• Much CNC Progress

    David Wheeler01/26/2020 at 20:08 0 comments

    It's been a while, let me catch you up.

    We left off failing at milling polycarbonate plates on the CNC machine.  After that I tried a different technique for securing the work: masking tape on the bed and workpiece, superglue holding those to each other.  It worked and I made the first base plate without any major flaws.  To do this, I made a temporary waste board that bolted to my previous one.  That was Sept 22, 2019.

    On Sept 29, 2019 I trammed the new wasteboard for more consistent depths of cut on these plates.

    Oct 6, 2019: I was able to make all the base plates and top plates for the four LED signs.

    With the plates pretty much done, I moved on to making the top and side pieces which are based on aluminum angle stock.  I've never milled aluminum before and the angle stock is going to require a vice mounted under the CNC.  So there's a lot of prep work before I can proceed.

    Oct 16, 2019: I started making a set of oak jaws to mount on a vice.  This will be for work holding aluminum angle stock.

    Oct 19, 2019: The jaws work but there isn't enough clearance in my CNC to use it!  I'll have to modify the CNC table to accomodate it

    Dec 31, 2019: I've moved up the CNC rails a couple inches and now the vice with oak jaws fits.  I trammed the top and sides of the jaws to square with the CNC machine's reference frame.

    Jan 1, 2020: New year, new maker skill -- I machined my first aluminum angle stock piece successfully.  The jaws worked nicely.

    Jan 18, 2020: I cut down some 8mm steel rod to use as alignment pins.  Now I can place stock in the CNC vice consistently for repeating operations.

    Jan 19, 2020: I made 6 of the 7 remaining side pieces.  On the last one, the CNC process went wrong.  I wasn't watching but it made a bang and then the workpiece was misaligned.  I'll have to finish these another time.

    Jan 26, 2020: I made the last side piece and all the remaining top/bottom pieces today.  The CNC operations were uneventful.  :-)

    The aluminum angle pieces that were milled so far are for the front half of each enclosure.  The back halves are the next thing on my hit list.  I'm not sure if the CNC is the best choice for making those or just use the drill press -- these parts are simpler to make manually but more complex to CNC.

  • Getting My CNC Ass Kicked

    David Wheeler05/27/2019 at 18:45 0 comments

    I designed this thing with three plates to be CNC machined.  All the LED segments mount on a base plate and there are similar plates for the front and back of the case.  I've been trying to make them with polycarbonate on my DIY MPCNC machine but I keep running into problems.

    On my first attempt, back on April 7, I made the first base plate with a few mistakes.  First, I didn't set the Z position properly and at one point, the CNC plunged through my workpiece and jig, gouging the work.  Second, I found the table was out of level so cuts on right side of the workpiece were too deep and results were rough.  Worst of all, though, on the last operation to cut around the perimeter of the piece, the machine missed some steps in Y and the toolpath at the end was off 2 or 3mm from where it started.  So, the beginning and end of the cut ended up in different places.

    Following that session, I leveled the table again and made a couple changes to the G-code.  (Instead of cutting a slot for the perimeter, I wanted to try adaptive milling a narrow band around the piece.)

    The second attempt to CNC a base plate was on May 3.  With the bed level and more care on homing the tool, things went well until I got to that perimeter operation again.  It skipped some steps and ended in the wrong position again but this time I saw what happened.  The router's power cable and the Z stepper cable shifted over the right side of the gantry while the machine worked on the left side of the perimeter.  Then, when it moved over to the right, those cables were pinched between the router head and the right side of the machine.  As it moved in -Y on the right edge of the workpiece, those cables went tight and didn't allow the gantry to move.  Eventually it unstuck itself but by then the machine was out of position.  So that's another day and another piece of stock wasted.

    Preparing for another run at this, I tied those troublesome cables in a way they can't get stuck again.  I also made some changes to the CAM setup in Fusion 360:

    • Changed tool setting to use 0.004 in chip size.  (Which I read was min for polycarb.)
    • Changed CNC operations a bit: shallower slots, higher clearance, eliminated redundant ops.
    • Found setting in repetier-host to NOT disable motors at end of operation.
    • Added a screw to Z lead screw nut.

    I noticed, though, that the table the CNC is mounted on has become very wobbly.  The legs aren't clamped tightly anymore.  Not sure if this will hurt accuracy.  (It can't help.)

    So then, on the third attempt today, I clamped a new piece of stock on the fixture, CNC cut the holes along the perimeter and noted that the holes in the workpiece don't line up with the holes in the jig now.  There seems to be a really bad skew in the X-Y axes that is different than it was when I made the jig.  This is a bit of a head-scratcher.  I tried to just move forward by shifting the workpiece and screwing it down but even that didn't really work.  (And, anyway, I'd have to solve this skew problem before cutting the remaining operations.)  So this workpiece is destined for the scrap heap, too.

    I'll need to re-do the square calibration on the machine before I can try again.  I also need to address the Z nut since the screws that fasten it down keep stripping.  And I'm not sure when I'll have time to do those so this project will probably end up on hold for a while now.

    One thing I need to consider is whether it makes sense to have these laser cut by Ponoko instead of using my CNC.  I'd have to choose a different material but that might not be bad.

  • What's New?

    David Wheeler03/19/2019 at 18:15 0 comments

    Yeah, I knew this would happen.  I haven't updated this since December.  Why?  Because robot season.  I mentor an FRC robotics team (4638 Jagbots) and in January the design and build season start.  That takes all my free time until competition season is over.  Our last event will be this weekend.  Unless, that is, we make it to our district championship which is two weeks after that.

    So in a week or three I should have some free time again.  I'll post some updates then. 

  • Corners and Sides (in CAD)

    David Wheeler12/23/2018 at 05:09 0 comments

    This evening I generated models for the corners and side pieces to close the edges of the sign.  I'm using angle aluminum for the side walls and it attaches to the top piece.  There are 3D printed pieces in the corners.

    One thing I struggled with: the whole thing got 1/8 inch taller and wider to accomodate the side walls.  I tried to find a way around that and eventually decided it wasn't worth it.  At exactly 10x16, I had a 20x32 sheet of acrylic I could cut into quarters for the face plates.  Now I can only get three plates from it.  Oh well.

  • 3D Prints Done (For Now)

    David Wheeler12/20/2018 at 18:21 1 comment

    I've been keeping our 3D printers at work busy.  Actually, some "real" work bumped me off our new RAISE3D machine so I finished the last few pieces using our old Makerbot Replicator 2.

    Here are all the pieces together for the first time.

    And last night I ordered some electronics:

    • Li Ion Battery
    • Charger
    • 5V boost regulator
    • Display and Joy featherwing
    • ESP32 Feather.

    It will be a fun weekend if I can find time to play with all these new toys.

  • Fit Testing

    David Wheeler12/18/2018 at 15:57 0 comments

    This morning, the first 3D prints of the LED segments are done.  It looks like they fit together pretty nicely!

    And the LED strips also fit into their little windows:

    I made the windows 0.25 inch, which is a very loose fit around the LEDs.  And the spacing was based on 60 LEDs per meter, which works out to 0.656 inch.  The adhesive on the strips isn't going to do me any good since I have them mounted on the wrong side.  I guess I'll just run some tape over the strips to attach them.

    Since I need to cut the LED strip down into small segments and re-wire them, I'm wondering if there's any special technique for soldering them?  I'm also thinking that I may wire up the power connections for all the segments in parallel so I don't have to use a heavy gauge wire.  (The data and clock wires will need to run in series through all the segments, of course.)

  • Base Plate and Top Plate

    David Wheeler12/18/2018 at 02:58 0 comments

    Tonight's progress: a base plate with mounting holes and access holes.  Also a top plate with mount holes.

    The design is coming together quickly.  I should probably pull the trigger on some electronics orders soon.

  • 3D Prints Underway

    David Wheeler12/17/2018 at 18:41 0 comments

    I printed the first piece this morning. It's the smallest part, the dot, and took just 2.5 hours.  Looks good.  The rectangular window even printed OK without supports.  First articles of the horizontal and vertical pieces are queued on the printer now.  18 hours until they finish so they'll be ready in the morning.

  • Draft

    David Wheeler12/17/2018 at 03:53 0 comments

    Instead of going to bed at a reasonable hour, I learned to use the draft tool to tilt the sides of the wells out by 15 degrees.  Now the LEDs should be less visible directly and will put more illumination on the face of each segment.

  • And So It Begins

    David Wheeler12/17/2018 at 01:28 0 comments

    This is a project I've been wanting to tackle for a long time.  The idea is to use these giant LED numerals for cheering on our robotics team at competitions.  And with the Christmas break nearly here and robotics build season starting in January, it seemed like this would be a good time to get started.

    Yesterday afternoon, I sketched out the basic design.  (See the gallery.)  Each of the seven segments is illuminated by neopixel LEDs shining in from the side.  The volume of each segment is actually a white 3D printed shell that's open on the top.  When viewed from the front, the LEDs shining on white plastic will hopefully give a nice uniform lighting effect.

    The neopixel string will be controlled by an Adafruit feather with BLE.  This should enable a group of digits to be controlled by a single cell phone via an app.  I also want the operator of each sign to be able to program simple sequences without needing a phone.  So there will be a simple OLED display on the back with a group of buttons.  I'll figure out some user interface later in the project.

    Today I picked this up again and worked out a rough Bill of Materials.  I'll post details on that later in the project, too.  For now I estimate each digit will cost about $150 buying all the components from Adafruit or Sparkfun.  So four digits will set me back about $600.  Ouch.  Still, I don't have to spend any money just to get started.  First I want to figure out if the design is feasible.

    Since the idea is to use 3D printed wells for each of the segments.  I thought I should try to CAD up the shapes in Fusion 360.  I'm a bit of a neophyte with CAD tools so I wasn't sure if I would get through this without hitting a stumbling block.  But I did succeed in developing the necessary shapes.  There's a vertical segment, a horizontal segment, and a "dot" segment that will form a colon if I want to display the time.  Each segment has "windows" for the neopixels to shine in from two sides.

    When assembled, the whole design looks pretty sharp!  That's helping motivate me to carry through.

    The 3D printing time for one set of segments is estimated at just over 60 hours.  (Ouch again.)  I'll get it started tomorrow morning when I get to the office.  (Yeah, I'm using company equipment to make the prints.)