Kb58's granite CNC router build

[kb58]

Been thinking about how to cut the linear rails when the time comes. There are six of them, two being of manageable size to cut with an abrasive disc, then two more that are "that's a bit much", with the last two being "I don't think so." There's a small water jet business in town that is happy to take nearly any job (their primary business is cutting stone inlays per rich home owners desires. They cut some stainless plate for me which came out great, so I'll pass this by them. The cuts would be to just shorten the rails some, so the water jet's issue with cutting with a bit of draft isn't an issue. Anyway, that's a ways off yet, I'm still finishing up getting it all into CAD first. Don't want to risk cutting the rails short...

 

[kb58]

I'm already a bit scared of the thing even before it's built. The gantry/spindle/rails/truck assembly will be around 80kgs, and the thought of it heading at full speed toward the end of travel is intimidating.

Speaking of limit switches, I've used many different types in other applications, and while solid state ones can be really accurate, I'm not sure that they'd be more reliable than mechanical switches in this application. I'm currently leaning toward either magnetically activated reed switches or lever-type micro switches due to their utter simplicity.

Speaking of travel limits: The base surface plate is 61cm (Y axis) x 91cm (X axis), but once setbacks and component sizes are factored in, the actual working space gets much smaller, ridiculously so. In the 61cm direction it gets reduced to 25cm of travel. Yup, 36cm gone due to the sheer size of everything. The main culprit is the length of the Y-axis trucks, at 19cm apiece. If I use two at each end like everyone does, the above limitations occur. There are three options: live with the reduced working space; remove one truck at each end and count on the sheer size of the remaining ones to handle alignment challenges or extend the Y travel.

The purchased rails are sufficiently long for extending the travel, so it is tempting to use their full length simply because they're here and may not have to be cut. The catch is that they'd extend beyond the back edge of the base surface plate by approximately 30cm. If that option is chosen, the question then becomes how to support the overhang. Again, more options: steel, or more granite. Thermally, more granite is best, because it would expand at the same rate as what it's bolted/glued to. Granite has a thermal expansion coefficient of around 8*10^-6 m/m deg C, while steel is about 50% more. The question is: how much do I care? The vertical dimension is the important one, at 0.3 meters. Say temperature swings by 30 deg C, so a piece of granite that size will grow/shrink by 0.12mm. Steel's expansion rate - over that of granite - would be about 0.06mm, and we're back to the original question about how much I should care. Unless someone convinces me otherwise, I think 0.06mm is probably acceptable*, but I know it increases the overall tolerance of the machine.

Aluminum is not being considered because its expansion rate is twice that of steel.

*Thinking about it again, I think it's even less than that since it would likely be a diagonal brace at 45 degrees. So the 0.06mm would get divided by square root of 2, reducing it down to 0.04mm.

 

[kb58]

Decided to go with granite, buying a small $50 surface plate and cutting it diagonally, forming two triangles with presumed accurate 90 degree corners*. It'll be drilled for studs and glued onto the existing side pieces. Yes, the right solution would be to buy longer side pieces to begin with, but bought them before that became apparent, plus I was not able to find 36" x 12" x 2" granite.

* This is an assumption, that the sides of the surface are sufficiently precise to be used as reference surfaces.

 

[kb58]

139 lbs of aluminum tooling plate, ranging from 6 - 25 mm thick. These are left-overs sold by the local industrial metal supplier, pieces that customers didn't want after their orders were cut to size. The pieces are random sizes, so bargain hunters sort through it to find what they need.

 

[kb58]

And so it begins, starting with drilling 12mm holes in the side pieces, then those holes will be used as guides for drilling into the base surface plate. Threaded inserts will be put in the surface plate, and the sides glued and attached. The yellow stuff is play dough forming a dam around the drill, and is filled with water.

 

[kb58]

Turns out that the two granite end pieces just fit onto the mill with a diamond grinder (left over from grinding the floor during the remodel). The shear weight of the two pieces was hoped sufficient (it was) to keep them stationary while grinding the tops flat. I admit that it was sketchy, worrying that the grinder might catch an edge and push them off the mill, so they were only ground 0.002" at a time. Next up is drilling the anchor holes in the main surface plate. It's nice to get this hard part of the build out of the way early, though the summer heat isn't helping.

 

[kb58]

Spent all day in the sauna that passes for a garage, too stubborn to leave until the current drilling phase is complete. At the end of the day, the side pieces and base surface plate are drilled, inserts added, and are ready to be bolted and glued together. Before that happens though, I want to drill the top of the side pieces for the linear rails while they're off the table and easier to position for drilling. It's more holes to drill but the holes are smaller so it should go faster.

 

[kb58]

Six hours of steady drilling to create 12 more holes. Mocking up the triangular pieces to the back side of the granite side pieces showed that either the triangles or granite side pieces are not cut at true right angles... I suspect the granite side pieces, which are ordinary garden steps, offering little in the way of precision, but they'll be made to work.

 

[kb58]

Glued on the granite end plates - it could have gone better. Somehow, "someone" screwed up and the carefully ground end pieces aren't level and will need correcting. To avoid more messy grinding, the rails will be shimmed. I bought a used Mitutoyo digital dial indicator that is very nice, measuring all the way down to microns. Of course, all those extra digits do is show just how far off from perfection everything is!

It's already clear that the toolbox is going to need leveling feet, as it's a bit concerning how it can be rocked, so a frame will be added around the base.

 

[kb58]

Carefully measured the top surface of the Y axis to determine what shims are needed and where, then ordered a set. McMaster is extremely quick and it'll be here tomorrow by 9am even though today is Sunday. The Y rails were placed over the freshly-glued-in studs with washers to serve as a fixture to keep the studs centered. I put the scale on the roller bearing housings to show just how huge they are, and it's why I briefly considered using just one roller assembly per side instead of two. After thinking it over, it made more sense to extend the travel and use what I have, plus, the rails don't need to be cut.

Next up is fabricating the two horizontal aluminum plates that the granite gantry beam sit on.

In other news, I was using the mill and had the X axis power feed on, but not engaged (the motor was not running). That made the smoke and sparks that shot out of it all that much more surprising. I bought the mill and power feeds back in 2019 and saw that they have a 5-yr warranty. Nice, but... that's only if they install it, so I was out of luck. I took it apart and found that the motor is fried, which again, is very strange given that it wasn't running. Annoyed that they don't have a better warranty, I decided not to reward them with another purchase and instead moved the Y power feed (that isn't used much) over to the X position.