Jump to content


Photo

Will 4 Point Cell Work?

  • Please log in to reply
9 replies to this topic

#1 RodShea

RodShea

    Sputnik

  • -----
  • Posts: 47
  • Joined: 01 Nov 2009

Posted 23 January 2013 - 11:01 PM

Hi,

I'm designing a telescope around a mirror I have. Looking at Mark Holm's PLOP pages, I'm guessing a 4 point cell might work, but he doesn't evaluate the exact mirror I have. Even reading the User Manual pointed to from David Lewis's PLOP page, I cannot figure out how to model the cell I want.

Could someone out there model it for me? I'm thinking of 4 points in a single circle, 2 are fixed, and the other 2 are on a single teeter-totter bar. The mirror is Pyrex, diameter is 332mm, thickness is 41mm, Fl is 1500mm, and the sec is 66mm in diameter.

Alternatively, is there other info out there about using the program? I can't seem to figure out how to describe this cell in the current version of PLOP.

Thanks very much,

Rod

#2 RodShea

RodShea

    Sputnik

  • -----
  • Posts: 47
  • Joined: 01 Nov 2009

Posted 23 January 2013 - 11:14 PM

Answered my own question... Looks likeI can just modify the inputs in 4point.gr, and run PLOP. Looks like I get an RMS error of 3.24663e-06, which looks like plenty good enough, yes?

Thanks,

Rod

#3 Nils Olof Carlin

Nils Olof Carlin

    Surveyor 1

  • *****
  • Posts: 1624
  • Joined: 26 Jul 2004

Posted 24 January 2013 - 04:14 AM

Here is one that I made, for a 6":
First from the mirror side

Attached Files



#4 Nils Olof Carlin

Nils Olof Carlin

    Surveyor 1

  • *****
  • Posts: 1624
  • Joined: 26 Jul 2004

Posted 24 January 2013 - 04:16 AM

From the bottom. The collimation screws are spring loaded against the frame.

Attached Files



#5 RodShea

RodShea

    Sputnik

  • -----
  • Posts: 47
  • Joined: 01 Nov 2009

Posted 24 January 2013 - 03:14 PM

Nils,

Thanks very much for the reply. Yours is similar to what I had in mind. I like the extended cutouts at the upper (in the picture) end in the lengthways bars.

Do you think it would matter if I put the fixed mirror supports on the left and the teeter-totter on the right? It would simplify my construction if I do it that way.

Put differently, is there an advantage to having the cell oriented the way yours is, or could I rotate the mirror supports 90 degrees, as long as the edge supports stay on the bottom?

Thanks,

Rod

#6 Nils Olof Carlin

Nils Olof Carlin

    Surveyor 1

  • *****
  • Posts: 1624
  • Joined: 26 Jul 2004

Posted 25 January 2013 - 02:13 PM

Put differently, is there an advantage to having the cell oriented the way yours is, or could I rotate the mirror supports 90 degrees, as long as the edge supports stay on the bottom?


I can't imagine it would make any difference in performance. Possibly I thought mounting the support bearings would be simpler the way I did it - I didn't consider the "sideways" alternative, but it would have been as easy reaching the collimation screws.

Nils Olof

#7 Nils Olof Carlin

Nils Olof Carlin

    Surveyor 1

  • *****
  • Posts: 1624
  • Joined: 26 Jul 2004

Posted 25 January 2013 - 02:19 PM

Come to think of it - one reason for a 4 pt support (over 3 pt) is that I can place the supports on a larger radius than strictly optimum. I have the idea (right or wrong) that this improves stability, and also decreases the distance between back and edge supports, giving a shorter lever arm in case of friction between support and edge (with ball bearings, friction is negligible anyway). No hard data behind this, it just felt better...

Nils Olof

#8 derangedhermit

derangedhermit

    Apollo

  • -----
  • Posts: 1158
  • Joined: 07 Oct 2009
  • Loc: USA

Posted 25 January 2013 - 03:53 PM

Come to think of it - one reason for a 4 pt support (over 3 pt) is that I can place the supports on a larger radius than strictly optimum. I have the idea (right or wrong) that this improves stability, and also decreases the distance between back and edge supports, giving a shorter lever arm in case of friction between support and edge (with ball bearings, friction is negligible anyway). No hard data behind this, it just felt better...

Nils Olof

That's why I've favored 6-point over 3 point. PLOP has always shown IIRC a location of about 0.4 radius for 3-point cells, and 0.6 radius for 6-point. It may not sound like much, but when you overlay the support points on a circle of the size of the mirror, the difference in the distribution of support points is clearer.

BTW, is there a rule of thumb for 4-point cells like those for 3 and 6 point? IIRC the 0.4 and 0.6 were rather insensitive to being a little off, across a wide range of mirror size and thickness.

#9 Nils Olof Carlin

Nils Olof Carlin

    Surveyor 1

  • *****
  • Posts: 1624
  • Joined: 26 Jul 2004

Posted 25 January 2013 - 04:26 PM

BTW, is there a rule of thumb for 4-point cells like those for 3 and 6 point? IIRC the 0.4 and 0.6 were rather insensitive to being a little off, across a wide range of mirror size and thickness.


I just ran a simple mirror with 4 points (dia 160 mm, thickness 15 mm - not so far from the one I made the cell for). Relative radius vs deformation in nm:

0.55 1.2 (optimized, others below fixed)
0.6 1.3
0.65 1.7
0.7 2.1
0.8 3.0
0.9 4.0
0.99 5.0

I wouldn't mind using 0.75 or 0.8 here, which I may well have done(was some years ago!) even though a bit from optimum.
You can always find the optimum, then try fixed values to see what is acceptable. 2.5 nm, or 4 nm RMS are reasonable goals.
BTW the same modeled mirror for 3 pt optimizes at 0.41, and for 6 pt at 0.58, so I guess 0.55 for 4 is a good rule of thumb.

#10 RodShea

RodShea

    Sputnik

  • -----
  • Posts: 47
  • Joined: 01 Nov 2009

Posted 25 January 2013 - 06:44 PM

Thanks for the thoughts. I'm planning a cell shaped like pi, that is, a T with 2 legs. It is a tailgate design to facilitate removing the mirror for travel. The axis of rotation will be thru the top/transverse part of pi. The mirror edge supports are on the same part. The two fixed mirror bottom supports are on one of the legs, and the teeter-totter is beside or inset into the other, and probably a fan will be inset between the legs. The collimation points are "outboard". They will move one end of the top, and a short bar connecting the two legs. The other end of the top will be fixed, or at least not so easily adjusted once its optimal position is found.

with this configuration, it seems a little easier to arrange the mirror supports with 2 fixed on one side, instead of adding crossmembers.

Thanks again for your thoughts on the design, and the "bigger square" calculations. It's interesting and very reassuring that the sweet spot is as big as it is.

Rod






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics