The Machinist’s Cube is an old exercise that machinists would do to learn how to mill. The story I’ve always heard is that students would be given a completed cube and asked to replicate it with no further instructions. I did mine on a Chevelier mill with some aluminum scrap.
Stock scrap I started with.
Other side, cut in half from a quick jog on the waterjet.
I used a horizontal band saw to remove as much material as I could so it was more square to save a bunch of passes on the mill.
One side faced.
First iteration. I used the pocket function of the mill, although it’s only 2D so I set the Z axis manually. Zero it carefully, because if you’re off by more than about .1 mm, it will look off. I also did all 6 sides for one iteration, then the next, and the next rather than cut all 3 pockets on a side in one go. I think it’s easier to spot mistakes early that way.
Second iteration done.
Here it is in the mill. I used parallels to keep it level. Make sure to blow out all the chips from the vice before you put down the parallels, because even a single chip will throw it off by quite a lot. I used a 0.5″ endmill and calculated my feedrate and depth with the CNCCookbook calculator. They also have a cube calculator, which basically gives you iteration dimensions that look aesthetically nice (there’s no magical math to pick them).
When you start getting into the 2nd and 3rd iteration, make sure to go easy on the vice pressure. If you look on the right of the cube here, you’ll see a tiny bulge from me clamping it down too hard. I found that by listening very carefully, you can start cranking it down and listening for ticks of the metal buckling. Then you’ll know to stop. I just tightened it as much as I dared, made sure I couldn’t wiggle it at all with my hand and then milled. I was careful to watch it to make sure it didn’t start vibrating at all.
Before and after.
I didn’t deburr any of the edges because I liked how sharp and clean they looked without a tiny chamfer. It’s not so sharp that you can’t pick it up and play with it though.
This project took me about an hour per iteration, so 3 hours once I squared the stock. Not too bad, and a fun way to refresh my milling skills.