Project Management and Computer Controlled Cutting

13 September 2017 | By Casey Evans

“You can't get much done in life if you only work on the days that you feel good.”

This is the final product from this week's project.


  • Group Assignment: Characterize your lasercutter, making test part(s) that vary cutting settings and dimensions.
  • Individual Assignment: Work through a git tutorial. Build a personal site in the class archive describing you and your final project. Cut something on the vinylcutter. Design, make, and document a parametric press-fit construction kit, accounting for the lasercutter kerf, which can be assembled in multiple ways.
  • Class Notes

    All taken in a notebook. Nothing transcribed yet.

    Training Summary

    We're in the EDS lab in the EECS department, which was kind of a stuggle to find, but with step 1 (find the room) accomplished we moved on to more important things like sign in, pick up safety glasses, remember that you can't go in unless you have close toed shoes, etc.

    Vinyl Cutter

    The vinyl cutter was introduced first. You just click on "EDS cut png" to get to mods, which is the user interface for the cutter driver. You may need to click on "modserver" first to stop the server. I don't remember exacty and would need to check, but that's in my notes. The vinyl cutter speed is adjustable, as is the pressure it uses on the material to be cut. There's a lever on the back to release and lock in rolls but it's pretty straightforward looking (haven't tried it myself yet though). When setting it up you need to make sure that the wheels are on the white patches so that it can measure your material properly. You can use the arrows on the cutter to roll in and out and reset the home location which is where it will start printing. You can adjust the knife by pulling it out and twisting it about if it's not working well. Note that you should barely be able to see the blade sticking out (vinyl is pretty thin). We've been told that you can also just use a piece instead of a roll but she wasn't able to demonstrate that (we think the piece was too small to work properly). You need to use a black and white image and it will cut around the black shape. If you want to scale the image you can change the dpi. Not really sure what that is but apparently it works. Make sure you're using a high resolution png.

    Laser Cutter

    CorelDRAW is the interface for the lasercutter. It accepts .dxf and .eps files. The size of EDS' lasercutters are 32x18 and 24x12, so keep your drawings within that range if you want to define the dimensions in your CAD program. Make sure you're working in RGB values. Red - vector cut, blue - vector engrave, black - raster. You hit File->Import. Check the colors then hit File->Print->Preferences (PLS675 or default selected)->Natural/MatBoard for cardboard, adjust any intensity/thickness settings->Apply->Okay->Print. Now it's been sent to the printer so you need to go to UPC which is the laser control panel. After checking the vents are open (one on the wall, one on the floor), the air assist is on (turn the power switch on the big one or the pump switch on the small one), the door is closed and all is adjusted, you are good to turn on the lasercutter power switch. Use focus view to move the laser and there is another control to move the image around. You can also replicate your image in this interface. Make sure that the laser head is just above the material using the focusing tool and adjusting z with the arrows on the machine (shown below).


    If the light is red, the laser isn't cutting. Hit play and it will start. DON'T LEAVE THE IMMEDIATE AREA WHILE CUTTING! You can hit pause and resume on the GUI, pause or open the case on the machine. So checklist items: compressor on, laser focused, GO, monitor job, wait ~30s before opening the case to collect your job (allows smelly air to leave). Other applications besides cardboard include cloth and plastic, but never cut plastic with a laser cutter unless you really need to (it's an "environmental plea" according to our TA). If your sticker has loads of little pieces you can use transfer tape to keep them all together. You can also make layered stickers if you want to use more than one color.


    Individual Assignments

    Assignment 1

    I've used git in the past and I have gone through a tutorial...the first semester of a microprocessing course we used GitLab, which required the command line stuff. Then we used BitBucket, which is compatible with SourceTree and very user friendly. I will set up/refresh myself with GitLab eventually probably but I'll want to speak with the TA to make sure that I won't copy everyone's repositories on accident like he suggested might happen. That would be the worst.

    I set up the link as a link to someone else's site who I knew was working first. Then I copied their code (and documented the source) to my website file and changed the link to that file. Once I got that working I actually tried to set up a site. I went to Wix because I thought it might be a good WYSISYG editor that I could copy the source code from but it doesn't seem to be that way. I'll keep using it as a way to record/compile my coursework but if need be I will craft up my own HTML (I took web design in 6th grade so I'm probably still a pro--actually though, that class is the main reason I don't look at my keys when I type, none of the keyboards had letters on them, it was awful at the time) or more likely I will find a better WYSIWYG that is more open source.

    Okay--downloaded git and it was being annoying so I just manually uploaded the files that I got from the internet (template for a Wikipedia style page). I've always like Wikipedia's organization style so it's a good opportunity for me to get a tiny bit under the hood. Also there are some good section ideas from this site

    It's kind of fun figuring it out. Since I'm not using git properly, I'm adding files one by one and I'm able to see how each file impacts the appearance of the website. I had to add the line to the header to get the favicon to show up properly. That was pretty exciting. This site taught me how to do that. It references this .ico editor which worked just fine: link.

    I used this to understand some more of the code (specifically linking down to further regions of a page).

    This has been an adventure and I'm obviously still in the works but it's a start!

    Assignment 2

    I've always wanted a Starfleet Academy car sticker (though I'd never put it on my car--maybe a laptop or something), so that's what I knew I wanted to do for the vinylcutter design. The image I found used two colors so I will try to do that as well, as explained by our lab tech. It was very straightforward except the lab didn't have the colors I wanted. Oh well. Also I am no good at aligning it so it prints where I want it to. The first time I printed over the edge of the side and the second time I printed over the first attempt. Here goes trial three. I take that back, it was just real close. Making a failure did give me a test case though which should come in handy. Using a ruler and a small pointy object is great for dealing with tiny pieces. I'm going to need tweezers. Below shows the slow progression from cut to actual sticker.


    And now it's on my laptop:


    For the press-fit, I will make boxes that can be made into dice, since my parents love the game 10K. The game requires 6 dice, and I will use the parameterized nature of the dice to make one large die to act as a box for the other die. Since the die sides will be pretty simple, it will be possible to get creative and assemble the kit in different ways (cheater die, other shapes, etc.) I plan to use Fusion 360 for the design and then export it as a .dxf file that I can edit the colors of in inkscape before finally outputting the file to the lasercutter.

    I started to get ideas for how to go about this by googling "press-fit construction kits" with image search. Someone cool named Martin Weiss made a breakdown of box faces.


    Very convenient. Since all the pieces are the same I'm guessing you could still get pretty creative with what you build, especially if there are 36 pieces as I plan to make. I will also raster on the dots...or, you know, maybe I'll just do numbers because symmetry is hard...who am I kidding, that'd be pretty lame. I'll give it a fair shot. But to do that I'll need another reference image.


    Image courtesy of this site.

    But then I also saw heart dice! Cool, but I don't really like hearts and I can't think of anything else cool enough to put there. Ah well. Maybe, you, potential reader, can think of something cool.


    Photo cred, which actually has a lot of cool ideas for different die numbering shapes (to include plain old numbers btw--so that's my fall back plan). Some designs are super intense and I will totally look into some depending on if I can get the basic shapes cut, lol.

    Fusion 360 is my weapon of choice. So I looked up some documentation to understand exporting .dxf (link) and parameterizing the design (link). I was having a lot of trouble over constraining things so I tried following along with the FreeCAD tutorial from class (link). The only difference is Fusion already considers the trimmed parts of your rectangle equal, so you over constrain if you try to call them equal again. Sheesh, it'd be nice if it noticed those things. But that would be a hard problem. To get the dimensions right I used this site as a reference:link. Of course it turns out that the reference image above is not very good for 3D. So I tried googling "press fit box" instead and found this link which has more reliable modelling examples on what pieces to use. But before I went to all the trouble of actually figuring it out, I decided to do some more modelling in Fusion to see how it would fit together. I was right! It would have little chunks taken out on the side, but when I showed my roommate, he liked them so I will probably try that design out first since it's very simple. Here's the 3D model:


    And here's the parameterized sketch:


    The important parameters are the sideWidth, which has been set to 0.67 in (the approx. size of a normal die), and materialThickness, currently set to 0.15 in. materialThickness is the variable I will be measuring in the lab before fabrication. Looking back to the 3D model, I realize I would only need one more "different" piece to make two copies of in order to close the gaps. So I'm going to do that. Here's the sketch.


    And its 3D model:


    This page was critical to getting this done: (link) It basically just taught me how to use upload in the data menu. Still having some issues with moving objects once they are in the design, but it's not a problem for the moment.And now I should add chamfers. I'm scared it's going to create terrible complications, but we'll see. Here's some brain storming on chamfer placement (red dots = chamfer), but I am going to go to sleep now...well, I will export the .dxl files and then go to sleep. I'm tired.


    Measured the cardboard - it's actually about 0.17in thick. But since our trials with similar cardboard suggested 0.13in makes the best joints, I will try that first. For the record, the dice are adorable without worrying about the top "monster" pieces (I say monster because they look kindof like monsters). For now the bigger issue is that it turns out that kerf is NOT negligible. My die fits together very loosely. I tried again with kerf incorporated (introduced a new parameter called laserKerf into my design, starting at 0.01 in) and it was closer but not perfect. It needs to be a bit smaller and I may start needing chamfers. Those I will just add in Inkscape because it's pretty easy. The second die also readjusted material thickness to 0.17 in which was slightly short in some areas and perfect everywhere else, so I'll keep it. I'll try 0.015 in for the next kerf estimate because just the 0.010 in change made a big difference and I don't want to overdo it. And we're still in the better, not best category, which would be fine for a static piece but these need to be pretty tight. Let's try 0.020 in now. Someone is using the lasercutter right now so I'll switch to vinyl. It was very straightforward except the lab didn't have the colors I wanted. Oh well. Also I am no good at aligning it so it prints where I want it to. The first time I printed over the edge of the side and the second time I printed over the first attempt. Here goes trial three. I take that back, it was just real close. Making a failure did give me a test case though which should come in handy. Using a ruler and a small pointy object is great for dealing with tiny pieces. I'm going to need tweezers. Back to the lasercutter...CorrelDraw hates me. It changes colors and it's hard to edit and meep. Very annoying. It also decided to create red circles under my black circles to bore holes through the die labels. I was thinking about trying that eventually but now was not it. On the plus side, I finally got the sizing right. Turns out it was a fluke so I went and printed 6 dice! Haha, I did a thing. They're beautiful.


    I was able to use the parameterization to easily scale up my box (changing sideLength and notchWidth but not materialThickness) to create a box for the dice. I did rescale the dots since those weren't part of the design but rather an add on.


    Group Assignments

    Our group characterized the lasercutter with a series of combs with different sized holes and keys. We also tested the effectiveness of chamfers on another comb. This process also helped us to get aquainted with the lasercutter software and hardware. The first combs we made were a test of concept. We didn't do the dimensions correctly and they ended up a weird size, but they were very pretty. It turned out we'd imported them as metric but they were originally imperial.


    We redid the comb with proper units here:


    We clearly got a little overzealous with the measurement labels but it was still able to demonstrate how great chamfers are. We found that although the width of the cardboard was about 0.15" the optimal cut was about 0.13" wide to create the appropriate pressure. Varying the intensity of the laser took it from no smoke to lots of smoke and black charring, so for the vector setting -20% was pretty good. 0% was a decent amount of smoke. Raster was good around 0% but could have used more. It depended a lot on the curve of the cardboard. The kerf of the laser was about 0.02in. Not a huge deal for this project, which accepted a range of ~0.01" pretty equally.