I have not been posting on the forum very much in the last couple years, but I still read occasionally, and it is astounding how fast all areas of this community are progressing. Not so many years ago, 3x3x3 WR had been stuck at 12 seconds for years, and now it is less than half that. I remember when I felt a bit proud to get 1st fewest moves on gelatinbrain's 3.3.3 (now called QuadX I think?) and a few other puzzles but now, others have solved puzzles so so much more complex. And I remember how amazed we were the first time someone built a Gigaminx or 6x6x6, but now with easy access to 3D printing, the world of custom puzzles has exploded.
I very much enjoy checking into the forum every couple of months to see the amazing things people have built and solved!
While reading the forum in April 2012, I stumbled across Timur's F-Skewb v2. Instantly, I loved this puzzle. It is simple, unique, and innovative all at once. By the way, thank you, Carl, for your long post in Sept. 2010 about what this puzzle is. The last puzzle I built was a rather low-quality Dino Cube in 2009, but the F-Skewb awakened my desire to build a puzzle again.
So I got to work and 8 months later, here is the result!
I concept-designed with POV-Ray, then made 3D models with OpenSCAD. It is a freeware program which can generate STL files using constructive solid geometry (CSG). The basic shapes are easy to generate but there is a severe drawback: no fillet tool! I enjoyed doing the math to write my own filleting code, though.
My mechanism is slightly different from Timur's. It only has one internal spherical level, but I think it contains more internal pieces. The puzzle contains 39 pieces, 11 of which are completely internal.
I printed the puzzle in unpolished WSF on Shapeways, then manually sanded the pieces and cut the stickers. Turning quality is good but not excellent. The peices are all solid, so it cost more than it could have, but I like the how the extra weight feels. The puzzle edge length is 57mm.
I'm also attaching a text file with some solution notes. It takes me 70-100 moves to solve using the Ultimate Magic Cube program. There's plenty of room for improvement, though.
AndrewG F-Skewb - 2 moves.JPG [ 142.09 KiB | Viewed 1910 times ]
AndrewG F-Skewb - Small Twist.JPG [ 144.15 KiB | Viewed 1910 times ]
AndrewG F-Skewb - Large Twist.JPG [ 144.68 KiB | Viewed 1910 times ]
AndrewG F-Skewb Piece Layout.JPG [ 185.04 KiB | Viewed 1910 times ]
Also, the tips on this forum have been extremely helpful when I've built puzzles, so in return, here are some details on this puzzle. Hopefully they can help someone.
1) When I built the Dino cube in 2009, I lengthened the spindle arms by 1 mm because I'd heard people talk about puzzles needing internal clearances. The Dino cube was much too lose. This puzzle was printed with zero internal clearances, and it doesn't need any. To loosen the puzzle slightly, I backed out the screws by fractions of a turn.
2) I sanded the pieces using 150 then 320 grit (and then 600 then 1000 grit on the external surfaces only), and I lapped (rubbed against each other) some of the internal grooves which were too small to sand. The external surfaces don't appear smooth but they feel smooth. In retrospect, it might have been worth my time to order the pieces in polished WSF.
3) I read elsewhere here a tip to fillet sticker corners by three small snips with scissors. That was very helpful!
4) For stickers, I used adhesive-backed vinyl that I bought from a local sign shop. It looks & feels great but the adhesive is somewhat weak.
5) A half-turn of a small layer on this puzzle makes it very easy to access the screws, but I didn't realize that until after I assembled it. Haha