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

600 mm f/3.3 Meniscus

This topic has been archived. This means that you cannot reply to this topic.
36 replies to this topic

#1 Stathis_Firstlight

Stathis_Firstlight

    Stathis Firstlight

  • *****
  • Vendors
  • topic starter
  • Posts: 74
  • Joined: 28 Dec 2019

Posted 10 April 2020 - 04:53 AM

All over sudden I have time, to continue my project, that I started some time ago.

It is a 600 mm f/3.3 mirror at 21 mm overall thickness (meniscus shape). The concave frontside and convex backside I generated with grinder machine on a jig out of a 600x34,5 mm Supremax blank:

 

flexomat1_k.jpg

 

Out of the 21.5 kg are now 13.3 kg left, 8 kg borosilicate glass tranformed into dust spread all overshocked.gif

 

Hey, all of us need to do some indoor training now. Instead of running on a treadmill or cycling on a bike Ergometer, see the alternative workout:

 

Video of grinding mirror backside on mirror frontside with grit K180 (11MB):

http://www.stathis-f...s/60cm_k180.mp4

 

Video grinding with a 40 cm granite tool on mirror frontside with grit K180 (10MB):

http://www.stathis-f...k180_granit.mp4

 

Some more pictures here:

http://www.stathis-f...f3.3_mirror.htm


Edited by Stathis_Firstlight, 10 April 2020 - 04:54 AM.


#2 Arjan

Arjan

    Surveyor 1

  • -----
  • Posts: 1,800
  • Joined: 21 Jan 2009

Posted 10 April 2020 - 09:20 AM

Stathis, nice challenge!
I hope you did not inhale any of the 8kg of borosilicate dust?

#3 Stathis_Firstlight

Stathis_Firstlight

    Stathis Firstlight

  • *****
  • Vendors
  • topic starter
  • Posts: 74
  • Joined: 28 Dec 2019

Posted 11 April 2020 - 07:04 AM

I hope you did not inhale any of the 8kg of borosilicate dust?

Some borosilicate molecules in the blood vessels should be a good nutritional supplement for a mirror makerlol.gif 

ok, I  used a protective mask, ear protection and a strong fan, that blowed the dust away from the body:

 

flexomat4_k.jpg

 

Here two videos showing the front an back generating: 

http://www.stathis-f...xomat_front.mp4

http://www.stathis-f...exomat_back.mp4

 

I just put togehter the English version:

http://www.stathis-f...3_mirroreng.htm



#4 eroyer

eroyer

    Vostok 1

  • -----
  • Posts: 114
  • Joined: 30 Aug 2014

Posted 12 April 2020 - 12:42 PM

Hello Stathis,

you seem very happy while grinding the mirrors. I hope you'll keep this smile until the end.

I have finished fine grinding the 2 blanks you sent me for my binoscope last year. The first is polished and I'm ready for parabolizing.

 

Eric



#5 Stathis_Firstlight

Stathis_Firstlight

    Stathis Firstlight

  • *****
  • Vendors
  • topic starter
  • Posts: 74
  • Joined: 28 Dec 2019

Posted 22 April 2020 - 02:47 AM

Hello Eric, nice to hear, that your ambitious project makes progress. Did you introduce it somewhere?

 

In the meantime I procceded to Microgrit 25 my, meditative smooth sound now.

See two new videos:

 http://www.stathis-f...inding_k320.mp4

 http://www.stathis-f...inding_25my.mp4

 

I found this interesting discussion about meniscus mirror stiffness from 2018:

https://www.cloudyni...t-back-mirrors/

where was sayed:

 

 "At constant weight, the stiffness of a thin meniscus mirror is smaller than that of a thin, flat-back mirror of the same size"
-Opto-Mechanical Systems Design, Fourth Edition, Volume 2 Pg 81
Paul Yoder, ‎Daniel Vukobratovich

This confuses me. I would expect, that a meniscus mirror is stiffer than one with a flat back of the same weight. Hey this is the reason, why I am doing this.

 

Approximating the vertical case (mirror looks to the zenith)  a 600x21 mm meniscus with PLOP automated cell design with: 

Diameter: 600 mm 

Edge thickness: 21 mm

Focal lenth: -400,000 mm (to get a sagitta of allmost 0) 

Central obstruction: 0

Cell: 18 point, allow angles to vary

Result: RMS = 4 nm which corresponds to ~1% Strehl loss

The mass is 13.2 kg

 

A mirror with a flat back of the same mass would have an edge thickness of 26.6 mm and would have a deflection of RMS= 6.7 nm, so 67% worse.

 

To get the same deflection of RMS=4 nm with flat back I have to chosse 32.8 mm edge thickness, this would have a mass of 17 kg, so 29% higher weight.

 

Is this way of using plop right? If so, why the above mentioned literature comes to a different result? What means "self-weight deflection" in this content?

 

How can we approximate the not vertical position? For example, how does a meniscus behave looking at 45° elevation? Here the lower support feet will apply some shear forces and thus some bending momentum to the glass. How to calculate or at least estimate this? 

 

How does a meniscus behave in the mirror test stand at 0° (looking to the horizon)? Here I would assume, that it bends more than one with a flat back, but hope, that the disadvantage is not too serious.


Edited by Stathis_Firstlight, 22 April 2020 - 02:57 AM.


#6 Benach

Benach

    Mercury-Atlas

  • *****
  • Posts: 2,643
  • Joined: 24 Jan 2008

Posted 22 April 2020 - 04:39 AM

Statthis: it requires a full FEA program to model a meniscus mirror. Deflection at different angles than 90° is impossible afaik in PLOP. It is fairly straightforward optomechanics though. The problem of using a flat mirror as an approach is that you make the wrong assumption that a meniscus shape and a flat have the same stiffness because the material and thicknesses are equal. This is wrong because stiffness also has a geometrical component, the bending moment of inertia and these are not equal.

Self-weight deflection means that one does not only minimizes the reaction forces in the support pivots, what PLOP does, but also the weight of the mirror and actually you should also take the stiffness of the support structure into account, PLOP does neither of the latter two.

For the DAST, Dutch Amateur Solar Telescope, I once had to make the FEA calculations about 6 years ago, for the 400mm f/3 meniscus shaped 30mm thick mirror. The results of PLOP quickly became nonsensical at best because of the reasons given above. But let me share the support structure as an inspiration:
http://www.zonnetele...dersteuning.jpg

http://www.zonnetele...ning-detail.jpg

#7 eroyer

eroyer

    Vostok 1

  • -----
  • Posts: 114
  • Joined: 30 Aug 2014

Posted 22 April 2020 - 05:22 AM

Hello Eric, nice to hear, that your ambitious project makes progress. Did you introduce it somewhere?

I've put some pictures here (the text is in french, sorry)

http://www.astrosurf...124758-bino600/

 

As for the deformation of meniscus mirrors, there is a study here by Pierre Strock (again in french) :

http://strock.pi.r2....t/Menisque.html

 



#8 Bob4BVM

Bob4BVM

    Fly Me to the Moon

  • *****
  • Posts: 5,124
  • Joined: 23 Mar 2015

Posted 22 April 2020 - 12:10 PM

Hello Stathis,

you seem very happy while grinding the mirrors. I hope you'll keep this smile until the end.

I have finished fine grinding the 2 blanks you sent me for my binoscope last year. The first is polished and I'm ready for parabolizing.

 

Eric

OK Eric, Don't keep me in suspense !   What is the new b-scope project ? :)

CS

Bob



#9 eroyer

eroyer

    Vostok 1

  • -----
  • Posts: 114
  • Joined: 30 Aug 2014

Posted 22 April 2020 - 03:30 PM

OK Eric, Don't keep me in suspense !   What is the new b-scope project ? smile.gif

CS

Bob

I started a 600m (24 inch) binoscope project. So far, I've finished polishing and figuring the secondary mirrors. I am curently polishing the primaries. So there's still a long way to go before first light.

 



#10 Bob4BVM

Bob4BVM

    Fly Me to the Moon

  • *****
  • Posts: 5,124
  • Joined: 23 Mar 2015

Posted 22 April 2020 - 07:29 PM

I started a 600m (24 inch) binoscope project. So far, I've finished polishing and figuring the secondary mirrors. I am curently polishing the primaries. So there's still a long way to go before first light.

Nice, a true monster in the making, can't wait to see it come to life !

You should make it a go-to / ride-aboard like that one in ^Japan (* I think)

Cheers

Bob 

PS -sorry for the OT, didn't mean to derail the OP's thread.



#11 Pierre Lemay

Pierre Lemay

    Surveyor 1

  • *****
  • Posts: 1,831
  • Joined: 30 Jan 2008

Posted 22 April 2020 - 10:13 PM

For example, how does a meniscus behave looking at 45° elevation? Here the lower support feet will apply some shear forces and thus some bending momentum to the glass. How to calculate or at least estimate this? 

 

How does a meniscus behave in the mirror test stand at 0° (looking to the horizon)? Here I would assume, that it bends more than one with a flat back, but hope, that the disadvantage is not too serious.

Have you read some of what Mel Bartels has been doing with thin, meniscus mirrors in the last few years? I had the opportunity to observe during several nights with his 25 inch f/2.6 last year at OSP. I can tell you, based on real observations, that the 15mm thick meniscus works fine at 45 deg elevation and all the other ones as well. The edge of the mirror is held on only two pegs and the back is supported on only a 9 point floatation cell but, the contact pads on the back of the glass are actually large rings, not small pads. The back of the slumped, plate glass mirror was ground to make sure the ultra thin blank would rest smoothly and regularly on the rings. There are no signs of astigmatism caused by the mirror support. 

 

However one thing Mel noticed about thin meniscus mirrors is that forced air cooling the back center of the mirror is NOT a good idea. When he tried this, the mirror showed undercorrection, the reverse of what one normally sees in thicker, regular mirrors when they are cooling down. When he stopped the fan, the mirror quickly went back to showing perfect star shapes. Keep this in mind when you are making the mirror box. 



#12 Stathis_Firstlight

Stathis_Firstlight

    Stathis Firstlight

  • *****
  • Vendors
  • topic starter
  • Posts: 74
  • Joined: 28 Dec 2019

Posted 23 April 2020 - 02:04 AM

Hello Eric, I had one year French in school looong time ago. But both of your links are so interesting, that I spend yesterday evening reading and trying to understand them for several hours. This is the cick with amateur astronomy and ATM. We are learning for life! In a few days I may even dream couramment le français grin.gif 

 

Thank you Pierre for your comment. Of course I know Mels Website about meniscus mirrors, I have read it several times.

 

But still there are open questions. Mel indicates, that a meniscus is stiffer, than an a flat disk of same thickness. But as I understand from Pierre Strocks FEM results, the arch effect is not pronounced enough even at f/3, to provide a substantial improvement. Similar was stated by Kai (Fraxinus) in the German Forum. He refers the the "Shallow Shell Theory". So, we can assume it bends like a flat disk and can calculate with PLOP putting a very long focal length. Right? Mel, are you reading here?

 

This would mean, that my PLOP calucations above are correct, right?

But then I still don't understand the statement in the book "Opto-Mechanical Systems Design" quoted obove, that a meniscus is weaker than a flat.

 

Who can solve my contradiction?

Who can FEM calculate the bending of a 600x21 mm f/3.3 meniscus on a floating 18- point mirror cell and compare with PLOP?   



#13 Stathis_Firstlight

Stathis_Firstlight

    Stathis Firstlight

  • *****
  • Vendors
  • topic starter
  • Posts: 74
  • Joined: 28 Dec 2019

Posted 23 April 2020 - 02:14 AM

Self-weight deflection means that one does not only minimizes the reaction forces in the support pivots, what PLOP does, but also the weight of the mirror and actually you should also take the stiffness of the support structure into account, PLOP does neither of the latter two.

Let's assume, that the cell has infinity stiffness. But even a real cell structure with free floating points will deforrm in zenith position enenly, so it will not change the Forces and thus not change the bending. 

 

Then PLOP does exactly this: Calculate the bending due to the own weight of the mirror. What am I missing here?



#14 Benach

Benach

    Mercury-Atlas

  • *****
  • Posts: 2,643
  • Joined: 24 Jan 2008

Posted 23 April 2020 - 03:19 AM

Statthis: if you give me a few days, I can try if my FEA program still works. However, it took me and my other professional optomechanical friend about a week to fully optimize the mirror and the support structure. It was fun, it took a huge amount of time, but I vowed to myself that I will not repeat this excersize again for free for third parties because it is too much of a hurdle. So I can check simple things, such as a PLOP vs. FEA analysis, but I am not going to design a full support structure, simply too much time.

I can guide you through our design process though.

What you are missing are three things:
1) PLOP only minimizes reaction forces. It does not take the weight of the mirror into account afaik. This arch effect is most definitely there at f/3. Normal deflections are in the order of microns or below, your sagitta alone is in the order of several millimeters.

2) Secondly, the support structure only deflects evenly if and only if the reaction forces per pivot are equal. However, in the case I mentioned above this was totally wrong. By heart, I do not have the data anymore, the reaction forces of the inner three pivots was about 2.5x lower than in the outer pivots.

3) Moreover, because the assumption of infinite stiffness is wrong, it will bend. To make it worse: In the commonly used isosceles triangles with the support in the middle you furthermore have the problem that the internal forces are not zero in the middle of the triangle. This would create a deflection wrt the hingle, purely because of the internal forces in the support structure. We deliberately added a fancy looking counterweight to reduce this effect.



#15 Pierre Lemay

Pierre Lemay

    Surveyor 1

  • *****
  • Posts: 1,831
  • Joined: 30 Jan 2008

Posted 23 April 2020 - 06:36 AM

But still there are open questions. Mel indicates, that a meniscus is stiffer, than an a flat disk of same thickness. But as I understand from Pierre Strocks FEM results, the arch effect is not pronounced enough even at f/3, to provide a substantial improvement. Similar was stated by Kai (Fraxinus) in the German Forum. He refers the the "Shallow Shell Theory". So, we can assume it bends like a flat disk and can calculate with PLOP putting a very long focal length. Right? Mel, are you reading here?

 

This would mean, that my PLOP calucations above are correct, right?

But then I still don't understand the statement in the book "Opto-Mechanical Systems Design" quoted obove, that a meniscus is weaker than a flat.

 

Who can solve my contradiction?

Who can FEM calculate the bending of a 600x21 mm f/3.3 meniscus on a floating 18- point mirror cell and compare with PLOP?   

Stathis,

With all due respect, you are thinking too much lol.gif . If someone has the time, the tools and the knowhow they can do a FEM of how a thin meniscus mirror behaves at different inclinations, but the reality is this: IT WORKS! All of us who have observed with Mel's 25 inch can confirm it. The stars are pinpoint from edge to edge, wherever the tube is pointed. In addition, Mel's 25 inch is thinner than the one you are making so your's should be even more robust.

 

So don't worry about it. As they say: "think less, grind more".

 

Looking forward to reading your adventures since I am 6 months away from starting the grinding on my own 28 inch f/2.8 meniscus:

 

28 inch slumped meniscus.jpg



#16 rik ter horst

rik ter horst

    Messenger

  • *****
  • Posts: 466
  • Joined: 01 Nov 2010

Posted 23 April 2020 - 07:51 AM

Stathis, theory versus practice...... Always interesting! A couple of years ago I have made a 16 inch F/3 telescope with a meniscus mirror of only 18 mm thickness. It is mounted on a cell with an 18-point support. In practice I never have seen astigmatism caused by the thin meniscus shape. Don't worry too much Stathis, I agree with Pierre Lemay... In fact, a friend of mine, Geert Kwast, and I have made a 24 inch F/3 meniscus mirror, which is just finished, with a thickness of 17mm! And guess what? It works too!! Enjoy this project, and knowing you a little from a disance, I'm sure you'll succeed!  Have fun and good luck!



#17 Benach

Benach

    Mercury-Atlas

  • *****
  • Posts: 2,643
  • Joined: 24 Jan 2008

Posted 23 April 2020 - 08:16 AM

Rik: where have you aluminized these mirrors? TNO?



#18 rik ter horst

rik ter horst

    Messenger

  • *****
  • Posts: 466
  • Joined: 01 Nov 2010

Posted 23 April 2020 - 08:41 AM

Benach, the 24 inch was done by Hamburg University, Germany. The 16 inch by Bart Postema, my former colleague ...



#19 Stathis_Firstlight

Stathis_Firstlight

    Stathis Firstlight

  • *****
  • Vendors
  • topic starter
  • Posts: 74
  • Joined: 28 Dec 2019

Posted 24 April 2020 - 01:54 AM

Stathis,

With all due respect, you are thinking too much lol.gif .

He, he, this is what I say to our novices in our mirror making group in Munich.

What a nice piece of meniscus! Looking forward to follow, how it develops.

 

Hello Rick, nice to hear from you. Greetings to Geert Kwast as well. Is the mechanics of that 24 inch x17 mm f/3 made as well already? What kind of mirror cell did you or Geert made?

 

Don´t get me wrong. I am not worried about the behavior in practical mirror making and use. I have made too many mirrors and have acompanied so many other large thin projects, so I am quite confident, to succeed with this one. Just wanted to understand the theory behind it a bit better. Benach, I didn't know, that it is such a big effort to run a FEM analysis on that. Don't spent too much much time on it. It's just for satisfiying curiosity.  

 

Ok, enough theory, just start with 15 my today.


Edited by Stathis_Firstlight, 24 April 2020 - 01:56 AM.


#20 Benach

Benach

    Mercury-Atlas

  • *****
  • Posts: 2,643
  • Joined: 24 Jan 2008

Posted 24 April 2020 - 03:46 AM

Statthis: been a professional FEA engineer for three years. Next time you see a VDL Citea bus (I believe they have them as well in Munich) it is not unlikely that I made the calculations for it. But to give you an idea: to model an entire bus, took me and two other colleagues about three months of work. The calculations were done within fifteen minutes on our server.

A solid mirror is relatively simple. And a single component as well, two to five components as well, but if the construction becomes increasingly large, the amount of work increases significantly.

And these were calculations for that bus were done with about 4 million elements. In certain situations one needs several billion elements. Note that calculation time goes with the third power approximately of the number of elements and you get the gist. A quick guestimation of the number of elements of the Extremely Large Telescope structure: probably in the order of 100 million elements.

PLOP works with, in the order of, thousand elements for one mirror only.

 

Then I will not even bother you with material norms, exotic materials like carbon fiber that have a directional strength, non-linearities, dynamical behavior, norms for welding, fatigue etc.



#21 Mel Bartels

Mel Bartels

    Ranger 4

  • *****
  • Posts: 389
  • Joined: 21 Sep 2007

Posted 24 April 2020 - 04:04 PM

Stathis, I admire and commend you for doing something significant - that is a nice looking piece of glass! And my astonishment and compliments to Rik and Geert for their beautiful telescope.


Years ago when I started grinding mirrors I would target a focal ratio, then, as the hours of rough grinding went by, would reconsider my choice. “I’ve got it down to F6, maybe that would be better than the F5 I originally projected. Longer focal length, easier time at higher magnifications. But it won’t fit in the car. Hmm”. I filled my mind with my thoughts. Later I began to think more of what the glass might have to say.


A thin meniscus mirror floating in space does not deform at all; it is quite happy. It’s only after we bring it back down to Earth and droop it over a support that it deforms. I once had a person yell at me that the cell and mirror were no good because PLOP said 1/100 wave deformation. 1/100 wave! The horror of it all. A couple of Oregon Scope Werks members got curious about verifying FEA with interferometry. To their surprise they found up to a 2x difference on very thin glass. Being a software developer who wrote some FEA code, I know that assumptions on how the software is to be used and interpreting the results is critical. Nonetheless, in this case, I suspect the mirror mounting.


Two people have asked to see the images in my 25 inch F2.6 scope without the coma corrector. It is quite something. Someone actually made fun of one of these people. But unless I look, how will I gain a feeling for the coma, like the feeling I have in my hands of what silky smooth final grinding feels like and in my ears of what it sounds like?


For the same experiential experience, I started my star testing of the 25 inch with a 3 point back support. Yes, indeed, that was quite an image - not. I experimented with “holding up the mirror’s edge”. I turned those weird blobs into long streaks. I anticipated needing at least an 18 point support, but when I went to the next level up, the 9 point support wasn’t bad at all. So I settled on a modified 9 point support where each support is actually a ring ground to shape against the mirror’s curved back. How many support points is that?


Another interesting observation is that we talk about the meniscus thickness differently than standard blanks. For the 30 inch F2.8 5/8 inch thick meniscus mirrors that I am finishing polishing and beginning figuring their equivalent thickness from center-back to forward-edge is actually 1.31 inches thick. That outer back side glass has been removed. Surely the rim is less steady but then all that weight reduction helps reduce the drooping over the cell’s support points. A nice mirror cell should hold the mirror’s edge steady, obviating the need for that thick edge. David Davis is experimenting with ‘pie pan’ shaped mirrors, where the meniscus shape is accentuated with an edge cylinder with exciting results.


We can go through the models (while a critical advance [I use PLOP by the way], models are ‘all wrong’ too) but maybe while grinding away for all those hours we might think not about FEA, not even about the mirror cell itself, but instead think about the process that we can use that will generate a good mirror cell design that matches the glass nicely.


Here is my current project, the twin 30 inch mirrors that I'm sharing on your thread with your permission... I am thinking of following the same empirical approach to generating a good mirror cell for them.

 

mirrors%20nearly%20polished%20out%20at%2

 

Mel Bartels


Edited by Mel Bartels, 25 April 2020 - 11:51 AM.


#22 Stathis_Firstlight

Stathis_Firstlight

    Stathis Firstlight

  • *****
  • Vendors
  • topic starter
  • Posts: 74
  • Joined: 28 Dec 2019

Posted 07 June 2020 - 06:17 AM

Over the last weeks I ground out the mirror up to Microgrit 3 my, polished it out with the 40 cm pitch lap, build a test stand and adopted the Foucault and interferometer testing set up.

 

See some pictures that I added onto the website:

http://www.stathis-f...3_mirroreng.htm

 

On the video it looks quite relaxed, but the pitch "bite" into the glass so firmly, that it was really exhausting pushing the tool around, uahhh! As a benefit of the strong action it was polished out after 7 hours net polishing only! 

http://www.stathis-f...s/polishing.mp4

 

Up to this point it looks fine: Nearly spherical (CC= -0,02), smooth, free of zones, sharp edge and - most importand- almost free of astigmatism.

 

foucault_edge_k.jpg

 

Now I will start the long journey to the parabola.



#23 eroyer

eroyer

    Vostok 1

  • -----
  • Posts: 114
  • Joined: 30 Aug 2014

Posted 07 June 2020 - 08:02 AM

Hello,

you did the grinding and polishing quite fast.

On your website, concerning the photo above, you say "The ruler test shows a good edge". I never heard of the ruler test before. Can you explain how this is done ?

 

Eric



#24 Stathis_Firstlight

Stathis_Firstlight

    Stathis Firstlight

  • *****
  • Vendors
  • topic starter
  • Posts: 74
  • Joined: 28 Dec 2019

Posted 07 June 2020 - 10:38 AM

Hello Eric,

 

you just compare the glow in brightness and extention of the mirror edge against the diffraction effect: 

 

Put a rouler or bar in upright position next to side opposite the knife edge and focus the Foucault tester at the edge zone. If the mirror is nearly spherical, this is just exactly at the ROC, like shown in the above picture. When moving in the knife edge the mirror surface darkens while the bright lines at the edges of the rouler are still glowing. This glow is caused caused by diffraction. If the mirror edge opposite the knife edge glows no brighter and is no wider than this diffraction glow, you have a perfect edge.

 

Example of a good edge:

http://www.astrotref.../20040624/5.jpg

 

Example of a bad edge:
http://www.astrotref.../20040624/6.jpg

 

The photos are from our beloved Alois Ortner. The whole thread in German:

http://www.astrotref...?TOPIC_ID=11206

 

In my case here the glow of the edge is a little brighter, than the diffracion of the rouler, but is as dimentionless fine as the diffraction lines, so it is still good (as to my own judgement while looking life through the Foucault tester with the eye). 


Edited by Stathis_Firstlight, 07 June 2020 - 10:53 AM.


#25 eroyer

eroyer

    Vostok 1

  • -----
  • Posts: 114
  • Joined: 30 Aug 2014

Posted 07 June 2020 - 11:02 AM

Ok thank you. I think there is a mismatch with the links but I understood "gute kante" is for good edge.




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