[DIY] SKATILE: Parametric Skadis Interlock Tiles - 3D Printing

To tame the clutter in my projects room I designed an Ikea Skadis-like slotboard system, with a few twists:

1. interlock-able tiles to build murals without having to screw each tile to the wall
2. option to secure the tiles to the wall via screws around the perimeter
3. option to use M4 inserts to mount tiles back to back 
4. option to back light each tile (in coming post, but the 15mm-skirt tiles herein are backlight ready)

Enter the SKATILE system. It is simply based on a backside skirt around the perimeter of each tile that provides a clean look (no gaping space between board & wall), clean spacing against the wall (no spacers) and its tapered shape allows to lock the tiles together via regular 19mm mini office binder clips, no screws into the walls needed:

And the tiles can be hung from shelves, again without needing screws in the walls:

But when screws are needed, the holes around the perimeter of each tile can either be uncovered with a drill / countersink bit, or by removing the covers in the CAD model. Like on this cabinet around my printers, which also includes front tiles with M4 inserts to mount tiles back to back:

The tapered skirt can also hold a backlight cover behind the tiles for lighting up an area or illuminating inlays in the tiles (coming post). Here is a peek at the light transparency tests:

The parametric Fusion 360 CAD model provides customization for the number of horizontal and vertical slots per tile, as well as custom skirt height, to adapt to each space's needs.

2 models available:

• Regular (single skirt, 4 corner screws/inserts)
• Large: includes design features to minimize "Benchy hull lines" / vertical wall line shifts (plus add'l vertical spine for rigidity and 8 screws/inserts)

The F360 model, STEP files for a dozen tile types (size: 180-360mm, skirt height: 11 or 15mm, with/out holes), and 3MF for H2S + P1S are all on Github

Feel free to copy, modify, publish, whatever.

Happy decluttering to all !

______________________________________________ 

 

Example Results

 

Regular vs Large Model

Both can be generated with any number of horizontal and vertical slots, but the Large model includes a couple of features to scale better with size:

1. center rib for rigidity
2. 8 peripheral screw / insert holes instead of 4 (each can still selectively be printed covered or not)
3. hidden wall-decoupling gap to reduce the "hull line" effect
4. special LARGE_shrink_factor parameter (details in Full-Bed Prints chapter below) 

 3D Printing

Warning: before printing, perform a filament shrinkage calibration if you intend to (1) print tiles larger than ~8 slots and (2) use Skadis accessories with a large hook to hook distance

• part shrinkage after cooldown leads to a cumulative slot-slot distance skew that can make it difficult to install large accessories like the Skadis Shelf that is 12 slots wide, as the slots can then be off by a couple of millimeters
• use either a shrinkage calibration test or simply print a ~200mm long thin bar and measure its length once cooled down. Then input the shrinkage factor in the slicer: 

For instance, on my H2S, PolyLite ABS was off by 2.5mm over 320mm. Resulting in a shrinkage compensation factor of 317.5 / 320 = 99.22%

Settings for PLA, PETG, ABS, tested on both Bambu P1S and H2S:

• 0.4mm standard flow nozzle, 0.24 layer height
• 3 walls, 2 top, 2 bottom
• can be bumped to 3 top/bottom for heavier applications, but 3-top leads to more line artifacts on the outside wall, and more light will be blocked when using the backlight option
• 15% Hilbert Curve infill (less warping with large tiles, more backlight)
• at least 10mm 'Length of Sparse Infill Anchor' (fewer boogers and gaps when printing the back surface)
• 'Only One Wall on 1st Layer' enabled (cleaner looking face)
• Seam on outside wall: paint it wherever it won't be seen once in the room
• Scarf on / off doesn't seem to make a difference in outside seam quality
• no support, no brim 

On an H2S it takes about 8 hours and ~270 grams to print a 16 x 15 slots (340x320mm) full-bed sized tile, with a 15mm high skirt. 

Tips to reduce "hull lines" on large tiles when using shrinkage prone filaments:

• the "hull lines" come from the ~8X higher layer time when printing the back surface on layers 19 & 20, resulting in more shrinkage
• use the Large model instead of Regular (has a hidden decoupling gap btwn the outside wall and the inside wall / surfaces)
• bumping chamber temp from 60°C to 65°C helped marginally
• limit the number of top layers to 2 as 3 can be worse
• Bambu Lab ABS filament printed the cleanest on H2S, but got PolyLite ABS very close with Large + 65°C + 2 top/bottom

 

Digression on nozzles: for large tiles a 0.6mm nozzle is theoretically better as it can cut hours out of a print. But all attempts to print a 340x320mm tile on an H2S with the 0.6 standard, or with the 0.6 / 0.4 high flow nozzles led to worse outside wall seam, boogers and hull lines than with the 0.4 standard nozzle. However, this is probably just a specific issue with the H2S and its poor 0.6 and HF profiles when printing at max bed size.

 

 

 CAD Model

Initially based on the Parametric SKADIS Board Generator v38 from Line Arc Line on Printables.

Only 2 parameters need to be changed to set the number of horizontal and vertical slots: columns and rows

skirt_height and starting_pattern can then be modified to customize a tile further:

 

List of parameters that can be modified by the user:

The hole covers can be selectively removed. Examples:

To do so, follow either the A or B method:

Binder Clips

Any 19mm standard mini paper binder clips should work. Once installed, remove the handles.

I used these from Amazon: 19mm Mini Binder Clips 3/4-Inch

Full-Bed Prints Conundrum

This section goes into the horrors-they-don't-tell-you-about of 3D printing. Skip it if you print tiles smaller than your printer's bed size.

Since Skadis boards are based on a 20mm grid, square board sizes are likely to match common maximum printer bed dimensions: 200x200, 300x300, etc. In that case issues can show up that I haven't seen addressed in 3D printed Skadis Board models out there. So here is my experience, along with the solutions I developed, in case it helps.

2 potential issues: 

1. Slicer complaining that the tile is larger than the bed size although it should exactly match (likely due to precision / floating point barely-over value in slicer or step file)
2. If the user sets a shrinkage factor lower than 100, the slicer automatically scales the tile up, which now goes over the bed dimensions. Obviously, that prevents from printing tiles that are both max-bed sized and shrinkage compensated

Fix for (1): in the slicer scale the Object in X and Y (not Z) by ~99.99% or so. The issue should disappear.

Now, achieving (2) is not possible without using a non-shrinking filament or giving up something. Since I really wanted to print tiles in ABS and as big as my H2S's 340x320mm bed and with slot-slot shrinkage compensation, I decided to give up in tile-to-tile spacing instead.

That's accomplished in the CAD model with the LARGE_shrink_factor parameter that reduces the distance between the perimeter of the tile and the nearest slots:

To set it, just divide the wanted length by the measured one. Example on my H2S: the 320mm large tile after cooldown measured 317.5mm, so I set LARGE_shrink_factor to 320/317.5 = 1.00787. That resulted in perfect slot-slot distances on the next printed tile. Of course, the cooled down ABS tile ended being 337x317.5mm instead of 340x320, but nothing can be done about that.

Anyway, with that we can now sell our croissant and eat it too, c'est fou non ? Though now the tile-to-tile spacing is no longer a multiple of 20mm. So, hanging large Skadis accessories across 2 adjacent tiles may be more difficult as the hooks may not line up perfectly with the slots anymore.

Obviously, when using LARGE_shrink_factor, the slicer's shrinkage parameter must be set back to 100.

End of the horrors-of-3DP-that-suck-80%-of-my-time rant. Back to the rainbows and unicorns we see on most not-a-corner-case-in-sight youtube 3DP shill channels... 😅