Jump to content

Banner.jpg.b83b14cd4142fe10848741bb2a14c66b.jpg

The title of a 70's British ATM book?

wurzelmike

By wurzelmike
in DIY Astronomer

Recommended Posts

Glasspusher

Glasspusher

Posted
Posted

Ajohn, you have raised many points here. I have used both the Dall and Ross null tests and would like to add some points. Both tests use a plano convex lens to introduce spherical aberration equal, but opposite in sign, to the mirror being tested. The lens used must be of high quality if the test is to be rigorous. The Dall test is better suited to longer focal length mirrors as is used off axis and as such introduces unwanted aberrations that make interpreting the mirors figure difficult. No such problem exists with the Ross test which is used on axis and therfore lends itself well to testing faster mirrors. With both tests accurate spacings are essential if you are to avoid a 'Hubble'! Fortunately, software is available for the Ross test which gives the spacing for the various components, see the link on this page..http://www.ceravolo.com/Ross_Null2.html

Here is an excellent account of how to use the Ross Null test...http://www.ceravolo.com/ross_null.pdf

For an overview of testing see my notes here...http://www.nicholoptical.co.uk/The%20Testing%20of%20Astronomical%20Telescope%20Optics.pdf

John

Link to comment
Share on other sites
  • Replies 66
  • Created
  • Last Reply
wurzelmike

wurzelmike

Posted
  • Author
Posted

Thanks for the advice Ajohn, plenty of stuff to think about. I'll probably start with tiles and see how things go. I've got a copy of Texereau's book, it looks like something you might wanna cuddle up to, on a wet, and windy winters eve. Something to really look forward to.

If anyone can point me in the direction of the Howard book I'll be much obliged :smiley:

Link to comment
Share on other sites
Ajohn

Ajohn

Posted
Posted

Howards book is on the internet archive as well. Handbook for telescope making. These and Texereau's books are way way out of copyright and one of the more recent get rich schemes is to sell them.

Dall's test is cleverer than Ross's as it's normally done. Have to get a bit technical here. With the Dall test the tester is position so that the image of the pin hole falls on the knife edge. This is rather important as the sphere can be converted to a paraboloid by removing glass from several different areas. The test accounts for that as the pin hole is positioned in relationship to the lens to suit a specific focal length mirror. This means that the tester can always be set up to the knife edge without any other measurements other than the one in the tester which is easy to set very accurately. Actually this is how the professional tend to work only they use testers with more lenses in it. The extra lenses are needed as they are chasing much higher levels of accuracy than any of us are likely to achieve and on very fast mirrors.

The Ross test as it is normally done involves setting the test lens at a very specific distance from the mirror. Any variation in that changes the accuracy of the test. People use long sticks etc to set it and even worry about them sagging. It doesn't self calibrate that distance itself like the Dall one does. On the Dall test the critical distance can be set with callipers digital or otherwise.

As to lens accuracy the Dall test will achieve good accuracy with a commercial quality lens. He suggests measuring the focal length of the lens used backwards as that ensures a reasonable degree of accuracy all by itself. Slow mirrors are often mentioned. The real problem is the focal length of the mirror. He suggests less than 2% of that as the max separation between the knife edge and the central axis of the tester. It can be used with bigger faster mirrors really. Also that the tester has to aimed to within a degree or two of the centre of the mirror. It's worth bearing in mind that he was a very conservative bloke. There is a note at the end of the article about some one who used the test and then checked it via auto collimation on a 24in F5 mirror and found that the results were the same. That is a size where Dall thought that the test wouldn't be as good as he would like. This mirror was being made by a professional for himself.

Where the Ross test may score is that allows a large lens to be used. The argument is that this means that only the very centre is being used so it's likely to be more accurate. The same can be less true of the Dall test. As to the Ceravalo link that's one of the ways he earns a living. For most people a lot depends on how much one wishes to do themselves eg polish up a lens with a lap to improve it's figure, very feasible on a Ross lens a little more difficult on one for Dall but frankly I trust his comments - measure the actual focal length etc.

I checked the Dall test using a lens out of a cheap 5x eye cup type loupe against the caustic test and the usual mask. Couldn't see any difference other than that the first 2 are easier to do. With mask I found a variation of another way of testing mirrors useful. Central hole, another at the 70% zone and another at the edge. Then checking that the shadows moved smoothly from the central portion to the edge. Also a mask as Texereau describes it. I made the Dall tester so that the 2 parts moved on a strip to avoid the use of tubes - :grin: bit of a mistake as the light from the pin hole illuminated the edge of the lens holder and gave some rather odd optical effects especially as the bits were made out of perspex. Had to black it out and cover it but still easier than using tube.

I've pointed out to some people that it should be possible to set up the Ross test in the same way as the Dall. I'm still not sure I would be happy using it though. Dall went through a whole load of lens and mirror combinations and I have never seen anything similar for Ross. People ray trace a very expensive precise lens to get round that. They need a precise lens to allow them to do that. Trying to measure a lens accurately enough for that at home is nigh on impossible. I would rather spend my money on a pellicle beam splitter.

John

-

Link to comment
Share on other sites
Ajohn

Ajohn

Posted
Posted

:evil: 6in is an interesting size for a tool for a 10in mirror. Never tried myself but have read that that relationship allows polished radii to be maintained at the same value if the stroke is correct were as usually it changes all the time. As far as rough grinding using a 6in tool it seems it can be use for up to about an 18in mirror. The whole idea behind mirror grinding is that there are only 2 surface that can rub together and maintain contact - a section of a sphere and a flat. The tool has to be moved around to ensure that's happening all over - eg centre over centre to start with and later a figure 8 etc and any other strokes you can think of. When the curve is deep enough best switch to a full sized plaster tool with tiles or what ever on it and later cast another to make a lap for polishing.

Actually when the final figuring is done on say a 12 or maybe 15in mirror by hand the 6in disc could come in handy for another lap. Texereau show a sketch of the "small tool" method. If some one wants to make a fast mirror it's probably the best bet along with some use of the full sized lap. The strokes are more or less normal in the centre which is deepened so it should remain central to the rest of the mirror. The edge of the mirror is then flattened a little. Again that shouldn't be too bad to do. There is a method where only the minimal amount of glass is removed, Fine on slow mirrors such as the usual 6in F8 etc but the chances of it working on a fast mirror is remote unless a lap with a petal like pitch areas is used. There are other methods as well but the kit needed gets more complicated. I think Texereau also mentions that a variety of laps can be useful. Once some one has done a bit of work like this they will soon realise why especially on faster mirrors. One option is to cut out paper shapes and press the lap onto the mirror with the shape in place. This depresses the lap locally so it removes glass from the other areas. It seems this is how some manufacturers work. They record the result of a test with a camera and a machine produces the paper shape.

:grin: The biggest problem with all of this really is experience. If some one is doing it all of the time they soon learn what to do to correct an error and for how long. Us well we keep correcting errors and hopefully as these move around get a better idea of what to do and maybe even by accident eventually produce an excellent mirror. I've made one and the initial stages of figuring taught me a lot very slowly. Hours of work and still no good. Eventually I managed it in a more sensible time. A very accurate mirror but I didn't really fully polish it out when it was a sphere. There were still a few very fines scratches left that would never polish out without going back to fine grinding. I suspect I didn't clean the grits off well enough as I went to a finer grade or maybe the grit was contaminated. The scratches showed up if the mirror was examined with a small cheap torch at at various angles. Nobody had suggested that.

John

-

Link to comment
Share on other sites
Ajohn

Ajohn

Posted
Posted

Hi John. As I am definitely going to try and finish an F3 mirror I am going to look more closely at the Ross test but there have been some discussion on mirror to lens spacing on cloudy nights very recently. The tolerance on the distance from the mirror to the lens is ridiculous. From memory Ceravolo mentions sticks with plastic buffers on the end. With this particular mirror even that would be difficult to achieve. What I don't understand on that aspect is that if a knife edge is used when the set up nulls the knife edge must be a fixed distance from the lens and that in turn should fix the distance from the lens to the mirror - much the same way as the Dall test. :evil: I mentioned this on the thread and it's gone silent. That may leave me with a problem. I can't ray trace it meaningfully without access to a very accurate lens. Dall has done that for me and I trust his conclusions on the quality of lens that is needed. Ceravolo's lens costs $500 odd without VAT. A suitable Dall lens from Linos is about £28. The best from them that might be suitable for Ross is over £60. It's 63mm dia. They are one of the few that make precision plano convex lenses. Dall also gives some guidance on what focal length lens to use in relationship to the focal length of the mirror. F3 is pushing it really but it's only a 220mm mirror. F3 because it's for a compound scope by the way.

The Dall spacing is simply multiplied by the conic constant if anything other than a paraboloid is needed. I suspect I can make a pellicle beam splitter. It wont be 50/50 but that doesn't matter really.

Testing the mirror is something I am still looking around at. Knife edge to pin hole distance in relationship to the focal length of a mirror is still a problem with either set up except Dall plus splitter. There is also the Wanio test which looks a lot better for a completely on axis set up. This uses a spherical mirror to do the same thing as the lens does in the other tests. Complete home brew set up with no chromatic problems. There is info here http://www.telescope-optics.net/waineo_null_test.htm also on the others but I would be careful with the calculated spacings for the Dall test. His is actually a graph obtained by ray tracing. There is also some comments on Waneo here.

http://www.cloudynig...1/o/all/fpart/2

:confused: One thing I should have said re the pin hole to lens spacing being the important thing in practice when I look at it properly I may find that the centre part of the mirror is the reference sphere that the tests are based on. Basically I have my doubts that a 1/10 wave null can be obtain with the knife edge reasonably close to the centre of curvature of the mirror other than when it's in exactly the right place. The knife edge can always be positioned there via a mask and a normal tester anyway. If anyone is interested in this sort of thing Olso EDU is free as is WinLens and both can be used to trace set up's like this.

John

-

Link to comment
Share on other sites
Ajohn

Ajohn

Posted
Posted

Hi John. As I am definitely going to try and finish an F3 mirror I am going to look more closely at the Ross test but there have been some discussion on mirror to lens spacing on cloudy nights very recently. The tolerance on the distance from the mirror to the lens is ridiculous. From memory Ceravolo mentions sticks with plastic buffers on the end. With this particular mirror even that would be difficult to achieve. What I don't understand on that aspect is that if a knife edge is used when the set up nulls the knife edge must be a fixed distance from the lens and that in turn should fix the distance from the lens to the mirror - much the same way as the Dall test. :evil: I mentioned this on the thread and it's gone silent. That may leave me with a problem. I can't ray trace it meaningfully without access to a very accurate lens. Dall has done that for me and I trust his conclusions on the quality of lens that is needed. Ceravolo's lens costs $500 odd without VAT. A suitable Dall lens from Linos is about £28. The best from them that might be suitable for Ross is over £60. It's 63mm dia. They are one of the few that make precision plano convex lenses. Dall also gives some guidance on what focal length lens to use in relationship to the focal length of the mirror. F3 is pushing it really but it's only a 220mm mirror. F3 because it's for a compound scope by the way.

The Dall spacing is simply multiplied by the conic constant if anything other than a paraboloid is needed. I suspect I can make a pellicle beam splitter. It wont be 50/50 but that doesn't matter really.

Testing the mirror is something I am still looking around at. Knife edge to pin hole distance in relationship to the focal length of a mirror is still a problem with either set up except Dall plus splitter. There is also the Wanio test which looks a lot better for a completely on axis set up. This uses a spherical mirror to do the same thing as the lens does in the other tests. Complete home brew set up with no chromatic problems. There is info here also on the others but I would be careful with the calculated spacings for the Dall test. His is actually a graph obtained by ray tracing.

http://www.telescope...o_null_test.htm

There is also some comments on Waneo here.

http://www.cloudynig...1/o/all/fpart/2

:confused: One thing I should have said re the pin hole to lens spacing being the important thing in practice when I look at it properly I may find that the centre part of the mirror is the reference sphere that the tests are based on. Basically I have my doubts that a 1/10 wave null can be obtain with the knife edge reasonably close to the centre of curvature of the mirror other than when it's in exactly the right place. The knife edge can always be positioned there via a mask and a normal tester anyway. If anyone is interested in this sort of thing Olso EDU is free as is WinLens and both can be used to trace set up's like this.

John

-

Edited twice due to problems with the links!

Link to comment
Share on other sites
Glasspusher

Glasspusher

Posted
Posted

Ajohn, Making an F3 mirror is not an easy task. When it comes to testing such a mirror the Dall null test is a none starter. The test is performed off axis and the amount of coma resulting from this would make the test unreliable. With the Ross test the spacings depend on the characteristics of the lens being used. My lens is 100mm in diameter and has a RoC of 230mm. Running the numbers through the software on Peter Ceravolo's site for a 220mm diameter F3 mirror shows the test to be quite useable. For a 1/4 wave null, the lens to mirror spacing can vary by about 5mm form the calculated value of 934.887mm. This kind of accuracy is easily obtainable with a standard measuring tape. The distance between the light source and the lens has to be a little tighter but it is only 194.896mm, a spacer of this length could be made to within 1mm with care. Incidentally, the working diameter of the lens in this case is 64.8mm.

Remember, The Ross lens is a plano convex lens and needs to be of proven quality. In general bigger longer radius lenses are preferable and will be easier to set up within tolerances.

John

Link to comment
Share on other sites
Ajohn

Ajohn

Posted
Posted

Thanks John. I have downloaded the software to look at that sort of thing.

I think you are missing the point on the Dall test. As he mentioned and as is often done in many forms of optical testing a beam splitter is needed. The layout is like this. It's even needed for testing a sphere or using the normal knife edge test on fast mirrors.

........................................................... Pin Hole and Source & Light

...................................................................................|

...................................................................................|

..............................................................................Lens

..................................................................................|

..................................................................................I

Mirror ------------------ Beam splitter at 45 degrees ---------- Knife edge etc

The usual advice for the beam splitter is a microscope cover slip that is within a few waves of flat. The other important thing is that it is very thin so that it doesn't introduce significant aberrations. If there are problems in a test the mirror can be ruled out by rotating it and seeing if the problem stays in the same place on the mirror. Same with the beam splitter. As a 50/50 spit beam splitter isn't ideal really a transparent one is a better option. I suspect I can make one out of cling film. 50/50 ones can be bought. They would send 50% of the light from the pinhole to the mirror and allow 50% back through to the knife edge. As it's easy to put lots of light into the pin hole only a small % needs to go to the mirror.

:grin: Reminds me of something. Once after spending a long time touching up the mirror and retesting I noticed that my vision was a bit strange. Something didn't feel right. I looked in a mirror and one pupil was wide open and the one I have used via the tester was like a pin [removed word]. Took several hours to settle down.

So as I hope you can see the Dall test isn't out of the question John. I also suspect when I look at the Ross test that light source / knife edge spacing will introduces problems beyond a certain point as well. Also using a test stated to be accurate to 1/4 wave doesn't help much because in the end that may be worse because out eyes and brain interpret the result. I do know that people who use the Ross test and discuss it seem to be interested in a 1/20 wave test set up. Having a test set up that has at least twice the accuracy needed make rather a lot of sense.

It's worth mentioning something about the 1/4 wave limit as well. The criteria is better than a 1/4 wave limit as thing get rapidly worse when it's larger. :eek: I have a little picture that illustrates this nicely but doesn't account for the additional effects of a central obstruction. The vertical axis is the fraction of the contrast that goes in that comes out of an optical system. The bottom axis is resolution but it's expressed in relationship to the point 2NA/wavelength where the contrast is 0 what ever is coming in. 0.1 for instance is where the resolution "going in" is 10 times "coarser". I have it on my machine because many people don't understand why they can't see things to Rayleigh's other limit - resolution with their microscope objectives. The contrast at that limit is about 7 1/2 %. This is physics so it's basically factual and assuming and even error on a mirror for instance defocus has the same effect. On a mirror though the location and size of the area of the error can have a further effect. A big hole in the middle for instance may have little effect when looked at this way as it's area is small. A serious error in the outer edges of a mirror can have a much worse effect as it's area is large piR^2 etc. Most mirrors are sold on P/V rating and often the RMS of that - just makes them look better than they really are. It's just like the other habit of specifying central obstructions as a % of area rather than dia - that is extremely common.

In my view if some one is going to make their own mirror they should at least try to get near Texereau's 1/10 wave limit in his terms or at least better than 1/4 wave but also in his terms.

post-2035-0-98148500-1359733452_thumb.jp

John

-

Link to comment
Share on other sites
Glasspusher

Glasspusher

Posted
Posted

Even with a beamsplitter arrangement you will still see coma with an F3 mirror using the off axis Dall null test. This would not be the case with the on axis Ross test. A 1/20 wave set up is unrealistic, obtaining a lens of the necessary quality would be difficult to say the least...not too mention the expense. I am just trying to suggest workable practical solutions based on my own experiences with both tests.

John

Link to comment
Share on other sites
Ajohn

Ajohn

Posted
Posted

I think John is missing something a little serious - the beam splitter keeps everything on axis. It basically acts like a mirror but only reflects so much while also letting some go straight through it.

So light leaves the pin hole goes through the lens and hits the beam splitter. Some passes straight though it the rest is reflected down to the mirror exactly as if the pin hole and lens was on the axis of the mirror. Light hits the mirror and is reflected back to the beam splitter, same thing happens, some is reflected back to the lens and pin hole unit and the rest goes through it exactly on axis to the knife edge.

If I bought a pellicle such as one of these which given the price is highly unlikely

http://www.thorlabs....ectgroup_id=898

and say it was 50/50 1/2 of the light is reflected as the little picture in the link shows and the other half goes straight through it. As the links states they are only a few 0.001mm think so optically have no effect at all other than there job to reflect t he beams and also let some go straight through. For testing it's generally better to have a beam splitter that lets a lot more through than it reflects. Hence the use of reasonably flat microscope cover slips which can be down to 0.1 mms thick, still an insignificant % of the optical path length which is the important thing. I think it can be done with cling film though. Don't see why not but then a I have a lathe etc.

Actually on very short focal ratio mirrors they can be needed with the Ross test as problems crop up as soon as the separation between the light source and the knife edge exceed a certain amount.

:embarassed: I read it on the web etc so it must be true. That sort of thing never ceases to amaze me.

John

-

Link to comment
Share on other sites
Ajohn

Ajohn

Posted
Posted

I have found and sorted out a design of a bolt on attachment to convert a normal parabolic mirror into an astrograph so things are moving. The original design is a bit big as I can't see myself finding a ccd camera with a 5in diagonal chip in it. Next thing to sort out is the cassegrain. Take the cassegrain attachment off and fit the astrograph attachment and ...............

post-2035-0-89286300-1360183321_thumb.jp

This is a 1/2 field spot diagram so it covers 3 1/2 degrees over a 4.8 in dia field. May come out more suitable with my 220mm mirror

John

-

Link to comment
Share on other sites
wurzelmike

wurzelmike

Posted
  • Author
Posted

Of course you could make your 6-inch disc into a secondary mirror and grind a larger primary mirror to power it....

I did think of this fwm891, but figured the primary would have to been around 18" :grin: .......a little more thought, and I came up with making a Cassigrain, in particular a Kletsov-Cassigrain one, with a six inch primary, a small meniscus lens and something called a Mangin mirror. Sounds like a fun project for a first attempt :smiley: and an interesting design too.

But after more reflection, I've decided to make two mirrors from the kit. A 6" f/5 from the pryrex disc. And another 6" ( this time a f/8) from what was originally intended to be the plate glass tool. I'm thinking that if I grind both mirrors at the same time, I should end up with a decent rich field scope, and a nice planetary one without much extra effort. Learning a little more along the way and making the most of the contents of the kit.

Link to comment
Share on other sites
wurzelmike

wurzelmike

Posted
  • Author
Posted

@John and John :smiley:,

I've been trying to follow your discussion as much as possible, not always succeeding :grin: But I was wondering, as a first time mirror maker what tests can i realistically use. Other than the knife edge/Foucault test are there others? I've come across something called the 'Coulder mask' and the 'pin stick' on the Stellafane site, and they seems to be ways of assessing the results produced from the Foucault test, is that correct?

So are there any other tests that can be easily employed by a newbie?

Also, what kind of accuracy can some one hope to get from different these methods?

all the best, Mike

Link to comment
Share on other sites
Glasspusher

Glasspusher

Posted
Posted

Wurzelmike, One of the simpler tests available to the newbe is the 'Matching Ronchi Test', with care this test can produce a diffraction limited mirror providing the mirror is not too fast, around F6 should be OK. Details of the test can be found in my notes on ronchi testing here: http://www.nicholoptical.co.uk/pdf/The%20Ronchi%20Test.pdf

History has taught us that it is probably best not to rely on a single test. For a first mirror the matching ronchi test could be used in combination with a simplified zonal test, details of which can be found here...http://www.nicholoptical.co.uk/pdf/A%20simple%20zonal%20mirror%20test.pdf

Hope you find this of help.

John

Link to comment
Share on other sites
wurzelmike

wurzelmike

Posted
  • Author
Posted

thanks for the suggestions Glasspusher. Interesting and useful stuff as always

It seems like my first lesson on 'how to make a telescope mirror' is that it should really be called 'how to test a telescope mirror'. Seems like that's going to be the difficult part of the job.

Link to comment
Share on other sites
fwm891

fwm891

Posted
Posted

Wurzelmike, you are spot on. The first question to ask when thinking about making a mirror is how can I test it. Testing may seem a little daunting at first but it is where the real fun in making your own optics is to be found.

John

It may seem OTT at first but a simple testing tunnel to remove thermals when testing makes a great deal of difference. A super bright green LED makes a great light source for testing...

Link to comment
Share on other sites
fwm891

fwm891

Posted
Posted

Oh dear, I came out of Maplins yesterday with a huge blue LED, thinking that, as we're using light to measure things the shorter the wavelength the better :rolleyes:

Your eye's are more sensitive to yellow/green area of the visible spectrum...

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

Go to topic listing

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue. By using this site, you agree to our Terms of Use.