Flatening the details on a cone as well as the shape.

[sreinmann]

My newest adventure in designing is TDRS-M. This image is actually from a pervious generation, but it'll suffice. You can see from this image that there are three dish antennas. Each has a lattice structure. Flattening these "cones" is relatively simple. And for the sake of the unfamiliar, I'll direct you to this wonderful site (Juex Et Mathematiqueswhich helps a great deal in calculating a flattened cone!

The problem comes to adding the details. How does one plan the flattening of the lattice so that it "connects" when the edges of the cone (annulus) are brought together?

[maurice]

Well this is a part of it, ignore the yellow they're just constructin aides.
How it could be done in MSPaint I've no idea.
Might be possible using https://www.turbocad.com/content/doublecad-xt-v5

[sreinmann]

So, how did you put this together Maurice? Did you measure the radials and then perform math to calculate their new position on the flattened part?

[maurice]

... and then perform math to calculate....

You know Scot i find that a extraordinary phrase but I've seen it used by others so I assume it's either a cultural thing or a sign of nervousness about mathematics in general. I would just settle for "... and then calculate.. "
No matter, in this case no intricate calculations were made or "math performed" to achieve the result.
It was done solely by using simple, very simple, geometric drawing in a competent CAD program.
And yes it can be done in Doublecad, have you downloaded your copy yet?

[sreinmann]

In my case its said with a tongue in cheek.

I've begun working with DoubleCAD and I'll freely admit I'm struggling up that first hurtle of learning a new program. Last time I had my hands in a cad program was in 1987, things have changed. I've surfed through youtube for some beginner's guides and googled importation of blender files (where the source of my graphic is coming from). The latter I haven't had much success with, so I'm guessing that I'll need to build the part in cad from scratch.

Anyway, as I've said before, learning new things is a worthy endeavor. At least the minute examination of my source has lead me to notice that its layout is not dimensionally true - meaning the lines don't line up. one step forward ...

[zubie]

For a regular lattice pattern, it shouldn't be that difficult. Once you know the number of longitudinal sections you can calculate the angle covered by same in the flattened cone section. For example a pattern that repeats over 15deg wedges yields 24 wedges. If the flattened cone covers 270 deg, then each wedge there should cover 11.25deg which is the result of 270/24. The concentric pattern is trickier only in that you have to determine the individual radii for the pattern - regular or mapping spherical to cone. In either case, the radii is centered at the flattened cone sections center so it's a fairly simple overlay like shown on maurice's picture. I don't know if this can be done in paint because to do this easily you need layer support and ability to rotate objects with some precision, ideally about user defined centers (thinking gimp, photoshop, inkscape etc)

A similar problem happened to me when trying to map a mercator projection onto an existing sphere that was flattened into petals. I did it in gimp using incremental selections and then using perspective distorts to fit into the petals in individual long×lat sections. There was no math involved, but it was time consuming. Of course if you can skin the shape before flattening, that tends to solve the problem right off the bat

[maurice]

Well I suppose that's a reasonable precis of what I did actually do but a quick shufti at source material would have shown that there are actually 32 segments.

Scot, never trust salesmen they are unlikely to fully disclose difficulties in their opening gambit.
Yes all Cad has a steep starting curve but the the benefits in capability and accuracy are well worth it (imffho). Also Doublecad is suprisingly competent as a graphics program.

Attached are paterns for the inside, outside and base of the antenna which might help along. It's on A4 but should /might print on letter.

[sreinmann]

Just a quick update for anyone still interested: Zubie nailed it for me. Segmenting the cone by dividing the number of points by the total angle has worked well.

Now, I'm working in MS Publisher which is not ideal. I'm sure any software may be capable, but this is what I have and know. Cons of constructing this way? Time. My completed sub-project, three conics with 32 lines and 32 curves to plot on each, angling and aligning (I had no perfect source images so had to build from scratch) took me 9 hours too the first prototype.

Pros? This exercise at least can show many of you that you don't need innate knowledge, software or skills (I sure don't) to design a model. Just a willingness to ask for advise from an audience of great people (thx again Maurice and Zubie) and an exchange of time.

Here's the flattened design. You'll have to take my word about the alignment being correct ... my kids exclaimed, "FLYING SAUCERS FOR ME!" and that was that. :D