Jump to content


Photo

wsDK - weak secondary Dall Kirkham telescope

  • Please log in to reply
101 replies to this topic

#51 dave brock

dave brock

    Viking 1

  • -----
  • Posts: 609
  • Joined: 06 Jun 2008
  • Loc: Hamilton, New Zealand

Posted 30 January 2013 - 12:44 AM

Photo's full credit Glenn Burgess, scope owner.(possibly watching here but may not be registered to post)

Dave

#52 MKV

MKV

    Vanguard

  • *****
  • Posts: 2476
  • Joined: 20 Jan 2011

Posted 30 January 2013 - 11:53 AM

Hi Dave, I was wrong in assuming the viisble diameter of M22 was its angular diameter. Apologies. Apparently the true size of M22 is about 3 times its visible size. This can be easily ascertained from this site. The visible cluster seems to subtend only about 7.5 arminutes. The reference star shown below is 10 arcmin from the visible center, which makes Glenn's entire frame 0.5 degrees across.

This makes the correcton seen (absence of coma) a lot more reasonable with even commercial coma corrector, and not so surprising after all for such a narrow field. But the picture is a telling example of what simplified Cassgerain (DK) optics together with commercial coma correctors can do for imaging narrow angular field, deep sky object.
Regards,
Mladen

Attached Files



#53 MKV

MKV

    Vanguard

  • *****
  • Posts: 2476
  • Joined: 20 Jan 2011

Posted 30 January 2013 - 12:15 PM

Using the new field 0.5 degree size, here is again (what I beileve is) Mike Jones' version of an f/6.5 DKC, with images shown 0.25 degrees off axis. Diffraciton-limited to the edge.

Attached Files



#54 mark1234

mark1234

    Sputnik

  • -----
  • Posts: 31
  • Joined: 15 Apr 2008

Posted 01 February 2013 - 05:16 PM

Mladen,
well made point on the image scale anamoly.


By the way, I've never come across any precision optics which didn't have provision for adjustment. I agree sensitivity to collimation can be an issue, but more often it is only really a major problem when the resolution of the means of adjustment is too crude or sloppy to cope with the precision required - for example cheap binos and blobs of superglue on the corners of prisms.
It's when the precision required of a lens in its' cell exceeds the clearance needed for thermal changes that one really runs into trouble, ie less than a gnats wing (a few microns). Of course seeing will spread all spots into jiggling blobs.

Mark

#55 mark1234

mark1234

    Sputnik

  • -----
  • Posts: 31
  • Joined: 15 Apr 2008

Posted 01 February 2013 - 05:40 PM

Hello Ed,
I looked up this thread because of another reference to it - reading from the beginning I thought this scope was up and running and then came across your status information. The secondary can really only be finished/figured in combination with the primary - at the correct separation/focal distances. It is normal to do it this way by any optician - which of your two opticians will take the blame if the SA is not correct!?
25" is a high-end piece of no doubt expensive optics and that figure on the primary needs to be integrated into an overall corrected system.

Mark

#56 Ed Jones

Ed Jones

    Mercury-Atlas

  • *****
  • Posts: 2673
  • Joined: 06 Apr 2004
  • Loc: Sin-sin-atti

Posted 01 February 2013 - 08:30 PM

Mark,

The secondary can really only be finished/figured in combination with the primary - at the correct separation/focal distances.



No in this design the secondary is spherical and can be checked against a test plate which can be tested with a KE. So the remaining problem is getting the airspaces and lenses right. It can't work without the lenses in place anyway. This isn't an easy scope to make at least for a beginner.

#57 MKV

MKV

    Vanguard

  • *****
  • Posts: 2476
  • Joined: 20 Jan 2011

Posted 01 February 2013 - 09:02 PM

Mark, the giggling blobs are much more likely in a 25" then smaller scopes. In fact, I seriously doubt you'll ever see an Airy disc, but wou can always use a 7.5 inch mask and have an f/20 unobstructed ("off axis") Cassegrain.

Up to probably 16 inches, you can see good Airy disks on good nights. After than it becomes progressively more difficult due to speckling.

Mladen

#58 siriusandthepup

siriusandthepup

    Mariner 2

  • *****
  • Posts: 212
  • Joined: 14 Feb 2006
  • Loc: Central Texas, USA

Posted 01 February 2013 - 09:14 PM

The secondary can really only be finished/figured in combination with the primary - at the correct separation/focal distances. It is normal to do it this way by any optician - which of your two opticians will take the blame if the SA is not correct!?
25" is a high-end piece of no doubt expensive optics and that figure on the primary needs to be integrated into an overall corrected system.

Mark

Hey Mark,
You are correct about the normal way to make a Cass. One optician carefully matches the primary and secondary. Why? Because the tolerances are so tight.

I purposely designed this scope for ease of manufacture. It is one of the prime features of the wsDK. I've run tolerancing studies to convince myself that it is ok. I did have to convince Mr. Royce that it was ok too.

The secondary has a design test plate fl of 91.7". For an 8" that's an f/11.46. So - how close does the focal length need to be to prevent an impact on the SA? Plus or minus 5". I have full faith that Mr. Royce can come closer than +-5 inches.

Who get the blame if the SA is unacceptable? That's an easy question to answer - ME. ;)

#59 MKV

MKV

    Vanguard

  • *****
  • Posts: 2476
  • Joined: 20 Jan 2011

Posted 01 February 2013 - 09:42 PM

So - how close does the focal length need to be to prevent an impact on the SA? Plus or minus 5". I have full faith that Mr. Royce can come closer than +-5 inches.

That's 5.5% on the radius! Seriousyl, that's not a tight tolerance at all. When you make compound optics, tolerances are often a fraction of a percent. I don't think you have to worry... :)

Mladen

#60 siriusandthepup

siriusandthepup

    Mariner 2

  • *****
  • Posts: 212
  • Joined: 14 Feb 2006
  • Loc: Central Texas, USA

Posted 01 February 2013 - 10:28 PM

Mladen,

In the interest of full disclosure, the tolerances are actually more generous than I have stated. I was just being conservative with the +-5 inches.

Here is the tolerance from my spreadsheets. Note that the primary, which is figured at 70.4%, is expected to be of good overall system figure at plus or minus 1% tolerance for this application.

Note the change in required primary correction over the secondary range. From 81 to 100 fl on the test plate will yield an acceptable match with the primary's existing correction. :grin:

Attached Files



#61 siriusandthepup

siriusandthepup

    Mariner 2

  • *****
  • Posts: 212
  • Joined: 14 Feb 2006
  • Loc: Central Texas, USA

Posted 01 February 2013 - 10:44 PM

I guess the moral of this story is to do your primary first on a wsDK and measure its correction carefully. Then you have tons of latitude with making your secondary.

#62 MKV

MKV

    Vanguard

  • *****
  • Posts: 2476
  • Joined: 20 Jan 2011

Posted 02 February 2013 - 09:26 AM

In the interest of full disclosure, the tolerances are actually more generous than I have stated. I was just being conservative with the +-5 inches.

Ok, thanks for clarifying this. But since we're on the design again, I looked at your original chart and noticed that you list your primary as having a positive curvature. Could you clarify that?'

Using standard optical annotation convention, both mirrors will have the curvatures of the same sign.

Mladen

#63 MKV

MKV

    Vanguard

  • *****
  • Posts: 2476
  • Joined: 20 Jan 2011

Posted 02 February 2013 - 10:03 AM

I guess the moral of this story is to do your primary first on a wsDK and measure its correction carefully. Then you have tons of latitude with making your secondary.

Ed the way it's done is:

1. make your primary as close as possible to the required figure
2. make your secondary spherical
3. assemble the telescope
4. test it at the focus (preferably by autocllimation rather than stars)
5. touch up the secondary until you have a perfect null

Hint 1: make your primary very, very accurately to the desired radius of curvature. Readjust your design parameters to match the final radius of curvature, then -- if you can -- perform a knife-edge test at the conjugate foci until you get a clean null. At that point you have the exact figure on your mirror. Otherwise get a good Foucault tester and try to get as close to the figure as possible. BW, conjugate foci testing requires lots of windless space! It's a nice theoretical method for small mirrors but not for anything over 10 inches in diameter. A 12-inch f/4 may require as much as 30 feet!

Hint 2: proceed to make the secondary to match the new adjusted figures based on the primary. Polish it and figure it spherical using a test plate. At this point the two optics should be close, which is why you still need to "tweak" the primary until you have a perfect null at the focus of the complete scope.

Mladen

#64 siriusandthepup

siriusandthepup

    Mariner 2

  • *****
  • Posts: 212
  • Joined: 14 Feb 2006
  • Loc: Central Texas, USA

Posted 02 February 2013 - 11:04 AM

That's my lack of optical design experience showing. I could have easily flopped the sign. In my way of thinking (wrong BTW) a concave mirror (primary) has a positive sign and a concave mirror (secondary) a negative sign. That is reflected in my spreadsheet. I understand now that in the conventions used in the optical design software the direction of the curves is the deciding factor and since the curves of both the primary and the secondary are in the same direction left to right then the signs should be the same. Please correct my rookie mistakes and I will try to learn and readjust my thinking.

Also (again my lack of experience) the way I thought about primary mirror production is referencing a parabola at conic -1. In using the Foucalt test I was thinking that it was useful to generate target numbers reference the parabola full correction numbers. Hence 70.4% of the full parabola reference numbers for the DK ellipse correction targets.

I am listening to your comments with the intent to learn the proper way to do optical design. At least that way I can learn and be on the same page with the rest of the optical design community.

Thanks for all your help!

#65 Mike I. Jones

Mike I. Jones

    Gemini

  • *****
  • Posts: 3206
  • Joined: 02 Jul 2006
  • Loc: Fort Worth TX

Posted 02 February 2013 - 11:07 AM

MAJOR correction to replied post, lest mirrors be made incorrectly!

Directly per Malacara, "Optical Shop Testing", Foucault/caustic test figuring strength is the negative of the conic constant!

Figuring strength is NOT eccentricity!

If the DK formulas calculate a primary eccentricity e=0.83904708, the conic constant is -e^2 = -0.704. Tested at COC with Foucault or caustic methods, the mirror figuring strength is 0.704, or 70.4% the correction an equivalent paraboloid would require.

The mirror correction is NOT 83.9%!!

All, please be careful when posting formulations here, and double-check with references and for accuracies and typos. A mistaken formula can mislead someone into spending a lot of time figuring a mirror incorrectly, only to find it out after expensive coatings have been applied, and during final assembly and initial star testing.

Thanks,
Mike

#66 siriusandthepup

siriusandthepup

    Mariner 2

  • *****
  • Posts: 212
  • Joined: 14 Feb 2006
  • Loc: Central Texas, USA

Posted 02 February 2013 - 11:33 AM

My primary is completed and coated at 70.4% of a parabola.

I sure hope that is right.

#67 MKV

MKV

    Vanguard

  • *****
  • Posts: 2476
  • Joined: 20 Jan 2011

Posted 02 February 2013 - 12:34 PM

Ok, thanks Mike. Apologies for mistaking the 70.4% meant to mean deviation from the sphere (which is what eccentricity, e, is). Just another reason to stick to conventional terminology i.e. conic constants, instead of %.

Mladen

#68 MKV

MKV

    Vanguard

  • *****
  • Posts: 2476
  • Joined: 20 Jan 2011

Posted 02 February 2013 - 12:49 PM

Sorry I misunderstood your terms, Ed. Your mirror has a conic constant = -0.704 and that is correct, spot on. Apparently this is what some call "correction". I just never thought of the conic constant as a simple percent of a parabola (because it's not).

Mladen

#69 siriusandthepup

siriusandthepup

    Mariner 2

  • *****
  • Posts: 212
  • Joined: 14 Feb 2006
  • Loc: Central Texas, USA

Posted 02 February 2013 - 01:06 PM

Mladen,

It's just a sign of my old school upbringing. When I started this project I had no ATM friends that were into Optical Design programs. I used ATMOS demo for my initial explorations and then generated my DK design spreadsheet based on the DK equations I found in a book. Again, no training or experience on my part. I checked my spreadsheet outputs against ATMOS results and all seemed well, so I proceeded from that point. It certainly would have been nice to have the benefit of Cloudy Nights "group think" and experience back then.

I evolved this project in a virtual design knowledge and experience vacuum, hence my slow pickup on the proper terminology and conventions.

I am enjoying the interactions with folks who do know what they are doing in optical design. It's a good learning experience for me, just please be patient with my slow pickup. I am persistent and I will eventually come up to speed. I do appreciate all the criticisms, questions and suggestions from everyone. I know that they are honest attempts to help. :jump:

Thanks All!

#70 MKV

MKV

    Vanguard

  • *****
  • Posts: 2476
  • Joined: 20 Jan 2011

Posted 02 February 2013 - 01:16 PM

Ed, thanks. You are very kind. It was an honest mistake. And thanks to Mike, again, for catching it in time. I altered the relevant post so as not to confuse anyone else.

Mladen.

#71 mark1234

mark1234

    Sputnik

  • -----
  • Posts: 31
  • Joined: 15 Apr 2008

Posted 02 February 2013 - 05:53 PM

Ed, my very old, maybe 20 years old? Dall Kirkham 16cm F3.8/13.5 was done from equations in James Muirden's book from the 1970's. He pointed out that a slight oblating figuring would be needed on the spherical secondary after assembly and he was spot on. I now can ray-trace retrospectively and understand that this corrects the higher orders which the basic conic equations and a slide rule omitted. I can also confirm that the tolerances on radius and spacing are very wide, relatively speaking, for diffraction limited optics.
This small scope is still going strong on the original coatings. Tried to post the Registax png of some webcam moon images - png file too big but I think you can visualise a 6" f13.5 image at or near diffraction without chromatic abberation.

Mark

#72 siriusandthepup

siriusandthepup

    Mariner 2

  • *****
  • Posts: 212
  • Joined: 14 Feb 2006
  • Loc: Central Texas, USA

Posted 02 February 2013 - 06:53 PM

Mark,

Do you remember the name of the Muirden book? I thought I'd read them all. I would love to know more about the touch ups not covered by the basic equations. If I remember correctly I got the equations from Stephen Tonkin's book "Amateur Telescope Making".

Also, do you have a pic of your telescope? I'm sure we'd all love to see it. Thanks!

#73 JohnH

JohnH

    Vanguard

  • ****-
  • Posts: 2135
  • Joined: 04 Oct 2005
  • Loc: Squamish BC Moved!!!!!

Posted 02 February 2013 - 10:48 PM

Someone was talking regarding centering the optics.

With a DK in mind, I gave a critical examination of the secondary cell/focuser and made these conclusions.

Since there is around 1 1/2" range from the stalk the secondary rides on, that spacing is not especially critical as it would help give a good range of back focus for various things like cameras, spectrographs etc.

There is NO provision for CENTERING the secondary, only adjusting its tilt, as the secondary is a spherical surface, it has no unique center of revolution

#74 Mike I. Jones

Mike I. Jones

    Gemini

  • *****
  • Posts: 3206
  • Joined: 02 Jul 2006
  • Loc: Fort Worth TX

Posted 03 February 2013 - 01:32 AM

Thanks for correcting the error, Mladen.

However, figuring strength actually can be expressed as relative to a paraboloid in percentage of correction. The negative of the conic constant >> IS << the figuring strength. Figuring strength directly and linearly gives the scale factor to be multiplied by theoretical Foucault r^2/R or r^2/2R knife-edge positions for non-paraboloidal mirrors.

An ellipsoidal primary for a Dall-Kirkham might have an eccentricity of 0.8390471, giving a conic constant of -(e^2) = -0.704 and thus a figuring strength of +0.704. This implies that if the primary were zonally tested at center of curvature, the zonal r^2/R or r^2/2R values giving theoretical KE shifts for an equivalent paraboloid would be scaled by 0.704, or 70.4% of the zonal shifts for the paraboloid. The mirror can accurately be said to be 70.4% corrected relative to a paraboloid.

Likewise, a hyperboloidal primary for a Ritchey-Chretien Cassegrain might have a conic constant of -1.076, and thus a figuring strength of +1.076. Zonal KE shifts for the equivalent paraboloid are all multiplied by 1.076, and the mirror can be zonally tested at COC in exactly the same fashion as the equivalent paraboloid, just with slightly stronger KE shifts. This RC primary could be said to be 107.6% corrected.

A Rosin hyperboloidal astrograph primary is an extreme example of figuring strength, as Rosin primaries often have conic constants of down to -1.5 and even -2.0 or less, depending on the overall system design.

A sphere is the other extreme of figuring strength. A sphere has zero eccentricity, thus both the conic constant and figuring strength are zero as well. Zonal knife edge positions for the equivalent paraboloid would all be multiplied by zero, and this is of course exactly what we know to be the case for testing a sphere.

The same principle can be applied when making the primary mirror for, say, a Wright-Schmidt telescope, which typically has an oblate spheroid figure with positive conic constant and negative figuring strength. Equivalent paraboloid zonal settings would then be multiplied by a negative figuring strength, implying that the center of the mirror should focus long relative to the edge, which is indeed the case for an oblate spheroid.

Hope that helps,
Mike

#75 MKV

MKV

    Vanguard

  • *****
  • Posts: 2476
  • Joined: 20 Jan 2011

Posted 03 February 2013 - 09:12 AM

Dall Kirkham 16cm F3.8/13.5 was done from equations in James Muirden's book from the 1970's. He pointed out that a slight oblating figuring would be needed on the spherical secondary after assembly and he was spot on.

Mark, if you undercorrect the primary, then slight oblating of the secondary would indeed correct the figure. However, if the primary is overcorrected, just the opposite would be true, you'd be creating a slight ellipsoid on the secondary.

That may seem like an easy fix mathematically, but from the shop technique point of view it's much easier to make all the corrections on the primary and leave the spherical secondary alone.

Spherical secondary also makes DKs easier to collimate and maintain collimation, so why throw that advantage away by aspherzying it when all the correction can be on the already aspherized primary?

Mladen






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics