Jump to content


Photo

Ross (et al) Null testing parabolas

  • Please log in to reply
82 replies to this topic

#1 Brian Engel

Brian Engel

    Vostok 1

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

Posted 07 January 2013 - 09:03 PM

Been a long time since I've posted but now after buying a new house and having a baby I am able to get back into mirror making once again :-).

So this has me thinking of building a null tester (again) and finish up work on my 20" f/4 (well, the hard part - figuring - has yet to begin).

I've read the rather long and interesting thread on Ed Jones' modified conjugate null test and just don't think that is going to work as I don't want to build the telescope 1st. So that leads me back to a Ross null tester.

Before I do, I have one nagging thing I want to make sure I understand (well, many things really :-). I had this discussion in a thread several years ago but.....

So much concern is made for very accurately measuring the "object" to lens and lens to mirror spacing. Measuring the object to lens distance very, very accurately is a piece of cake.... as you can get within several thousandths easily. So this is of no concern....

My question is why must the lens to mirror spacing be measured at all? Other than of course having a rough idea where to place your tester?

If you build a Ross Null tester that has a "stage" where by the source to lens spacing is measured and fixed; couldn't you just move that entire stage back and forth from the mirror until you get the best possible null (the object -lens spacing would NOT change)?

e.g. I would use the nifty "Ross Null Finder" software, plug in my lens data, mirror info etc... get my object to lens spacing, set that on my movable stage (as described above) and then just use a tape measure to measure my lens to mirror distance within 1/2". Then I would look across the KE or ronchi, and then precisely move the "fixed stage" back and forth until I get the best possible null. If I don't and see a raised zone etc... I would go back rework that area, then repeat.

As I understand it (which may be wrong), given a conic constant (-1 in this case) and using the object to lens spacing calculated by the software; there is only ONE conic I will null at and I can only get there when I have a parabola at the distance the software calculates.

To look at this the opposite way. For my understanding to be wrong, given that I set the object to lens distance the program tells me for a -1 conic, I can possibly achieve "false" nulls at other conics (hyperbolas etc...) by moving the "fixed" object/lens stage back and forth.... and thus end up with horrible under/over correction even though I see a perfect null through the Ross Null tester.


To make it more formulaic.... Object to lens distance (X) + Lens to mirror distance (Y) = -1 (Parabola).

If X is constant (as determined by the software), there is no Y value that will give me a null conic other than -1.

...and of course, I would be using my trusty Foucault tester as well and not rely solely on this.

Thoughts? Am I totally off base here?

Thanks...

#2 GlennLeDrew

GlennLeDrew

    Voyager 1

  • *****
  • Posts: 11008
  • Joined: 17 Jun 2008
  • Loc: Ottawa, Ontario, Canada

Posted 08 January 2013 - 01:25 AM

The mirror-to-null lens spacing is absolutely fixed for any given combination of mirror radius and null lens form. One cannot naively 'jiggle' things about, for then a Hubble-like incorrect figure will be imparted.

#3 MKV

MKV

    Mercury-Atlas

  • *****
  • Posts: 2623
  • Joined: 20 Jan 2011

Posted 08 January 2013 - 02:15 AM

Brian, the Ross null test doesn't work the way you describe it. For accurate distance, rather than a measuring tape or a ruler, you may consider using a radius bar, like the one pictured below. The bar has nylon pan head screws for a single point contact at each end, and its total length is easily measured with calipers to within ±0.001".

Posted Image

You will also need some sort of leveling meachanism for the sled. My sled (above) rides on another sled which has 3 leveling screws. Both sleds use t-slots for anchoring.

As for the test itself, you must set both distances (light source-to-lens and lens-to-mirror) correctly, then (1) lock the sled, and (2) unlock the light source in order to be able to pull back and forth a little (in and out of focus) in order to observe the fringes/shadows, since the calculated distances are for the exact focal position. If you have a true paraboloid, you's see a null. Otherwise you will see bowed Ronchi bands or and k-e "doughnut" shadows.

The idea that you can just slide the sled back and forth until you see a null, and then check if the lens to mirror distance is correct is not how it works.

If things worked the way you describe it, there would indeed be no reason to measure the lens-to-mirror distance. The only way you can be certain you have a true paraboloid is if both distances are correct (within the tolerance. chosen).

One more thing: I checked the spacings necessary for a 22 inch f/4 mirror, and 1/16 wave tolerance on the mirror (1/8 on the wavefront), and for that precision both of your distances must not vary by more than ±1.2 mm (0.047 inches), which is less than 1/20". You will need a rigid radius bar and a good measuring tape (preferably in mm) for the lens to mirror separation as well.

Also, since the Ross lens has trenmendous amount of chromatic ebrration, the test must be done either with lasers or by narrow band filters. Care must be taken to interpret the final data correctly in terms of the wavelength closest to the sensitivity of the eye close to 550 nm.

In the initial stages of figuring, this may not be as crucial as it is the final stage. Thus an optical null in the red (say 632.8 nm) will not be an optical null in the green (547 nm)., as can be seen in this quality Rohr example

http://rohr.aiax.de/ts-lomo03.jpg

Mladen

#4 Ajohn

Ajohn

    Mariner 2

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

Posted 08 January 2013 - 06:18 AM

I haven't made a mirror for a long time but used the Dall null test which is similar. Here the mirror rad is measured and the tester set up. It's then positioned so that the image of the pin hole is exactly on the centre of curvature of the mirror. No need for measurements just a case of accurately locating where it is and then moving the tester not the knife edge. Dall suggested a type of red filter which I couldn't find so I substituted a red led.

The colour problems are down to the wavelength the set up is designed for. I suppose it could be ray traced for blue really but red was used because dispersion of optical crown is low in that area.

I followed Texereau to the letter to start off with but found the shadows in the mask hard to read. His idea for testing a sphere for turned edge are really good. It's just a case of slit size and where the knife goes.

The Dall test had many curious artefacts in the view but eventually everything looked ok. I then went on to use a version of the Hartman test using a wire and 1in dia holes in the mask to get an accurate idea of what I had. I used a 10x loupe instead of an eyepiece. Hard to make an x-y set up at the time but with the availability of cheap Chinese lathe compound slides (top slides in the USA?) very easy to set up now.

Also tried Ronchi. Bits not available in the UK so made one by carefully winding 2 strands of fine wire and round a metal plate with a hole in it and then removing one strand. Made me wonder why people use it.

Given the quality of some of the Chines lathe parts I might use the same from Taig or Peatol in the UK. Depends on how much movement is needed.

John

#5 DAVIDG

DAVIDG

    Aurora

  • *****
  • Posts: 4915
  • Joined: 02 Dec 2004
  • Loc: Hockessin, De

Posted 08 January 2013 - 10:25 AM

We use the Ross Null as one of the main tests along with Double pass autocollimation at the Delmarva Mirror Making class which we have been doing for 12 years now. http://www.delmarvas...ive/mwhome.html
Over at least the last five years, one or more mirrors made at our class has won an award at Stellafane in the optical judging. I'm not bragging just saying that this is an independent verification of the test methods and proceduces we use and we have many years of experience in using this test to make quality optics.
You don't need to worry about the lens to object distance, what is critical is the lens to mirror distance. When this is set you do the test just like a Foucault or Ronchi test and move the knife edge or grating inside or outside of the exact focus position to examine the figure on the mirror. I recommend using a Ronchi screen with 100 to 133 LPI. Also if you use the Ross Null program it will give you the tolerance that you need to set the this distance for mirror to lens spacing.
Don't get hung up on the tolerance of setting the spacing. What your going to find is that it is MUCH more difficult to achieve a figure that shows a perfect null then being sure that the spacing is good to a few thousands of inch. A simple radius rod made from some pine works very well. You also need to know the radii on the lens and radii of the mirror to far degree of accuracy and you need to test in monochrome light of a known wavelength. Your tester needs to have the knife edge/grating coplanar with the light source.
Once you achieve a clean null, you can double check the results using other test methods and I HIGHLY recommend that you do use other methods to check the results. What your going to find is that each method has some uncertainty associated with it, but each test method should agree with the others to within the error tolerance of the test. If they don't, you need to determine why before you can having confidence in the results.
What a null test is going to show you is exactly how difficult it is going to be to polish a true 1/8 wave wavefront on a mirror of this size and F-ratio. As I said achieving anything close to a null is were the work is going to be and not having the spacing set exactly.

All the Best,
- Dave

#6 Brian Engel

Brian Engel

    Vostok 1

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

Posted 08 January 2013 - 10:26 AM

If things worked the way you describe it, there would indeed be no reason to measure the lens-to-mirror distance. The only way you can be certain you have a true paraboloid is if both distances are correct (within the tolerance. chosen).

One more thing: I checked the spacings necessary for a 22 inch f/4 mirror, and 1/16 wave tolerance on the mirror (1/8 on the wavefront), and for that precision both of your distances must not vary by more than ±1.2 mm (0.047 inches), which is less than 1/20". You will need a rigid radius bar and a good measuring tape (preferably in mm) for the lens to mirror separation as well.


If the object to lens distance remains fixed, can you get a "incorrect" null if you don't have the tester within the lens to mirror spacing the Ross Null program puts out?


Given a particular lens, and a particular spacing (as calculated by the Ross Null software), I thought there is only one condition where you will get a null.... 1) the surface is a parabola and 2) the spacing lens to mirror is exactly as calculated by the ross null software.

It is a 20" f/4 and as you have surmised, the lens to mirror spacing is quite tight....which is why I am concerned about it :-).


I guess what I am getting to is I am wondering what the consequences would be if I had my object to lens distance perfect but had my lens to mirror spacing off by, say, 10mm, then figured the mirror to a perfect null (if possible).

Also, BTW, I was planning to use a green LED for the source.

#7 Pinbout

Pinbout

    Cosmos

  • *****
  • Posts: 8293
  • Joined: 22 Feb 2010
  • Loc: You can't see me...

Posted 08 January 2013 - 10:38 AM

if you could make it in march, you should attend the Delmarva Mirror Making Seminar, hurry up and sign up.

Steve Swayze uses the rossnull to final figure the mirror. you'll see it first hand and get experience with it.

He uses red led's.

Attached Files



#8 DAVIDG

DAVIDG

    Aurora

  • *****
  • Posts: 4915
  • Joined: 02 Dec 2004
  • Loc: Hockessin, De

Posted 08 January 2013 - 10:40 AM

[quote name="Brian Engel"][quote]
I guess what I am getting to is I am wondering what the consequences would be if I had my object to lens distance perfect but had my lens to mirror spacing off by, say, 10mm, then figured the mirror to a perfect null (if possible).

Also, BTW, I was planning to use a green LED for the source. [/quote]

You don't want to set the lens to knife edge or ronchi distance. You need to be able to move the knife edge or Ronchi screen just like you do when your doing a typical Foucault test so you can see what the figure is. You set the lens to mirror distance and just do the test like your testing a spherical mirror.
If you use a green LED then your going to need to look thru a narrow band interference filter that has band pass of 10nm or less. This will define the wavelenth that your testing at and determine the spacing for that wavelength. You don't need the filter until you get close to a null, but once you get close you'll need to is sharpen up the Ronchi lines and/or improve the contrast of any faint zones you see with a knife edge.
I just use a Ronchi screen and position it so there are three lines showing and fairly quickly move it from inside to outside of focus and back again while looking for any slight bowing of the lines. When you can not detect any departure from dead straight, you have a good null.

- Dave

#9 Brian Engel

Brian Engel

    Vostok 1

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

Posted 08 January 2013 - 11:31 AM

Alright, I think I see what DAVIDG is saying....

As you described, the lens to mirror distance is fixed precisely. This has the effect of introducing aberration such that it allows a perfect parabola to focus at a single point. Then you vary the *KE (or ronchi) to lens spacing* to see the traditional shadow movements that you see with a standard foucault test. You move the knife edge back and forth through focus and make corrections as needed to the mirror surface to achive a null. So in effect, you really don't care about the KE to lens spacing. The "corrected" focus is where it is and you must move the KE to it.

What I am wondering is, why can't we fix the KE to lens distance and vary the lens to mirror spacing (by moving this stage back and forth)? Won't this have the same effect as above?

#10 MKV

MKV

    Mercury-Atlas

  • *****
  • Posts: 2623
  • Joined: 20 Jan 2011

Posted 08 January 2013 - 12:52 PM

Brian, the lens-to mirror distance is not critical for the Ronchi or Foucualt method. It is critical for interferometry because you you want to be exactly in the mid-focus position (i.e. between the marginal and paraxial focal points) in order tom get useful fringes.

The sled is useful for keeping everything lined up when you move the light source. Move the sled to the correct lens-to-mirror distance, then lock it in place. Then unlock the light source carriage and move where needed.

For lens-to-mirror distance use stiff radius bars with adjustable nylon tipped ends (a 1/4-20 screw is good enough). Tweak the ends to the exact distance needed and support the bar to avoid any sagging. Measure the length of the radius bar with a tape measure to the nearest mm.

Using autocllimation to check your Ross results is a good approach, but limited. At 22 inches, your mirror has pretty much exhausted that option unless you have access to a 24-inch flat. Foucault test is highly unreliable at f/4 and the Ronchi is not much more than "eyeing" the parabola - useless as far as determining the exact figure.

Your only other option is an interferometer, but again unless you have one that has a null lens built into it, such as the Zygo, you will have to depend on the Ross lens as your nulling lens, and therefore the correct lens-to-mirror distance, again, is the key factor. So, tailor your Ross test setup to meet the lens-to-mirror distance and the object-to-lens distance requirements to the "t".

Mladen

#11 MKV

MKV

    Mercury-Atlas

  • *****
  • Posts: 2623
  • Joined: 20 Jan 2011

Posted 08 January 2013 - 01:15 PM

What I am wondering is, why can't we fix the KE to lens distance and vary the lens to mirror spacing (by moving this stage back and forth)? Won't this have the same effect as above?

You should be able to, provided you use an eyepiece which is focused on the plane of the light source, set at the correct source-to-lens distance. Then, by sliding the sled until the image of the source is in sharp focus, the lens-to mirror distance should correspond to the theoretical value within the margins of error allowed.

Mladen

Attached Files



#12 Ajohn

Ajohn

    Mariner 2

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

Posted 08 January 2013 - 01:55 PM

If a knife edge can produce a null with the ross test it must be possible to set a lens to knife edge distance and then move the lot as a unit.

The yahoo atm_free group is heavily into bath interferometers. They are common path and do not need expensive parts and the analysis software is free. These use a none polarising cubic beam splitter, a small mirror and a lens plus a red laser diode run below the lasing point to obtain a small source.

Making an xy stage may be expensive or a pain so I have put a microscope xy stage to one side.

This is the groups wiki

http://starryridge.c..._Interferometer

They are a friendly lot on the yahoo group but following too many questions they will say best to build one and try it. There are plenty of reports by people who have used them on the group and for what you are doing it's likely to be a sensible way to go. Me too if I finish an F2.5 mirror.

John
-

#13 DAVIDG

DAVIDG

    Aurora

  • *****
  • Posts: 4915
  • Joined: 02 Dec 2004
  • Loc: Hockessin, De

Posted 08 January 2013 - 02:06 PM

Think of how you typically do a Foucault or Ronchi test by moving the knife edge/grating to different positions in the light path to examine different zones on the surface of the mirror. If the lens was moving along with the knife edge/grating, it would be varying the amount of spherical aberration and you again could find a position were it nulls but the conic is not what you want. So you couldn't examine a zone by looking a little inside or outside of focus. You really limit the usefulness of the test if you fixed the lens to knife edge distance.
This test is not difficult to do. One huge advantage is that you see how the whole mirror is departing from the parabola you want, just like one sees how a mirror is departing from a sphere. So a hole is a hole and a hill is hill and turned edge is turned edge but instead of sphere as the reference surface they are departing from it is a parabola. Even if the spacing isn't dead perfect and you achieve a null, the result is a slightly over or undercorrected mirror but that is smooth. That is much better then a mirror that is "zoney".
As I said, you'll discover that actually getting a mirror to show a clean null is going to be the real challenge. Once you get to that point you can start to worry about getting the spacing dead on and tweeking the figure. You'll also discover that the differences from a perfect null at very slight differences of lens spacing is not easy to detect. So it is going to take some practice to detect these small differences. Again I recommend using a Ronchi screen over a knife edge and by moving the screen from inside of focus to the outside and back again fairly quickly, you can detect very small chances in the straightness of the bands.
When you think you have a perfect null, move the lens spacing by 1/8" and do the test again and see if you can detect the slight bowing of the lines. When you have trained yourself to be able to detect the difference, then you can worry about getting the exact lens to mirror spacing.

All the Best,
- Dave

#14 Brian Engel

Brian Engel

    Vostok 1

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

Posted 08 January 2013 - 02:14 PM

You should be able to, provided you use an eyepiece which is focused on the plane of the light source, set at the correct source-to-lens distance. Then, by sliding the sled until the image of the source is in sharp focus, the lens-to mirror distance should correspond to the theoretical value within the margins of error allowed.


Why is the eyepiece needed? Guess I am missing something.... The source and the KE would be the same (slitless).

I am thinking that the same "effect" of looking across the KE with the the lens to mirror distance precisely fixed and varying the KE to lens difference would be exactly the same as fixing the source/KE to lens difference and varying the lens to mirror difference.

Guessing this is not the case? (obviously my understanding of lenses/optics is pretty limited :-) )

#15 MKV

MKV

    Mercury-Atlas

  • *****
  • Posts: 2623
  • Joined: 20 Jan 2011

Posted 08 January 2013 - 06:52 PM

If a knife edge can produce a null with the ross test it must be possible to set a lens to knife edge distance and then move the lot as a unit...The yahoo atm_free group is heavily into bath interferometers.... There are plenty of reports by people who have used them on the group and for what you are doing it's likely to be a sensible way to go. Me too if I finish an F2.5 mirror.

John, when I do the Ross test I set both distances. This way I know I am starting from the theoretical focus. To see the Ronchi bands, you have to move the source carriage a small distance until you see the bands. David G was only saying that for a Ronchi method you don't need a fixed object-lens distance, and he's right - because the Ronchi test doesn't work at the focus, but slightly in or out of the focal range.

Rather than using k-e at the focus,it's much simpler to use an eyepiece at the fixed source-to-lens distance. If the eyepiece shows a sharp image (a diffraction disc) of the object, but the lens-to-mirror distance is off, then your conic is off too. When both distances are correct and the image is in sharp focus, then the conic constant is correct within the tolerance limits.

For interferometry, you need to be in focus, so the source-to-lens distance needs to be fixed in addition to the lens-to-mirror distance.

Bath interferometers are very useful and easy to construct. I use them all the time. However, a Bath will not be of much use for an f/2.5 mirror. You're better of using an autocllimator or some other null test (such as conjugate null) with such short radius of curvature.

Mladen

Attached Files



#16 MKV

MKV

    Mercury-Atlas

  • *****
  • Posts: 2623
  • Joined: 20 Jan 2011

Posted 08 January 2013 - 06:59 PM

John, for an eyepiece test you should use a pinhole, not a slitless source. Slitless sources are not as sensitive. Anyone who's done a Ritchey-Common test for flats knows this from experience. The closer the light source is to the resolution limit the more sensitive it appears.

For the Ronchi, test you must be inside or outside of focus. That's why you should be able to move your object source in and out of the focal position. You will immediately see if you are overcorrected, undercorrected or just right.

You can then apply the k-e and see if you get a null. If you don't have a null, the k-e test is of little value here because the TA is different than at ROC to determine how much you're off.

Mladen

#17 Ed Jones

Ed Jones

    Mercury-Atlas

  • *****
  • Posts: 2731
  • Joined: 06 Apr 2004
  • Loc: Sin-sin-atti

Posted 08 January 2013 - 07:32 PM

Because the focal length of the lens is a lot shorter than the mirror then it's separation from the KE/source is by far more critical. This is assuming the lens focal length is known to a high degree of accuracy (which it should be) for the test wavelength and most likely the mirror focal length will be less accurately known. You should set the KE lens separation and use the lens/mirror separation as a check. Both should be good.

On the other hand if the lens is less accurately known and the mirror is known to high precision then it might make sense to do the opposite but then why would you be testing with a less qualified lens?

#18 MKV

MKV

    Mercury-Atlas

  • *****
  • Posts: 2623
  • Joined: 20 Jan 2011

Posted 08 January 2013 - 09:37 PM

On the other hand if the lens is less accurately known and the mirror is known to high precision then it might make sense to do the opposite but then why would you be testing with a less qualified lens?

That's a good observation. Industrial lens standard is not that high, so are you saying that most, if not all, amateur Ross tests are done with lenses of inferior quality compared to the mirrors?

Mladen

#19 Ed Jones

Ed Jones

    Mercury-Atlas

  • *****
  • Posts: 2731
  • Joined: 06 Apr 2004
  • Loc: Sin-sin-atti

Posted 08 January 2013 - 11:07 PM

Well they have a focal length tolerane of 1 or 2 percent. Off-the-shelf lenses may not be good enough and need to be checked.

#20 MKV

MKV

    Mercury-Atlas

  • *****
  • Posts: 2623
  • Joined: 20 Jan 2011

Posted 09 January 2013 - 01:49 AM

Well they have a focal length tolerane of 1 or 2 percent. Off-the-shelf lenses may not be good enough and need to be checked.

One to two percent focal length tolerance in a typical PCX lens (although it doesn't have to be PCX) used in the Ross is negligible. Try raytracing it and adjusting the bfl.

What's more important is the surface quality, but since you're typically using only the inner 30-40% of the CA, or even less in some cases, chances are your lens meets the 1/4 to 1/8 PV wave WFE, which, as DAVIDG observed, is not so easy to achieve, contrary to popular opinion, especially for big mirrors.

Perhaps if you're shooting for 1/16 wave WFE you may find yourself having to either make a better lens, or buy one for a hefty price, but surplus precision lenses, with documented data are not too expensive, and come in more than sufficient size needed for a Ross. For example SS lens PL1031 is 75 mm in diameter, fl = 185 mm, BK7, optimized at 546.1 nm (so it should be used with a correspodning NB filter), unused, and costs $8! Its data sheet number LPX 263 is from Melles Griot from which you can see the tolerances and precision, as well as other pertinenet data.

I am using a 115 mm diameter, precision positive meniscus from SS, and only its inner 35% central diameter is used.

Chances are such lenses meet the 1/4-1/16 WFE requirement within those parameters.

Mladen

#21 MKV

MKV

    Mercury-Atlas

  • *****
  • Posts: 2623
  • Joined: 20 Jan 2011

Posted 09 January 2013 - 02:17 AM

If the lens was moving along with the knife edge/grating, it would be varying the amount of spherical aberration and you again could find a position were it nulls but the conic is not what you want

Dave I don't know if you replied to me (my post is referenced in your reply), but I just wanted to say you're right. The only time the lens moves is to set the lens-to-mirror distance. From there on you can do a fixed object-to-lens distance "at focus test" (k-e looking for a null, or an eypeiece test looking for a sharp image of the pinhole), or you unlock the light source stage and move the light source and the Ronchi grating in an out to assess the correction, as you said.

The latter is absolutely the easiest Ross null method. With it, one can immediately see if the mirror is overcorrected or undercorrected, or "just right" (straight "jail bars").

A Ronchi null assures that the WFE is ~ 1/5 to 1/6 wave p-v, especially if using a single Ronchi bar and slowly panning it across. This by itself assures you that the mirror is good to 1/10-1/12 waves, and, in the absence of too many scratches, pits, hills, etc., a smooth mirror surface testing as an optical Null in a Ross will have an even more impressive RMS rating.

Mladen

#22 Ed Jones

Ed Jones

    Mercury-Atlas

  • *****
  • Posts: 2731
  • Joined: 06 Apr 2004
  • Loc: Sin-sin-atti

Posted 09 January 2013 - 08:14 AM

Most of the off-the-shelf lenses I've seen do not meet the 1/10 wave regularity requirement. Finding a good lens is one of the problems with the Ross test.

#23 MKV

MKV

    Mercury-Atlas

  • *****
  • Posts: 2623
  • Joined: 20 Jan 2011

Posted 09 January 2013 - 08:45 AM

Most of the off-the-shelf lenses I've seen do not meet the 1/10 wave regularity requirement. Finding a good lens is one of the problems with the Ross test.

Just curious, how do you test a lens for that quality? Aren't lens requirements less stringent than for mirrors?

Most lenses may indeed not meet the required quality, but precision lenses, especially if large, and used only in the 1/3 inner diameter, are usually deemed adequate.

Mladen

#24 Brian Engel

Brian Engel

    Vostok 1

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

Posted 09 January 2013 - 09:20 AM

Just curious, how do you test a lens for that quality? Aren't lens requirements less stringent than for mirrors?

Most lenses may indeed not meet the required quality, but precision lenses, especially if large, and used only in the 1/3 inner diameter, are usually deemed adequate.

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

#25 Brian Engel

Brian Engel

    Vostok 1

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

Posted 09 January 2013 - 09:34 AM

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.

Is this about right (I hope :-))?






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics