimate the primary mirror by adjusting the top primary mirror alignment screw (to minimize twisting the primary on its edge support(s) and constrain the motion to a simple “tilt” motion) until the reflections separate enough to reveal the upright and inverted reflections and the second inverted reflection. The “flanking” reflections (one upright, one inverted) should appear equidistant on either side of the bright primary mirror center spot reflection. What you’re looking for is the second, dimmer, inverted reflection. Limiting the alignment adjustments to the focuser or diagonal, the dim second inverted reflection should be positioned directly behind the bright primary mirror center spot reflection so that it appears like a “Star of David” (if you’re using a triangular center spot.) If you’re using a round ring center spot, the dim second inverted reflection will disappear behind the bright primary mirror center spot reflection--verify the focuser axis alignment with the laser. In this condition the visible distance between the two (equidistant!) flanking reflections magnifies the actual primary mirror axial error eight times! Finally, readjust the top primary mirror screw to stack the flanking reflections, and verify with the Cheshire.
I recommend using the tools at your disposal for each iterative cycle. In other words, if you’re going to use a laser, include it in the error reduction process. If you have a collimatible focuser, you can include it in the iterative procedure as well, assigning each tool to a collimatible element (i.e., laser for diagonal, Cheshire for primary, autocollimator for focuser.) If you prefer to leave the focuser collimation alone, you can alternate the laser and autocollimator when aligning the diagonal. Either procedure progressively reduces the alignment error(s).
When the images stack on the primary mirror center spot, the Cheshire will also indicate the center spot is accurately collimated—and the laser will be similarly precisely aligned. The primary mirror axis and the focuser axis are coincident. The full resolution of all of the available tools has been used to perfect the axial collimation.
"I didn't say it would be easy. I just said it would be the truth."
- Morpheus
ANALYSIS OF THE AUTOCOLLIMATOR REFLECTIONS
*Images from an 11-inch f/5.4 Newtonian. (see Movie Clips) Both images have been carefully decollimated to keep the observed reflections in a similar orientation. Infinity II 2-inch autocollimator. The slight offset error of the autocollimator reflection in the diagonal reflection does not affect the axial collimation. It does slightly affect the percentage illumination at the edge of the field of view.
Left Image--pure diagonal error--Note that the actual primary mirror center spot reflection (2nd from left) is centered, indicating the accurate alignment of the primary mirror axis. The focuser axis offset is indicated by the far left upright reflection (magnified two times via five reflections) and the far right inverted reflection (magnified four times via nine reflections). The second, dimmer inverted reflection can be seen between the actual primary mirror center spot reflection and the far right inverted reflection.
Right Image--pure primary error--Note that the actual primary mirror center spot reflection is no longer centered--it’s now slightly decollimated to the right. The primary axis offset is indicated by the far left reflection (magnified four times via five reflections) and the far right reflection (also magnified four times via nine reflections). The second, dimmer inverted reflection can be seen behind the actual primary mirror center spot reflection, indicating the accurate alignment of the focuser axis.
Once you’re comfortable with the various autocollimator protocols, it’s relatively simple to remove either the primary mirror axis error or the focuser axis error and then correct (or manipulate--as in the above images) the decollimation of the remaining axis.
Ray tracings and additional analysis of the reflections can be accessed at:
http://w1.411.telia.com/~u41105032/Acoll/Acoll.htm
OK, enough time for one last question...
Q. I’m still confused. Can I get you to collimate my ’scope?
A. Hmmm... it wouldn’t be the first, and it likely will not be the last. I enjoy demonstrating the procedures, interpreting image defects, commiserating when the seeing is miserable, and celebrating when it all comes together—the perfect image—the holy grail of the large aperture telescope owner.
Some will claim I’ve opened Pandora’s Box, in particular for those just getting started with their first ’scope—a 20-plus-inch, f/4-point-something, all-the-bells-and-whistles, high-performance really expensive model. I contend the box has been open for several years already, and it’s up to us to understand and master our sometimes possessed ’scopes (at least sometimes they seem that way!) I would also add, for the less experienced big Dob owner, that obsessing over perfect collimation should never preempt a good observing session. On the other hand, practice does make perfect...
Vic Menard
vmenard@tampabay.rr.com