11 replies to this topic

[Fivemileshigh]

I am planning for my next project, and just for testing, I put in the following data for a hypothetical mirror: 25" dia, f/3, 1" (!) thick, 7in secondary, no central hole and 27 point cell. It came back with an RMS error of 3.94e-06. That seems quite unbelievable. Any thoughts please?

 

Many thanks!

 

 

[Arjan]

Why do you consider this quite unbelievable?

[Fivemileshigh]

Because it seems an excessively thin mirror, it would be 12mm thick at the centre given a 13mm sagitta. Using the cruxis edge support calculator, with a 45deg whiffle tree, the Strehl reduction at the horizon is 0.09. All signs point to it being an impracticality thin mirror, which I already assumed, yet PLOP still says it’s fine with a 27 point cell. Is it because PLOP only calculates the case with the mirror pointing at the zenith? Thanks.

[triplemon]

Plop doesn't ask about side support or tilt angles, so yes, it obviously does not worry about any effects from tilting the mirror.

 

But 27 point supports are overkill for many cases. Look at Bartel's thin meniscus mirrors and see how simple cells he gets away with. So by my gutt feeling this result does not ring any alarm bells.

Edited by triplemon, 23 May 2024 - 04:15 AM.

[hamishbarker]

the thinness could also have increased risk in terms of astigmatism due to insufficiently even support during polishing and figuring, let alone the edge support issues already posted above.

[davidlewis]

Plop uses the plate model developed by Toshimi Taki and it will have somewhat larger error when the thickness varies greatly over the mirror. You can use Z88 to verify using a full 3D model that should be accurate. If you do let us know.

Also, Z88 lets you evaluate the effect of tilt using a glued cell or sling.

 

I just ran Z88; for my assumed 625 mm f/1875 with 100mm secondary Plop gives 7.55e-6 P-V and Z88 gives 4.51e-6. So Plop is pessimistic.

Edited by davidlewis, 23 May 2024 - 07:20 AM.

[duck]

remind me of the units

[MeridianStarGazer]

I don't know what numbers are good, but Dobson had fun with 25"x1", and if you are considering it too, then that means 3/8" glass is likely usable for 8" mirrors. On edge, the 25" is 15x as floppy.

Edited by MeridianStarGazer, 23 May 2024 - 12:18 PM.

[Lucullus]

remind me of the units

+1

What units is PLOP using? ...e-6...what? Millimeters, meters?
 

Edited by Lucullus, 23 May 2024 - 02:50 PM.

[CrazyPanda]

Millimeters. Here's the manual:

https://github.com/d...sers Manual.pdf

On page 32 it says "All units used in Plop are millimeters (mm.)"

So in the above number, it's 3.94 nanometers of error on the mirror surface, and you double it to get wavefront error, so about 8nm of error or 550/8 = error of 1/68.75th green light (550nm wavelength)

Edited by CrazyPanda, 23 May 2024 - 03:43 PM.

[Fivemileshigh]

Thank you all for your input. So it seems a 27 point cell can handle as thin a mirror as one dares to make (at zenith at least). Edge support is a different story.

[MeridianStarGazer]

Thank you all for your input. So it seems a 27 point cell can handle as thin a mirror as one dares to make (at zenith at least). Edge support is a different story.

Mel Bartels dared go thin at the edge. I think he said it gets some astigmatism though.

Some people do 24" flat back f5 3/4" thick edge. They say it potato chips, but is a cheap light bucket.