+- LDraw.org Discussion Forums (https://forums.ldraw.org)
+-- Forum: Models and Parts (https://forums.ldraw.org/forum-18.html)
+--- Forum: Parts Authoring (https://forums.ldraw.org/forum-19.html)
+--- Thread: Help with Duplo umbrella 40554 (/thread-28758.html)
Help with Duplo umbrella 40554 - Peter Blomberg - 2025-04-06
I started looking at the Duplo umbrella 40554 because of the nice mathematical shape. However, now I'm clueless of how to proceed.
I have the flat shape in a subfile and have scaled it to follow a hi-res arch of the right size for the outer skin of the umbrella.
The flat shape is just a bunch of lines; 96 line segments. Together they make a terrain type of height map.
umbrella1.png (Size: 9.21 KB / Downloads: 81)
The vertex density will need to be reduced (particularly in the middle) without loosing the rigde effect along the eight "spokes" of the umbrella. The lines (or equivalent point mesh) will need to be converted into triangles.
Any suggestions?
RE: Help with Duplo umbrella 40554 - Rene Rechthaler - 2025-04-06
the easiest way here to reduce filesize would be cutting it down to an eigth (0-45°), then use that subfile 8x in the main file...
(that also saves the amount of work)
also looks like the innermost one should be nearly round, the one farther one could be reduced to every second vertex.
(or go something like 12-12-8-8-6-6-4-2 to keep a constant distance)
this part also has an older variant without the stop ring (2164), but the top seems to be the same (except for the dome at the upper end)
did you compare it to the one from digital-bricks? http://www.digital-bricks.de/en/file.php?part=40554
http://www.digital-bricks.de/en/statistic4.php
that one looks way coarser than yours, but the dimensions may be better...
also, what about the orientation? tips/ribs to 0° or side to 0°?
RE: Help with Duplo umbrella 40554 - Gerald Lasser - 2025-04-06
I had a similar issue with the Zeppelin container 72490. To make it in a way that a render will put out a proper shape and shading.
I exaggaraterd the cross section slightly to achive the effect and both LDView and Studio will look ok.
I will post more tomorrow.
BTW I have attached the BI mesh, whcih is quite sub-par...
RE: Help with Duplo umbrella 40554 - Peter Blomberg - 2025-04-07
(2025-04-06, 22:07)Rene Rechthaler Wrote: looks like the innermost one should be nearly round...
did you compare it to the one from digital-bricks?
The ridgeness effect is visible on the physical part all the way to the dome on top. It should therefore not be round near the top. Quite the contrary, one should avoid having a circle at the top (except under the dome). This is why the digital bricks and BI mesh versions (perhaps the same) look so wrong. They correct the roundness too far out the shape. Exaggeration seems to be used also in the digital bricks version. They actually went with a line on the ridge.
umbrella3.png (Size: 22.17 KB / Downloads: 57)
umbrella4.png (Size: 25.25 KB / Downloads: 56)
Looking at the Zeppelin (awesome by the way), I see that quads are to be preferred over triangles and that spacing need not be even.
Here is a draft. Will see if it works out or not.
umbrella2.png (Size: 16.24 KB / Downloads: 57)
RE: Help with Duplo umbrella 40554 - Rene Rechthaler - 2025-04-07
I only have the old version (2164) without the top cone (very small hole instead), it is quite "round" at the innermost ring...
you could add one edge prim (for connecting only) just outside the dome, could make connecting easier...
is it already subfiled? can be used for both variants, the old one has a 5mm shaft (conical to 5.5mm at the top for unmoulding)
total lenght 55,8mm, top hole 2mm
RE: Help with Duplo umbrella 40554 - Peter Blomberg - 2025-04-07
How about this?
umbrella5.png (Size: 20.32 KB / Downloads: 39)
RE: Help with Duplo umbrella 40554 - Rene Rechthaler - 2025-04-07
UmbrellaAxis.png (Size: 27.68 KB / Downloads: 29)
where is the axis? along the ribs (red) or between them (blue)
what portion is the subpart?
RE: Help with Duplo umbrella 40554 - Gerald Lasser - 2025-04-07
(11 hours ago)Peter Blomberg Wrote: How about this?
That one looks really good!
Very nice work!
RE: Help with Duplo umbrella 40554 - Peter Blomberg - 2025-04-07
(2025-04-06, 22:07)Rene Rechthaler Wrote: also, what about the orientation? tips/ribs to 0° or side to 0°?
(9 hours ago)Rene Rechthaler Wrote: where is the axis? along the ribs (red) or between them (blue)
what portion is the subpart?
At this point it is possible to choose the angle of the axis versus the ridges/ribs. The current shape doesn't have a line of points at either red or blue angles. The smoothness of the shape at both angles requires that the lines of points are symmetrically on both sides of the angle. To make it worse, the spacing is different at these two angles. All of this makes the current shape unsuitable for placement with only cardinal rotations where every other segment is forward and every other in reverse. If only cardinal rotations are permitted (i.e. if 45 degree rotations wouldn't be), then the subfile should cover one quarter of the shape so that the angle offset at the borders are the same. Looking at this particular shape, which angle would you choose for the axis and which arch length would you choose for the subfile?
RE: Help with Duplo umbrella 40554 - Rene Rechthaler - 2025-04-07
(3 hours ago)Peter Blomberg Wrote: At this point it is possible to choose the angle of the axis versus the ridges/ribs. The current shape doesn't have a line of points at either red or blue angles. The smoothness of the shape at both angles requires that the lines of points are symmetrically on both sides of the angle. To make it worse, the spacing is different at these two angles. All of this makes the current shape unsuitable for placement with only cardinal rotations where every other segment is forward and every other in reverse. If only cardinal rotations are permitted (i.e. if 45 degree rotations wouldn't be), then the subfile should cover one quarter of the shape so that the angle offset at the borders are the same. Looking at this particular shape, which angle would you choose for the axis and which arch length would you choose for the subfile?
why no vertices exactly on the peak? it can still be a (round) curve (like a cyl prim)...
then it would be possible to use 1 subpart (each 45°) 8 times, without rotations...
red - main axis cut lines
blue - diagonal cut
yellow - scaling direction
Edit: theoretically, even a 1/16 cut would be possible (the rotated ones along the blue 45° curve need a high enough precision for that)