View unanswered posts | View active topics
|
Page 1 of 1
|
[ 14 posts ] |
|
| Author |
Message |
|
Joshua Bell
|
Post subject: Finished my Fisher 3x3x2 (a.k.a. custom Domino) Posted: Sat Mar 30, 2002 8:49 am |
|
Joined: Fri Dec 08, 2000 2:32 am Location: San Francisco, CA
|
One picture inline...  And others at: http://www.calormen.com/tmp/Cuboids1.jpghttp://www.calormen.com/tmp/Cuboids3.jpgDetails of the kernel at: http://www.calormen.com/tmp/Edges.jpgApologies - I had the pieces all made up and had mixed up a batch of epoxy when I remembered I needed to take pictures for Twisty. I photographed the black kernel close up on a white background with a flash and the result was incredibly blurry. So I highlighted the edges. ------- IMPORTANT BIT OF ADVICE: The 2x2x3 was MUCH more forgiving for mis-cut pieces. Be extremely paranoid in sizing the outer cubies for the 3x3x2. ------- Experiences: Making the pieces for this was really quite simple. Again, I used a 2-pack of regular/mini cubes from Walgreens, a two-part plastic epoxy, a hobby saw and stickers from Rubiks.com. Figuring out the kernel was the hardest part (or so I thought) - hopefully the picture will make it clear. One bit that the picture doesn't make clear - since the mini-cube's corner cubies have a cylindrical cut-out there's a choice - do you put the hole in the "bottom" of the corner cubie (so the angled edges of the kernel cubie are solid) or do you put the hole in one side (so one of the angled edges will be cut in half). I chose the latter, which makes the corner cubies asymmetrical but stronger since the floating "tower" in the corner is connected to the center of the kernel with a solid square of plastic. Once I had the kernel made it was a matter of mass-producing the edges and corners. For the corners you basically cut off one edge such that the corner covers half of one edge of the kernel. Be paranoid - since you're making 8 corners, you want to get the measurements right. I cut the edges with a "stair step" - a one step to sit on top and along the side of a kernel edge, another step to give the middle layer of the kernel some space when rotating a 3x3 layer around it. Frustratingly, once I had all of the outer cubies shaped I realized that while the feet and molded structure provided fairly nice guide-posts for where to cut off bits and pieces, the resulting 16 pieces I'd slaved over didn't quite fit on the kernel - they were all cut in a little too much. So be extremely careful. Once all of that was done I plunked on the center caps on "stalks" of 2mm styrene using the epoxy, then started decorating the outside. Failure! Since all of the pieces were just a smidge too deeply cut they went on at funny angles. From this point I figured out a useful trick - glop a substantial amount of epoxy on the cubie and stick it on the bottom of the kernel which is balancing on the central cap on your work surface. Do several at once (alternating between glued/non-glued and (next pass) dried/glued. Since they're all upside-down they lie flat on your work surface and the epoxy fills the gap. Let the epoxy dry (20 minutes+) then trim out any epoxy which oozed out. I did this for all but one last edge, only doing pieces on one layer of the puzzle at a time. Once I had all of the pieces but one made I put the whole puzzle together, glopped epoxy on the remaining edge and sliding it into place. ------- Results: It's stiffer than the 2x2x3 and seems more "finicky". I think I lucked out with my 2x2x3 - it feels as nice as a store-bought cube. The 3x3x2 is stiff on some turns due to slight mis-alignments in the edges which cause excess friction, and some of the corners wiggle, but overall I'm happy with it. With these customs I think there's a magic moment in time when it goes from a bunch of ill-fitting pieces which seem incredibly fragile to a functional robust puzzle. When turning an incomplete puzzle there is uneven pressure based on where I put my fingers. Once all of the pieces are in place, though, the mechanism takes over and it feels much smoother. Additionally, as soon as the stickers are on it looks finished and my paranoia goes away.
|
|
| Top |
|
 |
|
TM-Jeremiah
|
Post subject: Very nice Posted: Sat Mar 30, 2002 1:19 pm |
|
|
|
|
Is the picture of the kernel, the kernel used for the Domino? Did you use the same cuts to make both puzzles?
|
|
| Top |
|
 |
|
Carter
|
Post subject: Great Job! Some Questions Posted: Sun Mar 31, 2002 12:32 am |
|
Joined: Sat Dec 02, 2000 8:27 pm Location: Wilmington, NC, USA
|
|
Very nice puzzles! Great description.
That illustration of the kernel is really nice. That drawing really helped me understand these things.
Your description of globbing on the epoxy gave me the heeby-geebies. How do you keep from accidentally gluing adjacent pieces together?
It sounds like you disassemble the kernel to shave the pieces. Then assemble the kernel. Then glue on the outer pieces. Is that correct?
On the kernel for the 3x3x2, did you glue one of the top edges to the side center as Dieter Gebhardt described, to prevent the middle layer from jamming the puzzle?
Can the puzzles be disassembled once they are completed?
|
|
| Top |
|
 |
|
James East
|
Post subject: Very nice :) Posted: Sun Mar 31, 2002 8:49 am |
|
Joined: Sun May 27, 2001 7:03 pm
|
|
Well done! Its very satisfying to see those pictures.
|
|
| Top |
|
 |
|
Joshua Bell
|
Post subject: Nope, different kernels Posted: Sun Mar 31, 2002 1:19 pm |
|
Joined: Fri Dec 08, 2000 2:32 am Location: San Francisco, CA
|
|
The kernels for the 2x2x3 and the 3x3x2 are quite different.
The 2x2x3 kernel has to support two rotations, with corresponding trimmings - the top/bottom corner cubies need to be trimmed towards the center to allow the middle layer of the completed puzzle to rotate, and the edge cubies of the top/bottom layers of the kernel need to be trimmed to allow halves of the rotate.
The 3x3x2 kernel has almost exactly the opposite requirements - the edge cubies of the middle layer need to be trimmed to allow the top/bottom layers to rotate, and the corner cubies of the top/bottom laers need to be trimmed towards the center to allow the central layer to rotate past (in two axes).
|
|
| Top |
|
 |
|
Joshua Bell
|
Post subject: Answers Posted: Sun Mar 31, 2002 5:43 pm |
|
Joined: Fri Dec 08, 2000 2:32 am Location: San Francisco, CA
|
|
>> Your description of globbing on the epoxy gave me >> the heeby-geebies. How do you keep from >> accidentally gluing adjacent pieces together?
A few things - first, don't have uncured epoxy on two adjacent pieces. This will make it easier to separate them - by force or by knife. Second, since I ended up with the outer cubies not actually touching the inner cubies I was more willing to cut the pieces apart after the fact than risk them having poor joints. By doing only half of the puzzle at a time I was always able to see (and cut!) the overflowing epoxy - I did this after about 20 minutes, so the epoxy was not rock-solid and could easily be cut by an x-acto knife.
It wasn't pretty, but it worked.
>> It sounds like you disassemble the kernel to shave >> the pieces. Then assemble the kernel. Then glue on >> the outer pieces. Is that correct?
Actually, I cut the kernel down while it was whole. Mostly. The kernel cubies are so small that holding them in place is tricky. To make the cuts of the middle slide I turned it 45-degrees and then used a hobby saw to trim off the corners sticking out. To make each kernel corner cubie I followed these steps:
(For directions in the following description, assume I'm making the front/top/right corner cubie).
1) Figure out (by rotating a layer) which side of the corner cubie has a hole in it. Make sure you treat that as the side (not the bottom) of the corner. To put that another way, make sure the bottom of the cubie is solid. 2) Use a hobby saw and start a cut on the middle of the top/front edge of the cubie. Extend this cut until it completely bisects the top and front face. 3) Repeat on the top/right face. The top face should now be cut into quarters like a + sign. 4) Rotate the right face 45-degrees, then (like you did for the middle layer) cut down into the bit sticking out. Halfway through the cubie you should run into the cut from step 2 and 1/4 (by volume) of the cubie should come free. 5) Repeat by turning the front face 45-degrees and you're done.
Note that where you do the initial cuts in steps 2/3 depend on the relative sizes of your outer/inner cubies. Mine are almost exactly 2:1 so I just eyeball the mid point and use a file to make sure it's wide enough. Also, this ratio determines how much of the cubies actually needs to be removed while cutting with a face turned 45-degrees. I probably allowed 0.5mm of leeway to leave more plastic on the cut-down pieces, then used a file to ensure that the cut-outs left a smooth track.
I frequently disassembled and reassembled the kernel during the process to make sure the corners were oriented as I wanted (with a solid bottom face), but the cuts were made with it together.
>> On the kernel for the 3x3x2, did you glue one of >> the top edges to the side center as Dieter Gebhardt >> described, to prevent the middle layer from jamming >> the puzzle?
Yes. I was following those instructions for the most part, but given the relative size of the inner/outer cubies I was using, a more elaborate kernel was called for.
>> Can the puzzles be disassembled once they are >> completed?
I am not sure and honestly don't want to risk it unless I have to. It really depends on the strength of the epoxy bonds. For the 3x3x2 the usual trick of a half-rotation of one layer followed by prying up an edge should work if you know which layer is not coupled to the middle layer. I assembled the last piece by gluing it into place rather than snapping it in.
For the 2x2x3 the same trick should work but the outer edge cubies are very loosely connected to the inner edge cubies so it is more likely that the epoxy would snap unless you could work a knife through the gap into the kernel blindly. In this case, though, I did assemble the completed puzzle by snapping the last piece into place.
For that reason, when assembling the puzzles I was careful to make sure I could always disassemble it up until the very last step, and had a plan for the final assembly.
|
|
| Top |
|
 |
|
Aswapati
|
Post subject: Very nice :) Posted: Mon Apr 01, 2002 4:12 am |
|
Joined: Fri Jan 04, 2002 7:27 am Location: Munich, Germany
|
|
The two puzzle look great like. My best compliments. All further details are welcome (I hope I will be able to produce two of them).
|
|
| Top |
|
 |
|
Carter
|
Post subject: Great. Thanks. Posted: Mon Apr 01, 2002 5:43 pm |
|
Joined: Sat Dec 02, 2000 8:27 pm Location: Wilmington, NC, USA
|
|
I wanna make some of the these!
|
|
| Top |
|
 |
|
Sandy
|
Post subject: Congratulations, Joshua Posted: Mon Apr 01, 2002 9:50 pm |
|
Joined: Thu Jan 24, 2002 1:10 am Location: Toronto, Canada
|
Very nice additions to your collection. You must be proud. My 2x2x3 project has been on hold since about a week before I won the following auction (I broke a corner piece on my only keychain cube, leaving me without a kernel). http://cgi.ebay.com/ws/eBayISAPI.dll?MfcISAPICommand=ViewItem&item=1700870215They still haven't arrived! Seeing your puzzles is making me jealous and anxious to get going some more. Once again, well done. Hopefully we can learn from your experience when attempting our own! Sandy
|
|
| Top |
|
 |
|
Aswapati
|
Post subject: Dimensions of the kernel Posted: Tue Apr 02, 2002 12:47 am |
|
Joined: Fri Jan 04, 2002 7:27 am Location: Munich, Germany
|
|
Hi Joshua,
I have a question. What dimensions has the kernel that you used ? I have a normal Rubik's Keychain (30mm x 30mm) and a normal Cube (about 58mm x 58mm). Are they ok ? I know that there are keychain cubes that are smaller than 30mm.
Thanks,
Guido
|
|
| Top |
|
 |
|
sausage
|
Post subject: Edge pics Posted: Tue Apr 02, 2002 6:36 am |
|
Joined: Wed Nov 24, 1999 12:18 pm Location: Palerang Shire, NSW, Australia
|
|
Well done, Joshua, nice job. The edge pic for the 3x3x2 is a very good illustration. Can I update the article and use this one? Will it remain in your site space?
|
|
| Top |
|
 |
|
Tyler
|
Post subject: Well Done Joshua Posted: Wed Apr 03, 2002 1:34 am |
|
Joined: Sat Nov 24, 2001 8:31 am Location: New York
|
|
Hey, Just got back from Disney world. This looks great! The kernel looks just like the one i made... Seems fisher's drawings on twisty need some updating.
|
|
| Top |
|
 |
|
Joshua Bell
|
Post subject: Edge picture Posted: Wed Apr 03, 2002 6:36 am |
|
Joined: Fri Dec 08, 2000 2:32 am Location: San Francisco, CA
|
|
Rather than relying on that messy thing, I've sent Wayne a cleaned-up version that's just a B&W drawing with the critical A/B pieces joined. It should be easier to just add that to the Twisty site.
(In general, please don't link to anything in the /tmp/ directory of my site. If you need to I can move it to a permanent home.)
|
|
| Top |
|
 |
|
Juozas Granskas
|
Post subject: Article? Posted: Sat Apr 06, 2002 10:03 pm |
|
Joined: Thu Dec 16, 1999 9:31 am Location: Kaunas, Lithuania
|
|
Inspiring! Joshua, your great work, description, pictures, answers - everything - deserve an article. On Twisty, on your site, wherever. It would be an extremely valuable resource for custom puzzle makers.
|
|
| Top |
|
 |
|
Page 1 of 1
|
[ 14 posts ] |
|
Who is online |
Users browsing this forum: benpuzzles, Google [Bot], Google Adsense [Bot] and 11 guests |
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot post attachments in this forum
|
|
|