My first thought was: how the heck could you assemble a metal cube? Putting a plastic one together relies on the springiness of the material in order to place the final piece, but metal won't give in the same way. The solution:
> The face centers each receive a hollow stem that screws into the 5/16-24 thread. This assembly holds a captive spring-loaded screw, which screws into the core. The spring provides some give to the entire unit and allows the cube to be assembled and disassembled easily by snapping pieces in.
I had the same thought but just tested with a fairly standard "premium" plastic cube (not Rubick's brand) and I think it works in the same way. There are springs and that allows assembly/disassembly and I think it would work with a less flexible material. Not sure if the Rubick's brand ones are the same or if they really do rely on flexible plastic.
The original Rubiks brand cubes would rely on the springiness in the plastic. These tended to be rather hard to turn with even the slightest misalignment. Copy cats invented the improved mechanism. These will turn with a misalignment up to something silly like 40 degrees. Rubiks copied this mechanism for their own cubes eventually.
Long story short, a random chinese Rubiks cube knockoff is quite likely to be higher quality than an actual Rubiks brand cube these days.
That's very cool, and a beautiful piece in the end. Good choice to do it in 303. 316 really is a pain.
The ingenuity required to get things like this one on manual tools (especially pre-DRO, but even now) is amazing. Starting out machining, I expected the machining to be the hard part, but often a lot more time is spent on planning, fixturing, set-up, and so on. For example, the idea of making custom vise jaws to hold odd-shaped parts seemed crazy to me the first time I heard it, and now its just a completely normal way to behave. CNC makes things a whole lot easier, from being able to do complex shapes with CAD/CAM, to simple things like interpolated circles (when the mill cuts a circular shape by moving X and Y simultaneously, a seemingly trivial thing that's nearly impossible on a manual*). Having a 4th or 5th axis makes complex stuff even easier, and often completely avoids the need for wacky fixturing.
The lathe fixturing here is especially fun. A lot of old hobby books are written assuming you have only a lathe and no mill. Fun, often scary, fixturing was the order of the day.
* Ok, you can do crazy things with boring heads, but getting it right is really hard.
Yes, fixing stuff is one of the harder problems, as is avoiding toolstrikes or having to re-grip something because your carefully laid plans are messed up by the cutting depth or stability of your tools. Sometimes you spend more time figuring out how to safely grip a workpiece than you do on the actual machining.
This is so cool, stuff like this makes me wish I had studied mechanical engineering instead of computer science.
Does anyone know of a good route for learning the software stack associated with this stuff? I keep up with some maker YT channels like Stuff Made Here and they usually do an okay job of showing which machines were used and how each piece was made. But what they usually never show is which programs they used to model and mock up their creations before they started machining.
I'd really love to learn more about CAD modeling and designing these kinds of projects but I don't know how to get started.
I don't know of any great resources, but from my personal experience and understanding of the landscape is that Fusion360 is the most common real CAD program used by hobbyists, and small shops (especially those that are basically "hobbyist bought a bigger mill and moved into a larger garage"). Some of the bigger places will use SolidWorks (which is what 95% of my CAD experience is in), but the price there is way out of the range of someone just trying to learn.
In general, the aspect you're asking about is called an "assembly" - where you can bring in multiple parts that you have designed (and even design new parts "in place") to see how they go together and, to a minor extend, interact. I say "minor extent" there because most assembly systems aren't running full physics simulations or collision detection, at least most of the time - SolidWorks will happily let you design and assemble a model that is physically impossible to put together, while letting you rotate bodies through each other.
So yeah, I'd recommend starting with Fusion360. There are plenty of resources out there for learning it, but I do know that Grimsmo Knives and NYC CNC have videos showing how they specifically use it.
Here's what seems like a very in-depth video from Grimsmo on 5-axis machining (so definitely not applicable to just starting to learn, but this is the first I found - I know plenty of their other videos have details about fixturing and setup): https://www.youtube.com/watch?v=XqhctiVZtRU
NYC CNC has an entire playlist called "Fusion 360 for Beginners": https://www.youtube.com/nyccnc/playlists ; I haven't watched any of that, but I've watched plenty of their videos and enjoyed them, so think that playlist should be at least a bit helpful.
There's plenty more detail here, but I don't have the time at the moment to dig deeper - if you have any questions, feel free to leave them here and I'll see what I can dig up.
> Some of the bigger places will use SolidWorks (which is what 95% of my CAD experience is in), but the price there is way out of the range of someone just trying to learn.
FWIW, SolidWorks does have a low cost option ($99 a year):
And arguably there also features included in the above that are not available in the free version of Fusion 360 (available in the paid version, which has similar licensing costs to full blown SolidWorks).
This Old Tony, Blondihacks, Wintergatan, and others have done end-to-end videos from CAD to machining to assembly. I particularly recommend Blondihacks' beginner videos: well produced, fun to watch, and cover nearly everything you need to know to get started.
Also, face the reality you won't be doing work like this on day one. But you will make something fun.
So I'm a software engineer too, but also a cyclist. A weightweenie. At certain point, you begin to mix and match totally incompatible parts, solvable with a modest manufacturing capability. For me, the learning path was basically the same as with programming - lots of motivation, access to internet and tools (a machinist). Back in school, we were taught part sketching and machining basics, which helped a bit. I picked FreeCAD out of available open source solutions, watched a lot of great tutorials on YouTube and, though trial and error, managed to produce designs good enough to serve the purpose. Machinist services in Russia are incredibly cheap, especially if they have their own machines in a garage. With small beginner projects, the price of error is just your time and materials, don't be afraid to experiment.
On a side note, my mechanical engineer acquaintance complains that he should have followed the CS route because programmers earn substantially more here. The grass is greener.
I'm like your acquaintance. Did mechanical engineering, worked for a few years and realized the jobs around here doing mech eng kinda suck and the pays bad. In software now and sure, it's not hands on as much as I'd like but I earn enough money now to live comfortably and work on my own projects at home in my free time. For me it was a priority shift. Earn more so I can provide for my family better and then use the rest to work on modest projects as hobbies. Or earn less so I can work on mechanical projects but not be able to provide the quality of life for my family I'd like.
Fusion 360 has a free version and is very approachable. It's pretty popular among 3D printing hobbyists, but you can use it for all kinds of 2D and 3D designs.
I was Chris's housemate at the time, the victim of the hours of 'sssshhhh sssshhhh shhhhh'. Interestingly, he rendered this mostly by drawing boxes in pov-ray -- I'm not sure what he used to make dimensioned drawings, but it was definitely not any of the modern CAD tools, and they were mostly for his reference.
But these days, I do highly recommend Onshape -- it breaks down a lot of the 'rules' that I thought I knew about CAD software. I started using it about two months ago; one of my clients uses it for real industrial design of some IoT hardware, so it is powerful enough to do real things. Before I started using Onshape, I thought that 1) all CAD software was a million billion gigabytes, and required stupidly powerful hardware for no readily apparent reason, and 2) had an annoying licensing model that requires you to jump through hoops to get access to the free tier. Well, neither of these are true with Onshape: I went from 'hmm, maybe I should try this for my personal projects' to 'constraining a sketch' in about 90 seconds ... on Linux ... in Firefox ... on my Shenzhen ThinkPad ... with an Intel GPU. I was blown away at how much it failed to suck.
Anyway, my suggestion on choosing software is: it probably doesn't all that much matter. What you want to learn is the CAD mindset, not the software. An experienced MechE once told me that if you are not careful, you can end up writing 'spaghetti CAD'. These tools these days give you a lot of features that are, in theory, more expressive, but in practice, can result in unmanufacturable parts or unmaintainable designs: be careful!
> Before I started using Onshape, I thought that 1) all CAD software was a million billion gigabytes, and required stupidly powerful hardware for no readily apparent reason
Follow up idea, which I'm sure adds tons more complexity: have each face be a different material or finish. E.g. one face is stainless, another anodized aluminum, zinc plated for a third, brass on another. Each of the edge pieces would have to be a two-part assembly, and the corners composed of the three materials joined together (joinery, screws, brazing).
That surely would feel strange with the varying densities and thermal masses while solving, but would look very cool. To be fair, it appears from the pictures that each center piece's pattern is symmetric, which avoids the difficult part of aligning those when solving (center often doesn't line up when solving a cube with pictures on the faces in place of a solid color).
Probably far too niche for selling these to be profitable, but I'd love to have one.
I think that it would be difficult to keep everything clean enough to avoid galvanic corrosion at the metal interfaces once someone has touched it. I experienced surprisingly rapid corrosion when securing copper lugs using zinc-plated steel nuts and bolts until I started handling them exclusively with gloves.
The corner and edge pieces are guaranteed to have the right orientation, but the center pieces can be in any 90° orientation. Imagine an arrow printed on a center piece; a cube can be solved and still have that arrow pointing to any of the 4 neighboring sides.
Ah, interesting, thanks! So it's possible to reach a configuration where every side has the centre rotated but all the colours match up. I would not have believed that without seeing it.
I definately enjoyed the article, but apologies. I enjoyed the referrenced article https://kellycordes.com/2009/11/02/the-fun-scale/ much more. Perhaps it's just my mood at the moment, but it had me laughing out loud.
Regarding the Stainless Steel Rubik's Cube. What a work of art!
> Although it may be neat to look at, it weighs a massive three pounds (~1.4 kg) and is very tiring to manipulate. Solving this is more a test of the wrists than of the mind!
This was the punchline I was looking for. The aluminum version seem like a more desirable version at 500g.
With the same dimensions as the one in the article, tungsten would be around 7.3lb, and DU is only slightly lighter. If it was osmium, the densest known element, it would be around 8.5lb.
I saw a video (https://www.youtube.com/watch?v=4iB7kkCy1xM&t=160s) on EDM machining and the tolerances were magnificently tight. I had since wondered if such techniques would work for building a "one-move" Lemarchand Box (Hellraiser), assuming that engraving various designs would utterly obscure the actually disjoint pieces.
I am not sure what steel you would need. Something dimensionally stable, to be sure.
Wow, what a gorgeous piece of work. If you ever want to make something out of 304 and you're having problems: if you are too gentle with your cuts it will tend to 'smear', you really need to cut, probably a lot more aggressive than you might think if you are used to taking it easy on Aluminum and other soft metals which are pretty forgiving. 304 is not. The machinists joke is that it isn't hard, it's just hard to work.
My tendons are cowering in fear. How fast does it do finger tricks? It seems like there'd be a lot of resistance since so much surface-to-surface contact, where speedcubes typically flex on springs so that the shape of the cube changes to cut rotation on different curved planes.
Doing some quick image searches for teardowns, it seems that different models have been done both ways. Some with spherical bearings and some with intersecting cylinders. It makes some sense that either way works, as you're only ever rotating around one axis at a time.
He missed an opportunity to color the engravings on each side with the same colors from the original cube. I think it would have looked even better. But still incredible regardless.
I just bought a shower cassette removal tool (looks like Moen 104421), it's machined aluminum. I'm pretty sure I could make it myself...I couldn't make it for $15 and no time spent. This makes me sad.
I wonder if the relative similarity of the designs on the faces makes it particularly difficult to tell them apart, although if it's mostly for display that shouldn't matter much.
> The flat bottoms of the face centers do not allow the cube to turn as smoothly as it should; a problem that was rectified with a different design in the new cube.
just need an oven with good temp control. there is this lovely bronze color you can get on stainless.its strangely alot more stable than color you get from heat treating mild steel...but i bet it would rub off if you used it a nontrivial amount
The problem I imagine would be with the corner pieces. You'd need three different temperatures for the three different faces (with stable temperature across each face). I imagine that achieving this might be still technologically possible but almost certainly non-trivial.
The only way I could see of achieving that would be to heat each face very rapidly with something like a high-powered infrared laser, so it can reach oxide-forming temperature before the heat has time to spread. The low thermal conductivity of stainless steel compared to many other metals would actually be rather useful here.
> The face centers each receive a hollow stem that screws into the 5/16-24 thread. This assembly holds a captive spring-loaded screw, which screws into the core. The spring provides some give to the entire unit and allows the cube to be assembled and disassembled easily by snapping pieces in.
Brilliant!