3D Printing Dubrovnik Chessmen
Modeling the Knight
I wanted to 3D print the 1950 Dubrovnik chessmen that I had modeled and drawn for woodturning. Modeling the pieces with spun profiles was easy but I didn't have a decent model of the knight. I carve the knights from paper templates.
Knight Carving Template
|
|
I made several attempts to model the knight in CAD. I spent about 20 hrs solid modeling in Solidworks. While it was a great learning experience, I was not satisfied with the result.
Solidworks Modeling Attempt (Visible Edges)
|
Solidworks- Solid Modeling Attempt
|
Next, I thought surface modeling was the answer. I spent another 20 hrs learning surface modeling in Solidworks. The final result was smoother but the surfaces around the nose were troublesome. I didn't bother adding any final details before going down another avenue.
Solidworks surface model (Edges Visible)
|
Solidworks surface model
|
I thought to myself, sculpting this in wood is way easier. I finally searched for "scultpting software" and found Blender, a free mesh modeling software. Great, another software to learn... At the same time, I had been experimenting with photogrammetry scans to create 3D models of my dad's earliest chainsaw carvings. (Photogrammetry is a fancy term for "projection mapping". It uses photos taken at many angles of an object to create a 3D mesh model.)
I took 76 photos of a knight I carved by hand to perform the photogrammetry in Meshroom2021. I spent about (you guessed it) another 20 hrs learning Blender and refining the knight model. I cut off the base/pedestal portion of the photogrammetry model as that was easy to create in CAD and print separately. I mirrored the mesh across the centerline so the knight was perfectly symmetric. The overall shape was adjusted to an imported picture of an actual 1950 Dubrovnik knight. I redefined every cut to make the edges as crisp as possible. After 9 stages of refinement, this is the final result.
I took 76 photos of a knight I carved by hand to perform the photogrammetry in Meshroom2021. I spent about (you guessed it) another 20 hrs learning Blender and refining the knight model. I cut off the base/pedestal portion of the photogrammetry model as that was easy to create in CAD and print separately. I mirrored the mesh across the centerline so the knight was perfectly symmetric. The overall shape was adjusted to an imported picture of an actual 1950 Dubrovnik knight. I redefined every cut to make the edges as crisp as possible. After 9 stages of refinement, this is the final result.
|
|
I had already printed the other chessmen. They sat patiently at the board, waiting for the last members to arrive.
|
|
Resin Printing Experiment
As with any filament printer, the layers are visible in good lighting. A local library has a maker space with a Prusa SL1 resin printer (stereolithography) available to use for only the cost of the material. I had them print the knight in resin.
|
|
There is higher definition and no visible layer lines but the resin is much more brittle than the PLA. I'm not familiar with all the resin material options available to know if there is one with higher toughness, closer to PLA, but I don't think the pieces would hold up at the club. It's also trickier to make hollow cavities inside the part for balanced weighting as the resin needs somewhere to exit the finished part. Most stereolithography parts are printed at 100% density for this reason.
Scale to Fit
For people who love to debate which size of chessmen looks best on which size of board square, the pieces are easily scaled to try the options.
|
|
Weighting
Some weighting can be built into the print design by choosing the base/pedestal to have 100% infill and the top to have 7% infill. Some thin areas still need 100% infill (e.g. the pawn neck, king ball top) and can be done with local infill modifiers. This gives the pieces a nice balance and you could call them single-weighted.
|
|
Transparent PLA to show 100% base infill vs. 7% upper infill
|
Another way to weight the chessmen is to include a cavity and set the print to pause at the top of the cavity to insert the weights. This will capture the weight inside the print to prevent the chance of ever falling out. I did this technique to insert magnets in the magnetic travel chess set.
|
|
Support Geometry
The last challenge was the design of print supports. I didn't want to compromise the original piece geometry by angling the horizontal surfaces (i.e. under the collars) just to make them printable without supports. I tried different support design configurations in Prusa slicer until the support was easy to remove but still supported the low angle surfaces enough to have decent surface finish. These were the magic support settings I found:
Wild Color Schemes
Once those problems were solved, I started printing wild color schemes to make sets for friends and family.
|
|