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

Robotic mirror grinding?

  • Please log in to reply
37 replies to this topic

#1 bigdobsonfan

bigdobsonfan

    Lift Off

  • -----
  • topic starter
  • Posts: 16
  • Joined: 31 May 2021

Posted 20 June 2021 - 11:10 PM

This is hopefully not too out there, i'm just thinking out loud about something i'd wondered about for years, and was curious how feasible it is...

 

Many years ago I stumbled across Group 70 http://group70.org/ which was an attempt to turn an old 70 inch telescope blank into a usable amateur astronomy experiment.  I remember when there was more data on the status of the project when I first read about it, it's all from memory, but grinding of the mirror was all being done by hand with volunteers.  I remember something about being ground wrong accidentally and requiring a corrective grind then being needed...  As with many projects manpower was an issue along with funding...

 

It just had me wondering... how feasible would it be to design and build a scaleable bot to do the job of mirror grinding?

 

This is as interesting to me even if I were to grind some much smaller mirror say 10-12 inches as it would be to scale up to whatever size blanks could ever be found in the future.  I've glanced through books on telescope making and although the nature of the project interests me, spending many hours hand grinding a mirror is difficult to justify with a busy life and busy schedule.



#2 MeridianStarGazer

MeridianStarGazer

    Voyager 1

  • *****
  • Posts: 11,579
  • Joined: 01 Dec 2013
  • Loc: Southern Idaho

Posted 20 June 2021 - 11:30 PM

Grinding is typically done with a fixed post machine spun by a motor, not controlled by a robot. No work for you.

 

Do you mean robotic figuring, which is taking a sphere to a parabola? There is a professional on this forum who was working on robotic figuring. I don't know if he succeeded. I'm not saying who though.

 

I realize the mirror lab in Tucson AZ uses a true robot that wanders around the blank surface. I saw it in person on a tour. Very sub aperture.

 

Zambuto has a deterministic polishing algorithm that reliably takes a mirror to a parabola. It is not a robot, but just a pattern of polishing.

 

For big aperture though, spinning a large mirror and moving a large polisher might be too much work. That is why the mirror lab spun caste their glass to avoid hogging, and then had a robot go around grinding and polishing the high spots. Or maybe they spun ground it. It don't recall. That would be a huge piece of glass to spin, the it already spun during casting.


  • mark cowan likes this

#3 BGRE

BGRE

    Gemini

  • *****
  • Posts: 3,450
  • Joined: 21 Mar 2016
  • Loc: New Zealand

Posted 21 June 2021 - 12:57 AM

The mirror lab substrates after spin casting aren't accurate to a few microns so some coarse and fine grinding is still required.
NC grinding of aspheres has been done but a sufficiently stiff machine with adequate resolution encoders is required together with sufficiently accurate (sub micron) metrology so that achievement of the desired shape can be confirmed before polishing.
IIRC REOSC used computer controlled fine grinding on the VLT primary meniscii. The biggest problem is sufficiently accurate metrology once the grinding machine is tiff enough or at least its errors are sufficiently repeatable.
A process that corrects the error within 10% for each grind/polish and error measurement cycle ensures that the ground surface rapidly converges to the desired shape as long as the surface metrology is sufficiently accurate.

#4 Geo.

Geo.

    Vendor - Nexstar Parts

  • *****
  • Vendors
  • Posts: 5,711
  • Joined: 01 Oct 2008
  • Loc: Upstate NY

Posted 21 June 2021 - 09:51 AM

Sure, that's how Elmer-Perkins screwed up the Hubble primary. I've heard the mirror Kodak made for the Hubble in the traditional way still sits in a warehouse out in Rochester, NY. Said to have a great figure.


  • Augustus likes this

#5 KBHornblower

KBHornblower

    Viking 1

  • -----
  • Posts: 868
  • Joined: 01 Jul 2020
  • Loc: Falls Church, VA (Washington DC suburb)

Posted 21 June 2021 - 10:41 AM

Sure, that's how Elmer-Perkins screwed up the Hubble primary. I've heard the mirror Kodak made for the Hubble in the traditional way still sits in a warehouse out in Rochester, NY. Said to have a great figure.

My understanding is that a maladjusted test instrument was the reason for misfiguring the Hubble mirror, after cost-cutting decisions precluded using two different instruments as a sanity check.


  • PrestonE, gregj888, Benach and 1 other like this

#6 mark cowan

mark cowan

    Vendor (Veritas Optics)

  • *****
  • Vendors
  • Posts: 9,840
  • Joined: 03 Jun 2005
  • Loc: salem, OR

Posted 21 June 2021 - 11:43 AM

Hubble primary was manufactured quite accurately to the wrong specification provided by a misassembled test jig.

 

Spin grinding is very much an automatic process once dialed in on a machine (tool offset) but you still have to feed it grit somehow. :shrug:


  • gregj888, Benach, Pinbout and 2 others like this

#7 PeteDCard81

PeteDCard81

    Vostok 1

  • *****
  • Posts: 124
  • Joined: 30 Nov 2017
  • Loc: Twin Cities, Minnesota

Posted 21 June 2021 - 12:11 PM

The Hubble primary mirror was designed to the correct specification (parameters) but ended up misshaped because of an incorrectly assembled null optics test tool. This test tool was acquired from the Perkin Elmer side of the shop that did work on Defense projects. The null optics test tool assembly instructions were correct for the Hubble mirror but a small washer ended up in the wrong place.

 

The null optics test tool was used to guide the figuring and essentially was the only means to officially judge completion of the mirror.

 

There was some other unofficial tests done with tests devices that revealed a discrepancy but not followed up. Perkin Elmer was late in delivering the mirror to NASA and was under pressure.


  • Oberon and Augustus like this

#8 BGRE

BGRE

    Gemini

  • *****
  • Posts: 3,450
  • Joined: 21 Mar 2016
  • Loc: New Zealand

Posted 21 June 2021 - 04:52 PM

If the asphericity is large enough that it has to be ground in before fine correction during polishing the metrology methods employed can be come somewhat of a challenge especially during coarse grinding when "waxing" the mirror so optical testing is feasible.
If one has an NIR camera then "optical" testing of the fine ground surface at 1550nm is usually possible.
One advantage of grinding in most of the asphericity is that grinding is much faster than polishing.

The correction + error measurement cycle is the same process employed by amateurs when making mirrors albeit perhaps with different tests..

#9 Benach

Benach

    Mercury-Atlas

  • *****
  • Posts: 2,504
  • Joined: 24 Jan 2008

Posted 21 June 2021 - 05:23 PM

Well, I know there are some deterministic methods of producing optics, magnetorheological polishing and ITIRM, but these are way beyond the skill level of every amateur.



#10 BGRE

BGRE

    Gemini

  • *****
  • Posts: 3,450
  • Joined: 21 Mar 2016
  • Loc: New Zealand

Posted 21 June 2021 - 07:23 PM

Even hand figuring is somewhat deterministic, if it weren't the rub/polish-test cycle wouldn't converge. Its merely a somewhat intuitive application of Preston's law.
As long as the test can identify those areas that need to be worked on the process can converge.
  • PrestonE and Dale Eason like this

#11 mark cowan

mark cowan

    Vendor (Veritas Optics)

  • *****
  • Vendors
  • Posts: 9,840
  • Joined: 03 Jun 2005
  • Loc: salem, OR

Posted 21 June 2021 - 11:15 PM

Exactly.  The key is to keep the materials being worked together as consistent as possible, so that you CAN make predictions.



#12 555aaa

555aaa

    Vendor (Xerxes Scientific)

  • *****
  • Vendors
  • Posts: 2,080
  • Joined: 09 Aug 2016
  • Loc: Ellensburg, WA, USA

Posted 22 June 2021 - 08:21 AM

The in thing nowadays is that you use a general purpose robot to operate a special purpose machine, so I could see a kuka or fanuc in a slider, tending to a row of polishing machines.

Edited by 555aaa, 22 June 2021 - 08:21 AM.

  • PrestonE likes this

#13 rhetfield

rhetfield

    Surveyor 1

  • *****
  • Posts: 1,717
  • Joined: 14 Aug 2019
  • Loc: Suburban Chicago, IL, USA

Posted 22 June 2021 - 12:04 PM

I could see using a modern general purpose robot to polish a mirror.  However, the position sensors in the robot would not be adequate for measuring the figure of the mirror.  Frequent testing of the mirror using the proper testing equipment would be needed.  It would be the same process as hand polishing - just with automation involved so nobody gets tennis elbow.


  • Geo. and bigdobsonfan like this

#14 Oregon-raybender

Oregon-raybender

    Optical Research Engineer

  • *****
  • Vendors
  • Posts: 1,073
  • Joined: 13 May 2010
  • Loc: Oregon, South Western Coast

Posted 22 June 2021 - 02:01 PM

Robot or CNC machining of glass has been around since late 1960's, they were simple machines using punched tape

produced by a computer. It was used on free form optics. There was what is called profile grinding / polishing. Were a master is made of the shape needed, then it was traced. The size of the profiler varied base on the size of the optic. The cam follower would roll across the shape and diamond cutter of the same size or larger would cut the path. As young optician I ran few of these, it was dirty work, water, grit and polishing compound about.

 

A bit off topic, I attended the Hubble report at a international optical designers conference, it was the formal final review of what happen (2 hours long) It came down to politics, money, and time. The interferometer tester that was used was a all reflection one, using two mirrors to set the position of the tester to CC needed. There was Null lens tested that was a backup, but it was deemed not "good enough"due to use of refractive material (ie, strie, stress, and other glass issues) So the optician(s)  was "required" to use only the reflective tester. In defense of the optician, he did use other methods and noticed the error. But was overruled. The error that occurred was a basic human one (putting on my QA hat)  They not understand the basic use of this type of tester. The requirement was to focus the returning laser light from the primary mirror onto a spot just inside the tester. Since they could not see where the spot was, it was "assumed" that front face of the tester was the right position. The so call washer or spacer was in front of the tester was 2 mm forward. There was a strong push to get it done. Yes, the Kodak mirror was better of the two.     The repair in the end was to add lenses to correct. I guess glass refractive optics are not all bad.

 

Starry Nightswaytogo.gif


Edited by Oregon-raybender, 22 June 2021 - 02:02 PM.

  • Jon Isaacs, PrestonE, LarsMalmgren and 4 others like this

#15 KTAZ

KTAZ

    Surveyor 1

  • *****
  • Posts: 1,930
  • Joined: 09 Apr 2020
  • Loc: Scottsdale, AZ

Posted 22 June 2021 - 02:41 PM

The Hubble primary mirror was designed to the correct specification (parameters) but ended up misshaped because of an incorrectly assembled null optics test tool. This test tool was acquired from the Perkin Elmer side of the shop that did work on Defense projects. The null optics test tool assembly instructions were correct for the Hubble mirror but a small washer ended up in the wrong place.

 

The null optics test tool was used to guide the figuring and essentially was the only means to officially judge completion of the mirror.

 

There was some other unofficial tests done with tests devices that revealed a discrepancy but not followed up. Perkin Elmer was late in delivering the mirror to NASA and was under pressure.

I cannot even imagine, once the data started coming back from the first space images, the massive, dread induced, adrenaline surge among those at PE when they understood the immensity of one of the biggest f-ups in astronomical history.

 

I think my heart would have just stopped. Game over.



#16 PrestonE

PrestonE

    Vanguard

  • *****
  • Posts: 2,368
  • Joined: 29 Apr 2005
  • Loc: San Miguel de Allende,Mexico

Posted 22 June 2021 - 03:45 PM

Robot or CNC machining of glass has been around since late 1960's, they were simple machines using punched tape

produced by a computer. It was used on free form optics. There was what is called profile grinding / polishing. Were a master is made of the shape needed, then it was traced. The size of the profiler varied base on the size of the optic. The cam follower would roll across the shape and diamond cutter of the same size or larger would cut the path. As young optician I ran few of these, it was dirty work, water, grit and polishing compound about.

 

A bit off topic, I attended the Hubble report at a international optical designers conference, it was the formal final review of what happen (2 hours long) It came down to politics, money, and time. The interferometer tester that was used was a all reflection one, using two mirrors to set the position of the tester to CC needed. There was Null lens tested that was a backup, but it was deemed not "good enough"due to use of refractive material (ie, strie, stress, and other glass issues) So the optician(s)  was "required" to use only the reflective tester. In defense of the optician, he did use other methods and noticed the error. But was overruled. The error that occurred was a basic human one (putting on my QA hat)  They not understand the basic use of this type of tester. The requirement was to focus the returning laser light from the primary mirror onto a spot just inside the tester. Since they could not see where the spot was, it was "assumed" that front face of the tester was the right position. The so call washer or spacer was in front of the tester was 2 mm forward. There was a strong push to get it done. Yes, the Kodak mirror was better of the two.     The repair in the end was to add lenses to correct. I guess glass refractive optics are not all bad.

 

Starry Nightswaytogo.gif

I just finished reading, "The Hubble Wars" and found the arrogance incredible among many involved...the entire book was Spell Binding in what went on

during the commissioning...it was recommended here I believe by Mike Jones and is a Great Read...

 

Preston

 

ps...sorry for the off topic.


  • LarsMalmgren likes this

#17 Benach

Benach

    Mercury-Atlas

  • *****
  • Posts: 2,504
  • Joined: 24 Jan 2008

Posted 22 June 2021 - 04:35 PM

Preston: Do you have the author and ISBN for me?



#18 Oregon-raybender

Oregon-raybender

    Optical Research Engineer

  • *****
  • Vendors
  • Posts: 1,073
  • Joined: 13 May 2010
  • Loc: Oregon, South Western Coast

Posted 22 June 2021 - 04:35 PM

The book covered the politics of the mirror. What I heard was the inside info of what happen in the optical shop and how it came to be. The sad part was when it first came out, many of us designers made models using our optical software and had the results in a matter of days. It was the misplaced tester, several of us expressed this at a local optical meeting a month later.

It was not until 10 years later the inside scoop can out at the larger conference (all 100 of us)  They were lucky a repair could happen but at a cost. A good example of doing right the first time?  Off my QA soap box.

 

Starry Nightswaytogo.gif


  • Augustus and Tribe_Of_Dan like this

#19 Steve Dodds

Steve Dodds

    Apollo

  • *****
  • Posts: 1,019
  • Joined: 02 Apr 2007
  • Loc: Utah

Posted 22 June 2021 - 04:40 PM

Preston: Do you have the author and ISBN for me?

Here it is on Amazon

https://www.amazon.c...24397966&sr=8-1


  • PrestonE, LarsMalmgren and bigdobsonfan like this

#20 Benach

Benach

    Mercury-Atlas

  • *****
  • Posts: 2,504
  • Joined: 24 Jan 2008

Posted 22 June 2021 - 05:52 PM

Here it is on Amazon
https://www.amazon.c...24397966&sr=8-1

Thanks.

#21 bigdobsonfan

bigdobsonfan

    Lift Off

  • -----
  • topic starter
  • Posts: 16
  • Joined: 31 May 2021

Posted 29 June 2021 - 03:14 PM

I could see using a modern general purpose robot to polish a mirror.  However, the position sensors in the robot would not be adequate for measuring the figure of the mirror.  Frequent testing of the mirror using the proper testing equipment would be needed.  It would be the same process as hand polishing - just with automation involved so nobody gets tennis elbow.

Yeah I guess that was kind of my question.  (the responses here are SO far beyond my current understanding alot is over my head!)  I just remember reading (memory, not indisputable fact!) about people volunteering to hand-surface so many hours per week being a big bottleneck for finishing the group 70 mirror...  so I was wondering if that was something that could be robot-ified...  that the limitation could be more test equipment than the physical replacement of elbow grease makes alot of sense...  and potentially directs future questions! (though i'll withhold for now until I learn more about the subject that i'm asking :) )



#22 Pezdragon

Pezdragon

    Ranger 4

  • *****
  • Posts: 394
  • Joined: 10 Jun 2013
  • Loc: Bay Area, Ca.

Posted 29 June 2021 - 05:30 PM

I remember my early days at Tinsley in the early seventies and operating heavily modified 3 axis Bridgeport milling machines adapted to receive computer( punched tape) programs that worked highly local zones on 10” or smaller optics as Raybender alluded to. A analysis of the optic would be performed at intervals and a reiteration would be programmed in. I think that might be considered robotic polishing. Sometimes the polishing tools were a scant inch long and 1/4 inch wide.


  • Oregon-raybender likes this

#23 Oregon-raybender

Oregon-raybender

    Optical Research Engineer

  • *****
  • Vendors
  • Posts: 1,073
  • Joined: 13 May 2010
  • Loc: Oregon, South Western Coast

Posted 29 June 2021 - 10:03 PM

Yep, we did a lot of zonal polishing. But I think more about the TV lens production, where there was several machines running

They were free form lenses for production of TV color tubes. Real odd ball stuff.

 

Starry Nightswaytogo.gif



#24 Pezdragon

Pezdragon

    Ranger 4

  • *****
  • Posts: 394
  • Joined: 10 Jun 2013
  • Loc: Bay Area, Ca.

Posted 29 June 2021 - 10:22 PM

Oh yeah…computer profile ground then off to the caged rod fine grinder then lead bag/polishing dots ..did a summer on that line and …..I WAS THE ROBOT !

Back then and now I fantasize about a active real time interferometer built into a hollow spindle polisher that would actively interpret the lens surface and adjust stroke/ speed to obtain a preprogrammed curve/smoothness…now that would be a robotic polisher!


  • Oregon-raybender likes this

#25 tommm

tommm

    Apollo

  • *****
  • Posts: 1,007
  • Joined: 16 Dec 2015

Posted 30 June 2021 - 11:52 AM

How is an interferometer going to work with polishing slurry on the blank surface?  I don't see any way to do it other than as stated earlier, just use the machine to replace a human doing the polishing, but clean the mirror blank and test it as you would if a human were doing the work.  It still requires all the same care and attention to details like consistent slurry and polishing tool condition for consistent removal rate, and programming the machine to work the correct areas for the correct amount of time...It seems mainly of use for large production projects where it would aid in consistency due to humans getting bored and distracted - as long as the programmer isn't pushed too hard and screws up.




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