Vinyl Stickers with a 1610 CNC Mill/Engraver

The process for cutting out vinyl stickers with one of these cheap CNC mill/engravers involves mounting a drag-knife and figuring out the software.  It's not particularly hard.  I did it and it works surprisingly well. Of course, I do have a 3D printer to make the parts I needed.

A Drag-Knife is a cutting blade mounted in a holder such that it can freely rotate to follow the cut line.  The knife does not spin under power, it just turns to follow the cut. They go from big to very small. For vinyl stickers, you want the small ones.  With these, the knife is mounted in a bearing so it can rotate and is held in place with just a magnet.  The holder needs to be mounted in such a way that it can spring in the Z axis while being rigid in the X/Y plane.  The holder tip is designed to rest directly on the vinyl under the spring pressure that your mount provides.  The depth of cut is set by the knife stick-out which is set by the holder using a screw assembly.

I picked up a holder and 15 knives for $11 off Ebay (they're even cheaper now... search for vinyl cutter).  While you're ordering this, pick up some vinyl sticker as well.  As a hint, it turns out that people like sticking strips of vinyl at the top of their windshields and those strips are about the right width and very cheap too.

Anyway, about that mount and spring pressure.  There are various designs, some even involving linear bearings, but I decided to try out what's called a "Diaphragm Flexure" that I can 3D print.  It's a very interesting concept I came across looking for something else... such is life in the Information Age.

He're a youtube video of the basic concept:

Here's an academic paper on why it works:
http://www-personal.umich.edu/~awtar/PHD/Diaphragm.pdf

What we're aiming for is a 3D-printed precision linear bearing where the reciprocator is our drag-knife holder and the whole thing mounts in the CNC Z-axis carriage where the spindle motor usually goes.  Yes, a lot of constraints:  The drag-knife holder sets the internal bore, the motor mount sets the outer diameter, we need about 2mm of easy Z-axis movement, and it needs to be as rigid as possible in the X/Y plane.  Took me all of 8 prints to get it right.

The STLs are up on PinShape, there are Fusion 360 project files for the Diaphragm and Housing on A360, and you can download them all for free... though you're more than welcome to buy me a cup of coffee for my efforts.  The Fusion files use parameters for bar width and thickness and also the cut width.  With those, you can alter the springiness and rigidity easily (Modify - Change Parameters).  I started out prototyping in PLA, with surprisingly good results, then switched to nylon figuring it would give better life while springing.  However, given the space constraints, the nylon proved much too flexible.  I couldn't get enough rigidity in the X/Y plane without losing all flexibility in Z.  I considered going with ABS, but the PLA worked so I just went with that.  It's a very small amount of plastic and an easy assembly.  If it ever breaks, I'll just print another.  The 0.5mm test flexure took a surprising amount of abuse before it started breaking.

3D printing these diaphragm flexures is not hard but requires much attention to detail:

• Print on blue painters tape wiped with alcohol.  You'll never get the part off of glass or another solid surface without breaking the flexure bands.  Print, then peel the tape off the bed, part and all, then carefully peel the tape off the part.
• I use Cura and had to turn on quite a few advanced settings to make this work.  Realise that the springs in this are printed plastic bands only 0.9mm thick.  It takes a lot of slicer settings to get this right.
• The cut between these bands is very small, only 0.6mm and it can't be fouled with plastic or the spring won't work.  The model has a filet on the underside of each spring but you must still adjust how your printer lays down that first layer.  Too little clearance and the plastic will mush out and the cut will foul.  Too much clearance and the part won't stick to the bed.  Even with everything as good as I could get it, I still had to clear out some extra plastic under a magnifying glass while using a tiny scraper (an engraver bit actually).
• Make sure there is only one "body" in the STL and then print the top and bottom layers 'concentric' for maximum strength.  I went with a triangle fill pattern, but I doubt that matters much.
• I went with a layer height of 0.0875mm, which happens to match the step ratio on my printer, with an initial layer height of double that.  I did try an earlier one with a layer height of 0.04375mm (one step on my printer Z axis) but it doesn't seem to be needed.
• Set the top and bottom layer count to 2 each.
• I had to set my 'flow' to 85% (not fill, flow).  You will need to experiment with this to get the right amount of plastic laying down.
• Fill was also in the 85% range but I doubt it matters much.  Between that thicker first layer and then the top and bottom layers, there is very little room for fill in the spring bands.  The inner and outer bores are not particular.  It's the spring part that is important.
• The mount itself is not critical, just make sure the flexures tightly fit in the bore or you will need to glue.  Also, measure the distance between the 2 flexures when test fit on the holder and make sure this matches the length of the step inside the mount.  Your drag-knife holder may be a different length.  Note that the bottom step is actually a chamfer such that no support is required.

Print 2 diaphragm flexures and 1 mount.  To assemble: 

• Scrape/sand out the inner bore of one flexure until it just presses on the nose of the drag-knife holder.  It needs a decent interference fit, or you will have to glue.  I would avoid glue if I could, preferring to print again.  These diaphragm flexures can't last forever and glue will make replacement a lot harder.
• Do the same for the other flexure, sizing to match the other end of the drag-knife holder, but don't assemble this end yet. Yes, you will have to remove the drag-knife internals and locking nut.
• Verify that the mount internal spacer matches the distance between the flexures.  
• Press the drag-knife holder into the mount, seating the flexure that's around the nose.  Slide the second flexure down inside the mount from the other end until it seats on the other end of the drag-knife holder.
• If you've done everything right then you should have your mount.  The flexures should both be at about neutral, in the middle of their movement, and the drag-knife should easily slide up and down the bore a couple of millimetres each way while not moving side to side.  Put the drag-knife holder back together, insert a knife, and set your initial depth of cut.
• After that, pull the spindle motor out of the Z-axis of the CNC and slide this mount in.  It should be flush with the top and hang below the mount.  The bottom of the drag-knife holder must be lower than the lowest part of the Z-axis assembly.
• Block up the CNC table until you can get the label within range of the drag-knife.

To use:

• Make the top layer of the bed something that won't wreck your knife if the depth of cut is wrong.
• Tape some vinyl sticker down on the bed, masking tape along the edges works well.
• Pull the knife out of the holder.
• Lower your Z-axis until it just touches the vinyl, then another 1mm.  You should see the diaphragm flex a bit.  Set your zero.
• Raise the Z-axis and install the knife.
• Cut your sticker.  Make sure there's no spindle-on in the gcode, unless you've disconnected your motor entirely.  I just made a cradle to hold it on the top rail.
• It will take some experimenting to get the right depth of cut and adjusting said depth is not particularly easy with this setup, what with the adjuster and lock-ring down inside the mount.  But, it's a cheap conversion of a cheap mill.  Get over it.

Some random tips:

• The Fusion 360 '2D trace' toolpath works.  In the CAM setup, set your cut to zero, at the model top.
• Some fonts in Fusion 360 will generate 2 traces, both sides of each line.  Good for pocket milling out letters but very bad for cutting vinyl.  You can right-click on lettering and 'Explode Text'.  After that, you can easily delete one of the lines.
• Set multiple toolpaths if there's a cut inside another cut, as is typical in letters.  The first toolpath cuts the inside, then the next toolpath cuts the outside.  That way, if the depth of cut is too deep and the backing lets go, the inside will already be cut.
• Once cut, if the sticker is multi-part:

1. If the depth is off and the backing is partially cut then put some masking tape on the backside to hold things in place.
2. Then, peel off the waste vinyl from the backing (the part you don't want to be part of the sticker).
3. Then, take more masking tape but foul it up a bit first by tacking it on your pant leg or something.  It's an old painter's trick to stop strong tape from wrecking the work.  It only has to just hold things in place.  Put the tape on the front of the sticker in overlapping rows such that they go across parts of each sticker element.
4. Then, when you're ready to place the sticker, peel off the backing and any masking tape there.
5. Apply the sticker.
6. Remove the front masking tape, peeling row by row while holding the sticker parts in place.