In preparation for teaching a class called “How to Make Almost Anything” at Omaha’s Metropolitan Community College FabLab I thought it’d be fun, and a good idea, to do some experimenting with their Roland Modela MDX-40A desktop CNC mill. I had already done quite a lot of reading and research on CNC in general over the past few years, and with my experience in DIY 3D printing this felt like a pretty natural step.
Surfacing the spoilboard
In CNC-speak, a spoilboard is a piece of material that sits between the machine’s bed and the material you want to actually carve your design into. It’s primary purpose is to protect the machine’s bed from damage caused by the end mill going too far while trying to cut out your piece – a common scenario that happens when your material measurements aren’t perfect or you are slightly off when homing the machine.
To ensure the spoilboard is totally flat relative to the end mill it is smart to use a large end mill to cut a uniform layer off the top. This doesn’t need to be done terribly often, but since I was starting fresh and wanting to teach students who had never done this before, I thought the practice would be good.
For my first project I tried to create something that is both customizable and would be easy enough for beginners to learn how to do in one evening. Using the software that was bundled with the machine (Vectric VCarve Pro) I was able to quickly lay out my name, set up my workpiece and generate a toolpath with very little effort.
I used an old piece of scrap carving wood to test out my design, and got somewhat funny results. Despite being manufactured specifically for carving (by hand, not by machine, I think), the wood had quite a lot of tear-out caused by the end mill ripping splinter fragments while cutting. This can be minimized with different software settings and more specialized end mills, but I was hoping to not make the students do anything that wasn’t totally out-of-the-box for their first project.
Small foam tiles
Once I had gotten the hang of 2D designs I was super excited to move on to my favorite application of CNC milling – carving 3D forms. When working in 3D it is generally a good idea to do multiple carving passes using progressively smaller end mills, using the larger ones to remove large areas of unneeded materials and the smaller ones to carve progressively finer with little resistance.
To demonstrate this process I made a series of tiles in pink insulation foam using the same design, stopping at different points. The final piece uses three passes, which means I wanted to end up with three tiles, each with a different number of passes applied.
The design comes from user tc_fea on Thingiverse, and is called “ripples in time“.
Large foam tile
Next I wanted to show how large designs could, at least in the X and Y directions, so I set up a piece of pink insulation foam that was the same size as the machine’s bed (10×10″). To keep things simple I used the same designed as the previous foam tiles.
In theory this worked out perfectly (and very nearly did in reality), but I forgot one small thing – when changing end mills I need a patch of uncarved material left over so that I can home the machine to the exact same zero coordinate between passes. By carving my design out of the complete 10×10″ work envelope I had left no spot to put my auto-Z touch probe to get perfectly consistent readings between passes. An error, yes, but a very good one to share with students!
Now that I was feeling more comfortable with working with the machine and the software I felt ready to take on a more interesting material – wood! I had a stack of scrap 2×4″ pine stud cutoffs from prior projects laying around that seemed like a great place to start, so I popped on in the machine and picked a model.
In this case I chose to go with Davision3D‘s “Melt Face” model.
In my excitement to perform the second pass I accidentally forgot to zero the new end mill, causing the machine to go too far into the model and cut out a huge chunk of material from the chin and cheek area. Oops! It definitely messes up the piece, but its still another great lesson to pass on to students – a tangible, visual demonstration of what can (and probably will) happen when you skip a step.
Continuing into the theme of hyper-realistic 3D carvings I thought it would be fun to carve something that had been 3D scanned from a real-life object. I found this great 3D scan of a trilobite provided by the Museum Victoria in Australia that was actually created from a real fossil they hold in their collection (reference number P149623).
First I carved the model out of another scrap piece of 2×4″ pine stud, which yielded a very interesting result. Although all of the major details came through, there was quite a lot of “fuzz” throughout the model. This is actually the same issue I experienced with my very first experiment (the name block), and is caused when the end mill is moving upwards on the model, tearing out splinters as it goes. As I said earlier, software tweaks and specialized end mills can help with this, but the material itself does not lend itself well to carving due to its relatively large grain.
A colleague saw the result that I got and suggested I use a material with a finer grain, thinking that it may be less susceptible to visible splintering like the pine stud was showing. He gave me two small pieces of poplar, which I ran the same exact carving job on. The results this time were like night and day! Very sharp detail and little to no splintering. Sometimes you just have to use the right material for the job!
Finally, just for fun, I decided to return to the ripple simulation model that I created in form earlier and see how it would look in wood. Despite the fuzzies caused by tear-out, I chose to go with the 2×4″ pine stud again to see if I could observe some interesting detail in the colorful banding found throughout the grain. Sure enough, when this model was carved out of the pine stud it revealed some really beautiful waves of internal colored grain, creating a pretty unpredictable, yet awesome effect!