Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

New book on interferometry

  • Please log in to reply
72 replies to this topic

#51 BGRE

BGRE

    Gemini

  • *****
  • Posts: 3,005
  • Joined: 21 Mar 2016
  • Loc: New Zealand

Posted 15 April 2018 - 04:25 PM

I've never found it difficult to align either a Twyman-Green or a Fizeau.
Once the source images in the focal plane of the collimator (in the case of a Fizeau) or the focal plane of a lens in the case of a Twyman-Green overlap accurately, fringes are detectable once the viewing arm is reconfigured to image the test surface onto the detector, provided the OPD doesn't exceed the source coherence length.
Aligning a Williams interferometer should be just as easy.

#52 MKV

MKV

    Cosmos

  • *****
  • Posts: 9,080
  • Joined: 20 Jan 2011
  • Loc: Southeast, USA

Posted 15 April 2018 - 05:39 PM

It's not difficult; it's just a lot more difficult than it is with the Bath. With the Bath, you pan the mirror tot he side and image both beams on a screen, then move the whole IF with the screen until the test beam appears in focus, then pan back making sure the reference beam falls on top of the diverger. It takes about 30 seconds. With the T-G one needs two beam alignment screens, and if the test beam and the reference beam axes are not very closely perpendicular to each other you end up with something like this in case of a null (easier to illustrate) but very unlikely to observe in this case.

 

T-G alignment issue.jpg

 

Making sure that both beams return to a desired location makes the T-G a nighmare compared to the Bath. Also for the Btah you don't need optomechanical actuators for alignment. Night and day.



#53 BGRE

BGRE

    Gemini

  • *****
  • Posts: 3,005
  • Joined: 21 Mar 2016
  • Loc: New Zealand

Posted 15 April 2018 - 05:52 PM

It took all of 30 seconds to obtain fringes on a Twyman-Green when I first built one several decades ago.
I never encountered any issues with beam shear when the correct alignment method was used.

#54 MKV

MKV

    Cosmos

  • *****
  • Posts: 9,080
  • Joined: 20 Jan 2011
  • Loc: Southeast, USA

Posted 15 April 2018 - 07:54 PM

Depends what you mean by "the correct alignment method". And then the speed is meaningful only if you compare it to another IF. For a T-G, the reference beam has to be set first and fixed, and then the whole spherometer is moved in x,y, z to get the test beam to fall on top of the reference beam. Compared to the Bath, it's forever.

 

With the Bath you also don't need an HeNe laser, expensive reference surfaces,  or have to  worry about the coherence length of your signal. Aaaand, the Bath is minnisicule in size, weight, and cost compared to the T-G.

 

As I said, unless you need to test using an Offner/Ross type nulling compensator,an ATM should rarely if ever need to use anything more complex or expensive than a Bath. Last time I checked, ATMs who need and are willing to afford an Offner or a transmission flat are vewy, vewy few. :o)



#55 BGRE

BGRE

    Gemini

  • *****
  • Posts: 3,005
  • Joined: 21 Mar 2016
  • Loc: New Zealand

Posted 15 April 2018 - 09:28 PM

Done correctly, setting up a Williams interferometer should only take slightly longer than setting up a Foucault or Ronchi test.
The biggest drawback of the Williams interferometer is that the OPD cant be adjusted to maximise fringe contrast when using a cheap HeNe laser.
When an interferometer wavefront has a coma component common to the test and reference beams adding tilt fringes also produces astigmatism in the interferogram prortional to the number of tilt fringes.
Spherical aberration components common to the test and reference beams results in tilt dependent coma.
Common astigmatism components result in additional tilt dependent on the beam shear used to introduce tilt fringes.
The Bath, Fizeau, Twyman-Green and Williams interferometers all have this characteristic. Thus the wavefront errors of the laser (and collimator lens) used with these cannot be too large if accurate results are required.
Typically laser wavefront errors (especially SA, coma and astigmatism) should be 1 wave or preferably less if tilt fringes are to be used.

Sources of these aberrations include:

Laser diode itself - astigmatism
Beamspltter cube tilt (with non collimated light) - astigmatism, coma
diverger or collimator lens tilt - coma, astigmatism
diverger or collimator lens decentration - coma, astigmatism

The original Bath interferometer with the beamsplitter cube tilted at 45 degrees to the interferometer axis is particularly bad at producing coma and astigmatism common to both beams so particularly with fast test surfaces it should only be used with zero tilt fringes.
The modified Bath where the beasplitter has a tilt of less than 5 degrees produces an order of magnitude less coma and a couple of orders of magnitude less astigmatism.

Spatial filtering using a pinhole with a diameter <= 1/2 the Airy disk diameter of the collimator lens will virtually eliminate coma and astigmatism even in a laser module using a cheap plastic lens. Just use a high quality lens to recollimate the light transmitted by the spatial filter.
A single mode fiber can also be very effective in eliminating coma, spherical aberration and astigmatism.
Otherwise selecting laser modules for low (< 1 wave ptv) wavefront error is another option.
In a Bath (or other interferometer) one can test the setup for these effects by demonstrating that the coma and astigmatism do not vary significantly as the number of tilt fringes used is varied.
  • mark cowan likes this

#56 MKV

MKV

    Cosmos

  • *****
  • Posts: 9,080
  • Joined: 20 Jan 2011
  • Loc: Southeast, USA

Posted 16 April 2018 - 06:08 AM

Thank you, Bruce. I know this discussion has been a hot topic on on the interferometry forum lately, and is very interesting and informative, and relevant at some point, but none of it helps an amateur set up a Twyman-Green interferometer quickly.

 

This book (the topic itself) is intended for ATMs. That usually means low budget, limited tooling, skills and experience. What amateurs need are step-by-step instructions.

 

Here's my proposal: I can make a comparison chart but I don't want to prejudice the results. Why don't you list all the steps and necessary equipment needed to set up a T-G/William IF and I will then list steps on how I set up my Bath. Then we can tally, the time, the equipment and the money needed and I am quite confident that the Bath will turn out to be a more economical, easier to use and quicker to set up.

 

I built my T-G out of curiosity and in case I might need it for an Offner or Ross type compensator or for testing flats, but so far I found exactly zero-reason to use it. DFTFringe software has an artificial null so an Offner/Ross is not required, and since I am unlikely to ever test anything too big and too fast for the Bath to resolve, an  Offner will most likely never be needed; The same can be said for flats. I have optical flats and I can use a water test if I were ever to encounter one that's too big. So, I can safely conclude that an amateur is very unlikely to really need the complexity, the cost and the fuss of a T-G/William interferometer. 


  • SandyHouTex likes this

#57 BGRE

BGRE

    Gemini

  • *****
  • Posts: 3,005
  • Joined: 21 Mar 2016
  • Loc: New Zealand

Posted 16 April 2018 - 06:06 PM

The alignment requirements for all spherical wave interferometers are the same.

 

1) Center the test surface within the test beam (necessary even for a Bath to avoid test beam SA --> coma --> tilt dependent astigmatism). NB the reference surface NA should always be larger than that of the test surface. The test beam NA should also exceed that of the test surface.

 

2) Adjust  distance between interferometer and test surface and

 

3) Tilt of test surface

 

so that the test setup is autostigmatic (source and its reflected image are coincident)

 

Since both the test and reference beam return foci are physically accessible in a Williams interferometer alignment is in some ways easier than for a Bath interferometer.

 

In practice one would prealign the reference surface in a Williams interferometer and only the test surface would need alignment when testing.



#58 radicell2

radicell2

    Vostok 1

  • -----
  • Posts: 164
  • Joined: 15 Jan 2010
  • Loc: Toronto

Posted 17 April 2018 - 09:35 AM

I've never found it difficult to align either a Twyman-Green or a Fizeau.
Once the source images in the focal plane of the collimator (in the case of a Fizeau) or the focal plane of a lens in the case of a Twyman-Green overlap accurately, fringes are detectable once the viewing arm is reconfigured to image the test surface onto the detector, provided the OPD doesn't exceed the source coherence length.
Aligning a Williams interferometer should be just as easy.

I agree with Bruce.Did set up a LUPI  (as I mentioned on the IF forum years ago) using a smallish concave reference mirror and a red laser diode to test a 4 inch f/11 spherical mirror.Just barely got fringes that were ultra weak.One needs a green laser diode to work,as others have shown on the IF groups photo page and the older Yahoo groups page - check  - firstPDI .

 

Ric


Edited by radicell2, 18 April 2018 - 08:44 AM.


#59 MKV

MKV

    Cosmos

  • *****
  • Posts: 9,080
  • Joined: 20 Jan 2011
  • Loc: Southeast, USA

Posted 17 April 2018 - 04:08 PM

The alignment requirements for all spherical wave interferometers are the same.

 

1) Center the test surface within the test beam (necessary even for a Bath to avoid test beam SA --> coma --> tilt dependent astigmatism). NB the reference surface NA should always be larger than that of the test surface. The test beam NA should also exceed that of the test surface.

 

2) Adjust  distance between interferometer and test surface and

 

3) Tilt of test surface

 

so that the test setup is autostigmatic (source and its reflected image are coincident)

 

Since both the test and reference beam return foci are physically accessible in a Williams interferometer alignment is in some ways easier than for a Bath interferometer.

 

In practice one would prealign the reference surface in a Williams interferometer and only the test surface would need alignment when testing.

Thank you, Bruce. It goes without saying that each interferometer should be adjusted before attampting to test.  This means that for the T-G/Willams the reference beam should remain fixed. It should move only in one direction, to allow for focus.

 

For the Bath:

 

1) Point your Bath at the test mirror until the wide test beam is roughly centered on the mirror, center the reference beam on the mirror.

 

7879a.JPG

 

2) Pan the mirror on its stand until both beams appear on a screen placed roughly in the plane that bisects the beam splitter (the screen is placed there before the test setup  begins). The beam closest to the beam splitter is the test beam.

 

3) move (translate) the Bath and the screen together along the z axis (towards and away form the mirror) until the test beam appears in sharp focus.

 

DSC01551 (3).JPG

 

4) Pan the mirror in the opposite direction until the reference beam enters the diverging lens and roughly center it on the divegring lens (you can use a small piece of paper to do so but it's not necessary)

 

IMAG0684.jpg DSC01553 (3).JPG

 

5) The Bath is now aligned.

 

By replacing the alignment screen with a visual screen or a camera facing the beam splitter, one will observe two beams very close or on top of each other. Most often the fringes are visible. If not, a slight tilt brings them out.

______

 

As far as T-G/William is concerned. The best way to align in my experience has been to use the laser without a diverger. The test mirror is adjusted until the return beam and the reference beam coincide. This is often easier said than done.

 

In my  experience, the Bath is much, much easier to set up and align. It doesn't require optomechanical equipment, or high precision reference elements, GRIN lenses, and so on.



#60 Alan A.

Alan A.

    Viking 1

  • -----
  • topic starter
  • Posts: 571
  • Joined: 08 Feb 2007
  • Loc: San Francisco

Posted 17 April 2018 - 06:56 PM

Hi,

I believe I remember a previous post where someone had indicated the TG was better for testing a refractor than the Bath. Is this correct?

Best,

Alan

#61 BGRE

BGRE

    Gemini

  • *****
  • Posts: 3,005
  • Joined: 21 Mar 2016
  • Loc: New Zealand

Posted 17 April 2018 - 08:47 PM

Thank you, Bruce. It goes without saying that each interferometer should be adjusted before attampting to test.  This means that for the T-G/Willams the reference beam should remain fixed. It should move only in one direction, to allow for focus.

 


5) The Bath is now aligned.

 

By replacing the alignment screen with a visual screen or a camera facing the beam splitter, one will observe two beams very close or on top of each other. Most often the fringes are visible. If not, a slight tilt brings them out.

______

 

As far as T-G/William is concerned. The best way to align in my experience has been to use the laser without a diverger. The test mirror is adjusted until the return beam and the reference beam coincide. This is often easier said than done.

 

In my  experience, the Bath is much, much easier to set up and align. It doesn't require optomechanical equipment, or high precision reference elements, GRIN lenses, and so on.

You've made the Williams unnecessarily difficult to align by using a suboptimal alignment technique.

 

The first step is to block the test beam (with a suitable light trap eg velvet cloth or something more elaborate) center the reference surface in the reference beam and then adjust its tilt and focus so that the return beam forms an image of the source coincident with the source itself. Then block the reference beam and repeat for the test beam.

Once accurate autostigmatism is achieved for both beams fringes should appear if the source coherence length is great enough and both the source images are sufficiently close to each other. The separation between the test and reference return images for N fringes across the test surface is

s = N*(R/D)*Lambda.

where R is the ROC of the test surface and D its diameter.


Edited by BGRE, 17 April 2018 - 08:49 PM.


#62 hakann

hakann

    Vanguard

  • -----
  • Posts: 2,148
  • Joined: 06 Aug 2015

Posted 18 April 2018 - 03:30 AM

How much cost the system and the optical flat this book talk about ?
Is it a brand or home made unit.

—-

I know the Zygo stuff today can be at 1 m K + test room and fixtures, tables.
Pro normally do L8 or 10 nm RMS at the surface max in thoose, depending on optical flat tolerance, room enviroment, sag.

As ATM do and sell easy at home or in garage test and grinding to way over thoose tolerance ( works in the EP ) to very stiff wooden telescope design.
Why is IF really needed when much better test methods can be bought for 25 USD and are way better also.

As I got it, this become pricy vs some customer needs documents but in ATM this is done way better mirrors and can test them also better and more reliable.

Just my 2 cents.

#63 BGRE

BGRE

    Gemini

  • *****
  • Posts: 3,005
  • Joined: 21 Mar 2016
  • Loc: New Zealand

Posted 18 April 2018 - 03:45 AM

What $25 test do you have in mind??



#64 radicell2

radicell2

    Vostok 1

  • -----
  • Posts: 164
  • Joined: 15 Jan 2010
  • Loc: Toronto

Posted 18 April 2018 - 08:51 AM



5) The Bath is now aligned.

 

By replacing the alignment screen with a visual screen or a camera facing the beam splitter, one will observe two beams very close or on top of each other. Most often the fringes are visible. If not, a slight tilt brings them out.

______

 

A

You forgot to mention that one might see numerous pulsing "bulleyes" on the surface of the mirror.That's a good sign as one is now in the ballpark of getting fringes - just move in closer to the mirror.Small movements of all axis will get the correct fringes numbers

 

To answer a question of why use a IF - one sees the astig all at once on the surface.

 

Ric


Edited by radicell2, 18 April 2018 - 08:52 AM.


#65 MKV

MKV

    Cosmos

  • *****
  • Posts: 9,080
  • Joined: 20 Jan 2011
  • Loc: Southeast, USA

Posted 18 April 2018 - 09:57 AM

You've made the Williams unnecessarily difficult to align by using a suboptimal alignment technique.

 

The first step is to block the test beam (with a suitable light trap eg velvet cloth or something more elaborate) center the reference surface in the reference beam and then adjust its tilt and focus so that the return beam forms an image of the source coincident with the source itself. Then block the reference beam and repeat for the test beam.

Once accurate autostigmatism is achieved for both beams fringes should appear if the source coherence length is great enough and both the source images are sufficiently close to each other. The separation between the test and reference return images for N fringes across the test surface is

s = N*(R/D)*Lambda.

where R is the ROC of the test surface and D its diameter.

I don't know what's "suboptimal" about it, Bruce.

 

Here is a simple schematic of the TG/Williams alignment as I do it. It's still a lot more involved and "fussy" then aligning the Bath.

 

TG_alignment.jpg

 

Also, this usually results in some ghosting of the light source, as seen here.

 

_SAM2453DR.JPG

 

To avoid this, a larger beam splitter is used with the light source (which should really be as close to an airy disc as possibe) set closer to the beam splitter, and with the beam splitter slightly rotated relative to the optical axis.

 

beam split offset.jpg


Edited by MKV, 18 April 2018 - 01:17 PM.


#66 MKV

MKV

    Cosmos

  • *****
  • Posts: 9,080
  • Joined: 20 Jan 2011
  • Loc: Southeast, USA

Posted 18 April 2018 - 10:02 AM

I believe I remember a previous post where someone had indicated the TG was better for testing a refractor than the Bath. Is this correct?

Did they explain why?



#67 MKV

MKV

    Cosmos

  • *****
  • Posts: 9,080
  • Joined: 20 Jan 2011
  • Loc: Southeast, USA

Posted 18 April 2018 - 10:17 AM

You forgot to mention that one might see numerous pulsing "bulleyes" on the surface of the mirror.That's a good sign as one is now in the ballpark of getting fringes - just move in closer to the mirror.Small movements of all axis will get the correct fringes numbers

I didn't forget. I was describing how to align the Bath. Adjusting the fringes is not part of the alignment procedure. At any rate, if you dim the laser beam you can focus the test beam much better on the screen as shown in Fig 3 in #59. Once your Bath beams are aligned you're always in the ballpark and it's only a mater of slight refocusing that will bring out the fringes in case you don't see any. 



#68 Alan A.

Alan A.

    Viking 1

  • -----
  • topic starter
  • Posts: 571
  • Joined: 08 Feb 2007
  • Loc: San Francisco

Posted 18 April 2018 - 11:45 AM

Did they explain why?


Hi,

A specific explanation was not given.

So a Bath, TG, and Williams can all be used effectively to test refractors?

Best,

Alan

#69 MKV

MKV

    Cosmos

  • *****
  • Posts: 9,080
  • Joined: 20 Jan 2011
  • Loc: Southeast, USA

Posted 18 April 2018 - 01:05 PM

So a Bath, TG, and Williams can all be used effectively to test refractors?

Yes, provided you have an autocollimaiton flat.



#70 Arjan

Arjan

    Apollo

  • -----
  • Posts: 1,264
  • Joined: 21 Jan 2009
  • Loc: Netherlands

Posted 19 April 2018 - 08:48 AM

5) The Bath is now aligned.

I only align once, then I mark where it is on the bench and put it exactly there for each subsequent session. Mirror stand doesn't move between sessions.

After a few seconds of tweaking the XYZ controls I get good fringes and start shooting.

 

This way, aligning really isn't substantially adding to the total time spent taking igrams...



#71 MKV

MKV

    Cosmos

  • *****
  • Posts: 9,080
  • Joined: 20 Jan 2011
  • Loc: Southeast, USA

Posted 19 April 2018 - 08:50 AM

Another thing that needs to be mentioned is that with the T-G/Williams interfeormeter, everything has to be bolted down, and backlash-free. Otherwise it is really trying to get both beams to overalp at the focus. In comparison, the Bath is a breeze. I believe the author, Wlliam Zmek shortchanged himself by not actually building and using a Bath before his critical review of both interferometers. 



#72 MKV

MKV

    Cosmos

  • *****
  • Posts: 9,080
  • Joined: 20 Jan 2011
  • Loc: Southeast, USA

Posted 19 April 2018 - 08:55 AM

I only align once, then I mark where it is on the bench and put it exactly there for each subsequent session. Mirror stand doesn't move between sessions.

After a few seconds of tweaking the XYZ controls I get good fringes and start shooting.

 

This way, aligning really isn't substantially adding to the total time spent taking igrams...

Yes, that's the way to do it if you stay with the same equipment. However, I like to have the freedom to change the size of the beam splitter, the diverger and the laser diode. If I test lenses, I like the option of using three different lasers wavelengths, and for autocollimation setups it helps to use the smallest beam splitter possible, etc.



#73 BGRE

BGRE

    Gemini

  • *****
  • Posts: 3,005
  • Joined: 21 Mar 2016
  • Loc: New Zealand

Posted 26 April 2018 - 11:13 PM

Dust diffraction and other interferometer artifacts distort interferograms. The classical cure is to use an extended source that in effect averages out these distortions.

 

With a Bath using a real extended source other than a thin ring source can be problematic.

A Fizeau can only directly use a thin ring source or the fringe contrast suffers with long cavities.

A Twyman - Green used to measure a concave surface allows the reference surface to be adjusted to be conjugate to the test surface so that a large disk source can be used without loss of fringe contrast even for test surface ROCs of several meters as long as the source coherence length is great enough.

 

The general solution for virtually all interferometers is to use a synthesised extended source where the computed surface is averaged over a large number of source positions (corresponds to reference spot positions in the case of a Bath interferometer).




CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics