Date   

Re: Tilt dependent astigmatism

Franz
 

Thanks Jan and Bruce!


Re: Tilt dependent astigmatism

Bruce Griffiths
 

The test surface also has field astigmatism which is tilt dependent.

However this is usually insignificant except when  testing large test surfaces.

The field astigmatism of the test surface can easily be calculated from the beam separation (at the interferometer head), the test surface diameter and RoC. 

Similarly the field coma and astigmatism of the diverger lens can be found by raytracing.

Bruce

On 02 June 2021 at 03:58 janvangastel <jhm.vangastel@...> wrote:

If I remember correctly, the astig-tilt relationship is caused by the fact that the Bath is not *exactly* a common path interferometer, because depending on the amount of tilt the beam returning from the mirror will not go through exactly the same spot on the small lens as it went in opposite direction (going from Bath to mirror). Depening on characteristics of the actual lens this may cause astigmatism, or cancel it (partly) out. Therefore it is adviced to make - in each rotational position used - igrams with different amounts of tilt to average this astigmatism out as much as possible. 

Jan.


Op 1-6-2021 om 17:33 schreef franz.hagemann@...:
I bought  a new Laser module and did the same experiment. I got a very similar result.
In these first two experiments the laser was focused on the mirror. I then did a third experiment with the laser beam collimated / parallel (and only Y tilt). Again a very similar result.
To me it seems unlikely that both laser modules have the same coma, but it is possible.
Any thoughts?
Below is a copy of my spreadsheet with the results.

Roithner Laser: focused on mirror   
      
x Tilty TiltTiltAstig  
12,71116,800,055  
21,91928,990,143  
23,222,332,180,143  
2929,0541,050,21  
      
Picotronic Laser: focused on mirror   
      
x Tilty TiltTiltX AstigY AstigAstig
9,85,811,390,0250,090,09
14,410,217,650,0170,0880,09
23,315,327,8700,1380,14
36,621,742,550,0970,180,20
      
Picotronic Laser: parallel beam   
      
x Tilty TiltTiltX AstigY AstigAstig
012,712,70,020,0780,08
017170,010,1470,15
025,525,50,0290,1320,14
037,137,10,0680,1890,20
045,445,40,0640,2310,24


 


Re: Tilt dependent astigmatism

 

If I remember correctly, the astig-tilt relationship is caused by the fact that the Bath is not *exactly* a common path interferometer, because depending on the amount of tilt the beam returning from the mirror will not go through exactly the same spot on the small lens as it went in opposite direction (going from Bath to mirror). Depening on characteristics of the actual lens this may cause astigmatism, or cancel it (partly) out. Therefore it is adviced to make - in each rotational position used - igrams with different amounts of tilt to average this astigmatism out as much as possible. 

Jan.


Op 1-6-2021 om 17:33 schreef franz.hagemann@...:

I bought  a new Laser module and did the same experiment. I got a very similar result.
In these first two experiments the laser was focused on the mirror. I then did a third experiment with the laser beam collimated / parallel (and only Y tilt). Again a very similar result.
To me it seems unlikely that both laser modules have the same coma, but it is possible.
Any thoughts?
Below is a copy of my spreadsheet with the results.

Roithner Laser:   focused on mirror      
           
x Tilt y Tilt Tilt Astig    
12,7 11 16,80 0,055    
21,9 19 28,99 0,143    
23,2 22,3 32,18 0,143    
29 29,05 41,05 0,21    
           
Picotronic Laser:   focused on mirror      
           
x Tilt y Tilt Tilt X Astig Y Astig Astig
9,8 5,8 11,39 0,025 0,09 0,09
14,4 10,2 17,65 0,017 0,088 0,09
23,3 15,3 27,87 0 0,138 0,14
36,6 21,7 42,55 0,097 0,18 0,20
           
Picotronic Laser:   parallel beam      
           
x Tilt y Tilt Tilt X Astig Y Astig Astig
0 12,7 12,7 0,02 0,078 0,08
0 17 17 0,01 0,147 0,15
0 25,5 25,5 0,029 0,132 0,14
0 37,1 37,1 0,068 0,189 0,20
0 45,4 45,4 0,064 0,231 0,24
_._,_._,_



Re: Tilt dependent astigmatism

Franz
 

I bought  a new Laser module and did the same experiment. I got a very similar result.
In these first two experiments the laser was focused on the mirror. I then did a third experiment with the laser beam collimated / parallel (and only Y tilt). Again a very similar result.
To me it seems unlikely that both laser modules have the same coma, but it is possible.
Any thoughts?
Below is a copy of my spreadsheet with the results.

Roithner Laser:   focused on mirror      
           
x Tilt y Tilt Tilt Astig    
12,7 11 16,80 0,055    
21,9 19 28,99 0,143    
23,2 22,3 32,18 0,143    
29 29,05 41,05 0,21    
           
Picotronic Laser:   focused on mirror      
           
x Tilt y Tilt Tilt X Astig Y Astig Astig
9,8 5,8 11,39 0,025 0,09 0,09
14,4 10,2 17,65 0,017 0,088 0,09
23,3 15,3 27,87 0 0,138 0,14
36,6 21,7 42,55 0,097 0,18 0,20
           
Picotronic Laser:   parallel beam      
           
x Tilt y Tilt Tilt X Astig Y Astig Astig
0 12,7 12,7 0,02 0,078 0,08
0 17 17 0,01 0,147 0,15
0 25,5 25,5 0,029 0,132 0,14
0 37,1 37,1 0,068 0,189 0,20
0 45,4 45,4 0,064 0,231 0,24


locked Re: Bug in DFTFringe 5.0

Dale Eason
 

Bug fixes and enhancements to the 3D display.


locked Re: Bug in DFTFringe 5.0

vladimir galogaza
 

What else can we expect in the DFTFringe 5.1 ?

Vladimir

On Sun, 30 May 2021, 19:21 Dale Eason, <doeason@...> wrote:
I discovered that if you change the mirror diameter the profile plot will not display the profile correctly until dftfringe is restarted.  I have a fix for the next release coming soon.


--
Vladimir


locked Bug in DFTFringe 5.0

Dale Eason
 

I discovered that if you change the mirror diameter the profile plot will not display the profile correctly until dftfringe is restarted.  I have a fix for the next release coming soon.


Re: need advice: Parabolic primary -- d = 202mm RoC = 1562mm

Joe Babendreier
 

Extremely grateful for your analysis. Thanks,

Joe


Re: need advice: Parabolic primary -- d = 202mm RoC = 1562mm

Dale Eason
 

Here it is in DFTFringe64 next release and 5%.  Yes, I'm about to release 5.1.


Re: need advice: Parabolic primary -- d = 202mm RoC = 1562mm

Dale Eason
 

I think even 5% is  more accurate for your case.  That is because using your .wft file I can adjust the Foucault image so I can see that weave pattern that shows up when many igrams with different fringe orientation are taken.   Meaning that you have averaged out most of the igram noise and are left with remnants of fringe print through.   Also because your igram looks very good with little noise.   Thus you can reduce the blur radius.       So because the surface is not as smooth as a very good mirror and with a high edge the higher blur values are artificially smoothing the surface too much at even 10%.

Your red colored ronchi I think is an actual ronchi and it too shows the surface defects well.  I think at 5% DFTFringe simulated ronchi matches your actual better.  You can see a little more noise in the 5% but it captures the edge better than the 20%.

Dale


Re: need advice: Parabolic primary -- d = 202mm RoC = 1562mm

Joe Babendreier
 

Pictures:


need advice: Parabolic primary -- d = 202mm RoC = 1562mm

Joe Babendreier
 

I've been figuring a Pyrex 8" primary for a few months now. I use DFTFringe for testing. (Plus Foucault  & Rochigrams)

My question is about the proper setting for the Gaussian blur when trying to determine the wavefront.

In my case, should I set it to 10% or 20% (or some other number) ???

I attach some pictures, including an R-gram (outside of R0C) for the mirror as it stands now.

https://www.dropbox.com/s/ss5d9jiege3mlwn/BABO%20AG%20202-1564-10pc%20385%20s883.wft?dl=0

Joe


Re: Tilt dependent astigmatism

Franz
 

Thanks Michael!
That seems very likely as my laser beam moves in a circle when I turn the fucusing element.
I will try to improve the laser and report back.


Re: Tilt dependent astigmatism

Michael Peck
 

On 5/26/2021 7:47 AM, franz.hagemann@... wrote:

It seems like astigmatism is proportional to tilt. And it is substantial.

Does anyone know if this is normal?
It makes measuring astig really hard because it only cancels out for equal tilts.
My beam separation is 6.5 mm. DCX diverger with 9 mm focal lenght.

Cheers,
Franz

Tilt dependent astigmatism isn't exactly normal but it has been observed and discussed here a number of times. It's likely due to coma in the collimating lens of your laser, which in turn is likely due to decentering.

-- 
Michael Peck
http://wildlife-pix.com
mlpeck54 -at- earthlink.net


Tilt dependent astigmatism

Franz
 
Edited

Hey folks,
for some time now I had the impression, that more fringes (or more tilt) result in more astigmatism in the wavefront.
So I made a small experiment.
The mirror is an unfinished 6 inch f 2.8.
I made some interferograms with different amounts of tilt, without touching the mirror in between.
I got this result:
x Tilt      y Tilt     Astig (polar)
12.7         11           0.055
21.9         19           0.143 
23.2        22.3         0.143
29.0        29.0         0.21

It seems like astigmatism is proportional to tilt. And it is substantial.

Does anyone know if this is normal?
It makes measuring astig really hard because it only cancels out for equal tilts.
My beam separation is 6.5 mm. DCX diverger with 9 mm focal lenght.

Cheers,
Franz


Re: Updated Wiki Page: Diverger lens Residual Spherical aberration #wiki-notice

Bruce Griffiths
 

With an equibiconvex lens the fact that the test surface brings the collimated reference beam to a focus one test surface focal length in front of the test surface whereas the test beam from the laser is collimated results in a small mismatch in SA between the test and reference beams. As long as the test surface focal length is much greater than that of the diverger then this SA mismatch is typically negligibly small.

If a PCX lens is used it contributes an SA mismatch between test and reference beams.

The formula allows this SA mismatch to be estimated.

If the refractive index is unknown using a value of 1.5 allows a ballpark estimate of the SA mismatch so one can see if the resultant SA mismatch can be neglected when testing  a particular test surface when using a PCX diverger lens (a PCX lens may be all that's immediately available). If the resultant SA mismatch is significant it can be easily measured if the diverger lens can be reversed allowing measurements to be taken with both orientations of the diverger lens.

If for example, one were testing a slow test surface a long efl diverger lens that better matches (allows sorter exposures by using more of the laser beam) the test beam to the test surface may only be available as a PCX lens. In which case an estimate of the resultant SA mismatch can be very useful.

Bruce

On 13 May 2021 at 02:24 "George Roberts (Boston)" <bb@...> wrote:

Lol.  Basically in a Bath, there if you use a biconvex lens there is a tiny contribution of that lens to the spherical aberration term.  Typically it may add less than 1/1000 of a wave of error to the DFTFringe result.  No big deal.  It depends on things like how big and fast your mirror is and what type of glass is in the thin lens and so on.  Bruce has an equation that lets you calculate the error.  I might make a web page calculator where you enter the terms but the error is so small it's probably not worth the trouble except...

If you use a PCX lens instead of a BCX lens (PCX - plano convex lens - that means one side of the diverger lens is flat) then this spherical aberration grows.  It's still typically less than 1/50th of a wave but it could be significant if you have an unusual mirror.  I have a calculator here to see how much PCX will affect your mirror here - just plug in your mirror parameters:
gr5.org/bath/

However Bruce improved the equation to include the index of refraction of your PCX lens which most people probably won't know.  So maybe best to use the older version which just assumes 1.5 which is hopefully close enough.


Re: Updated Wiki Page: Diverger lens Residual Spherical aberration #wiki-notice

George Roberts (Boston)
 

Lol.  Basically in a Bath, there if you use a biconvex lens there is a tiny contribution of that lens to the spherical aberration term.  Typically it may add less than 1/1000 of a wave of error to the DFTFringe result.  No big deal.  It depends on things like how big and fast your mirror is and what type of glass is in the thin lens and so on.  Bruce has an equation that lets you calculate the error.  I might make a web page calculator where you enter the terms but the error is so small it's probably not worth the trouble except...

If you use a PCX lens instead of a BCX lens (PCX - plano convex lens - that means one side of the diverger lens is flat) then this spherical aberration grows.  It's still typically less than 1/50th of a wave but it could be significant if you have an unusual mirror.  I have a calculator here to see how much PCX will affect your mirror here - just plug in your mirror parameters:
gr5.org/bath/

However Bruce improved the equation to include the index of refraction of your PCX lens which most people probably won't know.  So maybe best to use the older version which just assumes 1.5 which is hopefully close enough.


Re: Fourier transforms of Zernike polynomials

Bruce Griffiths
 

I foolishly assumed it would be easy to recreate and didn't note them down as I was merely establishing that the program ran at the time.

I'll see if I can recreate the issue but it may take several days.

Bruce

On 12 May 2021 at 11:38 Dale Eason <doeason@...> wrote:

Bruce,
What where the parameters of the mirror and size of the simulated wavefront?


Re: Fourier transforms of Zernike polynomials

Dale Eason
 

Bruce,
What where the parameters of the mirror and size of the simulated wavefront?


Re: Fourier transforms of Zernike polynomials

Bruce Griffiths
 

Yes, I was using the 64 bit version.

I first created a synthetic interferogram with only a defocus term and analysed it using the default size Gaussian blur and noticed the artifact near the center. Reducing the Gaussian blur size made it smaller, vanishing for a blur size of zero.


Bruce

On 12 May 2021 at 08:13 Dale Eason <doeason@...> wrote:

On Tue, May 11, 2021 at 02:53 PM, Bruce Griffiths wrote:
I had noticed that using a Gaussian blur adds a non existent hollow to the centre of the computed surface when analysing an interferogram that only  has defocus.
Bruce, Where you using DFTFringe for that experiment?   It does not sound logical too me that that should happen.   So I tried it myself and at least for the test case I chose it was not the case.   For a nulled wavefront with a defocus of 1 the 20% blur turned the edge as I expected but did nothing to the middle that I could see on a 600 x 600 surface.  MIrrors was 95mm diameter F4.6 with expected conic of -1 and a perfect wavefront with a defocus of 1.   Defocus enabled in analysis. 

Dale

1061 - 1080 of 31808