43 replies to this topic

[Jamesarakaki]

Many years ago at the RIverside Telescope maker's meet in So Cal I ran into an older gentleman who was displaying some small fast newtonians that he had made by warping commercial spherical mirrors into a parabolic shape. This was probably in the early 1990s.

 

Does this sound familiar to anyone?

 

IIRC he epoxied a single hex head bolt to the back center of an aluminized, spherical 5 inch mirror. The mirror was supported mostly by the edges with a ring of carpet. The bolt protruded through the middle of a plywood 'cell' and was held in place with a washer and a wingnut.

He stated that applying tension on the central bolt would warp the mirror into a paraboloid.

 

Now, I can't recall anything else about this. Getting old sucks.

 

Apparently it's fairly cost effective commercially  to make a smooth spherical mirror, but parabolizing -not so much.

I was looking at my junk box, which included an f/8 spherical 114mm primary from a cheap Cometron reflector which needed recoating. I was thinking about making a pitch lap and parabolizing this thing before sending off for aluminizing when I remembered this mirror warping thing.

 

Now, it may not make that much difference with a small f/8 reflector, since my Celestron C4.5 apparently also has a spherical primary and it works reasonably well. But what about those cheap 5 inch short tube Jones-Bird scopes that have a 'corrector' in the focuser drawtube? They use a fast spherical primary and are common and cheap secondhand since they suck so badly. Maybe they would be a good candidate for mirror warping??

 

Has anyone heard of this and tried it?

 

 

[MitchAlsup]

As long as the spherical primary is "not too big" and "not to fast" the trick can work reasonably well.

 

By the time the primary is big enough to require 6-point cell, you probably can't do this effectively anymore.

 

Sky and Telescope mid 1980s

[Steve Dodds]

That was Bill Kelley, long since passed, but a lot of his spherical mirrors are still out there.

[davidc135]

Jamesarakaki. Alan Adler's article in S&T 'Flexing spheres into high quality telescope mirrors' is online and refers to Bill Kelley who was making paraboloid mirrors this way in the early 1990s and who may have been the gent that you met.

 

David

 

PS Posts crossed

Edited by davidc135, 22 November 2021 - 05:40 PM.

[gr5org]

There's a picture of bill kelley in this article:  https://skyandtelesc...-MirrorFlex.pdf

[Jamesarakaki]

Thank you all! That was the gentleman I met. Anyone here try this? Maybe I've not been paying attention but it seems like this isn't done much today if at all.

[Oregon-raybender]

This has been discussed before. I was at RTMC (backup judge) the year they brought the scope and data. The article in S&T is misleading since there where folks who invented this method long before. In 1972 R. A. Buchroeder and A. S. Leonard both wrote a OSA (Optica) paper on taking spherical mirrors and bending them in to conic shape. Dr. Humphrey also invented (patented) an improved method, using a curved back, which improved the optical performance. It was used in ophthalmic instrument in 1976. Finally Chromex was issued a patent 1990 on the very method they were using, IE using a bolt in the center of the optic and pulling or pushing the mirror into a conic shape. I pointed this to Alan and Billy, which they were not aware of, but their feeling was this was a new discovery.

 

Also one of the teachers at the Chabot Telescope Makers class (a former student of Frank Wright, who taught the class in 1930's) Made a bent mirror into a conic in 1980, using a 1/4 bolt and ring which it was pulled against at the 70.7 zone.  We tested and worked well. The tech issue with this method is you need to adjust the bending through out the night, it does not stay in shape. There is no question that they brought forward the idea and it is interesting. The requirement of having to adjust the mirror as one is using it caused the lack of interest., (also think about the cooling issues, and the others that many folks are focusing on)  Polishing the shape into glass is superior.

 

One patent of interest is from 1889, a bending of telescope mirror. You can look at the reference patents and see that there were many other ideas in that general direction of bending glass into shape.

 

Starry Nights

 

https://patents.goog...nt/US408511A/en

 

https://www.osapubli...ri=ao-11-7-1649

 

https://patents.goog...t/US4043644A/en

 

https://patents.goog...t/US4932768A/en

 

 

[Jamesarakaki]

Raybender- thank you for the clarification. I can see how a steel bolt would be frustrating. Invar is too expensive, and I guess a tension rod of fused quartz would be too expensive to cut a thread on. (Use a conventional steel tension screw attached to a plain quartz rod to set and then lock the rod axially with a collet arrangement? But then the mirror itself is changing shape as the temperature moves...

 

And likely an adhesive bond to the mirror would have to be cleaned off before the mirror could be recoated...

 

nevermind! You're right- easier to parabolize a mirror and be done with it.

[astrokeith]

The technique is still alive.

 

The JWST uses the technique to adjust focus of the primary mirror segments.

[Oregon-raybender]

The idea or thoughts about bending glass or metal into for use in optics may surly go back to Newton's time. I recall before prisms in lighthouses

that bending polished metal to focus light was thought of. So the question where this all started can go back 350 + years? It was not until better testing and manufacturing methods improved enough to now look into it's possible. While a Tinsley. I was one of opticians who helped with the bending of glass to make segmented mirrors, the method that helped make the Keck telescopes, that was in the mid 1970's Today's new designs for large scopes make use of pistons to adjust the mirrors through the night. Now there is tilted and free form optics to improve optical or change performance. I think it's a broad area with a long history, just that it was not well received by the purists and lost long ago, but now brought into the light (so to speak) with the use of better and more open publications and the web.  I will give credit to others who may have just discovered this, but I like to give some historical background, that a few walked this path before.  

 

Starry Nights

Edited by Oregon-raybender, 23 November 2021 - 10:41 PM.

[ccaissie]

Raybender said, "The tech issue with this method is you need to adjust the bending through out the night, it does not stay in shape."

 

I addressed this issue at Stellafane when Alan Adler was presenting this technique.  I had an undercorrected 10"f/6 and epoxied a 1/2" fine thread bolt at the rear center.  It was hard to keep the final excellent correction.  

Alan said  "use a spring to keep even tension".  Voila. 

 

Accidentally I whacked the bolt and it removed a lens of Pyrex from the back!!!  J.B. Weld is strong stuff!

 

The most exquisite planetary images from any of my scopes is from my 8f/7 flex scope. Bill Kelley looked through one of mine and we high-fived. His later project was short focus 6f5's with a special shaped back that pulled well with a central bolt vs. the puller/pusher of Adler's design.

 

If you somehow download the old DOS program FLEX, you can play with different diameters, f# , thicknesses and puller tension to see that large diameter and short f#'s really run out of tolerance room.   Superb technique for taking a super smooth sphere and microflexing it to a lovely paraboloid.  I've flexed 8f7, 6f8 and 6f/10.  I've seen larger mirrors flexed, but not successfully.  

 

I would think a 10 or 12" f/6-7 would be a tremendous planetary scope.

 

Edit:  Chriske here on CN has experience with flex mirrors.   I can see the usefulness in special aspheres...like in building a Yolo with long radii but high conic numbers, or a Schief with a -.55 conic.    Flexing in this slight non-paraboloid corrections from a sphere would be easily handled with a flex cell.

Edited by ccaissie, 24 November 2021 - 08:42 PM.

[ccaissie]

Raybender- thank you for the clarification. I can see how a steel bolt would be frustrating. Invar is too expensive, and I guess a tension rod of fused quartz would be too expensive to cut a thread on. (Use a conventional steel tension screw attached to a plain quartz rod to set and then lock the rod axially with a collet arrangement? But then the mirror itself is changing shape as the temperature moves...

 

And likely an adhesive bond to the mirror would have to be cleaned off before the mirror could be recoated...

 

nevermind! You're right- easier to parabolize a mirror and be done with it.

Don't give up that easy...  read Adler's approach...works.   I made modifications of that to simplify the cell design.  

[Oregon-raybender]

Yes, adding a spring solves the issue of the mirror shape during the night, up to a point. I knew of this, I was addressing the discovery, not the idea.

We looked at the issue on the mirror made at Chabot, it requires a good spring design, it has been to long for me to recall the solution what the mechanical engineer decided to use.  The idea works as noted in Alan's and in others design and build. The design did get a once over by a manufacturer, but having the owner adjust the mirror would open issues, What if the owner does not adjust it correctly and points to the fact it does not work? It would be a hard sell as bad press out does the good. For a ATMer, sure it's your mirror and design of the cell, if works for you great, go for it! But the idea for large scale commercial solution is just not likely, which Alan and Billy were hoping for (as noted by them at RTMC)  The Chabot design had ring at the 70.7 zone and 1/4 -20  glued to the back of the blank, I thing the spring may have been the bevel type?  So go for it if you have the will to try it. Be sure to let us know your process and results. Always look forward to new ideas.

 

Starry Nights

[ccaissie]

 The design did get a once over by a manufacturer, but having the owner adjust the mirror would open issues, What if the owner does not adjust it correctly and points to the fact it does not work? It would be a hard sell as bad press out does the good. For a ATMer, sure it's your mirror and design of the cell, if works for you great, go for it! But the idea for large scale commercial solution is just not likely, which Alan and Billy were hoping for (as noted by them at RTMC) 

I was in conversation with Dan Cassaro who had acquired a mountain of 7/8" thick beautiful plate glass and was waterjet cutting it into scope blanks.  I bought 8 pieces and finished some nice spheres for flexing.  Dan asked me if I was interested in producing flex mirror kits, but at the time I had enough projects.  

 

One drawback is the disassembly and rebuild of the cell when recoating time comes.   My 8f7 mirror is recoated and waiting for remount...Our club has so many scopes, and so few participants due to COVID, that I can't justify putting it together...yet.

[Jamesarakaki]

Hmmm.

I dug out yet another 114mm spherical mirror of f/8 ratio. This one has a decent aluminum coat.  Really, nothing to lose here. 

 

Kind of thought that it wouldn't be much different than having a second control along with focusing- ideally the microflexing adjustment would be doable while looking through the eyepiece. Perhaps a small dc gearmotor? (Microflexing sounds so much better than warping!)

 

Re the spring mentioned- is a 'bevel' spring a coil compression spring with ground ends like an automotive valve lifter spring? Not that stiff or heavy of course... or would a wave washer spring work? Might it be a good idea to hold it between hard flat machine washers, like the ones used with needle thrust bearings? We're not talking about a lot of axial travel here, so wave spring washers might work. (I've used these as an axial preload spring with along with three collimation push cup point setscrews in homemade refractor star diagonals).

 

Driving the flexing screw nut with a gearmotor would be a little like an electric focuser- hopefully it won't create too much vibration to evaluate the effect. An o ring drive belt might work instead of a cog belt for a little smoother drive, and would allow slip if hard stops are needed? 

 

Such a drive would probably be too expensive to market commercially but I do have a bunch of this electromechanical junk laying around doing nothing.

[Steve Dodds]

Hmmm.

I dug out yet another 114mm spherical mirror of f/8 ratio. This one has a decent aluminum coat.  Really, nothing to lose here. 

 

Kind of thought that it wouldn't be much different than having a second control along with focusing- ideally the microflexing adjustment would be doable while looking through the eyepiece. Perhaps a small dc gearmotor? (Microflexing sounds so much better than warping!)

 

Re the spring mentioned- is a 'bevel' spring a coil compression spring with ground ends like an automotive valve lifter spring? Not that stiff or heavy of course... or would a wave washer spring work? Might it be a good idea to hold it between hard flat machine washers, like the ones used with needle thrust bearings? We're not talking about a lot of axial travel here, so wave spring washers might work. (I've used these as an axial preload spring with along with three collimation push cup point setscrews in homemade refractor star diagonals).

 

Driving the flexing screw nut with a gearmotor would be a little like an electric focuser- hopefully it won't create too much vibration to evaluate the effect. An o ring drive belt might work instead of a cog belt for a little smoother drive, and would allow slip if hard stops are needed? 

 

Such a drive would probably be too expensive to market commercially but I do have a bunch of this electromechanical junk laying around doing nothing.

a 114mm f/8 the difference between a parabola and a sphere is much less than 1/4 wave, much better than diffraction limited.  You are not going to notice anything by pulling the back.  You need a bigger or faster mirror.

[Oregon-raybender]

I would suggest at least f/5 or F/6, less amount of pulling required to get good conic.

One would need to do a study on one type of mount using the same diameter mirror (8 inch) and material at various

f numbers (f4 to f 7) to make a graph of the results. Remove the variables and less chaos. I would suggest starting with a

1/10 wave sphere tested using a Bath or other type of interferometer, one that produces test results with less human interpretation. 

Then plot the amount of pull required to produce the right amount of conic, maybe even plot at -.7 to -1.5 to see if the amount

varies. It will also be worth checking the conic over time at a set value, pull of the spring and temp effecting it. Of course all

has to use calibrated equipment and allow for some else repeat the experiment to get the same results. 

 

Starry Nights

[davidlewis]

Plop actually supports designing a cell to warp a sphere into a parabola, but it is quite tricky to set up and can fail if something isn't perfect.

[brentelgeuse]

There's a picture of bill kelley in this article:  https://skyandtelesc...-MirrorFlex.pdf

Thanks for the article! I have a 6" f5 spherical I'd like to try this on. Any idea how many pounds of pull it takes to bend a mirror? Like 0.1? 100? I can't find the Flex software online now that's described in the article. 

[lukasik]

Hi everyone,

for those interested it looks like the link to flex.exe is still active off this Sky&Telescope page.

https://skyandtelesc...-sky-telescope/

 

Also here's a link to an old thread that may give you some food for thought.

https://www.cloudyni...id/?hl= flexing

 

Best Regards,

Bob

[brentelgeuse]

Hi everyone,

for those interested it looks like the link to flex.exe is still active off this Sky&Telescope page.

https://skyandtelesc...-sky-telescope/

 

Also here's a link to an old thread that may give you some food for thought.

https://www.cloudyni...id/?hl= flexing

 

Best Regards,

Bob

Thanks Bob!

[Oregon-raybender]

Here is some history on bending mirrors. Jerry Nelson interview. Page 35 and beyond is of interest.

I was working with Luis at this time.

 

Starry Nights

 

https://oralhistorie... OHO final.pdf

[ccaissie]

Thanks for the article! I have a 6" f5 spherical I'd like to try this on. Any idea how many pounds of pull it takes to bend a mirror? Like 0.1? 100? I can't find the Flex software online now that's described in the article. 

You need to have the thickness measurement.  If the mirror is full thickness, it will take a lot of force, like over 100#,  and one limit is the strength of the Polyurethane cement holding the puller on.  If you want an ok paraboloid, you can make the puller smaller but then the tension [psi] on the pad is higher. 

 

A larger puller will lower the psi , give a more accurate paraboloid, but total pull will be higher.

 

So, from 50-150#...guessing.  If you're serious, you could thin the mirror.  You'll have to grind the wedge out anyway, so making it thinner is in your best interest.

[ccaissie]

 I would suggest starting with a 1/10 wave sphere tested using a Bath or other type of interferometer, one that produces test results with less human interpretation. 

 

Foucault at ROC is a null for a sphere...nothing better. That's the beauty of flexing.

[Oberon]

Here is some history on bending mirrors. Jerry Nelson interview. Page 35 and beyond is of interest.

I was working with Luis at this time.

 

Starry Nights

 

https://oralhistorie... OHO final.pdf

Thank you. A very interesting read.