this project was inspired by the book “Soft Electronics” by Jaro Gielens and the beautiful renderings by Gijs here in the forum. The scans were made as part of a Uni project in which I explored different 3d scanning workflows and their use in design education and the preservation of old designs.
Besides this, I think their curvy, smooth 70s styling makes them a great basis for surfacing exercises. I’m planning to reconstruct some of these in Rhino to sharpen my surfacing skills and would like to compile my learnings in a short surfacing guide for new students.
Reference images including the logos and some detail of the inner construction of the appliances will follow soon.
The scans were created with the Revopoint POP2 3d scanner and are intended for educational use only. Meshing and further postprocessing was done in CloudCompare and AD Fusion.
These are some very nice scans! Many thanks for doing this! I can’t seem to find the Privileg H500 in the book and there doesn’t seem to be many pics of it online. Do you have any photos of that one?
Thanks – happy to hear!
Not yet, but I’ll upload photos as soon as I have put them back together. They are also still covered in scanning spray and have to be cleaned.
Hi @kevin.franz I’ve used the Rowenta scan as a basis for today’s webinar, where I’ve been showing some surfacing techniques and rendering tips. Again thanks for sharing these wonderful scans!
@Guillermo_Varela can you share a link from the recorded webinar to this thread and back? I’m sharing the model I’ve built, without the scan, so people can download a smaller file and the scan(s) separately.
The scan I’ve been using is Rowenta_EK_55_Complete.stl from the link in the first post of this thread.
In this model I’ve been using XNurbs. The surfaces have been created with history recording on, so you can run XNurbs and Edit the surfaces (in layer srfs) to see what settings I used.
Dear @Gijs
somehow it’s a pity to not see a pure / vanilla Rhino solution.
X-nurbs (as far as i know @XNurbs ?) is not available for mac
and I think this is a nice example, that shows some deficits in the rhino toolset regarding concept modelling with complex transitions.
(and i would love to see a single-span Class-A solution as well @Rhino_Bulgaria what about joining the party ?)
this is my rhino-only-tools apporach:
rework revolve like surface
reworked the initial pseudo-revolve to have less shear in the isocurves, set it up as approximation / deformable / rational
use an additional surfaces “cone like ring” (green) als helper.
1st blends
some _blendCrvs to find trim concept / positions / matching shapes.
initial blends
maybe my apporach is wrong - I would love to have more rules regarding how to find nice trims. But it would also be nice to have some interactive tools, that support trimming (position) + blending
I normally look at Curvature(graph / analysis), do a test filletSrf to see minimum required space - but the rest is more intuitive / visual. If there is more to learn - would be happy to see some resources.
especially for the 2nd blends _blendSrf did not help me.
I used some initial _loft and then some massage with _matchSrf, _changeDegree _insertKnot and some CV edit inbetween… _matchSrf with refine yes messed up with 1000s of knots
looking forward to see other pure rhino apporaches.
kind regards from a rainy afternoon ( i expected snow) - tom
I suggest to apply “RefitTrim” to both, the handle and the top cylindrical body, in order to heavily reduce the amount of control points needed for the nearby blend surfaces. And alternative solution is to manually move the control points so that both of those surfaces will remain untrimmed. I would use a deformable circle to extrude both cylinders, because Rhino does weird things with the default degree 2 cylindrical surfaces.
Another advice is to take advantage of the often overlooked “On surface” option of “Match surface”. It also could help you reduce the amount of control points for the blend surfaces (or lofted surfaces), because it takes into account the general shape of the simple target surface (and an optional target curve on surface) instead of a complex trimmed edge. Once you are happy with the final shape, you can trim the target surface by the matched surface.
The good thing about using that approach is that a History-enabled target curve will let you modify the matched surface by simply moving the control points of the former.
Alternatively, you can use the “On surface” option without a target curve. To do so, simply hit Enter when Rhino asks you to pick a target curve, resulting in skipping that particular step and directly asking you to pick a target surface instead.
Still… your is a really good solution.
Fairly simple.
I have a different , ad more complex solution, I will publish it soon just to share a different version.
I tried similar approaches to this (albeit in Solidworks…) and ended up settling on making the handle and outlet blends fully first, as separate exercises, then trimmed out the central area and tackled that blend last. My working theory was if the other blends are resolved, then there is less guess work involved to get the central area flowing nicely.
There’s a video of the layout here https://youtu.be/4ROS21Nzwyc in SW, only using 4 sided surfs and trims (no xnurbs or fill surfs in the final model) so the surf layout could be replicated in Rhino. I might have a crack at this in Rhino as well. Thanks Kevin for the great scans!
dear @andrew9
thanks for sharing your approach, just watched your video.
I like your result.
the 6 sided “fill” function you used for a first guess of the transition seams similar to x-nurbs/ patch. (something like this is really missing in native rhino)
I also like your orange zebra bubble.
I hope you don t mind, if i share a colored screenshot of your video,
showing the final surface-layout: Andrews Layout:
Hi Tom,
I used the fill surface to quickly check whether the surfaces surrounding the 6 sided hole were flowing suitably before I spent the time figuring out how to patch it with 4 sided surfaces. The fill tool is quite useful but you need to keep an eye on the output.
I’ve attached a Step import of the SW model (the main form). Don’t switch on the isoparm display, it is quite dense in places
