Jump to content


Photo

Thinking of making a flat...

  • Please log in to reply
139 replies to this topic

#51 DAVIDG

DAVIDG

    Aurora

  • *****
  • Posts: 4887
  • Joined: 02 Dec 2004
  • Loc: Hockessin, De

Posted 22 April 2014 - 02:08 PM

Mladen,
Both myself and Dick Parker have checked the results of the test and it works just like the article states. Dick has made at least 3, 12" flats and one 16" using this method to grind only one surface. I also belive a couple of students in his mirror making class that he teaches he year up in New England has used this method as well.
The purpose of the test is to allow one to check a ground surface that won't allow interference testing and to get it flat enough that once you start to polishing you can be close enough to make it optically flat. It is not to determine the exact flatness. So as I stated it allows you to grind only one surface and check it to determine if it is concave or convex as your grinding it and adjust the grinding technique accordingly. To me this save both time and money since I'm only working one surface. Once you get it mechanically flat ins which the surface is within a few waves you then polish it and start the typical optical testing, via contact interference against a Master flat and/or the water test or using spherical mirror.

- Dave

#52 MKV

MKV

    Mercury-Atlas

  • *****
  • Posts: 2610
  • Joined: 20 Jan 2011

Posted 22 April 2014 - 03:08 PM

Hi Dave, I agree, the acetone test serves well its purpose for non-polished surfaces. So does a spherometer, but it's much easier to get acetone than a spherometer. And, I think just grinding the tool and the flat, and alternating them top-to-bottom, will produce a sufficiently flat surface. Progress can be checked with a good machinist's straightedge and a bright light. One can be pretty sure that when no light is showing underneath the straightedge, the surface is more than flat enough for autocllimation purposes.

The way I look at it, making tools is a one-time deal in a long while. So why not go the extra mile and go with three disks, and get one spare, one master and one coated out of it, or sell two for a good profit. A set of flats will last you a liftetime. I don't see a need to cut corners when setting up a well tooled workshop. And the experience is another thing. If we look only at how quickly we can start using optics, or how much work is involved in making optics, why not just buy a flat, or a telescope? But, as always, let everyone decide what suits them best. :) This is just informational to show different aspects of different approaches.

regards,
Mladen

#53 Pinbout

Pinbout

    Cosmos

  • *****
  • Posts: 8034
  • Joined: 22 Feb 2010
  • Loc: nj

Posted 22 April 2014 - 04:31 PM

 This is just informational to show different aspects of different approaches. 



And that's why I posted the vid. It doesn't get talked about that much. Most go right to the three disc method. And with the price of supremax33, why spend money...I know I don't have.

#54 DAVIDG

DAVIDG

    Aurora

  • *****
  • Posts: 4887
  • Joined: 02 Dec 2004
  • Loc: Hockessin, De

Posted 22 April 2014 - 05:23 PM

Mladen,
As the name states, the three disks method requires three disks and it requires grinding and polishing all of them if you want to make three good flats. If you only want to make one flat surface it still requires grinding three surface and two of which won't be used.
I would rather save both the time and money by only grinding and figuring one surface. The time and money I save can be used to make a much larger flat, which to me is more useful then 3 smaller ones.
If one is going to go through the time to make three flats via the three disk method then one should use three disks of glass of the correct material ie low expansion glass of the correct thickness. With the cost of 10" disks of low expansion material costing about $250 each, and 12" disks cost around $500 each that it fair amount of money when you need to purchased three of them. I personally find having one large autocollimation flat much more useful then three smaller flats so I would spend the money to get one large piece of glass and the acetone test will allow me to grind that piece to a mechanical flatness that then can be optically polished flat.
To each his own.

- Dave

#55 Ed Jones

Ed Jones

    Mercury-Atlas

  • *****
  • Posts: 2713
  • Joined: 06 Apr 2004
  • Loc: Sin-sin-atti

Posted 22 April 2014 - 08:13 PM

The other method that I have used is to use a spherometer as well. I zero the spherometer on the piece that is going to be the flat and then test the tool.


I've used this method and it's quite sensitive with a good spherometer gauge. I zero on the tool an read the flat which doubles the difference as well. It's better and cheaper than the 3 part method IMHO.

#56 Pinbout

Pinbout

    Cosmos

  • *****
  • Posts: 8034
  • Joined: 22 Feb 2010
  • Loc: nj

Posted 22 April 2014 - 08:25 PM

for my 10", I used a 4.75" dia spherometer and I guess I didn't move it around enough to see the side rising up. it would've helped to have one the same dia.

with the acetone you can see the shape and where it starts to rise. no measuring and then guessing where it deviates from flat.

#57 MKV

MKV

    Mercury-Atlas

  • *****
  • Posts: 2610
  • Joined: 20 Jan 2011

Posted 22 April 2014 - 11:34 PM

for my 10", I used a 4.75" dia spherometer and I guess I didn't move it around enough to see the side rising up. it would've helped to have one the same dia.


Danny, if you're going to use a spherometer to assess the flatness of a disc, not only should it be nearly as large as the disk itself, but it should read to 0.0001" or preferably better (0.0001" is about 5 waves or 10 fringes)!

If the acetone test is truly 5 times more accurate, as claimed, then it beats the spherometer hands down. I would like to read more about this test.

regards,
Mladen

#58 Mark Harry

Mark Harry

    Vendor

  • *****
  • Vendors
  • Posts: 6189
  • Joined: 05 Sep 2005
  • Loc: Northeast USA

Posted 23 April 2014 - 07:13 AM

I get to within a couple fringes, and my sphero is 5" and graduated to only .001" between divisions. -Flat- lies exactly a half-mark (.0005") above zero on this particular tool.
I see that the acetone test requires use of a reasonable flat. That was my original question.
But I have various plate glass in various thicknesses. Blue tinted, and water white; along with a couple pieces of B270.
Absolutely -NONE- of it is what I would call REASONABLY flat. Not by any stretch. I can get a far better indication of flatness with the sphero, period. My 2 cents. (Of course, I'm totally familiar with it!)
M.

#59 DAVIDG

DAVIDG

    Aurora

  • *****
  • Posts: 4887
  • Joined: 02 Dec 2004
  • Loc: Hockessin, De

Posted 23 April 2014 - 08:42 AM

I agree that one can get a more accurate reading of the exact flatness using spherometer but that is not the point of the acetone test. It will allow one to grind only one surface and get it mechanically flat enough to be polished optically flat. It doesn't require fabricating a precision spherometer and it doesn't require an optical flat of known flatness to zero the spherometer on. A cheap piece of relativity flat glass and $5 worth of Acetone from the hardward store is all that is needed. It is a "close enough" test.

- Dave

#60 MKV

MKV

    Mercury-Atlas

  • *****
  • Posts: 2610
  • Joined: 20 Jan 2011

Posted 23 April 2014 - 09:31 AM

I get to within a couple fringes, and my sphero is 5" and graduated to only .001" between divisions. -Flat- lies exactly a half-mark (.0005") above zero on this particular tool.

Mark, a surface that departs flatness by 0.0005" is 25 waves or 50 fringes deep. In order to read flatness to a couple of fringes, your spherometer should be able to read to one wave, or 0.00002".

regards,
Mladen

#61 Ed Jones

Ed Jones

    Mercury-Atlas

  • *****
  • Posts: 2713
  • Joined: 06 Apr 2004
  • Loc: Sin-sin-atti

Posted 23 April 2014 - 10:07 AM

to use a spherometer to assess the flatness of a disc, not only should it be nearly as large as the disk itself, but it should read to 0.0001" or preferably better (0.0001" is about 5 waves or 10 fringes)


Quite right. I use a Mahr gauge that resolves 1/2 micron with a large base. That's about 2 fringes.

#62 MKV

MKV

    Mercury-Atlas

  • *****
  • Posts: 2610
  • Joined: 20 Jan 2011

Posted 23 April 2014 - 10:11 AM

As the name states, the three disks method requires three disks and it requires grinding and polishing all of them if you want to make three good flats.

Dave, the main advantage of a 3-disk method is that it favors flatness. You don't have to measure or test during grinding and lapping. Of the three disks, at least one should be a stable substrate that will be used as a flat; the others can be plate glass disks affixed to cement tools. They can be used later for making pitch laps for polishing and figuring of the flat itself.

So, the cost doesn't have to be excessive. And neither does one have to polish and figure all three disks! You can finish only one of the three, but if all three are stable substance, you can make parabolic mirrors out of them later on.

The stable substrate flat (for autcollimaiton flats, a Pyrex disk is stable enough) can be figured using using alternate methods other than the ABC reduction. You can use the Ritchey-Common test, with a long spherical mirror, as you hinted, or an interferometer, etc. but in either case, when you're figuring a single flat you need additional optics and/or equipment, such as a really good (1/20 wave RMS) long focal length sphere, which most people don't just have lying around.

So, short of borrowing one (chances are if someone has such a quality long focus sphere, he may also have a master flat...), you'll have to make a sphere, and that involves grinding and figuring another set of optics to a high precision -- and by that time you could have finished your three flats! The ABC method is costlier if all three disks are stable substrate, but if you make 3 good large flats you will more than recuperate the cost by selling just one of them.

regards,
Mladen

#63 MKV

MKV

    Mercury-Atlas

  • *****
  • Posts: 2610
  • Joined: 20 Jan 2011

Posted 23 April 2014 - 10:39 AM

Also, for those who're interested in testing unpolished surfaces for flatness, the pros use grazing interferometry, but there is also a simpler version called grazing Ronchi test.

Again, this is just informational, not a critique of other tests. There is no doubt that the acetone test is cheap, and easily performed. I have not found any source, outside of those already mentioned, that actually prove the accuracy of the acetone test being about 1 wave -- in which case it beats the spherometer hands down, except for the Mahr gauge; and a Mahr gauge, such as the one used by Ed Jones, is not cheap.

I was hoping that someone would offer an empirical proof of that claim, namely a test of a known 1 wave concave/convex surface tested by acetone and showing measurable curvature. But since an autocollimaiton flat can be many fringes concave or convex as long a sit has a smooth figure of revolution, it really doesn't matter if the acetone test has such accuracy. It may as well be 10 fringes off, it's still good enough for starters.

Mladen

#64 DAVIDG

DAVIDG

    Aurora

  • *****
  • Posts: 4887
  • Joined: 02 Dec 2004
  • Loc: Hockessin, De

Posted 23 April 2014 - 10:46 AM

Mladen,
One can use the water test which is cheap and very accurate vs making a sphere to test a flat. You don't have to a use collimated light source that is as large as the surface your testing when using the water test and scan over the surface without any loss of accuracy. I have a copy of an article from the Journal of the British Astronomical Society from the 70's that goes into the details and since that article was written I know of others that have used the water test using this method with excellent results. So between the acetone test and the water test I can make a larger flat in less time and less money then the three disk method.
I know that it has been done with excellent results because Dick Parker and Alan Hall made a 16" autocollimation flats to test their optics in their award winning twin 16" Cassegrains using these methods.
There are many ways to accomplish the same task. I'm just trying to shows a different method that others have used successfully.

- Dave

Attached Files



#65 MKV

MKV

    Mercury-Atlas

  • *****
  • Posts: 2610
  • Joined: 20 Jan 2011

Posted 23 April 2014 - 11:38 AM

Dave, Ed Jones has a Youtube video which shows how the test looks. Personally, I would find that test very difficult to use, especially in the final figuring stages of a flat, when you do short figuring spells, long cooling spells, and short testing spells.

The tests seems to be a nice cheap alternative to testing finished surfaces, and the sensitivity seems to be about 3 fringes (judging by the video) or 1.5 waves. That's plenty accurate for an autocllimation flat, but it seems to take much longer to obtain the results then by other tests.

By the way, you can use the Ricthey-Common test with a very small mirror compared to the size of the flat; it's just that the test is most sensitive when the two surfaces are angled at 54°45' -- which dictates the optimum sphere size.

regards,
Mladen

#66 Mark Harry

Mark Harry

    Vendor

  • *****
  • Vendors
  • Posts: 6189
  • Joined: 05 Sep 2005
  • Loc: Northeast USA

Posted 23 April 2014 - 03:30 PM

I get to within a couple fringes, and my sphero is 5" and graduated to only .001" between divisions. -Flat- lies exactly a half-mark (.0005") above zero on this particular tool.

Mark, a surface that departs flatness by 0.0005" is 25 waves or 50 fringes deep. In order to read flatness to a couple of fringes, your spherometer should be able to read to one wave, or 0.00002".

regards,
Mladen

***********
No it doesn't. With visual acuity (paying attention to what your eye resolves) and a magnifier, I can get well down in the single # fringes rather easily. Of course I've made thousands of them, so I'm sorta familiar in how to do such expeditiously. (practice makes perfect!)
M.

#67 Mark Harry

Mark Harry

    Vendor

  • *****
  • Vendors
  • Posts: 6189
  • Joined: 05 Sep 2005
  • Loc: Northeast USA

Posted 23 April 2014 - 03:33 PM

If you re-read what I said, -FLAT- is EXACTLY @ +.00050; not +.00049 or +.00051" This difference can be easily seen with a moderately strong glass. (if your eyes are up to it)
M.

#68 kfrederick

kfrederick

    Gemini

  • *****
  • Posts: 3018
  • Joined: 01 Feb 2008

Posted 23 April 2014 - 03:55 PM

Mercury flats and water flats are very flat. Wonder how flat they are frozen ? https://www.youtube....ntJaLJYKIO1FoYw

#69 Ed Jones

Ed Jones

    Mercury-Atlas

  • *****
  • Posts: 2713
  • Joined: 06 Apr 2004
  • Loc: Sin-sin-atti

Posted 23 April 2014 - 04:25 PM

The tests seems to be a nice cheap alternative to testing finished surfaces, and the sensitivity seems to be about 3 fringes (judging by the video) or 1.5 waves. That's plenty accurate for an autocllimation flat, but it seems to take much longer to obtain the results then by other tests.


Mladen,
The water test is actually more sensitive than testing with a master flat since a fringe in water is only 3/8 wave rather than 1/2 wave in air. When using a master flat you still need the part and master to equilibrate to the same temperature and using the water test the part needs to set in water to equilibrate so there isn't much time difference. If you are making an autocollimation flat you can get away from needing a collimating lens and use a light box provided you observe as far away as possible and always observe from the same distance. It will be a few fringes from absolute flat however. You still need a collimator for making a master flat.

#70 Pinbout

Pinbout

    Cosmos

  • *****
  • Posts: 8034
  • Joined: 22 Feb 2010
  • Loc: nj

Posted 23 April 2014 - 05:38 PM

With no collimating lens, at what viewing angle do ya need to be when your far away :question:

#71 Ed Jones

Ed Jones

    Mercury-Atlas

  • *****
  • Posts: 2713
  • Joined: 06 Apr 2004
  • Loc: Sin-sin-atti

Posted 23 April 2014 - 06:16 PM

at what viewing angle do ya need to be?

That's a ray trace problem. It depends on the size of the part and water thickness and how many fringes you can be off.

#72 Ed Jones

Ed Jones

    Mercury-Atlas

  • *****
  • Posts: 2713
  • Joined: 06 Apr 2004
  • Loc: Sin-sin-atti

Posted 23 April 2014 - 08:22 PM

I forgot I had this graph, it shows how far away you need to be for a given number of fringes given a water thickness of 1 mm.

Attached Files



#73 MKV

MKV

    Mercury-Atlas

  • *****
  • Posts: 2610
  • Joined: 20 Jan 2011

Posted 23 April 2014 - 11:51 PM

If you are making an autocollimation flat you can get away from needing a collimating lens and use a light box provided you observe as far away as possible and always observe from the same distance. It will be a few fringes from absolute flat however

I have always tested flat sets (ABC) using the standard viewer (pictured below), from 25 feet, and with a telephoto lens. I am surprised this would still be a few fringes off due to parallax.

Edmund Scientific has 12x12 inch Fresnel lenses which may make satisfactory collimating devices.

regards,
Mladen.

Attached Files



#74 MKV

MKV

    Mercury-Atlas

  • *****
  • Posts: 2610
  • Joined: 20 Jan 2011

Posted 24 April 2014 - 01:29 AM

With visual acuity (paying attention to what your eye resolves) and a magnifier, I can get well down in the single # fringes rather easily....If you re-read what I said, -FLAT- is EXACTLY @ +.00050; not +.00049 or +.00051" This difference can be easily seen with a moderately strong glass. (if your eyes are up to it)

That's not possible with digital readouts Mark. And one cannot go beyond manufacturer's accuracy rating and still claim accuracy. I don't know what equipment you have, so I'll take your word for it, but it's a little more than just visual acuity.

For those who don't have 0.001" dial indicator (0.001" = about 50 waves, or 100 fringes), each division is 1/16" (0.625") wide and the needle tip is 1/80" (or 0.0125") wide, so the needle tip fits inside the division exactly five times. That means each 1/5 of that division represents 10 wave or 20 fringes (letting wavelength = 0.00002" for practical purposes).

regards,
Mladen

#75 Mark Harry

Mark Harry

    Vendor

  • *****
  • Vendors
  • Posts: 6189
  • Joined: 05 Sep 2005
  • Loc: Northeast USA

Posted 24 April 2014 - 06:22 AM

I do -NOT- use a digital readout.(It's a vernier; a very expensive one !!!) You're ASSUMING too much, and modify my statements to suit your own point of view, while trying to "preach to the choir". This "trolling" is an annoyance, and it clouds the facts.
***
ED JONES is precisely right; if you want to make a master, you HAVE to use a collimated light source, or make allowances for viewing angles not perfectly normal to the surface under test. His chart posted here should give you a clue.
Visualise looking straight down at the test. The thickness of the water/substrate will be minimal, and significantly less than if viewed at an angle.(being at an angle, the water/substrate will have an actual thicker layer for the light to travel through, which makes it act like a negative lens.) The thicker water/substrate distorts the fringe pattern and makes direct visual assessment difficult; particularly if you LACK EXPERIENCE with interpretation. To me this distortion makes the optic look as though it has a down edge, or oblate shape. With small pieces/references, I use a positive PCX lens placed right on top to get rid of this distortion.
If you're just trying to make an AC flat for typical Newt mirrors, or to check a telescope as a system, a fringe or two off won't hurt anything. But this isn't the case for -ALL- optical elements.
And if one is going to go through the trouble to MAKE a flat, and not just buy one to check so as to avoid the necessary work, it doesn't require much effort to make the piece truely flat instead of a couple or 10 or 100 fringes off. Visual assessment of regularity with a +- 10 fringe flat is coarse and crude, and +- 100 fringe flat is essentially impossible.
I'm fairly sure that most who have made any amount of flatwork will agree with this. I may have forgotten a point or two.
M.






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics