Computer-Controlled Machining

Using a feed rate of 80mm/s being a bit messier than the default 60mm/s. Comparing the plywood with the OSB, there was a much bigger difference in fraying (see figure 9). We could conclude this feed rate difference matters less than what the wood used. Even in a single slab of plywood, cuts in the same direction can be frayed on one part of the board and be fine on another.

A second test to include the spindle speed did not really reveal any sweet spot for that particular piece of wood. Although, I this case I would use 14k rpm and a feed rate of 40 mm/s.

Apart from the fraying OSB the results we pretty similar between the OSB and the plywood. But, milling the OSB was a lot louder, probably because it contains more glue.

After the speed tests we ran a runout test by projecting the shadow of the rotating bit on a piece of paper filming that.

At 1k rpm it is a bit wobbly.

The alignment between the X and Y-axis seem pretty good (see figure 10).

Take the following steps when using the ShopBot:

1. Turn on the computer.
2. Turn on the ShopBot.
3. Remove the milling bit for inspection. See also: changing the milling bit.
4. Clean the collet.
5. Install the milling bit.
6. Clean the sacrificial layer and remove any unevenness.
7. Find and use the least damaged part of the sacrificial layer, which will be covered by the material, to do Z-leveling. See also: Zeroing the axes.
8. Place the material on the bed.
9. Open a new job in VCarve and enter the material dimensions. See also: Job setup.
10. If the material is large enough to stay in place during dilling, create a drill path in VCarve for screwing down the material onto the sacrificial layer. Keep them a few centimeters away from the paths you will be cutting and make them a couple of millimeters deep. See also: Drilling toolpath.
11. Check that nothing is left on the bed or leaning agains the machine.
12. Zero the machine to the bottom-right edged of the material using XY in ShopBot Console toolbar. See also step about Z-leveling.
13. Take a photo of the X and Y coordinates.
14. Start the dust collector.
15. Start the spindle. See also Starting the job.
16. Select the drilling job as exported from VCarve.
17. Run the job.
18. Turn off the spindle and move the head away from the material.
19. Screw down the material in the drilled holes.
20. Start the spindle and run the job.
21. Done? Stop the machine and vacuum, remove work from the bed, and clean up the mess!

Before milling the project, I wanted to do some press-fit tests. I used the model below to make profile toolpaths with different Clearance offset values.

bit_diameter = 5;
thickness = 18;

$fn = 200;

module slot_mask(coordinates, bits = [true, true, true, true]) {
  r = bit_diameter * .75;
  square(coordinates);
  if (bits[0]) { translate([r / 2, r / 2, 0]) circle(r); }
  if (bits[1]) { translate([coordinates[0] - r / 2, r / 2, 0]) circle(r); }
  if (bits[2]) { translate([coordinates[0] - r / 2, coordinates[1] - r / 2, 0]) circle(r); }
  if (bits[3]) { translate([r / 2, coordinates[1] - r / 2, 0]) circle(r); }
}

slot_mask([thickness, thickness * 3]);

translate([thickness * 2, 0, 0]) difference() {
  union() {
    square([thickness, thickness * 3]);
    translate([thickness, -.5 * thickness])
      square([thickness * 1.5, thickness * 4]);
  }
  translate([0, -thickness, 0]) slot_mask([thickness, thickness], [false, false, true, false]);
  translate([0, thickness * 3, 0]) slot_mask([thickness, thickness], [false, true, false, false]);
}

Former student Paola Zanchetta documented her work on doing a clearance test in her week 7 documentation. I started by trying her values.

VCarve Pro can import PDF-files (not SVG). OpenSCAD allows exporting to PDF, but it only exports to paper sizes like A4, so it will not work for "something big". Fortunately, exporting to SVG, opening it in Inkscape and exporting there to PDF does produce a PDF with the correct dimensions.

Before creating a new job in VCarve, I measured the piece of plywood I was going to use: 2000mm by 1000mm and 18mm (varying between 17.9 and 18.1) thick (see figure 13). For the zero position on the Z-axis, I selected the bottom of the material.

Next, I imported the clearance test and table PDFs I created. While importing I noticed the dog bones fillets VCarve creates are much wider than the bit diameter (see figure 14). I corrected my design to make the holes a bit bigger (because maybe I am too optimistic?), and re-imported them.

I had a bit of a struggle defining toolpaths and creating SPB-files. The icons in the toolbars are a bit confusing so it felt like feeling my way through the dark, but after a while, I got the hang of it.

I started with drill holes (5mm circles), which I placed around the clearance test shapes. For the toolpath I selected drilling using the already selected "Fab Academy End Mill (5mm)" option and make sure it was what I expected it to be (see figure 15).

Ready to mill! I removed the bit to inspect it (see figure 16), reattached it, cleared the bed, z-zeroed on a good part of the bed, put the plywood on it (eye-ing the alignment with the bed), x-y-zeroed the bit on the bottom-right corner of the material, turned on the vacuum, started the spindle, loaded the drilling job into ShopBot Console, and ran it.

All went as expected. After the machine was done, I moved the head to the back (along the X-axis) to make some room to screw the board down using "woodies" (see figure 17).

Then, once it was screwed down, I copied the slot part of the test four times and gave them clearance offsets: 0, 0.1mm, 0.2mm, and 0.3mm. Satisfied, I started this job, but after a couple of passes on the first element of the test, I realized I forgot to add tabs. Afraid I would be catapulted to death by a loose piece of the wood, I stopped the job, adjusted the toolpaths to include tabs.

While it was cutting the last slot, I used calipers to measure the slot width; it was too narrow… I should have used negative offsets. And the skirt came loose (tighten that wing nut!).

I added new slots with negative offsets and ran it again.

After unscrewing the board and popping out the spoils, I tested the fit. Clearance offset -0.3mm gave the perfect fit (see figure 19).

Next, I reinstalled the plywood on the bed (without screws!). I prepared a lot of drill holes around the table shapes, ran the job, and screwed the board down again (using only the new holes!).

After 37 minutes of loud milling (and Così fan tutte on noise-cancelling headphones), it was done. I was at the Waag at 8:00, the job finished at 12:00. Henk was not kidding when he said it would take at least half a day with the CNC!