Drawing a blank credit card is an easy task. The fun comes when trying to add the company name and logo: the extrusion of the company name needed just a resizing and positioning, while the gearwheel logo has been more tricky. My idea was to include the core of the gearwheel design in the id badge, keeping it “low relief”. As in the business card the gear appears tilted and turned, I tried to reproduce the same positioning in 3D, I cut out from the badge card the tilted and turned extrusion of the full twisted gear and then I inserted in the cutout hole a thin slice, cut symmetrically around the plane where the inside cut changes shape, as well positioned and tilted. After adding a cutout hole for the clip badge holder, I sent it to the printer. For this first print run I choose to engrave the company name (as an alternative of having it also as low relief) and to use white resin.
According to the prototype results, I will add the name of the card holder and have it printed in metal.
The basic shape of the logo is a twisted gear, which was generated with one code line with OpenSCAD. The original object is too big to be used directly in the 3d badge. As FreeCAD does not offer so many options to manipulate stl objects (as, for example, Blender does), I decided to generate the same gear using FreeCAD tools (and to consider it for me a challenge to learn more FreeCAD features tipps and tricks, instead of a “not so skilled” way to complicate my life as a designer when a clean and easy way as software developer would be at hand 😉 ). I had available the section of the gear in svg format, easy to be imported inside FreeCAD, but as FreeCAD does not offer yet the possibility to twist a surface while extruding it, I had to find a different way to build the twisted gear.
I based my construction on FreeCAD arrays: I drew a sketch of the single element of the gearwheel rim (the “tooth”) on the x-y plane and generated the gearwheel section as a polar array of the “tooth” sketch. The basic idea was to place a copy of the gearwheel section at a different height along the z axis, rotate it, and connect the two with the “loft” tool. As there is no possibility to define a correspondence between points of the two curves, I had to play a little bit with the relative rotation of the two gear sections, to see how the “loft” tool behaved, according to the rotation angle (if it is equal to the sector occupied by a “tooth”, then the extrusion is vertical connecting to corresponding subsequent “teeth”on the two sections, not twisted to connect a “tooth”on the bottom the corresponding rotated “tooth” on the top profile).
one “tooth” sector rotation
5 deg rotation
At the end I found out that I could have the ascending anticlockwise twist with a rotation angle up to 5º.
The solution automatically became: an “ortho” array of 6 slice elements of the gear, rotated one w.r.t. the other of 5º. Considering that the gear section is a polar array of 18 “teeth”, merged in a single sketch, I created an “ortho” array of “lofts” elements connecting two copies of the same “polar” array of the single “tooth” of the gear. This is really building a whole from the single element 😀 . (Yes, I’m aware of the one code line with OpenSCAD, but where would be the 3d designer fun? I would have been the same proud as a coder, had I written myself that one code line…). No need to enter details about the inside cavity: a cylinder and a cube used as cutout tool for a Boolean “difference” with the gear.
Now that our company is starting, we use the Fablab facility for professionals. It would be nice to wear a corporate id badge. As we are three associates, using our living rooms respectively as R&D office and test lab, we have no need of an usual id badge to check in and out at the company gate (we have the key for both the main building entrance and of our apartment doors). Then the idea came to my mind to design a 3d corporate badge, to gain more company visibility when we are at the FabLab workshop, printing our prototypes, or we take part to workshops, sharing ideas with other professionals. Wanting to keep the same appearance of the business cards, I asked Lorenzo for the original objects he used to draw them: a path from Inkscape for the company name and an stl object generated and exported with OpenSCAD for the twisted gear. I needed to reproduce the same shapes in 3D with Freecad, with the obvious size limits (especially in the overall thickness) of a “nearly flat” credit card like badge. If I may allow myself the comparison, without being too arrogant with respect to real sculptors, I needed to create a “low relief” of the business card objects. Before importing the Inkscape path of the company name inside FreeCAD, I flattened the path’s Bezier’s curve turning into an approximation composed of straight-line paths and reduced to only one node the agglomeration of nodes sometimes created by the conversion: in this way once imported the svg file inside FreeCAD, it can be handled as a set of lines (or “edges”) and there is no risk to convert an Inkscape closed path into a FreeCAD open path if the wrong node in the agglomeration is kept in the import. This made a lot easier to have all the letters as surfaces and extract the internal cutout surfaces (for letters like “P” or “D”) to create by boolean difference the exact text-shape to be extruded.
How I dealt with the twisted gear deserves a dedicated post. Just wait and read…
The class 43 cabin right out from the printer: in the picture in the middle of the right column it is next to the first print I made at Shapeways (the public transportation card placed below -“credit card” format- is just as reference for the size), while the two on the bottom right show the piece with the support structure generated by the printer, then easily removed with a clipper. Both the upper part and the lower valance are much more “curved” and close to the original than in the first model.
The outline of the chassis must match the open bottom line of the model. Being the model symmetrical with respect to the longitudinal axis, it’s enough to draw half of the outline, which is easily done by intersecting a plane with half of the locomotive model. For the cabin, only the lower valance is needed.
Once the section is available, some hand working with the “Draft” workbench is needed, basically to downgrade the intersection object into lower level objects: first from a surface to a wire, then a wire into single edges. Once the edges are available, a wire closing the internal profile must be drawn, to be selected together with the the edges making the internal profile to be joined together again into a new “wire”object which will then be upgraded to a surface.
This is the outer surface to be then extruded and cut through for housing the engine and letting the bogies to be connected (the whole object being mirrored to have it complete).
The class 43 redesign is finished and ready for the 3D printer.
It’s the right time to test the brand new printer at home, let’s print the cabin in white resin, with the highest precision available. Just started… it will take three hours, it will be ready after lunch 😉 .
The first geometrical challenge puzzling my mind was how to combine together two ore more knot patterns, oriented at special angles one respect to the other, making it look like a continuous pattern, with no weird crossings or overlaps. As my first pattern had all crossovers at 45 deg, I thought the feasible choice would be to link two patterns at a right angle. The diagonal parts were already properly oriented, but it has been a little bit tricky to find the right distance between the two patterns to preserve the regularity of cross overs. After a few attempts I realised the obvious: they should have been placed at 90 deg one respect to the other, each pattern starting at the same coordinate of the side of the other, leaving a free square where they would intersect if extended, the side of which corresponds to the width of each thread.
The next step is to design the angular connecting block, joining the threads of both patterns, in way that geometrically could represent a 3D physically possible continuity (my goal is to create 3D printable objects, not 2D “impossible” paintings). To be noticed that the angular block must match the geometry of the crossovers and the width of the threads: threads with longer or shorter elements, of the same width can use the same angular block.
As an alternative, the horizontal thread can be moved up an left, to match it’s upper left diagonal side with the lower right diagonal side of the vertical thread, shrinking the angular part as in the picture. This is more compact but optically more tricky to imagine the vertical ribbon bending with a 45 deg angle and such a spiky external corner. It could be easier imagined as sheet of paper either folded as an origami or with the shapes cutout and then interwoven by hand.
I’ve seen my nine years old son drawing this geometry on a squared notebook. He saw my interest and simply told me “mummy, I can show you how to draw it” and repeated the sequence of segments just for me. I started being captured by the sequence, the 2D geometry, the symmetry and I imagined it as a 2D projection of an “8” shaped spiky knot. I have always been fascinated by knots, especially as patterns, like the celtic motifs. This triggered me now as a 2D/3D designer and I came up with decomposing the shape in simple repeated elements to be reproduced as an array (copied, mirrored, turned).
With FreeCAD, first I drew the basic geometry elements as two sketches
and repeated them in arrays with the same distance between elements.
Then, mirroring the left array symmetrically on the other side of the “S” shaped one , I could easily reproduce a sequence of “two crossing ribbons”, without closing the ends as in the initial “8” knot.
The three arrays all together compose the pattern I was looking for, made of the two simple basic sketch objects.
It is still a 2D drawing, the 3D alternate overlapping of the two threads is left to the imagination of the observer, while it could be helped with colors.
As a fan of the works of a geometry genius as M.C.Escher, I also started trying to imagine how this shape could be turned and placed in 2D space to auto intersect with its own copies… see my next post.