•

# Why do we collimate by tilting the primary?

51 replies to this topic

Vanguard

• topic starter
• Posts: 2052
• Joined: 15 Sep 2010
• Loc: Basel, Switzerland

Posted 07 May 2013 - 08:09 AM

I have a question about collimation. The way that we all do it involves tilting the primary and secondary mirrors with respect to a fixed focuser. The goal being to get the objective's image plane perpendicular to the focuser axis. However, wouldn't it be easier for the user to perform the alignment with respect to a fixed primary mirror? So adjusting only the tilt of the secondary and the tilt of the focuser? I'm sure there's some good reason for not doing this, but I don't know what it is.

### #2 Thomas Karpf

Thomas Karpf

Vanguard

• Posts: 2080
• Joined: 09 Feb 2009
• Loc: Newington, CT

Posted 07 May 2013 - 08:16 AM

If the primary was fixed and wasn't in the correct position, how would you collimate the scope?

Vanguard

• topic starter
• Posts: 2052
• Joined: 15 Sep 2010
• Loc: Basel, Switzerland

Posted 07 May 2013 - 08:20 AM

As I say: by tilting the focuser. That wouldn't work?

### #4 Thomas Karpf

Thomas Karpf

Vanguard

• Posts: 2080
• Joined: 09 Feb 2009
• Loc: Newington, CT

Posted 07 May 2013 - 08:38 AM

Imagine a primary mirror that is HIDEOUSLY mis-collimated; the mirror is in the tube at a 45 degree angle.

You look through the focuser.

The focuser is square to the tube. The secondary is in the correct position; it is centered under the focuser.

The secondary is aimed correctly. The center of the primary mirror is centered as seen in the secondary.

The primary mirror is badly misaligned.

What you see looking through the focuser into the secondary mirror off the primary mirror is the side of the tube, not the sky.

Vanguard

• topic starter
• Posts: 2052
• Joined: 15 Sep 2010
• Loc: Basel, Switzerland

Posted 07 May 2013 - 08:47 AM

I'm not sure, but I think we may be talking cross-purposes; forget the squaring of the focuser in the tube: it's unnecessary. The point is to adjust the focuser and secondary as a pair. With that in mind, shouldn't it be possible to get everything configured as you'd like?

Perhaps the sticking point is that this approach might be more likely to require translation of the focuser or secondary in addition to tilt. The current approach we employ usually only involves tilt adjustments.

### #6 Dick Jacobson

Dick Jacobson

Viking 1

• Posts: 958
• Joined: 22 Dec 2006
• Loc: Plymouth, Minnesota, USA

Posted 07 May 2013 - 08:54 AM

The primary purpose of collimation is to put the optical axis of the main mirror (at the center of the image plane) down the center of the focuser. If the main mirror optical axis was misaligned, you would have to laterally move both the secondary (so the axis hits the center of the secondary) and the focuser (so the beam from the secondary passes through the center of the drawtube). Much easier to just adjust the main mirror.

### #7 Thomas Karpf

Thomas Karpf

Vanguard

• Posts: 2080
• Joined: 09 Feb 2009
• Loc: Newington, CT

Posted 07 May 2013 - 09:00 AM

Because tilting the components is easier than translating the focuser, secondary, and/or primary mirror a couple of millimeters on some direction.

Most focusers are bolted to the tube; there's no way to easily nudge the focuser a couple of millimeters up the tube or across the surface of the tube.

Translating the secondary is slightly more possible as you can adjust the screws holding the spider in place.

Translating the primary is similarly possible.

Tilting is easier and faster. Tilting the primary mirror (by lifting one edge a millimeter) moves the axis of the mirror AT THE SECONDARY by the 1mm x focal ratio of the primary mirror.

Vanguard

• topic starter
• Posts: 2052
• Joined: 15 Sep 2010
• Loc: Basel, Switzerland

Posted 07 May 2013 - 09:01 AM

Tom, I was thinking theoretically so obviously I was imagining a highly modified focuser. I wasn't envisaging something you could easily do in a stock scope. I see what you mean about the translations, though. It likely would be more confusing to have to contend with tilting the secondary and the focuser and also translating one or both of those

### #9 Thomas Karpf

Thomas Karpf

Vanguard

• Posts: 2080
• Joined: 09 Feb 2009
• Loc: Newington, CT

Posted 07 May 2013 - 09:13 AM

Tom, I was thinking theoretically so obviously I was imagining a highly modified focuser. I wasn't envisaging something you could easily do in a stock scope. I see what you mean about the translations, though. It likely would be more confusing to have to contend with tilting the secondary and the focuser and also translating one or both of those

You would still need the center of the cone of light from the primary coming cleanly and squarely up the focuser to tye eyepiece. That's going to be MUCH easier to do by eye by tilting the primary than by trying to figure out where exactly the focuser needs to be and how it needs to be tilted to 'catch' the light correctly.

### #10 careysub

careysub

Vanguard

• Posts: 2114
• Joined: 18 Feb 2011
• Loc: Rancho Cucamonga, CA

Posted 07 May 2013 - 09:47 AM

I'm not sure, but I think we may be talking cross-purposes; forget the squaring of the focuser in the tube: it's unnecessary. The point is to adjust the focuser and secondary as a pair. With that in mind, shouldn't it be possible to get everything configured as you'd like?

Perhaps the sticking point is that this approach might be more likely to require translation of the focuser or secondary in addition to tilt. The current approach we employ usually only involves tilt adjustments.

Recall that a focuser is required to support up to several pounds of optical equipment (an ES100 25mm plus a Paracorr weighs about 3.5 pounds) in a cantilever arrangement (they stick out and are supported only at one end) which is mechanically disadvantageous. This requires the focuser to be very rigidly mounted. Also, keeping the focuser plus focuser mounting board light in the UTA is an important design goal. A system that is strong and rigid, yet allows tilting and translating the focuser sounds complex and heavy.

### #11 Jon Isaacs

Jon Isaacs

ISS

• Posts: 51285
• Joined: 16 Jun 2004
• Loc: San Diego and Boulevard, CA

Posted 07 May 2013 - 09:56 AM

I have a question about collimation. The way that we all do it involves tilting the primary and secondary mirrors with respect to a fixed focuser. The goal being to get the objective's image plane perpendicular to the focuser axis. However, wouldn't it be easier for the user to perform the alignment with respect to a fixed primary mirror? So adjusting only the tilt of the secondary and the tilt of the focuser? I'm sure there's some good reason for not doing this, but I don't know what it is.

The goal of collimation is to have the optical axes in alignment and it is desirable to have the secondary mirror properly centered/offset. I believe that if the primary were fixed, this would require adjusting not only the tilt of the secondary and focuser but also could require some lateral translation.

The focuser must not only be parallel to the optical axis but it also must be concentric, meeting both these critera means that were the position of the primary fixed, the position of the focuser or the secondary might need be moved.

Jon

Fly Me to the Moon

• Posts: 6392
• Joined: 28 Apr 2003
• Loc: Atlanta, GA

Posted 07 May 2013 - 10:06 AM

The only way I could see this working is in a truss design, where you could adjust the truss lengths in small increments to move the whole UTA to center it on the axis of the primary. In a tube it won't work because if the primary is off-center the tub itself will cut off light.

But even in the truss example, that would be a lot more complex than simply tilting the primary.

### #13 GaryS

GaryS

Ranger 4

• Posts: 386
• Joined: 30 Oct 2006

Posted 07 May 2013 - 10:38 AM

I have a question about collimation. The way that we all do it involves tilting the primary and secondary mirrors with respect to a fixed focuser. The goal being to get the objective's image plane perpendicular to the focuser axis. However, wouldn't it be easier for the user to perform the alignment with respect to a fixed primary mirror? So adjusting only the tilt of the secondary and the tilt of the focuser? I'm sure there's some good reason for not doing this, but I don't know what it is.

I have to think about this a bit more, but it's not obvious to me why it wouldn't work. The advantage would be that you could make your adjustments while looking in the collimating eyepiece -- and that's a significant plus.

Regards,
Gary

Vanguard

• topic starter
• Posts: 2052
• Joined: 15 Sep 2010
• Loc: Basel, Switzerland

Posted 07 May 2013 - 10:45 AM

Yes, that's how I saw it. Perhaps for tweaking the collimation, replacing primary tilt with focuser tilt might work. But for larger errors one might well have to translate the focuser and that would be awkward for the reasons mentioned above. I reckon it should be fairly easy to make a light, low-profile, and rigid tilt adjustment system at the focuser. Perhaps it would work as a viable option in combination with primary tilt adjustment. Then again, may be more trouble than it's worth. It was a just a passing thought I had...

### #15 Thomas Karpf

Thomas Karpf

Vanguard

• Posts: 2080
• Joined: 09 Feb 2009
• Loc: Newington, CT

Posted 07 May 2013 - 11:24 AM

Frankly, if you want to tweak primary collimation from your eyepiece, consider how you would directly tweak the primary.

Possibilities:

1) threaded rods that run down the length of the scope from the primary cell to the vicinity of the focuser so you can adjust the primary while looking through the focuser,
2) electrical controls at the focuser end that allow you to tweak remotely operated actuators at the primary end.

### #16 Jason D

Jason D

Fly Me to the Moon

• Posts: 6971
• Joined: 21 Oct 2006
• Loc: California

Posted 07 May 2013 - 12:35 PM

Rob,
Assuming the primary mirror optical axes is not too far off, you will need to adjust both the secondary mirror and the focuser to achieve the following two alignments:
1- Focuser axis has to intercept the optical axis
2- Above interception point has to be located on the secondary mirror surface
3- Secondary surface plane angle has to be positioned to reflect the focuser axis to the primary center and vice versa
4- Secondary mirror has to appear centered and rounded from the focuser end
Too many alignment to get things right
Below would be the proposed steps:
1- Insert a cheshire then adjust the focuser until the primary mirror looks aligned (cheshire reflection vs. center spot reflection)
2- Assess secondary mirror appearance. If not centered, then adjust the secondary mirror appropriately but that will disrupt step#1. Bear in mind these secondary adjustments will involve the center bolt – not fun.
3- Repeat steps 1 & 2 until:
a. Chesire alignment is met
b. Secondary mirror appearance looks centered and rounded
Actually, there is a unique benefit to the above system. You can shift the focal plane position with respect to the focuser by ~-10mm by pointing the focuser higher or ~+10mm by pointing it lower.
Ummm, if you use a laser collimator with a holographic attachment then the process might be easier than above but it will still be more involved than the well-established collimation steps.

Jason

### #17 Jason D

Jason D

Fly Me to the Moon

• Posts: 6971
• Joined: 21 Oct 2006
• Loc: California

Posted 07 May 2013 - 12:44 PM

Actually, there is a unique benefit to the above system. You can shift the focal plane position with respect to the focuser by ~-10mm by pointing the focuser higher or ~+10mm by pointing it lower.

Well, that could also cause nightmares. Imagine after you are done you find out some of your EPs can no longer come to focus.
Jason

Vanguard

• topic starter
• Posts: 2052
• Joined: 15 Sep 2010
• Loc: Basel, Switzerland

Posted 07 May 2013 - 02:54 PM

Thanks for the detailed description, Jason.

### #19 Vic Menard

Vic Menard

Fly Me to the Moon

• Posts: 5225
• Joined: 21 Jul 2004

Posted 07 May 2013 - 03:10 PM

...wouldn't it be easier for the user to perform the alignment with respect to a fixed primary mirror? So adjusting only the tilt of the secondary and the tilt of the focuser? I'm sure there's some good reason for not doing this, but I don't know what it is.

Of the three tilt elements (primary mirror, secondary mirror, and focuser) the primary mirror is unique in that even fairly significant tilt adjustments have little or no impact on the alignments of the other two elements (secondary mirror and focuser). This means one can sort out the focuser/secondary mirror geometry/alignments without consideration of the critical primary mirror tilt alignment, which can be corrected as the final step. Fixing the primary mirror alignment essentially reverses the procedure, which requires more mechanical accommodations with respect to the focuser/secondary mirror geometry/alignments and significantly complicates the alignment procedure.

FWIW, a similar problem would entail fixing the optical axial alignments to the mechanical OTA axis for precise DSC performance, assuming of course that the OTA axis is indeed orthogonal to the mount axes. In such a situation, if the OTA axis is the reference axis, (and the primary mirror is precisely centered in the OTA) then the primary mirror axis by definition is fixed. On the surface, this sounds to be a reasonable goal. But getting there is far more complicated than basic optical alignment...

### #20 Nils Olof Carlin

Nils Olof Carlin

Surveyor 1

• Posts: 1860
• Joined: 26 Jul 2004

Posted 08 May 2013 - 03:45 AM

However, wouldn't it be easier for the user to perform the alignment with respect to a fixed primary mirror? So adjusting only the tilt of the secondary and the tilt of the focuser?

One method that has been advocated (and possibly even used) is to start by removing the secondary and aligning the primary to the mounting hole in the spider - this done, the primary is left fixed. The secondary is replaced, and moved/tilted as needed to finish collimation - even with a non-adjustable focuser. This has the presumptive advantages of getting the offset right, provided it is included in the secondary mount, and having the optical axis centered at the level of the spider (not highly critical anyway).
I've never tried it, but the adjustment of the secondary sounds tedious and iterative to me, and I can't see that the result would be better.
Anyway, the critical adjustment in optical collimation is that of the primary's axis, not the focuser axis - some have even advocated doing the final centering of the primary's axis by fine-adjusting the secondary tilt (this will center the sweet spot, at the cost of de-centering the focuser axis by a much larger amount, but you may still find it acceptable in terms of image quality).

To me, starting with the focuser axis seems the easiest way, not needing iterations (at least if the miscollimation is moderate to start with).

Nils Olof

### #21 Thomas Karpf

Thomas Karpf

Vanguard

• Posts: 2080
• Joined: 09 Feb 2009
• Loc: Newington, CT

Posted 08 May 2013 - 05:05 AM

One method that has been advocated (and possibly even used) is to start by removing the secondary and aligning the primary to the mounting hole in the spider - this done, the primary is left fixed.

Okay, so how EXACTLY does one do 'FIX' the primary so it does not move? Isn't that the main job of current mirror cells (all of which require occasional collimation)?

Primary cells do not hold the mirror firmly. ABSOLUTELY preventing the mirror from moving at all under any circumstances ABSOLUTELY requires that something be touching the surface of the mirror, thus holding the mirror firmly against the mirror cell.

Until/unless this is done, there is no way to guarantee that the mirror will not move. Consequently, you will need to collimate the primary.

As the primary mirror is a huge honkin' chunk of glass which WILL move against its constraints as you're driving and hitting potholes, speed bumps and vague dips and bumps in the road, you're going to need to collimate the primary every so often.

Vanguard

• topic starter
• Posts: 2052
• Joined: 15 Sep 2010
• Loc: Basel, Switzerland

Posted 08 May 2013 - 07:37 AM

Tom, I think you're taking all this a bit too literally. We're just tossing ideas about. Clearly the consensus method for collimating a reflector has been arrived at because it is the most effective. Thinking about the alternatives, as Nils just did, is an interesting way of showing why. The tone of Nils' writing (e.g. the statement "and possibly even used") suggests that he doesn't consider what he describes to be worth the effort. The concern of stabilising the primary is the obvious issue, particularly at larger mirror sizes.

### #23 Jon Isaacs

Jon Isaacs

ISS

• Posts: 51285
• Joined: 16 Jun 2004
• Loc: San Diego and Boulevard, CA

Posted 08 May 2013 - 09:34 AM

Tom, I think you're taking all this a bit too literally. We're just tossing ideas about. Clearly the consensus method for collimating a reflector has been arrived at because it is the most effective. Thinking about the alternatives, as Nils just did, is an interesting way of showing why. The tone of Nils' writing (e.g. the statement "and possibly even used") suggests that he doesn't consider what he describes to be worth the effort. The concern of stabilising the primary is the obvious issue, particularly at larger mirror sizes.

I thought Tom's point was a very good one. It is all well and good to consider issues like this in their abstract forms but Tom added a dose of reality...

The primary cannot be tightly held, it can shift. It is unlikely that the focuser angle shifts, it is also difficult to adjust..

As an engineer/mechanic, ideally you adjust things that do shift, that makes for a stable mechanical system. If you are continually adjusting Y to make up for shifts in X, that can be an unstable system...

Myself, I have a hard time thinking about collimating a Truss scope after assembly without adjusting the tilt of the primary mirror.

Jon

### #24 FirstSight

FirstSight

Duke of Deneb

• Posts: 11107
• Joined: 26 Dec 2005
• Loc: Raleigh, NC

Posted 08 May 2013 - 09:44 AM

I have a question about collimation. The way that we all do it involves tilting the primary and secondary mirrors with respect to a fixed focuser. The goal being to get the objective's image plane perpendicular to the focuser axis. However, wouldn't it be easier for the user to perform the alignment with respect to a fixed primary mirror? So adjusting only the tilt of the secondary and the tilt of the focuser? I'm sure there's some good reason for not doing this, but I don't know what it is.

Because the primary mirror cell has collimation knobs, the focuser doesn't!

### #25 Jason D

Jason D

Fly Me to the Moon

• Posts: 6971
• Joined: 21 Oct 2006
• Loc: California

Posted 08 May 2013 - 12:08 PM

Tom, I think you're taking all this a bit too literally. We're just tossing ideas about. Clearly the consensus method for collimating a reflector has been arrived at because it is the most effective. Thinking about the alternatives, as Nils just did, is an interesting way of showing why. The tone of Nils' writing (e.g. the statement "and possibly even used") suggests that he doesn't consider what he describes to be worth the effort. The concern of stabilising the primary is the obvious issue, particularly at larger mirror sizes.

I thought Tom's point was a very good one. It is all well and good to consider issues like this in their abstract forms but Tom added a dose of reality...

The primary cannot be tightly held, it can shift. It is unlikely that the focuser angle shifts, it is also difficult to adjust..

As an engineer/mechanic, ideally you adjust things that do shift, that makes for a stable mechanical system. If you are continually adjusting Y to make up for shifts in X, that can be an unstable system...

Myself, I have a hard time thinking about collimating a Truss scope after assembly without adjusting the tilt of the primary mirror.

Jon

But what does the primary mirror shift have to do with Rob's question? If the primary mirror shifts then it will have the same negative impact on both a typical scope and the one Rob described.

Jason

# Imaging/Sketching Contest

 Cloudy Nights LLC Cloudy Nights Sponsor: Astronomics