I have another project lined up that I've been sort of dragging my feet on, because it involves removing quite a lot of aluminum — basically pocketing out the inside of a 4" x 4" x 1" piece of aluminum. The last time I cut something out of aluminum was a second spindle bracket for the Makita, out of 12mm plate. That was a three step operation:
- Roughing out the inside of the bracket: pocketing, 1/4" end mill, 0.6mm depth of cut at 900mm/min
- Roughing out the outside of the bracket: slotting, 1/4" end mill, 0.5mm depth of cut at 600mm/min
- Finishing the entire thing: outside profile with a little bit of slotting, 1/8" end mill, 0.5mm depth of cut 500mm/min
This morning I thought I'd set up a little bit of a challenge to apply some of the stuff I've been thinking about with regard to tool engagement to cutting aluminum. So my challenge was this: cut a pocket in 6061, 25mm x 25mm square with 6.35mm radiused corners, 7.5mm deep, as fast as possible.
To keep everything else constant, I worked up a toolpath in SolidWorks and HSMXpress, and kept it the same for each experiment: adaptive clearing (trochoidal path) with a helical entry at 350mm/min.
The way I would normally do it has a 8m40s runtime: 2.54mm tool load (think "stepover," but held around corners), 0.6mm depth of cut, 900mm/min.
So my first experiment was to run a 0.5mm tool load, 2.5mm depth of cut, 1,700mm/min at 10K RPM. Here's what happened.
This is definitely not a happy Shapeoko... but it didn't break anything either. The spindle is noticeably deflecting (it's especially bad at the extremes in Y). The finish at the bottom of the pocket wasn't acceptable to me — it had a fair bit of distortion as a result of the spindle twisting.
I then did a whole bunch of stupid stuff that isn't worth talking about, like rapiding 10mm straight down into the piece of stock....
I tried adjusting the tool load down to 0.25mm, but that didn't help very much. Decreasing the feed rate down to 1,200mm/min wasn't very helpful either. After a little bit more experimenting, I figured two things out (maybe obvious):
- The only way I could control the spindle deflection was with reducing the feed rate. The existing gantry carriage system just does not have enough rigidity to drive the spindle quickly through aluminum. I think....
- The Makita RT0701CX speed control is limited, in that the motor can't apply full power at low RPM. Tweaking the spindle speed up to around 15K RPM made things work out better.
A 2.5mm pocket on the left and a 5.0mm pocket on the right, same cutting parameters: There is one bit of deflection fairly early in the expansion of the 5mm pocket, which I think is due to chip buildup. (It is very hard to get all the chips out of the small pocket, as the video makes very clear!) Surface finish was absolutely acceptable to me for a roughing pass. This path would take 5m19s for the pocket... but maybe I can do better.
0.5mm tool load, 7.5mm depth of cut (yup, full depth), 450mm/min at 17.5K RPM: let's make some chips.
And what a lot of chips it is: The surface finish is, again, totally acceptable. I'd want to run a finishing end mill over it, but it's completely acceptable for a roughing pass. This would complete my challenge in 3m51s, which is half the time I started with.
I think a 3/16" 3-helix aluminum rougher would provide even better performance. I could more efficiently use the higher RPM range with the spindle, which should let me increase the feed rate (at the expense of decreasing tool load, but it's a net win). A end mill with chip breakers would help deal with chip build-up.
I think that given how most people seem to use Shapeokos (basically CNC routers, with mostly slotting operations) this is not a useful technique. Chip clearing is a major problem, and deep slots are very challenging. However, if you are doing an operation where you can get the chips out of the way (pocketing, outside contouring, etc.), and you can generate a suitable toolpath, this method shows some promise.