Jump to content


Photo

Ross (et al) Null testing parabolas

  • Please log in to reply
82 replies to this topic

#26 Ajohn

Ajohn

    Mariner 2

  • *****
  • Posts: 283
  • Joined: 03 Dec 2007

Posted 09 January 2013 - 01:41 PM

The Bath seems to have a diameter limit at shorter focal lengths. There is a brief thread on the subject here

http://tech.dir.grou...e/message/11257

In this case it's a 210mm mirror that when I measure it is more likely to be a bit over F2.5 rather than under.

Still not sure I get this fixed distance from lens to mirror. What ever that distance is the mirror when figured will only show a null when the knife edge is at a certain distance from the lens. I do see a problem though. So maybe I do get it. When glass is removed to create the conic it can be removed from several different places to create it. I assume that can have an effect on the final true focal length of the mirror. Dall seems to get round this by placing the knife edge at the mirrors centre of curvature but makes no mention of what zone should look high etc when the tester is in position.

One thought. Lenses made for red laser work may be suitable for making either test set up. Also I had wondered about stretching cling film over a lapped ring to make a pellicle for the Dall test.

John
-

#27 MKV

MKV

    Mercury-Atlas

  • *****
  • Posts: 2622
  • Joined: 20 Jan 2011

Posted 09 January 2013 - 01:49 PM

Certainly hope this is the case... the ross null tester I am about to build is using about 1/2 the lens diameter.

Precision lens? I hope so.

Mladen

#28 MKV

MKV

    Mercury-Atlas

  • *****
  • Posts: 2622
  • Joined: 20 Jan 2011

Posted 09 January 2013 - 02:10 PM

Just to circle back, I think I am getting it. Given lens X, the lens to mirror spacing defines the amount of "offsetting " spherical aberration that is introduced.... i.e. the conic that it will null at.

It seems to me the KE/source to lens distance is really irrelevant. The focus is where the focus is, you move the KE to it. That is, it behaves just like testing a sphere at its RoC. Now, theoretically if you have the lens to mirror spacing exactly right and you have a perfect null; then the KE to lens spacing should be what has been calculated.

Yes. If you get a null with KE, you better double check the lens to mirror as well as lens to light source distance. They should both match the the theoretical figures within tolerances given.

Use a higher frequency Ronchi instead of KE, as the Ronchi immediately tell you if you are under, over or jsut about spot on, and you can follow the progress as the bands become straighter. Much easier to interpret than KE shadows. Once you are close to a null then the KE can give you a lot more detail - provided the light source is small enoug.

Also, the light to lens distance is not irrelevant for interferometry, so you should still be able to check and set your light to lens distance accurately and repeatedly.

Remember, the correct lens to mirror distance must remain fixed at all times.

Mladen

#29 Brian Engel

Brian Engel

    Vostok 1

  • -----
  • Posts: 122
  • Joined: 25 Aug 2009
  • Loc: Cincinnati,Oh

Posted 09 January 2013 - 02:27 PM

Certainly hope this is the case... the ross null tester I am about to build is using about 1/2 the lens diameter.

Precision lens? I hope so.

Well, it is a 4.5 inch Jaeger lens from surplus shed.

As a side note, I find interesting is Surplus Shed sells SO many lens but they do not have to means (sphereometer) to measure the radius on them :-). So I am going to have to derive it from the index of refraction and focal length.

#30 MKV

MKV

    Mercury-Atlas

  • *****
  • Posts: 2622
  • Joined: 20 Jan 2011

Posted 09 January 2013 - 02:53 PM

The Bath seems to have a diameter limit at shorter focal lengths. There is a brief thread on the subject here

http://tech.dir.grou...e/message/11257

John, thaty individual doesn't go into any details, and nothing he reports is conclusive. he's speculating. The problem with faster focal ratios than f/3 is full illumination of the mirror.

In theory the focal length of the lens needed is f = d*R/D, where f is the lens fl, d is the laser beam diameter and R is the mirror ROC and D the mirror diameter.

For an f/2.5 you need a lens with f = 0.59 inches (~15 mm). You can find surplus lenses even of 10 mm fl but they are not ideal for lasers.

Your best bet is to cannibalize some cheaper laser pointers and extract their lenses, but even these tend to produce elliptical, rather than circular, beams thus failing to illuminate evenly in both x and y direction, resulting in uneven illumination, as shown below.

I have an 8-inch f/3 sphere and I know that the illumination is a problem with most lenses. Your best bet is to purchase laser-quality, lenses but they are not cheap. Also, don't skimp on laser diodes. The higher end ones have perfect circular beams and 3 mm beam diameter.

The quality of your Bath device will depend on quality components. You get what you pay for.

Mladen

Attached Files



#31 MKV

MKV

    Mercury-Atlas

  • *****
  • Posts: 2622
  • Joined: 20 Jan 2011

Posted 09 January 2013 - 03:04 PM

Well, it is a 4.5 inch Jaeger lens from surplus shed.

Is that a Jaegers achromat?

#32 MKV

MKV

    Mercury-Atlas

  • *****
  • Posts: 2622
  • Joined: 20 Jan 2011

Posted 09 January 2013 - 03:07 PM

Still not sure I get this fixed distance from lens to mirror

That determines which conic will be nulled.

#33 Brian Engel

Brian Engel

    Vostok 1

  • -----
  • Posts: 122
  • Joined: 25 Aug 2009
  • Loc: Cincinnati,Oh

Posted 09 January 2013 - 05:00 PM

Is that a Jaegers achromat?


Not an achromat.... it is this one actually.
http://www.surplussh...tem/l10266.html

I am thinking it is crown glass, could be BK7. Perhaps part of an achromat set????

#34 MKV

MKV

    Mercury-Atlas

  • *****
  • Posts: 2622
  • Joined: 20 Jan 2011

Posted 10 January 2013 - 12:02 AM

Not an achromat.... it is this one actually.
http://www.surplussh...tem/l10266.html. I am thinking it is crown glass, could be BK7. Perhaps part of an achromat set????

Brian, the lens is BK7. When you plug in BK7 glass you get a fl = 274.999 mm. The radius of curvature of the lens is 142.12 mm.

Surplus Shed lists the lens as L10266, which means it's not a precision lens (otherwise it would be PL10266).

Besides the fact that the lens is not a precision lens, the Ross null setup for your 22 inch f/4 mirror with this lens gives distance tolerances of only 0.001 inches (or 0.025 mm, or 25 microns!) for a 1/8 wave maximum error. That's a heck of a positioning requirement.

Mladen

Attached Files



#35 Ajohn

Ajohn

    Mariner 2

  • *****
  • Posts: 283
  • Joined: 03 Dec 2007

Posted 10 January 2013 - 07:42 AM

Hi Maiden.
I did ask around this sort of thing on the yahoo group a while ago now. :question: The mirror is a sort of on going problem. So far I have collected an edmunds laser diode and that's about it. My impression was that I would need a 10mm or maybe even smaller diameter biconvex with a focal length of 10mm plus a rather small cubic beam splitter. Everything needs to be small to keep the beam path separation small.

I suspect the oval illumination is probably down to the shape of the source which might be why they seem to specify shorter focal lengths lenses than should ideally be needed. It's interesting to note they even mention LED's now. It was suggested that the source size of those was too big when I asked.

It's interesting to hear from some one who uses them off group. I asked too many questions and eventually got the answers are in the posts on the group response which of course in my terms aren't. I spent a day going through them.

1st thing before I do any more on the mirror is a polishing machine. I used hard tiles for the tool and these don't give the same level of finish I could get with glass on glass. I do have the parts needed to make that now and a couple of ideas that may improve on the finish the tiles give.

John
-

#36 Brian Engel

Brian Engel

    Vostok 1

  • -----
  • Posts: 122
  • Joined: 25 Aug 2009
  • Loc: Cincinnati,Oh

Posted 10 January 2013 - 09:41 AM

Surplus Shed lists the lens as L10266, which means it's not a precision lens (otherwise it would be PL10266).

Besides the fact that the lens is not a precision lens, the Ross null setup for your 22 inch f/4 mirror with this lens gives distance tolerances of only 0.001 inches (or 0.025 mm, or 25 microns!) for a 1/8 wave maximum error. That's a heck of a positioning requirement.


I never really got much clarification from them on what is "precision" and what is not. I'm only using the center of it, so hoping it is good (eyeball Ronchi test against a star doesn't show anything horrible with it).

Hmmm... The precision from Ross XP software I get is about 1 mm....not much better :-). But it is not so bad for < 1 wave.... I am using the Ross null as more of a qualitative than quantitative analysis of the mirror. I plan to rely on the Foucault test for that (the tolerances for that are pretty frightening as well :-) ).

In order to get more relaxed lens to mirror spacing (you can see my concern now :-) ), you have to go with a longer FL lens..... @ 20" f/4, that means you have to use more of the lens and thus the accuracy of the lens across its entire surface becomes more critical.

Honestly, I totally realize that using this lens (or others of its ilk) for a Ross test on my mirror is really "ambitious". I really can't plop down $500 for a certified one, so it is kind of a tinkering around thing. The Foucault test (which I have done many, many tests over the years on fast mirrors) is going to have to be my qualitative test.

Of course, if anyone has a better suggestion for a reasonably priced (<$100) lens, I would be glad to hear it!

I'll probably get the final,final test on a interferometer. I was hoping to hit Ed Jones up for that but I think he sold it (just kidding)....


I have probably 5 more hours of polishing to do, but so far my mirror is a good sphere (small hill in the middle) with no stig.....

#37 MKV

MKV

    Mercury-Atlas

  • *****
  • Posts: 2622
  • Joined: 20 Jan 2011

Posted 10 January 2013 - 09:53 AM

Hello John! First, that would be MLADEN, not MAIDEN.

The Bath beam separation is important in faster mirrors because of relatively larger astigmatism caused by the separation. This astigmatism can be mathematically removed in fringe analysis, so it's not absolutely necessary to have a very small beam-to-beam separation to account for astigmatism.

The amount of astigmatism expected for a given beam separation (d) is given in terms of OPD waves as OPD = D^2*d^2/(16*w*R^3), where D = the diameter of the mirror, R = the radius of the mirrors, d = beam-to-beam separation, and w = wavelength used.

For example, for a given w = 0.00055 mm, and d = 10 mm, a 200 mm f/4 will produce astigmatism equivalent to 1/9 or 0.11 waves of optical path difference. By comaprison, a 200 mm f/8 mirror will have only 1/72 or 0.014 waves of OPD due to artificially induced beam-tobeam astigmatism.

You do need a biconvex lens of relatively short focal length. You can get them by cannibalizing inexpensive laser pointers. At least you'll know they are laser-quality, unlike some of the lenses you can buy second hand.

You can use super bright LEDs in conjunction with a narrow-band filter, a collimating lens and diverging lens. I don't see any reason to use LEDs in the Bath.

Regards,
Mladen

#38 MKV

MKV

    Mercury-Atlas

  • *****
  • Posts: 2622
  • Joined: 20 Jan 2011

Posted 10 January 2013 - 10:19 AM

Brian "precision" refers to tolerances. For example, precision lenses guarantee that your lens focal length will not vary by more than 1 to 2%, your central thickness to 0.2 mm, etc. Non-precision lenses have much more relaxed tolerance range.

As for the Ross precision, I entered your 22 inch f/4 (in metric units) and your Jaegers lens as your Ross null lens, and the distance precision for 1/8 wave tolerance is 0.025 mm (0.001").

The Ross null using Ronchi or KE is always qualitative. If you use it in conjunction with an interferometer you'll get a qualitative result. The Foucault is unreliable when it comes to f/4! It's also undersampled.

Your lens as it is uses over 55% of the CA. You have to find a compromise between a "doable" distance tolerance (i.e. several mm) and the working lens diameter (a function of the focal length). In your case, if you went from 275 mm to 300 mm fl. your tolerance would change from 0.0025 mm (0.001 inch) to more than 3 mm (plus or minus 1/8 inch).

In your case it's the f/4 that's at the heart of your problem. That's why even some top mirror makers will not produce f/4 but will make mirrors no faster than f/4.5, even though they have test equipment.

You can look for precision surplus lenses on Surplus Shed of sufficient size and focal length. That will cost you a fraction of the $100 limit you have. Likewise, making a working (not necessarily top rated!) Bath interferometer will allow you quantitative analysis of your optics easily under $100.

Mladen

Attached Files



#39 Brian Engel

Brian Engel

    Vostok 1

  • -----
  • Posts: 122
  • Joined: 25 Aug 2009
  • Loc: Cincinnati,Oh

Posted 10 January 2013 - 11:39 AM

22 inch f/4

It's a 20 inch f/4 (actually ~4.3 at the moment)

#40 Brian Engel

Brian Engel

    Vostok 1

  • -----
  • Posts: 122
  • Joined: 25 Aug 2009
  • Loc: Cincinnati,Oh

Posted 10 January 2013 - 11:50 AM

In your case, if you went from 275 mm to 300 mm fl. your tolerance would change from 0.0025 mm (0.001 inch) to more than 3 mm (plus or minus 1/8 inch).

I believe I do have a PL lens from surplus shed that is 85mm X 300mm FL. I went with the 117mm lens as, if I remember correctly, most of the 85mm lens would be used... and I am not sure I trust it enough. Maybe I will take another look it.

I have looked at their current inventory and they don't have what I really want... > 80mm with a FL between 300mm - 500mm.

#41 MKV

MKV

    Mercury-Atlas

  • *****
  • Posts: 2622
  • Joined: 20 Jan 2011

Posted 10 January 2013 - 12:03 PM

Okay, sorry about that. Your margin of error is 0.356 mm (0.014"), still a heck of tight requirement. You'd have to devise a radius bar with micrometric extensions at the end for that kind of accuracy.

Mladen

#42 MKV

MKV

    Mercury-Atlas

  • *****
  • Posts: 2622
  • Joined: 20 Jan 2011

Posted 10 January 2013 - 12:06 PM

Brian, with precision lenses you can use more of the lens precisely because it is a better corrected, more accurate surface. With ordinary lenses you want to keep the working diameter as small as possible.

Mladen

#43 Mike Lockwood

Mike Lockwood

    Vendor, Lockwood Custom Optics

  • *****
  • Vendors
  • Posts: 499
  • Joined: 01 Oct 2007
  • Loc: Usually in my optical shop

Posted 11 January 2013 - 12:31 AM

In your case it's the f/4 that's at the heart of your problem. That's why even some top mirror makers will not produce f/4 but will make mirrors no faster than f/4.5, even though they have test equipment.

f/4 is no problem.

Brian, if your mirror is fairly smooth and a good figure of revolution, then Foucault testing alone can be used to make a superb mirror.

If you have the telescope built, you have a decent figure of revolution test. A Foucault test will show the smoothness of the surface.

The only lens I would recommend buying is a small achromat to make a low-power telescope for your Foucault tester. That makes testing much more enjoyable.

#44 MKV

MKV

    Mercury-Atlas

  • *****
  • Posts: 2622
  • Joined: 20 Jan 2011

Posted 11 January 2013 - 12:31 PM

f/4 is no problem...

Well, coming from someone who professionally specializes in under-f/4 mirrors of up to 60-inch diameters (and won't even touch anything smaller than 14-inch disks), I'd say that for your experience and tooling it's no problem for sure! :)

For most "mortal" ATMs, an f/4 paraboloid is, if not a problem, then a lot more work than anything even a wee bit slower. In fact, even some whose mirrors are their bread and butter, like Zambuto,will charge up to 20% more for an f/4 vs. f/4.5, simply because of the extra work required!

Anyway, the beauty of null tests is that you measure once and not dozens of times. There are no shadows to chase, and interpret, or estimate. The Foucault does give a nice picture of the surface quality and a KE null shows a great deal of detail. And, a Ross null can be used in conjunction with an interferometer for quantitative results.

On the subject of the Foucault, I am curious, since you deal with much larger optics than commonly encountered among amateurs, why do you think the professional community abandoned the Foucault way back in the 30's for their large observatory mirrors and bothered to devise alternative testing methods?

Regards,
Mladen

#45 Brian Engel

Brian Engel

    Vostok 1

  • -----
  • Posts: 122
  • Joined: 25 Aug 2009
  • Loc: Cincinnati,Oh

Posted 11 January 2013 - 02:55 PM

No doubt a cheap, simple, easy to use null test for parabolids is the holy grail :-).

I would completely trust Mike ability/skill to Foucault test fast mirrors.... comes with extensive experience. The thing is, I doubt *mine* :-). Can I tell the difference in a zone null with the steep slopes of a f/4, within 1-2 thousandths?.... not so sure... :-).

I like the Ross null as a gut check and as a qualitative way to put the initial 5 or so waves(!) of correction in. In conjunction of course with the Foucault.

I think the Ross test is a fantastic tester for more "reasonable" ATM mirrors (e.g. 14 f/5 and such).

#46 Ajohn

Ajohn

    Mariner 2

  • *****
  • Posts: 283
  • Joined: 03 Dec 2007

Posted 11 January 2013 - 05:26 PM

I made a post earlier about a another null test that is probably more appropriate for a largish F4 mirror if real accuracy is needed. Seems to have got lost so briefly.

Ofner null, can be read about in Reflecting Telescope Optics II, manufacture, testing ......... on google books. It hardly more complicate than the Ross and just needs another lens but does account for higher order aberrations so is suitable for faster mirrors.

John
-

#47 MKV

MKV

    Mercury-Atlas

  • *****
  • Posts: 2622
  • Joined: 20 Jan 2011

Posted 11 January 2013 - 07:51 PM

No doubt a cheap, simple, easy to use null test for parabolids is the holy grail :-).

I would completely trust Mike ability/skill to Foucault test fast mirrors.... comes with extensive experience. The thing is, I doubt *mine* :-). Can I tell the difference in a zone null with the steep slopes of a f/4, within 1-2 thousandths?.... not so sure... :-).

I like the Ross null as a gut check and as a qualitative way to put the initial 5 or so waves(!) of correction in. In conjunction of course with the Foucault.

I think the Ross test is a fantastic tester for more "reasonable" ATM mirrors (e.g. 14 f/5 and such).

First, Mike's reputation speaks volumes of his skills. There was never any doubt that whatever tests he uses he puts them to a good use. (heck, even my "tag line" is his - because it true!)

But for an inexperienced person (relatively speaking), i.e. someone who's more likely to produce a one-time 20-inch mirror, the Ross null is, as you cheap and best suited to tackle the delicate surface tolerances of an f/4 configuration.

The only requirement is to set the lens to mirro distance very, very accurately, and keep it that way. One measurement, versus hundreds for a Foucault test.

Ross null using a KE shows all the intricate minute detail normally seen when a good spehre is observed - as "flat", evenly illuminated disk with barely perceptible surface features of very low profile.

Straight Ronchi bands can also bi qualitatively analyzed using ordinary graphic tools - such as ruler or Photoshop. If the band remains straight across aperture in a null test chances are the mirror at least meets the Rayleigh's limit, and for a 20-inch that's darn good, given that atmospheric scintillation precludes seeing sharp diffraction images most of the time at that aperture, and that cooling issues are also very long unless you can produce an astigmatism-free wafer-thin large mirror, as Mike Lockwood does.

Autocollimation is out of the question for most amateurs at that size. I am not sure if Mike has a 22-inch flat but can make one. I can't. So, no matter how you turn it around, the Ross null test, considering the simplicity cost and no need for additional software or technology, is a blessing.

Mladen

#48 ccaissie

ccaissie

    Viking 1

  • -----
  • Posts: 634
  • Joined: 13 Sep 2010
  • Loc: Whitefield, Maine

Posted 11 January 2013 - 07:51 PM

Ofner null, can be read about in Reflecting Telescope Optics II, manufacture, testing ......... on google books. It hardly more complicate than the Ross and just needs another lens but does account for higher order aberrations so is suitable for faster mirrors.

You bet. And it can be made all reflecting!
With a residual OPD of .009 waves, and that's for a goofy 1 meter RC primary, who wouldn't build one of these puppies?

#49 ccaissie

ccaissie

    Viking 1

  • -----
  • Posts: 634
  • Joined: 13 Sep 2010
  • Loc: Whitefield, Maine

Posted 11 January 2013 - 08:39 PM

Now that the construction and other issues of the Ross are handled, one of my favorite bench tests to analyze a null, is the bench star test, and that should be useful in the Ross. Anyone tried it? With a monochromatic pinpoint source, the nulled image can be observed with an ocular and analyzed as we would a star at infinity. Ref: Suiter.

this works good on spheres I've produced, and once the zones get to be hard to see, the star test faithfully shows any slight spherical abberration. In a shop, could be good to 1/60 wave. (Welford) But it is quite helpful in locating zones also, as the zones reliably show in the diffraction disk.

The point source I've most often used is a lensless laser pointer diode, which with lowered voltage does not "lase". I've also used a laser pointer aimed at a shined ball bearing next to the eyepiece to create a pointsource "glint" (Texereau). This latter source worked for me on a short focus concave, the light cone was ample for an autocollimated f/3, and yielded much info in the bench star test.

#50 Brian Engel

Brian Engel

    Vostok 1

  • -----
  • Posts: 122
  • Joined: 25 Aug 2009
  • Loc: Cincinnati,Oh

Posted 11 January 2013 - 10:42 PM

The only requirement is to set the lens to mirror distance very, very accurately, and keep it that way. One measurement, versus hundreds for a Foucault test.

That is the "rub". But this too is problematic for ATM's that want to make large fast mirrors. As Ed has said over and over, getting a lens of sufficient quality is usually the biggest problem. You have to pay big bucks to get one :-). Otherwise you are left to find compromises and end up having to measure the mirrors to lens spacing within a few millimeters.

Given your average tape measure is lucky to have a tolerance of < 2mm, this presents a challenge. Which is what kicked off this thread :-). I was hoping to find a way "around" it by fixing the KE to lens distance and then move the whole stage back and forth to get the best null I can as I work the mirror (just like if it is was a sphere). Figuring the with a given lens and with the KE lens distance fixed at its calculated spacing for a parabola, the only conic it would null at is -1.... In my mind, it was kind of like "moving" the parabola through the sphere kinda like in figureXP.

It appears I misunderstood that and it will not work. So I am back to 1)find a more forgiving, large lens or 2) measure the distance with 1mm or so accuracy over that long distance.

I love the notion of a null test as you can look at it and instantly have a very good idea of what needs to be done to the mirror's surface - without tedious measurements and a "calculated" profile in figureXP (et al).

Guess I will just have to fool around with it and see how it works out :-).....






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics