Research, research, research.

August, 2014.

The failed cut I tried by hand helped me understand just what I was attempting to do, and why a CNC could do it many times better than I ever could.  Not only would I get accuracy, I would also get repeatability as I go to cut multiple copies of something. I would also get cleaner cuts, as it would cut layers rather than the whole thickness all in one go.  So now the question was, how do I go about getting one?  I had a bonus from my company in August, of which I could use $900 for a machine.  I needed to do research.  Luckily, I have five friends that could help me, as they all have CNC experience.  Four of them I met at Solid State Depot, where I’m a member.

I started looking at designs: ones I could build out of MDF/Plywood, ones I could buy the parts for, and ones that were kits.  My friend Kevin helped me decide to go with a kit when he mentioned that my goal was to cut projects with the CNC rather than the CNC being the project itself.  My objectives were:

  • Needed to cut at least a 20″ x 30″ area, larger would be nicer.  This meant a minimum of 1000mm x 1000mm if I were to go the shapeoko route.
  • Needed to be somewhat quiet, i.e. not using a dremel tool.  My son’s room is right above where I’ll be cutting, and I don’t need it waking him up.  This meant a quiet cut spindle.
  • I wanted to use LinuxCNC instead of the GRBL setup that the stock shapeoko came with.  I was influenced in this decision by my friends at the hackerspace (one of which is a maintainer of LinuxCNC).  This meant a different controller board than the Arduino and motor shield that the stock shapeoko comes with.
  • I wanted it to be quick to assemble, as I wanted to get cutting with it right away.

I decided to do some research to find out if I could build the kit myself.  I made myself sick with research.  I finally came to the conclusion that the Shapeoko 2 kit was the best combo of convenience and cost, that I should go that route.  However, before I spent the money on the full “the works” kit, I asked my friends what they thought.  Most of them have gone the route of buying the mechanical kit instead of the full, and filling in with the parts they wanted rather than what comes stock.  An example would be the motors:  the stock shapeoko 2 comes with NEMA 17 motors, and they instead used NEMA 23 motors.  Another example is the controller: stock uses GRBL on an Arduino, and they use a parallel port optoisolator card with Pololu 8825 drivers, hooked to a PC running LinuxCNC.

I was ambivalent; on one side, I would have more work sourcing parts, but would get what I wanted ultimately, and cheaper than buying the kit which I would then have to upgrade.  On the other side, if I bought the Shapeoko 2 kit, it would be much quicker to get up and running, but I would have to upgrade it at significant cost.  My decision was made easier by Sparkfun throwing a moving sale where I was able to purchase the NEMA 23 motors at about half the regular cost.  At this point I remembered thinking to myself that with a deal like that on motors, I might as well try the blended approach rather than the kit, because then I would get what I really wanted.  I had friends that had done this approach, so I wasn’t wading into alligator infested waters.  Well OK – I was – but I wasn’t alone.

The sale at Sparkfun was on August 25th, 2014, and kicked off a buying spree to get all of the other components.  The research I had done was extremely helpful, and allowed me to buy a good many of the parts at great prices.  As for timelines,  I had to be realistic.  I wouldn’t get all of the parts until somewhere mid-September.  But I wanted so much to have it done, at least by mid-November.


And so it begins…

My CNC journey is a great example of how small projects avalanche into larger ones.  There I was, at a well known hobby store with my wife on the day before Christmas Eve, looking at the clearance items.  I happened across a piece of artwork that was unique in that it was a cutout piece that stood off of the wall for depth.  There was no question what I’d do with it: backlight it with LEDs!  The day after Christmas, I decided to use some of my gift money to buy it, heck, it was 50% off so why not?


It only took until April to actually wire up the LEDs.   I had some WS2801 LEDs on the polyimide tape strip that I bought to light up another project as well as some code to drive them.  I cut ten LEDs off of the strip and wired them up behind the openings on the tree.

IMG_20140415_220805463 IMG_20140415_220819483 IMG_20140415_220911154 IMG_20140415_220953240

When I was done, it looked awesome.  I decided to hang it in my dining room, and we turn it on whenever I have guests.  But the response I got from family and friends was very positive, enough so to get me thinking… maybe I could do these as art, and possibly have a sideline business that would put money BACK into my lab rather than it being a black hole…

Here’s a youtube video showing it in action:

I started looking into what it would take to cut such pieces. I had assumed it was laser cut, but after talking to a lady that owns a laser cutter business in this area, she mentioned that MDF would be horrible to attempt, as it’s full of glue and doesn’t burn well.  That’s when one of my good friends mentioned that if I looked closely, it looked more CNC routed than laser cut.  And he was right, you could see that it didn’t have sharp edges, only rounded ones where they used a 1/8″ cutter.  This was within reach!  This was do-able.  A laser cutter that could cut a 20″x30″ area would be way too expensive, and wouldn’t be right for cutting MDF.  However, a CNC machine to do so would be affordable, and could do it without the burning problems that the laser has.

I decided that I needed to test out cutting by hand, just to see if I could do it without spending the money on a CNC machine.  While the rotozip worked, it didn’t really want to be precise.  Nor did it forgive me for any wavering, as it would catch and drag the zip bit through the MDF and ruin a track.  You can see it has rough edges everywhere.


There was one other catch to doing this by hand:  as I was sketching out the shuttle, I realized that I had already drawn it too big for the board I was using, which could easily be handled if it was digital and I was scaling it.  Chalk up another reason to go with a CNC.

With that last question answered, a whole new series opened up.  How do I get a CNC machine big enough to carve 20″x30″ pieces?