How measure collimation error

[SteveBz]

Hi All,

I imagine there must be some numeric measure of collimation.  Eg it might be the angular separation of the central image ray from the occular axis.

So my question is what is the measure and how to determine it?

Regards,

Steve.

[laowhoo]

We use these 2 in conjunction, that of the optical axis/mechanical axis error types of which there are 4

http://web.telia.com/~u41105032/kolli/kolli.html

and that of the allowable error as per focal ratio, for example,

http://www.telescope-optics.net/newtonian_off_axis_aberrations.htm

http://www.telescope-optics.net/induced2.htm

http://www.telescope-optics.net/two-mirror2.htm

This is another excellent site for info

http://www.rfroyce.com/mirror_performance.htm

Legault is good too, whether for imaging or a treatment of collimation effects as per MTF

http://www.astrophoto.fr/collim.html

There are tables for focal ratio allowances too. We've just torn down our scope after 5 years and are beginning over according to these, including checking the center spots on primary/secondary, offset, etc.

Here's a thread giving the formula for allowable PAE (Primary Axis Error)

https://www.cloudynights.com/topic/514681-my-collimation-ring-is-off-center-should-i-worry/

We built for f/6 partly b/c the allowable PAE is 1.4mm, nearly double that for f/5. I just can't find the table.

Of course, a star test will solve for any of these (visual) and Suiter is the bible

http://www.willbell.com/tm/tm5.htm

and is an excellent read.

How could I forget? Vic Menard worked with Nils Olof Carlin and is our collimation master now that Nils has passed the baton

http://www.vicmenard.com/telescopes/addendum-to-perspectives.html

so anything by him. He has lotsa posts at CN.

If you're imaging there's this program

http://www.goldastro.com/goldfocus/collimation_techniques.php

with calculator

http://www.goldastro.com/goldfocus/focus_calculator.php

but we steer well clear of such.

[SteveBz]

Amazing Iaowhoo,

Fantastic set of links. I haven't read them all yet. But I even might buy one of those books!

Thanks a lot

Steve.

[SteveBz]

6 hours ago, laowhoo said:

We use these 2 in conjunction, that of the optical axis/mechanical axis error types of which there are 4

http://web.telia.com/~u41105032/kolli/kolli.html

and that of the allowable error as per focal ratio, for example,

http://www.telescope-optics.net/newtonian_off_axis_aberrations.htm

http://www.telescope-optics.net/induced2.htm

http://www.telescope-optics.net/two-mirror2.htm

This is another excellent site for info

http://www.rfroyce.com/mirror_performance.htm

Legault is good too, whether for imaging or a treatment of collimation effects as per MTF

http://www.astrophoto.fr/collim.html

There are tables for focal ratio allowances too. We've just torn down our scope after 5 years and are beginning over according to these, including checking the center spots on primary/secondary, offset, etc.

Here's a thread giving the formula for allowable PAE (Primary Axis Error)

https://www.cloudynights.com/topic/514681-my-collimation-ring-is-off-center-should-i-worry/

We built for f/6 partly b/c the allowable PAE is 1.4mm, nearly double that for f/5. I just can't find the table.

Of course, a star test will solve for any of these (visual) and Suiter is the bible

http://www.willbell.com/tm/tm5.htm

and is an excellent read.

How could I forget? Vic Menard worked with Nils Olof Carlin and is our collimation master now that Nils has passed the baton

http://www.vicmenard.com/telescopes/addendum-to-perspectives.html

so anything by him. He has lotsa posts at CN.

If you're imaging there's this program

http://www.goldastro.com/goldfocus/collimation_techniques.php

with calculator

http://www.goldastro.com/goldfocus/focus_calculator.php

but we steer well clear of such.

Ps, do you think there is a metric on those rings like HFD is for focus, maybe something about the eccentricity of the rings.

Steve

[laowhoo]

I see you have an f/5 on go-to so I'm assuming you're fully offset. But you image too(?...your sig), wh/ would rattle me too much to even delve into. But if I were imaging, I'd sure be using that calculator at Goldastro and would've bought the book I've seen mentioned here, Making Every Photon Count. But that's just me--I even research products after I've bought them (won't say the word for this). I like to have too much info--figure whatever settles in the bottom of the glass is useful, at least for me.

By rings are you talking about the diffraction rings of a simple star test? The spot diagram, and if so, at pixel dimension? Royce is treating the spot for visual optimization:

The following spot diagrams were generated for a 10 inch paraboloidal mirror having focal ratios of  f/4, f/4.5, f/5, f/6 and f/8. The indicated off-axis amount in degrees shown above each spot is the half angle, so one should multiply times 2 to get the total field. As an aid in analyzing the spot diagrams, one should understand that the ideal is that all the light be contained within the Airy disk, or as it is sometimes known, the spurious disk. This is the area where, in a perfect optical system, approximately 82 percent of all the light should be concentrated. The remaining light is distributed around the disk in the form of concentric rings, the so-called diffraction circles. (On the very best of nights these rings can be observed with a fine optical system.) Practically speaking, the quarter wave Rayleigh criterion can be achieved with a blur spot approximately twice the size of the Airy disk. After this, the image will begin to appear degraded.

Failing suitable conditions for a star test, you can set up an artificial test. But this isn't going to help you for imaging purposes I'm guessing. In our case, I try to tune for an f/5 or f/4 primary, and then hope I've gotten the most out of our more forgiving f/6. I aim high to hit the middle, but I also like to collimate. Realistically, I'm sure my efforts aren't even detectable. Do a quick search of Rayleigh/Dawes, but again, for imaging.......I mean, when I saw that program reading the spot diagram at half/pixel scales, well, color me impressed. You imagers are a different breed. I just tried to offer a few resources.

p.s. If, OTOH, by ring eccentricity you mean these

http://www.telescope-optics.net/diffraction_pattern_and_aberrations.htm

then you're maybe looking at primary aberrations of various sorts.

 

[SteveBz]

9 hours ago, laowhoo said:

bought the book I've seen mentioned here, Making Every Photon Count.

Yes I did read that, it was very useful, in fact I wished I'd read it earlier.

9 hours ago, laowhoo said:

I see you have an f/5 on go-to so I'm assuming you're fully offset.

It's not really goto, more Arduino. What do you mean by fully offset?

 

9 hours ago, laowhoo said:

p.s. If, OTOH, by ring eccentricity you mean these

http://www.telescope-optics.net/diffraction_pattern_and_aberrations.htm

then you're maybe looking at primary aberrations of various sorts.

Well, sort of, I just think there should be a numeric calculation that can be automated.

Regards

Steve.

[laowhoo]

Afraid I don't know Arduino. When I've come across anything go-to I move on, but I've seen/heard that if your secondary isn't fully offset, your go-to will lose its bearings above certain latitudes (altitude of the scope, and where we maybe spend more time b/c seeing gets better for us w/ less atmosphere to have to cut through, less skyglow, etc.). The full offset of the secondary is to move it down (toward primary) and back (away from the focuser) on the mechanical axis, allowing the light cone to properly fold on the secondary such that it maximally illuminates at the EP field stop. The modern method, for visual only, is simply to sight tube through the focuser and adjust the secondary until it's fully concentric w/ the sight tube/focuser. When we do that we've actually offset the secondary, by tilting it down slightly and wh/ sort of mimics a true mechanical offset. But yeah, logic or whatever intimates that we would just create a perfectly geometric mechanical axis from primary to secondary to EP, but b/c of the 45* tilt of the secondary, the light cone off the primary hits the bottom of the secondary first, the top of it secondarily, and the resultant fold spills some of the area of 75% illumination of the cone over the top of the field stop of the EP. Hence the need to offset. For visual we didn't need it (beyond the modern tilt method), but I built it in anyway b/c we were already there. It came to about 2mm for f/6, an amount we wouldn't detect visually, but wh/ would show up in imaging and in any go-to. Just wanted to be safe, given the compounding of the possible collimation errors in the optical train, and b/c we use an additional star diagonal. At f/5 your offset will be more.

As for the numeric calculation that can be automated, that's what the various calculators are for, whether they be the simplest ray trace program or that amazing spot diagram program calculating at half-pixel scales. And if in this you're referencing the diffraction rings/Airy disk of collimation accuracy, a simple PAE calculation (primary axis error of the optical axis) will give you the amount allowed for visual, as given by Royce above. At f/6 we can be off center by 1.4mm w/o noticing, but I think that at f/5 you have about half that allowable error before you notice any image degradation. I imagine that for imaging, though, your constraints are much tighter. You'll also have to decide whether your "ring eccentricities" are of the sort shown by miscollimation in wh/ the star test shows the coma aberration inherent in all paraboloidal primaries, or are of the primary figure sort as indicated by Suiter et al., the most common being a turned down edge. So where/how are you getting these ring eccentricities if not by a star test? Are you just getting coma at the center of field of view under prime focus? I hope somebody jumps in to help b/c I'm feeling a little out of my depths.

[SteveBz]

On 08/01/2018 at 16:55, laowhoo said:

Afraid I don't know Arduino. When I've come across anything go-to I move on, but I've seen/heard that if your secondary isn't fully offset, your go-to will lose its bearings above certain latitudes (altitude of the scope, and where we maybe spend more time b/c seeing gets better for us w/ less atmosphere to have to cut through, less skyglow, etc.). The full offset of the secondary is to move it down (toward primary) and back (away from the focuser) on the mechanical axis, allowing the light cone to properly fold on the secondary such that it maximally illuminates at the EP field stop. The modern method, for visual only, is simply to sight tube through the focuser and adjust the secondary until it's fully concentric w/ the sight tube/focuser. When we do that we've actually offset the secondary, by tilting it down slightly and wh/ sort of mimics a true mechanical offset. But yeah, logic or whatever intimates that we would just create a perfectly geometric mechanical axis from primary to secondary to EP, but b/c of the 45* tilt of the secondary, the light cone off the primary hits the bottom of the secondary first, the top of it secondarily, and the resultant fold spills some of the area of 75% illumination of the cone over the top of the field stop of the EP. Hence the need to offset. For visual we didn't need it (beyond the modern tilt method), but I built it in anyway b/c we were already there. It came to about 2mm for f/6, an amount we wouldn't detect visually, but wh/ would show up in imaging and in any go-to. Just wanted to be safe, given the compounding of the possible collimation errors in the optical train, and b/c we use an additional star diagonal. At f/5 your offset will be more.

As for the numeric calculation that can be automated, that's what the various calculators are for, whether they be the simplest ray trace program or that amazing spot diagram program calculating at half-pixel scales. And if in this you're referencing the diffraction rings/Airy disk of collimation accuracy, a simple PAE calculation (primary axis error of the optical axis) will give you the amount allowed for visual, as given by Royce above. At f/6 we can be off center by 1.4mm w/o noticing, but I think that at f/5 you have about half that allowable error before you notice any image degradation. I imagine that for imaging, though, your constraints are much tighter. You'll also have to decide whether your "ring eccentricities" are of the sort shown by miscollimation in wh/ the star test shows the coma aberration inherent in all paraboloidal primaries, or are of the primary figure sort as indicated by Suiter et al., the most common being a turned down edge. So where/how are you getting these ring eccentricities if not by a star test? Are you just getting coma at the center of field of view under prime focus? I hope somebody jumps in to help b/c I'm feeling a little out of my depths.

Wow. Great answer. I'll need to read it two or three times to fully grasp it, preferably when the kids are in bed.

Steve.