Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

Telescope mirror geometry

Please login with a confirmed email address before reporting spam

Hi folks I am new to COMSOL. Need to draw a geometry of telescope mirror of paraboloid shape in COMSOL. It should be a short cylinder of thickness 15cm and diameter 1m. The top face of the cylinder should have parabolic/spherical shape with radius of curvature of 10m. Means the reflecting surface of the mirror should be part of sphere whose diameter 5m. I know it must be pretty straightforward and perhaps trivial but I still don't have the hang of COMSOL. Please respond.

Great Thanks in advance
-Simon

8 Replies Last Post 20 janv. 2010, 22:16 UTC−5
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 14 déc. 2009, 11:09 UTC−5
Hi

I assume you men a a paraboloid of revolution of focal length F=5[m], hence a central radius of R=2*F=10[m].
Now in COMSOL the geometry curves are based on Weighted Bezier Splines, the 2nd order one for weight = 1 are in fact a parabola expression (see command.pdf V3.5a Ch4 "Advanced Geometry topics" "Rational Bezier Curves", pp516-524, for more on Bezier cures try
wiki or
www.cs.mtu.edu/~shene/COURSES/cs3621/NOTES/spline/NURBS/RB-conics.html or jtauber.com/blog/2008/11/08/from_focus_and_directrix_to_bezier_curve_parameters/ or mathworld.wolfram.com/BezierCurve.html or ...)

a 3 point (P0,P1,P2) 2nd order Bezier curve with weight factors=1 are in fact the startingpoint P0 and ending point P2 of a parabola with tangeant of P0 along the line P1-P0 and at P3 along the line P2-P1 if you align these points correctly. Typically if the apex of the prabola is P0=0,0, then the next two points could naturally be (1,0) and (2,4) for a paraboly y=x^2 passing by (0,0) (1,1), (2,4) with a tangeant direction at (0,0) of (1,0) and a tangeant at (2,4) passing by (1,0) or slope of 1/4, hence with a parabolic axis along (1,0) and a focal length of "1/2". Now you need to scale this correctly for your case, see the article of "jtauber" above I believe it's the closest to your demand.

So for COMSOL I would make a 2D axial-symmetric shape, and the revolve it into 3D. Now, where to put the points?
You want a base of 0.5m in radius along r and 0.15m along z (?) and a parabola of focal length 5[m].
This means 2*5[m] hight (z) at r=10[m] or 0.1[m] at r=1[m] or 0.1/4=0.025[m] at r=1/2=0.5[m] hat is the rim height difference of the paraboly is some 25 mm. The slope at r=0.5m is 1:10 (2*0.5/10) this corresponds then to the three points Apex: (0,0) centre control point (0.25,0) and rim: (0.5,0.025), all with weights = 1. and the remaining 0.125[m] of your cylinder extending in -Z direction

Pls check carefully I might have missed a scale factor, as I went through it rather quickly

Good luck
Ivar
Hi I assume you men a a paraboloid of revolution of focal length F=5[m], hence a central radius of R=2*F=10[m]. Now in COMSOL the geometry curves are based on Weighted Bezier Splines, the 2nd order one for weight = 1 are in fact a parabola expression (see command.pdf V3.5a Ch4 "Advanced Geometry topics" "Rational Bezier Curves", pp516-524, for more on Bezier cures try wiki or http://www.cs.mtu.edu/~shene/COURSES/cs3621/NOTES/spline/NURBS/RB-conics.html or http://jtauber.com/blog/2008/11/08/from_focus_and_directrix_to_bezier_curve_parameters/ or http://mathworld.wolfram.com/BezierCurve.html or ...) a 3 point (P0,P1,P2) 2nd order Bezier curve with weight factors=1 are in fact the startingpoint P0 and ending point P2 of a parabola with tangeant of P0 along the line P1-P0 and at P3 along the line P2-P1 if you align these points correctly. Typically if the apex of the prabola is P0=0,0, then the next two points could naturally be (1,0) and (2,4) for a paraboly y=x^2 passing by (0,0) (1,1), (2,4) with a tangeant direction at (0,0) of (1,0) and a tangeant at (2,4) passing by (1,0) or slope of 1/4, hence with a parabolic axis along (1,0) and a focal length of "1/2". Now you need to scale this correctly for your case, see the article of "jtauber" above I believe it's the closest to your demand. So for COMSOL I would make a 2D axial-symmetric shape, and the revolve it into 3D. Now, where to put the points? You want a base of 0.5m in radius along r and 0.15m along z (?) and a parabola of focal length 5[m]. This means 2*5[m] hight (z) at r=10[m] or 0.1[m] at r=1[m] or 0.1/4=0.025[m] at r=1/2=0.5[m] hat is the rim height difference of the paraboly is some 25 mm. The slope at r=0.5m is 1:10 (2*0.5/10) this corresponds then to the three points Apex: (0,0) centre control point (0.25,0) and rim: (0.5,0.025), all with weights = 1. and the remaining 0.125[m] of your cylinder extending in -Z direction Pls check carefully I might have missed a scale factor, as I went through it rather quickly Good luck Ivar


Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 15 déc. 2009, 02:16 UTC−5
Thanks a lot Ivar. It was really that simple. In fact, simpler than drawing a spherical mirror of similar dimensions. I wish I had paid more attention to conic section in my high school-:)

-Simon
Thanks a lot Ivar. It was really that simple. In fact, simpler than drawing a spherical mirror of similar dimensions. I wish I had paid more attention to conic section in my high school-:) -Simon

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 12 janv. 2010, 03:26 UTC−5
Hello Ivar,

I just created a a parabola, like you suggested before: Bezier 2nd degree, with the points (0/0), (1/0), (2/4), and the weight of 1. Now i have y(x)=x^2. How can i create a parabola with the formula: y(x)=ax^2, e.g. with a=1/44!

I could really need some help on this. Thanks a lot!

Moritz
Hello Ivar, I just created a a parabola, like you suggested before: Bezier 2nd degree, with the points (0/0), (1/0), (2/4), and the weight of 1. Now i have y(x)=x^2. How can i create a parabola with the formula: y(x)=ax^2, e.g. with a=1/44! I could really need some help on this. Thanks a lot! Moritz

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 12 janv. 2010, 12:25 UTC−5
Hi,
I used the scale function to scale it up/down, else you need to calculate the derivatives and user the correct point values, the three chosen were the first natural number I could thik of
Good luck
Ivar
Hi, I used the scale function to scale it up/down, else you need to calculate the derivatives and user the correct point values, the three chosen were the first natural number I could thik of Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 18 janv. 2010, 10:28 UTC−5
Hi Ivar,

I think the construction of the parable succeded, but now I have another problem. Have look in the attachment! I want to create a surface, which looks like the one in the attachment. One of the curves is the parabola, the other a curve, which is quasi parallel to the parabola. This surface needs to be extruded.

Do you have an idea, how to construct the second curve? In a normal CAD-Programm I would use the "offset" or "wall-thickness"-Function.

Thanks for your help!
Regards,
Moritz
Hi Ivar, I think the construction of the parable succeded, but now I have another problem. Have look in the attachment! I want to create a surface, which looks like the one in the attachment. One of the curves is the parabola, the other a curve, which is quasi parallel to the parabola. This surface needs to be extruded. Do you have an idea, how to construct the second curve? In a normal CAD-Programm I would use the "offset" or "wall-thickness"-Function. Thanks for your help! Regards, Moritz


Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 19 janv. 2010, 14:31 UTC−5
Hi

It depends on the final precision you want, I would suppose, but if you make the two sides of the wall with a line of the desired wall thickness and then link the two points with a 3 point spline (using the two defined) I beleive you can drag a reasonable shape with your mouse. Or you could try to select your parabola, doa Cntrl-C Cntr-V (or Cut&Paste) and then define the offset vector at a given angle (in one or two steps with further "move") And then finally reconnect the sides.

Or import it from a CAD tool ;)

Good Luck
Ivar
Hi It depends on the final precision you want, I would suppose, but if you make the two sides of the wall with a line of the desired wall thickness and then link the two points with a 3 point spline (using the two defined) I beleive you can drag a reasonable shape with your mouse. Or you could try to select your parabola, doa Cntrl-C Cntr-V (or Cut&Paste) and then define the offset vector at a given angle (in one or two steps with further "move") And then finally reconnect the sides. Or import it from a CAD tool ;) Good Luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 20 janv. 2010, 04:53 UTC−5
Thanks Ivar,

I think I try importing a .dxf file. This might be the easiest way!

Ragards,
Moritz
Thanks Ivar, I think I try importing a .dxf file. This might be the easiest way! Ragards, Moritz

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 20 janv. 2010, 22:16 UTC−5
Hi

understand well your choice, but if you are working in 2D note that the COMSOL geometrical primitives are rather powerfull and not that complex to learn to use via matlab, in 3D it becomes more combersome and length to write out, but still possible for simple geometries.

Furthermore in 2D you can run parametrised geometries within the GUI

Good luck
Ivar
Hi understand well your choice, but if you are working in 2D note that the COMSOL geometrical primitives are rather powerfull and not that complex to learn to use via matlab, in 3D it becomes more combersome and length to write out, but still possible for simple geometries. Furthermore in 2D you can run parametrised geometries within the GUI Good luck Ivar

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.