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

Fast 25" in S&T

  • Please log in to reply
102 replies to this topic

#26 hakann

hakann

    Surveyor 1

  • -----
  • topic starter
  • Posts: 1743
  • Joined: 06 Aug 2015

Posted 10 January 2019 - 12:49 PM

Tomo,

I'll just hope I could look into the EP more as I was invited, but I has full understanding of other projects.

Time will come.
I say it again, it’s a very inspiring telescope worth travel a long distance for.



#27 Pierre Lemay

Pierre Lemay

    Apollo

  • *****
  • Posts: 1330
  • Joined: 30 Jan 2008
  • Loc: Montréal, Canada

Posted 10 January 2019 - 01:16 PM

Of course you cannot just slump a flat piece of glass, but you could grind the back to a smaller radius than the front such the edge is say 1/3rd the thickness of the center (with a large diamond coring drill and rotating sine table) 

You can slump a flat piece of glass and have different radii of curvature on the front and back. That's kind of what Mel did on his 25 inch and the approach I will be using on my 28 inch, but in reverse to what you are suggesting. Both the 25 and the 28 inch blanks were slumped on a mold that had a much shorter radius of curvature than the intended final RoC. Mel's ¾ inch thick x 25 inch diameter plate glass blank was slumped by David Davis to an f/2 curve whereas my ¾ inch thick x 28 inch diameter Borofloat blank was slumped by Normand Fullum to an f/2.15 curve.

 

Here's a picture of my 28 inch blank on its slumping mold just before going in Norm's annealing kiln:

 

28 inch on slumping mold.jpg

 

And a picture of the 28 inch blank as it came out:

 

Slumped 28 inch photo2.jpg

 

The grinding procedure is then to lengthen the RoC on the front (concave) surface by grinding it back to the intended focal length (there are several ways of doing this, including a diamond cup. I haven't decided yet how I will do it). Mel stopped at f/2.5 and I'm aiming for f/2.8 or, maybe, f/2.7. This is done by thinning the periphery of the front surface by about 3/16 inch near the edge (for my 28 inch) and gradually less towards the center. The result is that the center of the blank more or less maintains its original ¾ inch thickness and the edge is now down to near ½ inch thick. The back of the blank maintains its strongly curved f/2 shape which helps in providing the structural strength of the meniscus.

 

The other thing that contributes to a rigid mirror blank is the lightness of it. My 28 inch meniscus mirror blank currently wheigs about 37 pounds. By the time I've finished grinding in the longer RoC it will have lost almost five pounds. Most of that weight loss will be NEAR THE EDGE. The final weight should be around 32 pounds, which is the same weight my 20 inch conical blank weighs. Combine this lightness with the strongly curved glass and you have a rigid combination that can better maintain it's shape in various positions. That probably explains why Mel was able to get away with a simple 9 point floatation cell for the back and two simple pegs at 90 degrees for edge support!

 

A 1.25 inch thick f/5 25 inch, constant thickness meniscus might have a much harder time maintaining it's shape in those conditions. But it works for very thin, strongly curved blanks because you are using the benefit of both those characteristics to your advantage.


Edited by Pierre Lemay, 10 January 2019 - 01:22 PM.

  • PrestonE, brave_ulysses, LarsMalmgren and 3 others like this

#28 hakann

hakann

    Surveyor 1

  • -----
  • topic starter
  • Posts: 1743
  • Joined: 06 Aug 2015

Posted 10 January 2019 - 03:30 PM

Ok, so bigger strongly curved might make this working and a faster curve on backside ( say f/2 ) and say parabola to ex f/3 = thicker centre.

-I thought Mel's was equal 1/2" thick.

It do sounds to me that it should not hurt use a centre thickness atleast 1.5" and machined in Quartz.



#29 hakann

hakann

    Surveyor 1

  • -----
  • topic starter
  • Posts: 1743
  • Joined: 06 Aug 2015

Posted 11 January 2019 - 12:59 PM

On the other side, as equal thick mean equal cool-off, so why not go thicker and in Quartz as ppm/C at 0.5 from 9.0.

A f/3 @25" and 1.5" thick means pretty stiff if now the meniscus shape get the stiffer 'effect' at ex f/3 curve.



#30 mark cowan

mark cowan

    Vendor (Veritas Optics)

  • *****
  • Vendors
  • Posts: 9419
  • Joined: 03 Jun 2005
  • Loc: salem, OR

Posted 11 January 2019 - 02:48 PM

No it's less "effect" the thicker the meniscus.  The improvement for actual performance only makes it equivalent to what you would get from adding back the curve on the back side to make a flat back mirror.  This amount decreases the thicker the meniscus and in the case you mention is small - 1.75' vs 1.5" (for a flat back mirror).



#31 hakann

hakann

    Surveyor 1

  • -----
  • topic starter
  • Posts: 1743
  • Joined: 06 Aug 2015

Posted 11 January 2019 - 03:04 PM

Mark, so you mean a equal curved and thickness ( front/back at say f/3 ) and 1" or 1.5" thickness is less stiff than a 1/2" ?

Ok, say vs a back at f/2 and parabola to f/3 and edge is 1/2" ( so centre is little thicker )

 

I did CAD it up so on a 25" and the back is at f/2 and parabola at f/3 and edge at 1/2" ( 12.7 mm ) the centre was 20.21 mm.

So centree is 7.5 mm thicker than edge.

 

On ex Mel's f/2.6 the CAD showed a centre thickness of 18.6 mm, so it is 5.9 mm thicker than the edge if that is at 12.7 mm.

 

-If this is how his mirror was done at, I don't know.

I thought I heard it was equal thick as that was the idea of equal coolings.


Edited by hakann, 11 January 2019 - 03:25 PM.


#32 mark cowan

mark cowan

    Vendor (Veritas Optics)

  • *****
  • Vendors
  • Posts: 9419
  • Joined: 03 Jun 2005
  • Loc: salem, OR

Posted 11 January 2019 - 03:48 PM

I didn't say that.  Work the math out if you want to understand it better, or run PLOP as described earlier and see what happens for various thicknesses of menisci compared to the flat back mirror they would be cut out of.  The thicker the meniscus or the slower the mirror the less effect there is, but the effect is never all that large.



#33 ckh

ckh

    Surveyor 1

  • -----
  • Posts: 1916
  • Joined: 21 Feb 2015
  • Loc: Arizona

Posted 11 January 2019 - 07:41 PM

What do you consider large?  1.6 is pretty good. 

 

Note that Mel polished (or at least figured) with MOT.  That way only the overhanging part of the mirror causes bending forces (if it's polished using only its own weight).  

 

You can alway use more support points to compensate for a thinner mirror. The problem is in polishing. How do you apply the pressure necessary to polish without bending the glass?  One promising solution I've seen is using vacuum at the polisher to apply the load. The polisher is counterbalanced so it applies no weight. Then you are just squeezing the glass against the polisher with ambient air pressure. The glass can be supported in the mirror cell. It seems that it could work well with a thin meniscus (with constant thickness along the radius).  

 

Someone (an amateur) tried this (with a cellular I think) and gave up. The way he was moving the polisher was a bit unusual.  In his design the polisher had an unchanneled outer rim to contain the vacuum. That creates a couple of problems. One is that the outer part of the lap is not channeled and another is that it can only go so far over the edge before leaking the vacuum.  But what if each facet had a vacuum hole in the middle and a ball check to keep it from sucking too much air when off the edge?  Then, the force per facet would be uniform (theoretically). The pitch is going to squish so how do you maintain such facets?

 

Sounds difficult but maybe there's a chance.  I believe it has been done by professionals but I have no details.

 

https://www.youtube....h?v=Hx3T9LThI8U


Edited by ckh, 11 January 2019 - 07:45 PM.


#34 mark cowan

mark cowan

    Vendor (Veritas Optics)

  • *****
  • Vendors
  • Posts: 9419
  • Joined: 03 Jun 2005
  • Loc: salem, OR

Posted 11 January 2019 - 08:14 PM

I was writing a reply but hit the wrong button.  1.6 would be a fairly special case.  You can see this with PLOP simulations as described already.  They tend to be around 1.4 though.  If you make the meniscus slower or thicker the advantage decreases.

 

Take an example 20" f/3 at 1" thick, PLOP returns 1.48nm RMS for an 18 point cell (modeled as an f/1000).  You have to use a 37mm flat back mirror to match that error.  The ratio there is 1.46.

 

For a 1/2" thick meniscus of otherwise same specs the error is 4.15nm RMS and the matching flat back case is 21.8mm, that's a ratio of 1.7,  more interesting territory.   (This is from PLOP without the lengthy Z-80 3D analysis which is more accurate).

 

 

Fabrication is another story and there's lots of ways to do this correctly, none of which I'll describe (for lack of time mostly). wink.gif

 

 

 

edit:  Re. the above examples, note that the flat back equivalent is well predicted by just adding the meniscus sagitta to it's thickness.  Sagitta for 20" f/3 is 10.6mm, 25.4+10.6mm = 36.2 mm (compare 37), and 12.7+10.6mm = 23.3 mm (compare 21.8).  As a rule of thumb it works pretty well.


Edited by mark cowan, 11 January 2019 - 09:34 PM.


#35 radicell2

radicell2

    Vostok 1

  • -----
  • Posts: 152
  • Joined: 15 Jan 2010
  • Loc: Toronto

Posted 12 January 2019 - 11:07 AM


 

T The result is that the center of the blank more or less maintains its original ¾ inch thickness and the edge is now down to near ½ inch thick. The back of the blank maintains its strongly curved f/2 shape which helps in providing the structural strength of the meniscus.

 

Having worked a 24 inch cellular with a 5/8 inch edge that 1/2 inch of yours will be a big problem if not supported properly.Even then the techniques used on thicker blanks wont work here due to deflection.

 

One reason to slump to a faster F -ratio is due to the distortion induced on the mold itself while at temp and in the annealing cycle.You grind out the wedge distortion in the resluting slumped blank.

 

Ric


Edited by radicell2, 12 January 2019 - 11:12 AM.

  • PrestonE likes this

#36 ckh

ckh

    Surveyor 1

  • -----
  • Posts: 1916
  • Joined: 21 Feb 2015
  • Loc: Arizona

Posted 12 January 2019 - 11:40 AM

 

Fabrication is another story and there's lots of ways to do this correctly, none of which I'll describe (for lack of time mostly). wink.gif

 

 

Figuring without bending is the hard part.  I remember you were working on a thin one (was it a meniscus?) supported by pitch. How did that work out?



#37 hakann

hakann

    Surveyor 1

  • -----
  • topic starter
  • Posts: 1743
  • Joined: 06 Aug 2015

Posted 12 January 2019 - 12:01 PM

I read on Mel's web and back was told if I read it correct, just polished from the slumped f/2. ( not shore )

 

I did some Plop test.

My plano 18” f/4 mirror in Quartz whit a 4” diagonal and 33 mm edge thickness got this numbers whit a 18 pt cell and sling ;
6.25232e-06 PV
1.08961e-06 RMS

 

Now I did run Mel’s plate-glass 25” at 34 mm edge and at f/2.6, but as this is a meniscus mirror so sigatta 15.4 mm + centre at 18.6 mm ( f/2 vs f/2.6  from CAD ) an told of a edge at 12.7 mm ( 1/2” ) whit a 5.5” diagonal and FL at 65” (1651 mm ) so whit a 9 pts cell and I set in ‘sling’, but it’s on a 2 pt edge ( tuff deal for a sling on this design )

 

-I'm not shore if this is or can be a correct way to say a 1/2" meniscus is equal at 34 mm, but Plop has only plano mirrors in the system.
Numbers I got was ;

6.56678e-05 PV

1.70197e-05 RMS

 

If I added twice the centre thickness = 18.6 x 2 = 37.2 mm and sigatta at 15.4 mm = 52.6 mm, and still a 9 pts cell and still in plate glass.

Numbers was ;

4.62211e-05 PV

8.99874e-06 RMS

 

If added pts to 27 in cell ;

6.20409e-06 PV

1.1181e-06 RMS

 

-Now the numbers get equal to my 18" mirror.

 

But now I changed back to the 34 mm but added up to 27 pts in cell ;

1.57628e-05 PV

2.91721e-06 RMS

 

In this test is show number of pts in the mirror cell do has a BIG effect, but it is also the thickness that get the lower readings possible.

But to get decent readings at 34 mm thickness one do need more pts, atleast according to Plop.

 

-Correct me if I made some mistakes in analyse !

 

Note.

It do seems like this design whit a 9 pts cell and a 3.7 mm EP = 525X and 20X/inch or little over 1 mm pupil a good image or high standard star-test.

 

Again, amazing telescope.


Edited by hakann, 12 January 2019 - 12:12 PM.


#38 Pierre Lemay

Pierre Lemay

    Apollo

  • *****
  • Posts: 1330
  • Joined: 30 Jan 2008
  • Loc: Montréal, Canada

Posted 12 January 2019 - 02:52 PM

Having worked a 24 inch cellular with a 5/8 inch edge, that 1/2 inch of yours will be a big problem if not supported properly. Even then the techniques used on thicker blanks wont work here due to deflection.

Ric,

I agree, the parabolization techniques used on thicker blanks must be adapted for this type of extremely thin blank. Normand Fullum had exactly the same comment when I first told him about this crazy project.

 

Grinding, after the f/2.8 curve has been generated in the current f/2.15 curve, and polishing will both be underken MOT. I might do as Mel did and push the mirror through a frame surrounding the glass or, I might core a 3 or 4 inch hole in the center and push using that, which is what I did when I made my 20 inch conical mirror. By grinding and polishing CoC this way, you are almost sure of making it to a fully polished sphere that has no astigmatism. 

 

Parabolizing this thing will be the real challenge. As you noted, the classical method of using sub-diameter tools with the mirror on the bottom is very difficult to do well due to the thin glass and the ease with which it will bend. Here I will carefully follow in Mel’s footsteps, as if I was following someone who knows his way through a mine field. In Mel’s technique, gross parabolization is also undertaken MoT, using a specially designed petal lap on the bottom and CoC strokes. This should bring the mirror to within a wave or two of the final shape. 

 

For the moment I don’t know how I will tackle the last bits of correction to bring the surface to an acceptable parabola. Mel likes to use a full size tool, I feel more confortable with sub diameter ones. I will cross that bridge when I get to it. In the mean time I will practice the parabolization method described above on my 20 inch mirror. I’m not happy withe the f3.9 parabola I produced seven years ago so it’s going back on the polishing machine this spring to be redone. This time I will practice with the MoT/petal polishing lap approach to experiment this method which Mel has successfully used on several fast mirrors. 

 

I could of approached this project from the more rational and proven method of using a thicker blank or a fused cellular blank. Normand Fullum’s shop is only 45 minute drive away. I could have asked him to make a 28 inch technofusion blank, which would have made my life easier.

 

I don’t know if I will succeed or fail. I’m not Mel Bartels. I don’t have his experience, nor his skills in mirror making. But I like a challenge and the prize, if I’m successfull, is an under 100 pound, flat footed, 28 inch telescope. I think it’s worth a try. 

 

If I fail, the birds will enjoy a very expensive bird bath 🐥🐥🐥🐥.


Edited by Pierre Lemay, 12 January 2019 - 02:54 PM.

  • jtsenghas likes this

#39 hakann

hakann

    Surveyor 1

  • -----
  • topic starter
  • Posts: 1743
  • Joined: 06 Aug 2015

Posted 12 January 2019 - 03:06 PM

Pierre,

 

Cool project !

( 28" f/2.8 whit a centre thickness at 3/4" )

 

-Has you look in some of Mel's scopes ?

 

Do you like observe at 1-2 mm pupil, or most lower powers ?

 

A question.

The idea and what you did on your conical mirror.

Will a 28" whit a centre cell in the mid hole at ex 4" and glued hold up ?


Edited by hakann, 12 January 2019 - 03:35 PM.


#40 Pierre Lemay

Pierre Lemay

    Apollo

  • *****
  • Posts: 1330
  • Joined: 30 Jan 2008
  • Loc: Montréal, Canada

Posted 12 January 2019 - 09:58 PM

Do you like observe at 1-2 mm pupil, or most lower powers ?

 

A question.

The idea and what you did on your conical mirror.

Will a 28" whit a centre cell in the mid hole at ex 4" and glued hold up ?

As I mentioned earlier, my 20 inch has optical issues at the moment so I rarely go above 200x. Nevertheless I will observe up to 350X (2mm) with the 20 inch, and a little more with other scopes when the sky here in Quebec allows it but that’s kind of near the upper limit of what we get around here. I haven't tried observing much above 450X, which is a waste of time. I'm convinced my rework of the mirror in the coming months will yield a much better mirror but, alas, the atmosphere in my part of the world will remain unfavourable to high power observing.

 

As for putting in a central perforation in the 28 inch, it's something I'm considering for two reasons:

  • To provide a place on the convex side for my grinding machine to push during grinding/polishing/rough parabolizing with MOT
  • To provide "edge" support during observing. This would not be done exactly like I do on my 20 inch conical mirror. The 28 inch meniscus glass is too thin at its center to hold the whole mirror, contrary to the 20 inch which has a big, central 2 inch thick x 7 inch diameter chunk of glass, after which the mirror starts to taper off to a 1/2 inch edge thickness.

Instead, a central support would keep the meniscus from moving "vertically" as the scope would move away from zenith. A conventional mirror floatation support system (9? 18? contact points) would support the back to spread and support the glass weight against deformations for views closer to zenith, as is done on all, large Newtonian telescopes.

 

I would not glue the meniscus to the central perforation. The "peg" holding the mirror through the perforation would be just a bit smaller than the perforation, small just enough to take into account the worst case differences in the coefficient of thermal expansion between the "peg" material and the low expansion Borosilicate glass (you don't want the "peg" to split the glass on a hot day!). Instead of gluing I would use some kind of clips or a retaining ring, to keep the mirror from falling out of its cell if the mirror box is placed upside down.

 

By replacing the edge supports with a central support, any strain developed around the roughly 3.5 inch central perforation, would find itself in the shadow of the 7 inch minor axis diagonal. This approach was used a few years ago by Mike Linolt for his 20 inch ball scope, apparently quite successfully. I'm sure others have done this as well.


Edited by Pierre Lemay, 12 January 2019 - 10:04 PM.

  • jtsenghas and ckh like this

#41 mark cowan

mark cowan

    Vendor (Veritas Optics)

  • *****
  • Vendors
  • Posts: 9419
  • Joined: 03 Jun 2005
  • Loc: salem, OR

Posted 12 January 2019 - 10:08 PM

Figuring without bending is the hard part.  I remember you were working on a thin one (was it a meniscus?) supported by pitch. How did that work out?

No it's not, but of course it depends on how you do it.  Figuring is the hardest part, support is just the use of some method with mechanical integrity for the duration of that stage.  

 

FWIW the thin meniscus pitched to a thicker carrier works fine, since it's a very straightforward technique,  You just have to get the pitch at the right hardness to relieve stress over time yet not soften under ordinary polishing loads, for whatever temperature you're working at.


Edited by mark cowan, 12 January 2019 - 10:12 PM.


#42 hakann

hakann

    Surveyor 1

  • -----
  • topic starter
  • Posts: 1743
  • Joined: 06 Aug 2015

Posted 13 January 2019 - 05:31 AM

Pierre,

Maybe a centre perforation can be a idea but ex Mel did not do that.
I’m amazed he get away whit the 9 pts cell ( has you visit him and look in the EP at 1-2 mm pupil ? )

 

One question of thickness.
You both will has a thicker centre but both a end thickness around 1/2” or 12.7 mm.
If you has a near 19 mm central thickness it is around 6 mm thicker and Mel has around 3 mm thicker, but you has also 3 more inch on the diameter.

 

A question, the 2 different curves might be a good idea here but why go this thin ?
Even in plate-glass that has over 9 in ppm/C will cool-off pretty fast even if it was twice the thickness ex Mel made and you has Pyrex that has around 2.5 in ppm/C.
-So why did you go that thin instead of at least a centre at 1.5” ( or thicker ) ?

 

As I see it, ”trick” here might be the fast curves, but it’s not the thin thickness.
Little thicker means still pretty light and fast cool-off but more stiff, as that is natural.

My calculations I did ( see above ) from Plop ( sigatta + centre thickness ) was at 25” size and 2.6 focal.
That number did not come off that great, but if I added twice the thickness and added up for 27 pts - I got 1 nm RMS.
Thickness was then over 50 mm ( or 2” ) for a plano at f/2.6 and 27 pts in plate-glass ( use a sling )

As I see it, if Plop is correct and we has a 2” thickness but remove mass at edge ( make it meniscus and equal thickness ) we will NOT has a stiffer structure than a plano.
Maybe ’as’ stiff - but not stiffer !
I would like see a FEM on this if it really can be as stiff really as a plano.

 

What make Plop say 1 nm RMS was 2” thickness and numbers of cell pts = 27.

If only 9 pts and 34 mm thickness instead of 52.6 mm thickness the RMS readings was = 17 nm RMS.

I know 1 nm RMS is real nice, but I don’t know how far one can go in higher power whit 17 nm RMS.

 

So either this design will do some kind of magic out of a plano just to remove the end mass, but if not - we need 2” equal thickness and 27 pts to get 1 nm RMS.

 

Maybe some has inputs.


Edited by hakann, 13 January 2019 - 08:52 AM.


#43 Pierre Lemay

Pierre Lemay

    Apollo

  • *****
  • Posts: 1330
  • Joined: 30 Jan 2008
  • Loc: Montréal, Canada

Posted 13 January 2019 - 07:32 AM

A question, the 2 different curves might be a good idea here but why go this thin ?
Even in plate-glass that has over 9 in ppm/C will cool-off pretty fast even if it was twice the thickness ex Mel made and you has Pyrex that has around 2.5 in ppm/C.
-So why did you go that thin instead of atleast a centre at 1.5” ( or thicker ) ?

Weight. 



#44 hakann

hakann

    Surveyor 1

  • -----
  • topic starter
  • Posts: 1743
  • Joined: 06 Aug 2015

Posted 13 January 2019 - 08:46 AM

Ok, that’s a factor.
Ex a Dream cast 6;1 cellular ( very stiff & will change very little on the gravity effect ) is around 16 -17 kg on a 25”.
They can do them lighter but even in this double conical design has a limit to retain real ’stiff’.
I guess a 1/2” meniscus are around 11-12 kg.
So say a 1.5” - 2” is of course more mass to carry.
But is weight more important than nm RMS, yes it can be my guess.

 

I did run your mirror on Plop and based it out from a plano of 35 mm ( sigatta 16 mm + 19 mm centre thickness = 35 mm ) and got near 7 nm RMS error whit a 18 pt cell.

Whit a 9 pts cell it is 23 nm RMS error.

 

As I said above, maybe the meniscus shape do some kind of 'magic' over the full plano in design. ( at the same total thickness vs the Plop program )


Edited by hakann, 13 January 2019 - 09:08 AM.


#45 tommm

tommm

    Messenger

  • -----
  • Posts: 463
  • Joined: 16 Dec 2015

Posted 13 January 2019 - 10:43 AM

You can slump a flat piece of glass and have different radii of curvature on the front and back. That's kind of what Mel did on his 25 inch and the approach I will be using on my 28 inch, but in reverse to what you are suggesting. Both the 25 and the 28 inch blanks were slumped on a mold that had a much shorter radius of curvature than the intended final RoC. Mel's ¾ inch thick x 25 inch diameter plate glass blank was slumped by David Davis to an f/2 curve whereas my ¾ inch thick x 28 inch diameter Borofloat blank was slumped by Normand Fullum to an f/2.15 curve.

 

Here's a picture of my 28 inch blank on its slumping mold just before going in Norm's annealing kiln:

 

attachicon.gif 28 inch on slumping mold.jpg

 

And a picture of the 28 inch blank as it came out:

 

attachicon.gif Slumped 28 inch photo2.jpg

 

The grinding procedure is then to lengthen the RoC on the front (concave) surface by grinding it back to the intended focal length (there are several ways of doing this, including a diamond cup. I haven't decided yet how I will do it). Mel stopped at f/2.5 and I'm aiming for f/2.8 or, maybe, f/2.7. This is done by thinning the periphery of the front surface by about 3/16 inch near the edge (for my 28 inch) and gradually less towards the center. The result is that the center of the blank more or less maintains its original ¾ inch thickness and the edge is now down to near ½ inch thick. The back of the blank maintains its strongly curved f/2 shape which helps in providing the structural strength of the meniscus.

 

The other thing that contributes to a rigid mirror blank is the lightness of it. My 28 inch meniscus mirror blank currently wheigs about 37 pounds. By the time I've finished grinding in the longer RoC it will have lost almost five pounds. Most of that weight loss will be NEAR THE EDGE. The final weight should be around 32 pounds, which is the same weight my 20 inch conical blank weighs. Combine this lightness with the strongly curved glass and you have a rigid combination that can better maintain it's shape in various positions. That probably explains why Mel was able to get away with a simple 9 point floatation cell for the back and two simple pegs at 90 degrees for edge support!

 

A 1.25 inch thick f/5 25 inch, constant thickness meniscus might have a much harder time maintaining it's shape in those conditions. But it works for very thin, strongly curved blanks because you are using the benefit of both those characteristics to your advantage.

Pierre, how do you plan to support the blank for grinding/figuring, or do you plan to only do MOT? Mel commented on how difficult it is to figure these large fast mirrors, and struggled for a long time figuring his 25".



#46 mark cowan

mark cowan

    Vendor (Veritas Optics)

  • *****
  • Vendors
  • Posts: 9419
  • Joined: 03 Jun 2005
  • Loc: salem, OR

Posted 13 January 2019 - 11:58 AM

 But is weight more important than nm RMS, yes it can be my guess.

It may be your "guess" but it is a wrong analysis, even though you can run PLOP yourself for more complex cells. You may be running PLOP wrong as well.  I get 4.5nm RMS error for a flat back Pyrex 635x35mm f/2.6 and 18-point cell in the quick analysis (checking "allow angles to vary") and 3,6nm for Z88 3D. (And it's possible to optimize the cell beyond that if you know how.)

 

As I said earlier, Mel can "get away" with a 9-point cell thanks to the apparent fact that the meniscus responds to change in support angles with refocusing to a new optimal figure, not a distortion as in a flat back mirror.  This is strictly down to the geometry, and though it may also be due to Mel's 25"'s particular cross section it has been seen elsewhere for various sorts of support. Which it is remains to seen but Mel is a very picky mirror maker and observer, so if he is seeing it it's actually happening, for some reason.  shrug.gif

 

Even if that weren't the case, a more complex cell is quite capable of supporting a very thin meniscus without significant RMS error.

 

It's never quite as simple as a direct trade off between "weight...and RMS" because the cell support is quite complicated in how it functions in terms of the actual optical figure and how that trade off can be improved - in the case of the thin menisci mirrors some of these aren't obvious at all, though the considerable difficulty remains in fabrication and testing, not support in eventual use.


Edited by mark cowan, 13 January 2019 - 12:21 PM.


#47 Pierre Lemay

Pierre Lemay

    Apollo

  • *****
  • Posts: 1330
  • Joined: 30 Jan 2008
  • Loc: Montréal, Canada

Posted 13 January 2019 - 12:14 PM

Pierre, how do you plan to support the blank for grinding/figuring, or do you plan to only do MOT? Mel commented on how difficult it is to figure these large fast mirrors, and struggled for a long time figuring his 25".

Like I mention in post 40, I intend to do 95% of the work with MOT. I hope to get to about 1 wave of the final parabola that way. Then, I will do TOT using sub diameter tools with very little pressure. I will have to find an acceptable way of supporting the glass without it bending for this last parabolization step.



#48 tommm

tommm

    Messenger

  • -----
  • Posts: 463
  • Joined: 16 Dec 2015

Posted 13 January 2019 - 12:44 PM

Thanks Pierre. 

 

The link below gives some further info on the process of slumping glass that the OR folks are using.  Note the date on the report - 2012 - Mel had his 25" blank "in hand" at that time and others had even larger ones. Lot of learning there.

 

https://www.eugeneas.../Io_2012-12.pdf



#49 hakann

hakann

    Surveyor 1

  • -----
  • topic starter
  • Posts: 1743
  • Joined: 06 Aug 2015

Posted 13 January 2019 - 12:48 PM

Mark,

Yes you helped me on my 18" cell to modified from the automatic mode at 1 nm RMS error to 0.8 nm RMS.
So yes, I guess it can be better from what I calculated in Plop at automatic setting, but newer less it was not much.

( I will try to get the low number you got, as 4.5 nm RMS on a 635 x 35 mm is ok )

 

Now the cell was made from a FEW/FEA calculation so it can hold twice diffraction limit.
I has no clue on mirror specs, but I hope the cell will hold up.
The arms went to be standing 40 x 12.7 mm thick, and equal side 6.3 mm thick triangles in a pocket milled 7075 cell whit ceramic bearings at both side of the arms ( wiffle tree edge )

 

You said the shape is strictly down to geometry but where is the math behind it ?

Plop can't do the meniscus more than do what I did here converting it to a plano.
From what I think anyway vs sigatta plus centre thickness gets a plano thickness, but when remove the end mass it will be less stiff, but ok a good cell can handle a thin mirror, but then we has the edge support to.
Mel get away whit a 9 pts cell and as I said, it’s kind of amazing no matter how we see it.

 

But it should be real nice see a FEM on that one.
In the real world the EP will tell.
-I saw it at 92X, 3.68X per inch of power. ( just great )


Edited by hakann, 13 January 2019 - 03:36 PM.


#50 ckh

ckh

    Surveyor 1

  • -----
  • Posts: 1916
  • Joined: 21 Feb 2015
  • Loc: Arizona

Posted 13 January 2019 - 02:52 PM

No it's not, but of course it depends on how you do it.  Figuring is the hardest part, support is just the use of some method with mechanical integrity for the duration of that stage.  

 

FWIW the thin meniscus pitched to a thicker carrier works fine, since it's a very straightforward technique,  You just have to get the pitch at the right hardness to relieve stress over time yet not soften under ordinary polishing loads, for whatever temperature you're working at.

That's the technique they use to make ultra thin (~2 mm thick) secondardies for adaptive optics. For the MMT adaptive secondary they grind the mirror to the support surface and then used a 100 micron layer of Gugolz #73.  They determined that such as layer was thin enough to provide the rigidity they required under polishing forces. That's pretty thin, 0.1 mm. Since the mirror was so thin, it had virtually no self-support.

 

It's not pratical for an amatuer to use a thick glass support due to cost, but maybe hydrostone would work. 


Edited by ckh, 13 January 2019 - 02:55 PM.

  • Augustus likes this


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