Kb58's granite CNC router build

[kb58]

Spent the entire day installing/removing/tightening/loosing the shims, about a million times it seemed. At the end of the day, the Y rails are within about 10 micron. Trying for anything better than that is on the slippery slope to insanity!

 

[kb58]

The chips have started flying with the two horizontal plates that ride on the Y bearings, and they're what the granite gantry sits on. They were rough-cut on the band saw first, then milled them as one. They probably should have been milled separately because my mill is a bit small to handle 5cm-thick material.

One issue I've been putting off thinking about is dust/chip collection on the router. Aluminum will (probably) be machined only rarely, with wood and plastic being used much more often. I suppose a spindle surround/hose/vacuum will be needed, but I'm already dreading the expected noise level.

 

[kb58]

The horizontal Y plates are done other than the mounting holes for the gantry end plates. Test-bolted them onto the linear bearings and confirmed that: I didn't screw up (the two are mirror images of each other); the hole spacing for the bearings is correct, and nothing binds as the assembly is rolled back and forth. The goofy cutouts allow the moving gantry assembly to get closer to the sides, providing an additional 10cm of lateral workspace.

Regarding the granite gantry, it's a bit of an unknown, being three 2-cm pieces epoxied together. The first trick will be getting the 100-lb thing onto the mill for measurements to see how straight and square it is, then grinding if necessary. I might have to ask for help if I can't come up with a clever way of getting it up there by myself.

 

[kb58]

Turns out that the gantry is 73lbs, so still manageable. It's too long to grind along the long axis in one go, and had to be turned around, throwing off the geometry. It was bowed in the center by about 0.3mm, but it was tough because laying it flat caused the bow to reduce from what it would be had it been standing on edge (like it will when installed). Ground both sides but I'm unsure how much better it is. I've resigned myself to "get on with things" and will use shims if required. Next up is drilling 28 holes to mount the gantry rails, whoo hoo!

 

[kb58]

Spent all day doing endless drilling but got it finished.

 

[kb58]

So... yeah. Shown is the largest of what's now three pieces of the gantry... Four full days of work down the drain and I have to start all over again. If there's a good side to this, it's that I didn't end up with crushed fingers, because I went down with it in-hand. I'm in a very foul mood right now.

 

[kb58]

After I calmed down, I approached a different stone supplier about making another gantry piece. It was kinda disappointing dealing with the young salesperson and having to explain what I want - about half a dozen times. I explained that the finished piece should be 50-60mm thick, so either three 2cm pieces, or two 3cm pieces would work. She said they have 2cm granite and showed me their remnants. I saw a thicker pieces and asked about it. "Oh that one is 3cm." Right... The funny/sad part was that when I suggested using two pieces of the 3cm, she had to use her calculator to convert that to millimeters. Ugh... Anyway, they'll get back to me on what the whole thing would run, including the drilled holes.

 

[kb58]

With the stone fabricator mulling over the granite gantry project, I'm cutting the aluminum for the Y drive subassemblies. That got sidetracked into wanting the Y-axis ball screw shaft to be the actual length, meaning the surplus parts on-hand have to be shortened. As expected, they're hardened steel, but an abrasive cut-off wheel took care of that. Much tougher was machining the shortened end to fit a bearing support. I got the bright idea of using the grinder to remove most of the material, then use the lathe to clean it up. That was really messy and inaccurate, so for the second one, a carbide cutter will likely give its life, but be far cleaner.

Putting the shaft into the lathe meant getting sidetracked - again - because the ball nut had to be removed, and we know what that means. I have some experience with them and suspected what I was in for... yup. Gave up at the end of the day because it kept binding up, meaning I didn't know how it worked (I can almost hear Jeremy Clarkson's "how hard can it be"). Somehow, I'd reassembled it wrong, with its 328 ball bearings (no kidding). Unfortunately, YouTube doesn't have assembly video on this particular type of anti-backlash ball nut - it's more complicated than most - but existing videos showed enough - I think.

 

[kb58]

Figuring out how this ball screw operates was quite the education. When first disassembled, it appeared as though balls circulate the entire length of the assembly, but in hindsight it was clear that this wasn't so due to how the crossover tubes are routed. Anyway, after many false reassemblies, it's back together and working smoothly. The only issue is that it's stiffer than before. That appears due to it being an anti-backlash type, where two ball screws are attached together such that they preload against each other. The two sub-assemblies spin freely, and it only gets stiff when they're connected. The attachment arrangement is very precise and not easy to relax a bit, so I'm puzzled where the additional loading is coming from.

In related news, it seems like Every. Single. Part. has to be modified to some extent. When the bearing carrier was inserted into the motor/screw shaft adapter, the mounting holes don't like up. Of course not. Either custom bearing carriers can be made, or modify what I already have. I chose the latter. It won't be pretty but it's function over form.

 

[kb58]

The left-side Y servo subassembly is done, other than modifying the motor-to-screw shaft adapter. It's pretty darn heavy - around 20kg - but is modular and will remain aligned on the machine or off. It's shown oriented as it'll be installed, up just below the Y rail. There will be four holes to drill in the side granite to mount it, then all the above gets to be done again for the right-side unit.