39 replies to this topic

[junkbum35]

The Hopewell Observatory (in Va) has a classical 12" Cassegrain optical tube and optics that were manufactured about 50 years ago and were originally mounted on an Ealing mount, for the University of Maryland, but they discarded it during the 1970s or 1980s and it eventually made its way to us (long before my time with the observatory).

 

The optics were seen as rather disappointing, and a cardboard mask was made at one point, to reduce the optics to about a 10" diameter, but it apparently didn't help much. The OTA was replaced with an orange-tube Celestron 14" SCT which works reasonably well. 

 

Recently, I was asked to check out the optics on this original classical Cass, which is supposed to have a parabolic primary and a hyperbolic secondary. I did Ronchi testing, Couder-Foucault zonal testing, and double-pass autocollimation testing, and I found that the primary is way over-corrected, veering into hyperbolic territory. In fact, Figure XP claims that the conic section of best fit has a Schwartzschild constant of about -1.1, but if it is supposed to be parabolic, then it has a wavefront error of about 5/9 lambda.

 

I have not yet tested the secondary or been successful at running a test of the whole telescope with an artificial star. For the indoor star test, it appears that it only comes to a focus maybe a meter or two behind the primary!

 

It had a very interesting DC motor - cum - potentiometer assembly to help in moving the secondary mirror in and out, for focusing.

 

My question is this: Has anybody else done any optical testing on these classical Cassegrains by Ealing et al?

 

Guy Brandenburg

Washington, DC

 

[Chuck Hards]

Hi Guy,

 

Our club has an Ealing 16" Cassegrain that I restored some years ago, and was originally installed on the Physics Building at the University of Utah for decades.  It tests very well, has the original Ealing optics set.

At one time, a student at the University had attached a motor for moving the secondary, but that was removed and the telescope has been restored to very close to it's original condition, again focusing with a conventional focuser behind the primary.

 

One of the more popular telescopes at the club observatory.

[DAVIDG]

 Hi Guy,

    I have refigured a 24"  Tinsley Cass and we meet a number of times at the Delmarva mirror making class. With the primary being over corrected it sounds like you have a Ritchey–Chrétien vs a classic cass. It also sounds like the secondary is nowhere near the correct position. If you can measure the radii of the primary and secondary I'll be glad to key into OLSO so we can have look at the design and see what is going on. 

  If it was me  and I needed to refigure the system and it is RC, I would use the Ross Null to figure the primary to the  needed hyberbola, then use a Hindle sphere to figure the hyberbolic secondary, then tweak the correction via double pass autocollimation. 

  In my experience I have seen very few commercial cassigrains of any the typeical flavors that have been figured correctly. Most optical companies that made them  did not seem to have the means to test and figure the secondary correctly and there seems to be a misunderstand of how to calculate the knife edge values for any conic besides a parabola so the primary are not figured correctly. 

 

                - Dave 

[DAVIDG]

 I took a guess at the design based on the fact that the other person that has a  similar scope said it was f/13 and I'll assume that it was meant to be an RC design. I set the back focal length to 16" behind the primary. So  I get the 12" f/4 primary with conic of -1.0901 which is very close to what you measured, the radius on the secondary is  44.189"  with the conic of -4.503, and the  separation of 32.8".  Here is the design and spot diagram for 1/2 degree view of view. The black circle is the size of the  Airy disk.  

   With the amplification of the  secondary being 3.2X it doesn't  take much of  an error in it's position to move the focal plane way back and also make the  correction terrible. So it just might be a mechanical error in the placement of the optics that is the problem. 

 

                        - Dave 

 

 

    

Edited by DAVIDG, 07 September 2017 - 09:19 AM.

[apfever]

Hi Dave,

 

A lot here caught my eye. One thing to start was the nature of the string, scope type, size, then Tinsley, then....lol.

 so I have a Tinsley 12" clear (slight plus blank size) F16 classic Cass., F4 primary, 4X Hyper secondary, typical Tinsley, massive curved secondary cast unit holder..thing...

 

Apparently the spacing from primary to secondary affects the image enough to where that seems to be the goal, then leave focusing to the back.  It's the longest focal length scope I have, even beats quirky big dobs.

 

I've never had a mount for it.

[apfever]

ohh, to the op

 

Guy,

 

Everybody likes to start with radius of curvature, this is the spherical thing, or how big is the ball that fits the curve of your mirror, and they are always big.  I check this all the time, quick and easy. It seems all these other wild things from paranoidbolic to hyperactbolic gets put in from the sphere.  If you want check it out pm me.  If you are already into dissection of the point spread function then never mind.  That's cool too.  One of the best readings on that is Vlad's work on optics, primarily astonomy related, but I never can remember that site.  It helps if you like math.

[Ben H]

 One of the best readings on that is Vlad's work on optics, primarily astonomy related, but I never can remember that site.  It helps if you like math.

http://www.telescope-optics.net/ ? 

[DAVIDG]

The OP, Guy Brandenburg has been running a mirror making class for years down in DC.  He is very knowledgeable about optics, testing and theory.  http://home.earthlin..._Home_Page.html

 

              - Dave 

[junkbum35]

Thanks for all the suggestions.

 

The idea that this was in fact intended to be a Ritchey-Chretien system with both mirrors hyperbolic did occur to us. If it is, I'm not looking forward to testing a convex hyperboloid, but we do have a spherical mirror that we can perforate in order to do the Hindle test, which I've never done before, and looks like it needs to be aligned with extreme accuracy for success. 

 

Even setting up an indoor star test is more than a little bit of a PITA because we have to haul two very heavy tables into a large playroom at a city community center, set up the various flat mirrors to bounce an artificial-star reflection across the room numerous times, then align the telescope, try to bring it to focus .... and about then, the staff chase us out and we have to put it all away because they are closing the center for the night, and no, we can't leave the setup alone.
 
BTW, the focal length of the mirror is 47.5 inches, ROC 95 inches, thus the primary is a tad under f/4.

 

In case you are interested, the zonal measurements are as follows (digested & nulled and averaged from six sets of readings) with a fixed light source, all measurements in inches. Yes, the first zone was quite narrow (long story)...

 

Zone     Effective Radius   Ideal Reading  Actual reading.

1                 2.340                   0.000              0.000

2                 3.235                   0.052              0.060

3                 4.215                   0.129              0.151

4                 5.005                   0.206              0.236

5                 5.685                   0.282              0.304

 

I am not succeeding in posting any images. Not sure why.

 

Guy

[junkbum35]

Here are some images on my blog:

https://guysmathastr...gran-telescope/

 

Guy

[junkbum35]

Oh, yeah, I taught secondary mathematics for about 30 years, and it is quite a pleasure for me to actually put all that stuff about conic sections to actual, practical use!

 

Guy

[Ben H]

Hi Guy,

The DC motorized secondary seems to point all the more at this being an RC. I'm sure the motor was added because of the critical spacing due to the nature of the design. 

Do you really need the Hindle test to verify the secondary is hyperbolic? 

[junkbum35]

I'm not sure of much of anything, and Ben, you may be right about this being an RC.. But, supposedly the optics suck, so something needs to be checked out, and in general, testing convex surfaces is quite diffricult. Without a matching surface with which to do interference test with Newton's rings, the Hindle seems required, no?

 

Guy

[DAVIDG]

 Guy,

    You said you tested the primary via double pass so if you have aflat that is large enough for that , then I would set the telescope up, pointed at the flat and test the complete system. Since the back focal length is way behind the primary, that tells me the  secondary is much to close. So in double pass you would start moving the secondary away from the primary until you find the best correction. A simple Ronchi screen mounted in film canister with a LED attached and placed  in the  focuser  is all you need. ( see the this thread for the design https://www.cloudyni...m-dpac/page-6) 

 If you can't find a position were the Ronchi lines are dead straight then the  system will need to be refigured. If the primary is optically smooth you can figure the secondary to null the system. So you place the secondary at the position were the system  showing the least amount of spherical aberration in double pass and then figure the secondary to null the system.

  If you don't have  a glass flat that is large enough you  can use a pan of oil since the surface of a  liquid is flat to something like 1/300 of wave.

 

                                 - Dave  

Edited by DAVIDG, 08 September 2017 - 02:15 PM.

[junkbum35]

Sounds very doable, Dave. Got the flat, which is lucky, because I've had an awfully hard time getting the surface of a bucket of water to stay vertical.unless it was frozen.     Or else I'd have to test while standing on a ladder, which is not much fun  either.

[DAVIDG]

 Guy,

   Glad you have  a glass flat so it shouldn't be to difficult to setup and test.  My guess at the design puts the secondary around 32" from the primary. So that is close to were it most likely should be for  a reasonable back focal length. If it is a couple of inches closer then that I can understand why the image is so bad and so far behind the primary. 

   I could understand if  someone in the past may have wanted to use some detector on the back of the scope and needed  more back focal length.  So the position of the secondary and the focusing system was moved without the understand of how it would effect the image quality.

 

                     - Dave 

[photoracer18]

I have a Cave 12.5" Classical Cass/Newt convertible from about 1962 that had the primary replaced with a mirror made by Dick Brandt while he was in charge of the optical shop at Optical Craftsman sometime in the late 60's. I am near the beginning of restoration so the mirror is currently out of the tube. FL is 60.15" according to Dick's mark on the back (F4.8).The Classical Cass secondary has some mirror issues including spots in the aluminized coatings. Might be a reason the scope was left as a Newtonian for something like 30 years before it was rescued. I need to get it tested to see if the Cass secondary is recoverable or not and whether I should mount it as a Newtonian also. I moved from Maryland a few years ago. I was an NCA member back in the early to mid-1970's.

Edited by photoracer18, 08 September 2017 - 04:27 PM.

[junkbum35]

I finally made a testing rig for the cass as Dave Groski suggested, with our flat in front of the telescope and a ronchi grating with LED at the exit. It turned out that the focal point could be made as close to the primary as we wanted, by moving the secondary closer to, or farther away from, thje primary.

 

MAN! Those are about the worst ronchi images I've ever seen! The astigmatism and complete lack of symmetry were astounding! We tried playing with the alignment of the secondary to see if it made any difference, and moved the focal point in and out, but all that did was just create somewhat different, but still horrible, images.

 

Unfortunately, in the time we had, I was not able to make a decent photograph of the ronchigram (too bright). But trust me -- it was horrendous.

 

I am guessing that's because the secondary was so poorly figured. The primary does not appear to have any trace of astigmatism. Next step, I guess, is testing the secondary separately. Bob Bolster has kindly lent us the primary mirror he made for (I think) a Schmidt-Cassegrain back when he was in high school. It's an f/0.9 (!!) sphere. I know a place that does really nice trepanning, etc...

 

Guy

[DAVIDG]

 Guy,

    I would save the time of testing the secondary with Hindle sphere. If you just rotate the secondary while testing it in double pass you'll know were the astigmatism is coming from. You can also the test the primary by itself in double pass. Since it is over corrected you'll see that but the Ronchi lines should be smoothly bowing. You can test for astigmatism by placing the Ronchi screen inside of focus and adjusting it so the pattern is perfectly vertical. Then move the grating to the outside of focus and see if the pattern rotates even by the smallest amount. If it does the primary has astigmatism.

   If the primary tests smooth, then I would reassemble the scope and test it via double pass. You know the secondary is bad and testing the whole scope via double pass is better then figuring the secondary with a Hindle sphere and then having to touch it up using double pass so just go with double pass from the start. 

   I would polish the secondary back to a sphere and keep testing the system until the astigmatism is gone. Of course when the secondary is spherical or close to it you  will see a large amount of under corrrection when you test the whole system. Now you can use a peddle shaped lap to put the hyperbolic figure back onto the secondary and you keep figuring the secondary until the complete system shows a clean null.

 

                  - Dave 

[junkbum35]

I already did look for astigmatism with the primary - as you described - and found none. I also did the double-pass autocollimation test on the primary and found it to be smoothly bowing outwards (a bit) in the middle when inside focus and bending smoothly inwards (a bit) towards the middle when outside focus.

 

Rotating the secondary isn't easy - in fact, even getting to the secondary at all isn't easy! It seems to me that figuring and testing the secondary all by itself with the Hindle might be easier. OTOH, if it is supposed to be a R-C, as opposed to a classical Cass, the conic constants for the primary and the secondary have to match with extraordinary precision, and I'm not sure I can do that.

 

Guy

[DAVIDG]

 Guy.

  "OTOH, if it is supposed to be a R-C, as opposed to a classical Cass, the conic constants for the primary and the secondary have to match with extraordinary precision, and I'm not sure I can do that."

  That is why you figure the secondary using double pass on the complete system. You don't have to know what the conic is. When the system nulls you have achieved the conic needed to perfectly correct the system.

  The Hindle test is great test but it is also a source of errors since it is an  independent setup and it is testing only the secondary. Double pass is testing the complete system so the errors your seeing are what the system is producing. So when the system nulls all the errors are fixed. 

   Maybe you can modify the secondary hold so the secondary can be quickly removed and also installed. Maybe a metal plate glued to the back of the secondary and magnet in the holder ? 

 

                     - Dave 

[junkbum35]

My attempt to determine the foci of the apparent hyperbolic primary on the Hopewell Ealing 12inch cassegrain.

 

My given information is that the mirror has a radius of curvature of 95 inches by my direct measurement, and its Schwarzschild constant (indicated by K) of best fit, according to FigureXP using my six sets of Couder-mask Foucault readings, is -1.112. I prefer to use the letter p, which equals K + 1. Thus, p = -0.112. I use R for the radius of curvature, and decided it should be negative, that is, off to the left (I think).

 

One can obtain the equation of any conic by using the formula

 

Y^2 – 2Rx + px^2 = 0.

 

When I plug in my values, I get

 

Y^2 + 190x -0.112x^2 = 0.

 

I then used ordinary completing-the-square techniques to find the values of a, b, and c when putting this equation in standard form, that is something like, y^2/a^2 – x^2/b^2 = 1

 

Omitting some of the steps because they are a pain to type, and rounding large values on this paper to the nearest integer (but not in my calculator), I get

 

y^2 – 0.112(x – 848)^2 = - 80540 and eventually

 

(x – 848)^2 / 848^2 – y^2 / 248^2 = 1

 

Which means that a is 848 inches, which is over 70 feet, and b is 284 inches, or almost 24 feet. Since a^2 + b^2 = c^2, then c is about 894. And the focal points are 894 inches from the center of the double-knapped hyperboloid, which is located at (848, 0), so it looks a lot like this graph which apparently I can't paste.

Which of the two naps of this conic section is the location of the actual mirror? I suppose it doesn’t make a big difference.

 

Making that assumption that means that the foci of this hyperbolic mirror are about 894 - 848 = 46 inches from the center of the primary mirror. I don’t have the exact measurement from the center of the primary to the center of the secondary, but this at least gives me a start. That measurement will need to be made very, very carefully and the location of the secondary checked in three dimensions so that the ronchi lines are as straight as possible.

[DAVIDG]

 Guy,

   You don't need to go thru all that math. All you do is set the scope up in double pass and move the secondary until you get the straightest lines. At that point hopefully you''ll be at the back focal length that you want. If not you position the secondary to give the back focal length that you want, the Ronchi band will show issues  and using double pass figure it until you achieve a null.

   This is exact same procedure one uses when you test and figure the secondary using a Hindle sphere. The Hindle sphere doesn't have to have the radius of curvature that exactly matches the focal length of primary. There is range that will work. You position the secondary to give you the back focal length you want and figure the secondary until it nulls. So your doing the same in double pass but it is a  better setup since your matching the secondary to the primary it will be used with.

 

                    - Dave 

[junkbum35]

Perhaps. Actually, doing the math was great fun, for me. Guess why: all those years of teaching students how to complete the square in algebra 1 or 2, and now I'm doing it for an extremely good reason!

 

But recall this: the primary mirror actually looked to me, and to others, to be very well-figured. However, when I tested the entire R-C system in autocollimation with that flat, the ronchi lines looked like they were drawn by spiders tripping on psychedelics - lines crossing all over the place. Literally the worst ronchi lines I have ever seen. No matter what changes we made to the primary-secondary distance, or to the tilt of the secondary, seemed to provide any improvement at all.

 

Which makes me think that making a Hindle to test just the secondary, to see if it indeed has got a completely insane figure, would give some valuable information, and either confirm or rebut my conclusion about the primary. (I don't trust my own measurements!! I've fooled myself too many times!!)

 

Guy

[Mirzam]

Hello Guy,

Have you removed the secondary from its mount?  Your description of the ronchi lines reminds me of a thin primary mirror that was attached to a wooden cell using excessive amounts of glue.  After I released the mirror from the cell the spider web of ronchi lines reverted to straight lines indicating a perfect sphere.

 

JimC