Hello,

In this post, Oskar challenged me to design and prototype this kit. He provided the gears and some basic 3x3x3 pieces. I rebuilt the 3x3x3 pieces from scratch in Solidworks and used TomZ's tolerances. I also need to thank Verdes for the V-Cube design which is used in this puzzle. This kit wouldn't exist in this form if it weren't for these people. I have a deal with Oskar where I can sell this puzzle on Shapeways. However I want the mods to verify also that its ok with Verdes. I know Verdes has given the ok to use his V-Cube design to Oskar and Mefferts for their Gear Cubes however with mine you can actually leave the gears out and if you do you have a V-Cube 3x3x3. As such I plan on redoing the models I used to print this puzzle for myself. One had just the 3x3x3 parts and one had the gears. If one were to order just the 3x3x3 parts you'd have what I believe would be considered a KO V-Cube 3x3x3. That wasn't the intent here and I now know a few tricks and I believe I can get this down to just one model so you wouldn't be able to buy the 3x3x3 parts without buying the gears.

I've worked on this basically all weekend and I'm running short of good light now so I hope to have some better pictures and maybe a few more videos later but here is what I have at the moment.





And here is a video of my idea of a Fused Cube. This puzzle solves the exact same way however should it become illegal to actually glue cubies together in the future this puzzle will still be legal.

Video

With the pieces here you can build all 35 of the puzzles detailed in this table.



I could also add stickers to the center to give the face centers an orientation and then the total jumps to 70 different puzzles. I'm still toying with how best to design the face centers so I can get this effect with tiles printed in SWF which can be dyed with the other tiles.

Enjoy,
Carl

Wow! That looks awesome, Carl!

This is pretty incredible Carl! Proper gear design is quite a challenge.

When you configure this puzzle with the 1:1 gear ratios like the Gear Cube, how do the middle edges twist as you're turning them? It doesn't seem like there is the clearance for them to spin. The only thing I can think of is that you have to pull the halves of the puzzle apart like Oskar's Gear Shift in order to make a move. Can you explain how it works?

bmenrigh wrote:This is pretty incredible Carl! Proper gear design is quite a challenge.

This was pure joy. However now that I can design a gear puzzle I still have no idea how to actually design the gears themselves in Solidworks. So I imported the gears Oskar gave me in his challenge post. At some point in the future I really need to figure out how to make these.

bmenrigh wrote:When you configure this puzzle with the 1:1 gear ratios like the Gear Cube, how do the middle edges twist as you're turning them? It doesn't seem like there is the clearance for them to spin. The only thing I can think of is that you have to pull the halves of the puzzle apart like Oskar's Gear Shift in order to make a move. Can you explain how it works?

The easy answer is the middle edges don't twist. All the twisting is done by the blue gears inside the puzzle. On the outside it looks and move just as Oskar's Geary Cube seen here:

http://www.youtube.com/watch?v=cf2OnAKUuZg

So there is no need to pull the puzzle apart while turning.

Carl

Are the 1 to 4 gears just included so that the mechanics of other setups work? Great job!

Absolutely incredible! Brilliant achievement Carl!

PuzzleMaster6262 wrote:Are the 1 to 4 gears just included so that the mechanics of other setups work? Great job!

Not sure this answers the question or not but I've included 1:1, 1:2, 1:3, and 1:4 gear ratios as that is all the types that give unique states to the puzzle after they are applied.

You can think of a slice-turn-only 3x3x3 as one with a 1:-1 gear ratio but this produces the same states as 1:3.
You can think of a typical fused cube as having a 1:0 gear ratio but this produces the same states as 1:4.

Similiary I haven't included 1:5 as this produces the same states as 1:1.

All of what I will call "clasical types" of gear bandaging are possible with this puzzle. There are certainly more exotic types of gear bandaging that one can dream up. See Oskar's Variomatic Cube for an example.

Carl

Carl,

Wow, that looks great! Thank you for taking the challenge. Here are some more.
-Please show demo videos of all (35?) different constellations. As Tony Fisher once said: "Every puzzle needs to made at least once".
-Please show as assembly video of the 1:2+1:3+1:4
-Sell the kit

Again congratulations!

Oskar

Oskar wrote:Wow, that looks great! Thank you for taking the challenge.

You are VERY welcome. Thank you for the gears. You really should make a video showing how you make the gears in the software you use. I still reference TomZ's video about making sticker templates every time I make a new puzzle.

As for looking great... there are aspects of this that I feel should look better. While gluing the tiles on there are several places where there is now white on the back surface near the tile where the super glue made it out onto the surface. Also the CRC silicon spray leachesd some color out of the black parts and now my white tiles are no longer minty white and I think the green gears which are still in the puzzle now also have some black areas on them. After spending hours over the weekend dying all the parts I really wish it would have stayed looking much nicer. I'm tempted to take a second shot and leave the 3x3x3 parts white and then only dye one set of stickers black. I think this would help in two areas. The white from the super glue wouldn't show up on white parts. And I think white would also make the colors of the gears inside the puzzle pop. That and there would be no black parts in contact with the gears. Is there a better glue other then super glue which I could use and which would dry clear?

And one oddity I still need to look at more. One of the screw holes is either just a hair too big or I have a bad screw. There is a screw which I can screw in and out but it never tightens. I suspect a slight printing error on Shapeways part. I used the diameters that TomZ said to use for M3-0.50 screws and 5 of the 6 work perfectly. I'm wondering if I should use just a slightly smaller diameter hole... but before I go change the model I'll try a few different screws.

Oskar wrote:Here are some more.
-Please show demo videos of all (35?) different constellations. As Tony Fisher once said: "Every puzzle needs to made at least once".

35 videos of the same puzzle kit? That's far more then the total number of videos I have on YouTube now. So if I make all 35 versions am I supposed to name all the currently unnamed ones? That may be harder then making all the videos. I'll give this a shot but I think I'll need to learn some video editing skills so I can put more then 1 puzzle in a video.

Oskar wrote:-Please show as assembly video of the 1:2+1:3+1:4

This won't be too hard however I'll be out of town this coming weekend so give me a bit. Something key to knowing how to put the puzzle together is knowing how I marked the gears and if I redo things I'll make these marks bigger. They can be hard to see on the white parts before they are dyed as they are quite small but after they are dyed they aren't to hard to see in good light. Not sure how well they'll show up on camera.

Oskar wrote:-Sell the kit

I will offer the kit on Shapeways soon. Can you comment on your Paso Doble puzzle? Can it be put together without the gears and function as a V-Cube 3x3x3? I'm a bit worried about Verdes being upset that this kit can be assembled as a KO V-Cube.

If someone wants this Gear Cube kit PM me. After the amount of time I've put into I really don't think I can charge much for my time or it would be like a $1000 puzzle. Even without my time its still going to be my most expensize puzzle but I'll try to keep it reasonable. I don't plan of letting this one go until after IPP (or maybe at IPP if I haven't worked out something before then.)

Thanks,
Carl

wwwmwww wrote:

PuzzleMaster6262 wrote:Are the 1 to 4 gears just included so that the mechanics of other setups work? Great job!

Not sure this answers the question or not but I've included 1:1, 1:2, 1:3, and 1:4 gear ratios as that is all the types that give unique states to the puzzle after they are applied.

You can think of a slice-turn-only 3x3x3 as one with a 1:-1 gear ratio but this produces the same states as 1:3.
You can think of a typical fused cube as having a 1:0 gear ratio but this produces the same states as 1:4.

Similiary I haven't included 1:5 as this produces the same states as 1:1.

All of what I will call "clasical types" of gear bandaging are possible with this puzzle. There are certainly more exotic types of gear bandaging that one can dream up. See Oskar's Variomatic Cube for an example.

Carl

I'm meaning the puzzle with all 1:4 gears as shown in the video functions just like a normal 3x3x3, or at least acts like one after a full turn. Correct?

PuzzleMaster6262 wrote:I'm meaning the puzzle with all 1:4 gears as shown in the video functions just like a normal 3x3x3, or at least acts like one after a full turn. Correct?

No. The puzzle in the video does not function like a normal 3x3x3. A normal 3x3x3 has two independant layers per axis of rotation. The puzzle in the video only has one. If you rotate one face relative to the slice layer it determines where the opposite face will go. In the case of the puzzle in the video, it functions like a Fused Cube. The white, red, and blue faces can only be turned by 360 degrees. Anything other then 360 degrees and the opposite faces aren't in a position where they can be scrambled. The end result is the red/white/blue corner can never be moved. The red/white, red/blue, and white/blue edges can never be moved. And the red, white, and blue face centers are always in the same orientation. You have in effect fuzed them all together as they are in the Fuzed Cube yet they aren't glued together.

Carl

This is awesome! I would love to see a kit like this get mass produced!

Looks great and seems to turn great too! I'm kind of sad that I won't be able to afford this (or any other Shapeways puzzle for that matter)

Now let's have a Multi Gear Megaminx Kit!

(If you ask me [which you probably won't], the minx should gear together non-adjacent face pairs [or a mix of opposite and non-adjacent] so we could have a megaminx that jumbles!)

wwwmwww wrote:You really should make a video showing how you make the gears in the software you use.

Regarding gear design Timur has made this article and these videos.

Also, about the vmech copyright issue, couldn't one for example buy three Gear Cube Extremes, shuffle the pieces to end up with a vmech 3x3 with no gearing?

Are you going to redesign the cubies, or for example move the core to the file with the gears?

Volitar Prime wrote:This is awesome! I would love to see a kit like this get mass produced!

I'm certainly open to that. Mefferts and Oskar have worked together before on gear puzzles so I think its a possibility.

Carl

wwwmwww wrote:Mefferts and Oskar have worked together before on gear puzzles so I think its a possibility.

But we don't have 35 planets to name the puzzles after!

Absolutely brilliant work Carl. I really do love the idea and the execution of this. Well done.

Luke wrote:But we don't have 35 planets to name the puzzles after!

I believe there are currently 786 known planets, so we should be OK

But ideally it would just be a single kit with exchangeable parts, like this.

Coaster1235 wrote:Looks great and seems to turn great too! I'm kind of sad that I won't be able to afford this (or any other Shapeways puzzle for that matter)

I'm very happy with the function. I still need to test all the combinations but the puzzle works great.

Coaster1235 wrote:Now let's have a Multi Gear Megaminx Kit!

(If you ask me [which you probably won't], the minx should gear together non-adjacent face pairs [or a mix of opposite and non-adjacent] so we could have a megaminx that jumbles!)

That is a whole 'nother can-o-worms. But regardless of how its geared it still wouldn't jumble.

Coaster1235 wrote:Regarding gear design Timur has made this article and these videos.

Thanks for these. I think I've seen one of these before but I can't pull up YouTube at the moment. As I recall these were a great resource and I need to study them more, that is for sure. Still I'd love to see Oskar's take on it.

Coaster1235 wrote:Also, about the vmech copyright issue, couldn't one for example buy three Gear Cube Extremes, shuffle the pieces to end up with a vmech 3x3 with no gearing?

True but Verdes has given his permission to Oskar and Mefferts to use his V-mech for those puzzles. I'd still like to me sure he's ok with it before I offer this forsale. That why I ask the mods this in my first post.

wwwmwww wrote:However I want the mods to verify also that its ok with Verdes.

They have been in direct contact with him before and I haven't. Plus I suspect they are much more likely to get a response. I'm in no rush as it will be a few weeks before I have time to make the final touches to the model..

Coaster1235 wrote:Are you going to redesign the cubies, or for example move the core to the file with the gears?

My plan is to put all the parts in one model. This will make the cost as low as possible as there won't be a need to pay Shapeways multiple start-up fees. I'm also still toying with how best to give the face centers orientation information so I may make a change to the face centers. And I also need to pull up a drawing of an M3 screw to see if I have any room to make the screw holes smaller.

Carl

Luke wrote:But we don't have 35 planets to name the puzzles after!

Oh if I have to name each one it would probably be something like CubeABC where A is either F, 1, 2, 3, or 4 for the gear ration on the x-axis. B is F, 1, 2, 3, or 4 for the gear ration on the y-axis. And C is F, 1, 2, 3, or 4 for the gear ration on the z-axis. This would make the normal 3x3x3 equal to CubeFFF.

Luke wrote:Absolutely brilliant work Carl. I really do love the idea and the execution of this. Well done.

Thanks!!!

Carl

As far as I know, Tom uses 2.75mm (diameter) screw holes for his puzzles. I go slightly smaller, with 2.65mm. Slight fluctuation with the screw size can sometimes causes looseness, and 2.65 seems to be a good balance between being too large and too small. Alternatively you could go midway and use 2.7mm. Anything between 2.75 and 2.65 should be fine.

wwwmwww wrote:

PuzzleMaster6262 wrote:I'm meaning the puzzle with all 1:4 gears as shown in the video functions just like a normal 3x3x3, or at least acts like one after a full turn. Correct?

No. The puzzle in the video does not function like a normal 3x3x3. A normal 3x3x3 has two independant layers per axis of rotation. The puzzle in the video only has one. If you rotate one face relative to the slice layer it determines where the opposite face will go. In the case of the puzzle in the video, it functions like a Fused Cube. The white, red, and blue faces can only be turned by 360 degrees. Anything other then 360 degrees and the opposite faces aren't in a position where they can be scrambled. The end result is the red/white/blue corner can never be moved. The red/white, red/blue, and white/blue edges can never be moved. And the red, white, and blue face centers are always in the same orientation. You have in effect fuzed them all together as they are in the Fuzed Cube yet they aren't glued together.

Carl

All I was thinking was how only one layer was rotated 90 degrees. Even though you directly showed in the video this only occurs once per axis.

But that is actually pretty cool. Can the kit create any other non geared but bandaged 3x3x3s?

PuzzleMaster6262 wrote:Can the kit create any other non geared but bandaged 3x3x3s?

Yes, it can. If we use the notation I mention above then we have:

Cube444 = Fused Cube
CubeF44 = Cube with two adjacent faces bandaged to the edge inbetween them.
CubeFF4 = Cube with one face center bandaged to the core, i.e. it can't turn.
Cube333 = Slice-turn-only 3x3x3, i.e. all the corners are bandaged together.
CubeF33 = Cube with the corners bandaged into 2 groups of 4 corners each.
CubeFF3 = Cube with the opposite face centers bandaged together.

And you may be able to mix some 3's and 4's and make a few more but I can't keep those strait long enough in my head to be sure.

By the way, here is a puzzle for you (or anyone who wants to play along)... Prove that these 4 ratios cover all the posibilities. For example, why do I not have the gear ratio 2:3 represented between opposite faces?

Carl

wwwmwww wrote:why do I not have the gear ratio 2:3 represented between opposite faces?

Because I did not provide you with the appropriate gears? (Or did I?)
Oskar

Luke wrote:As far as I know, Tom uses 2.75mm (diameter) screw holes for his puzzles. I go slightly smaller, with 2.65mm. Slight fluctuation with the screw size can sometimes causes looseness, and 2.65 seems to be a good balance between being too large and too small. Alternatively you could go midway and use 2.7mm. Anything between 2.75 and 2.65 should be fine.

Nice!!!! This is EXACTLY what I was after.

Thanks,
Carl

Oskar wrote:

wwwmwww wrote:why do I not have the gear ratio 2:3 represented between opposite faces?

Because I did not provide you with the appropriate gears? (Or did I?)
Oskar

Actually you did. You gave me gears which make use of the 2:3 gear ratio which when used with the 1:2 gear ratio gears make the 1:3 gear ratio parts that ARE in the puzzle. If I paired them up with the 1:1 gear ratio gears that you gave me I could easly make a set of parts which would pair opposite faces together with a 2:3 gear ratio. However my point is that isn't necessary... if you did you would duplicate one of the "end puzzles" already on the list.

By "end puzzle" I mean in the way my Cube444 is a Fuzed Cube even though they turn very differently. Same with my Cube113 and your Slice Gear Cube. They don't turn exactly the same but they are the same puzzle to solve.

So what puzzle on the list is the same as the Cube with x=2:3, y=2:3, z=2:3? It is one of the 35 listed in the first post.

Carl

P.S. Here is a hint. Its one of the 35 that has already been built.

This is really cool! I would love to buy one.
I just want to make sure Oskar is okay with you selling these(I know he has commented on this thread, but it’s always good to be safe.)

2:3 is equal to 1:2, making it the Cube222 or Paso Doble.

wwwmwww wrote:That is a whole 'nother can-o-worms. But regardless of how its geared it still wouldn't jumble.

Suppose we have two non-adjacent, non-opposite faces geared together with gearing ratio 1:2. Now let's turn the '2' face (face that turns 2 steps when other face turns one step) once, leaving the it's counterpart face halfway in midturn. Now we can turn some face adjacent to the '2' face, but can't turn any faces adjacent to the '1' face in midturn. Unbandaging is required, and it jumbles. To my reasoning, at least. Hopefully I was clear enough.

eye2eye wrote:This is really cool! I would love to buy one.
I just want to make sure Oskar is okay with you selling these(I know he has commented on this thread, but it’s always good to be safe.)

Oskar, you are welcome to say what you want to say here. I'll simply say that yes Oskar and I have worked out a deal which will allow me to sell these. And yes that deal involves money. And yes that cost will be passed on to you... the end customer. Isn't that the way business normally works? I don't want to disclose the full details without knowing he wants those details out there... I'll leave that up to him. But I think his requests are more then reasonable. Without the gears he provided this would still just be a dream and I'm honored he trusted me with this project.

Now I realize that is all just a lot of hot air coming from me... so Oskar please comment.

Thanks,
Carl

Coaster1235 wrote:2:3 is equal to 1:2, making it the Cube222 or Paso Doble.

Correct.

Coaster1235 wrote: Suppose we have two non-adjacent, non-opposite faces geared together with gearing ratio 1:2. Now let's turn the '2' face (face that turns 2 steps when other face turns one step) once, leaving the it's counterpart face halfway in midturn. Now we can turn some face adjacent to the '2' face, but can't turn any faces adjacent to the '1' face in midturn. Unbandaging is required, and it jumbles. To my reasoning, at least. Hopefully I was clear enough.

Interesting... I hadn't though about leaving one face in mid-turn like that. And yes I think you could make a puzzle this way which would be almost impossible to unbandage but as the gearing ratio would have to be rational I think in theory it could be unbandaged with a finite number of cuts. Granted that number may easily become so big that it wouldn't be practical but I don't think I'd call this jumbling.

Hmmm... thinking a bit more... there would be an easy way to unbandage the puzzle. You could just cut the teeth off the gears.

Carl

wwwmwww wrote:...but I don't think I'd call this jumbling.

I think the jumbling is because of the geometry itself. The Bermuda cubes' faces turn by increments of 45 degrees, half of the 90 degree fourfold symmetry, and they jumble (after I've slept I can add a picture of attempted unbandaging). My intuition, and many puzzles, tell that everything that can turn less than 360/n-fold symmetry degrees jumbles, and so does the geary megaminx. I don't have a formal proof though.

Should this be diverted to PMs or another thread?

Coaster1235 wrote:Should this be diverted to PMs or another thread?

Hmmm... you may be right. But let's start a new thread. I guess the biggest question in my mind is rather jumbling should be applied to a gear puzzle. I view gears as a form of bandaging and I have a concern about the implication that you can bandage a non-jumbling puzzle into a jumbling puzzle. It would seem that the simple way to unbandage it would be just to remove the gears. But yes, lets start a new thread and I'm curious to see what others think.

Carl

I've now got a video which shows how to assemble Cube234.

VIDEO

I tried to film this outside but the heat kept causing my iPhone to shut off half way through the video. So I filmed this inside but at the moment I have my mother living with me. We have 2 cats and a dog. She has 3 cats and two dogs. So please forgive all the background noises. If you watch closely you'll notice one of the cats was even able to get part of itself on camera.

And I also made some head way on other fronts:

(1) The screw holes are now 2.65mm.
(2) I've redone the face centers so they now have 4 small tiles which can be added (or left off) which can be used to give them orientation.
(3) And I've now got all the parts in a single STL file. Through triangle optimization I was able to get the count low enough that I can upload it to shapeways. This allows you to avoid paying the startup/setup costs twice.

However this brings me to my next problem. The two models I've purchased are seen in the Gear1 and Gear2 pictures attached. The model I just uploaded with the above changes is seen in the Gear3 picture. However to print the combined model costs 15% MORE then the cost of printing the 2 seperate models even including paying the startup costs twice. I assume this is due to the inefficient packing of the parts. I need to do some digging on Shapeways as I don't fulling understand their pricing structure. If anyone could point me to that it would be appreciated.

But considering that Shapeways really flubbed their so called solution to the composite model support coupled with my strong desire to offer this to you at the best price I can has me seriously rething my standing on their backdoor approach. That method would give you the best price possible but would require anyone that purchased this kit to contact me so I could verify that shapeways printed the correct model. I'm very curious to hear your thoughts.

As for getting the best packing of the parts possible does anyone have tips on that front. I saw a post a while back where TomZ was working on something that might help with this. I'll check with him too.

Thanks,
Carl

I hate the backdoor too but it's the only way to get a reasonable price from Shapeways. You could do like Eric did with his Master Starminx and use the co-creator feature.

http://www.shapeways.com/model/493484/m ... rminx.html

Instead of somebody ordering a cube and hoping they contact you in time to prevent them from getting a nylon brick, the order stops at you for "customization" which allows you to always handle the model swap on your own timeline.

This doesn't solve the "Carl gets hit by a bus" issue but that'd be a bigger loss to the community that the inability to order a puzzle.

EDIT:
Here is the link: http://www.shapeways.com/Co-Creator_Pla ... ow_to_join

Edge Length is 60mm not counting the tiles. The tiles stick up 0.75mm above the surface of the cube so you'd need a box of edge length 61.5mm to contain the entire puzzle with the tiles attached.

As for weight. The Cube234 weighs in at 88.1 grams.



If you count all the parts in the kit, it comes in at 150.9 grams.





Carl

Thank you Carl for the measurements.