Monday, May 30, 2011
Boot-strappable Open Laser Cutter
I thought I'd take a second to make a quick post about a bootstrappable open laser cutter design that I've been working on. I know I haven't posted relating to the Open Selective Laser Sintering (SLS) Project much after the Reciprocating Laser Cutter. Much of my lack of work on the project was due to our local hackerspace moving over the course of several months (when I had to help construct the new space), before quickly having to finish my PhD and begin a postdoc.
I love having access to a laser cutter. One can construct rather large items fairly quickly, and the materials available (like acrylic) mean that not only will one's creations be structurally sound, but they'll often look wonderful too. While my dad and I have been tinkering with our own 3d printers for a couple of years now, I still feel very much a beginner in constructing 3d models, unless using a tool more geared towards programmers (such as OpenSCAD). That being said, when I was introduced to laser cutting, the relative ease of designing and working in 2.5D as well as the (even more rapid) time to go from design to physical prototype simply made my creative self very happy. I remember several times waiting impatiently at the laser cutter for it to cut out the next sheet of some design I'd made that afternoon -- and upon assembling it, realizing how silly it was to be impatient that I could design, fabricate, and assemble some fantastic machine in only a few short hours. Not only for the OpenSLS project, but also for other creative projects, art, and tinkering, I greatly miss having access to a laser cutter, and decided to investigate how expensive it would be to construct one.
Several other folks have already been in this exact place, including the lasersaur folks ( http://labs.nortd.com/lasersaur/ ), who aim to create a fairly large format open laser cutter, as well as Bart Dring's excellent work on the open buildlog 2X laser ( http://www.buildlog.net ), a smaller format cutter that uses mostly off-the-shelf components combined with a few custom CNC-made parts. When I started thinking about this a month or two ago, Bart had just released his new design and had begin shipping the first kits to folks, and had quite a waiting list built up. I decided to get on his waiting list, but that I'd also get ahold of some of the parts that I could (such as the laser and power supply) to get a head start.
When the laser arrived, I ended up realizing just how big the tube is. With a tube approximately 70cm long (not including the mirrors to steer the beam), I learned that most of the length of most laser cutter designs is taken up by the tube. For example, with the Buildlog 2X design, the tube sits along the entire back of the machine, while the front includes both the x/y axis in the cut chamber (about 2/3 of the volume), as well as a separate area on the right that holds the electronics (about the remaining 1/3 of the volume). This is similar to a bunch of other designs, including the Chinese laser cutter we were using at my old hackerspace. I decided that while the ~20x12" cut area of the Buildlog 2X laser was great and much larger than what I had used before, that if the tube plus optics was nearly a meter long, I might as well make the cut area as much of that space as I could, and just stuff the electronics (which are relatively small) somewhere else -- say, under the tube. Making one axis a meter long meant the other might as well be a meter long too, so that the cut area would be square. (Put another way, I realized that from a cost standpoint, increasing the size of the machine doesn't really add a great deal of expense -- but it increases the utility by expanding the size of what you can build. I remember hearing Bre Pettis say that one of his thoughts, after having used a laser cutter for so long now, was that he had wished they had found a larger one).
Using these rough guidelines and inspired by Bart's design, I worked on designing and constructing my own laser cutter. Bart's design is really great, and he's put some wonderful work into it. Not only that, his v-rail design looks to make creating even large machines relatively easy. Unfortunately, with such a large demand for his parts, I decided that I might have to try an alternate design, and went with a more traditional bearings-on-linear-shafts design. The buildlog folks have amazing solidworks skills, but being new to building laser cutters, I decided to just sketch something out on a piece of paper, order a bunch of t-slot parts, and figure it out as I went.
The general design keeps very much with what other folks have been up to -- a t-slot frame, where the laser sits at the back, and a 2-axis CNC system sits in the front. Most of the difference is that I've added some room at the back under the laser to hide the electronics. Keeping SLS in mind, I'm also trying to keep a bit of height in the build chamber itself for future tinkering.
The "top" axis consists of two shaft mounts at either end (that mount atop the pillow blocks), one containing a NEMA17 motor mount, and the other an idler pulley mount.
The carriage for the Z mirror and focusing lens mounts on several smaller pillow blocks as well. Like the buildlog design, I'm using the inexpensive mirrors, lens, and the respective holders from lightobject. The idler and carriage have mounts for these parts, but still need some work. I also still have to add the tension mounts for the "bottom" axis belts to the shaft holders.
The "bottom" axis is driven at each end by a NEMA17 stepper. I thought this might be easier than using a single dual-shaft stepper in the middle, and designing my own drive system. It does require a bit more electronics (namely a second stepper controller on that axis), but otherwise should simplify things mechanically quite a bit.
The steppers are the standard, torquey steppers that Adafruit sells. I think the dual-shafted steppers used on the Buildlog 2X laser have nearly twice the torque, but there is very little friction in this design -- so these should be more than okay.
I'm on iteration 2 or 3 on most of the parts so far (since this is a learning experience for me, too). The second iteration of the NEMA17 holder is pictured above -- it has mounts to fit in standard 20mm t-slot, and is reversible, so you can flip it around (for the other side).
The idler design is very simple and low-cost, but I'm not sure it will end up working in the end. Here, instead of a bearing, the idler cog simply slides upon an M5 machine screw, while two jam nuts ensure that the screw holds in place and doesn't move. This seemed to work great before I started tensioning the belts, but now there appears to be quite a bit of friction, so I'll likely have another look at this design, and try to find some small (inexpensive) bearings.
That's my quick update! While the design is still a work-in-progress and I'm refining and printing out the custom bits, the rest of the parts are all sitting on the workbench waiting to be added in as soon as the axis design works pretty well. It's kind of been my hope that using these 3D prints I could get the system to work well enough to then design and cut out a more precision set of axis parts using the cutter. That being said, as I learn how to design for strength and resolution limits the quality of parts is getting better, and the latest set of parts is remarkably rigid -- so much so, that I'm hopeful they might not need to be replaced, and could serve as a long-term set of functioning axis parts for the machine.
I'm very much planning to make the source files available on thingiverse when I have a working design, in the hopes that folks with a reprap, cupcake, or thingomatic who are also interested in using their 3D printer to bootstrap their own laser cutter might find them helpful.
thanks for reading!
Where did you find those timing pulleys, by the way? They look quite nice.
I look foward to following your progress.
Thanks forest :). I'm really happy with how it's turning out, too! I hope to post the parts to thingiverse as soon as they're verified working, too.
The linear rods are 12mm dia at about 1m length, and this seems to be enough to prevent much sagging. There is definitely some sagging if you put a lot of weight on one of the axes with your hand, but the weight of the pillow blocks and printed parts is pretty low, so I'm hopeful this won't be an issue. If I were to do it again and the price difference between 12mm and 16mm wasn't too bad, I think I'd err on the side of caution and get the larger diameter, though.
The shoulder screw with a bit of grease is definitely a good idea (and one my dad had suggested), but given how well the pulleys spun without the belt tensioned on just the raw bolt, I decided to give it a try. With the belt tensioned there's a great deal more friction coming from them, so I think I might skip the shoulder bolt altogether and try to find a small 5mm bore bearing -- just to ensure that there's as little friction as possible. (There's already such low friction with the ball bearings on the linear axes, it seems a shame to add so much with the pulley if you don't have to).
Most of the mechanical components (including the t-slot, rods, and bearings) are from Misumi -- they custom cut the sizes to a very high tolerance, so it's great and there's /very/ little alignment issues right out of the box. The timing pulleys and belts are from Stock Drive Products, and the NEMA17 motors are from Adafruit. The optics are from Light Object (they have a special bundle right now), and the laser/power supply is from love-happyshopping on eBay. I still have to find an inexpensive source for the aluminum sides, polycarbonate window, and (parallel port?) controller.
♦ People who cut fabric want a width of say 1.5m and a length up to the length of the roll. Depending on how the most economic nesting of parts can be done, they are probably happy with less but long is good.
♦ Space is a problem, but most people have space under the mattrass they sleep on. I expect there are standard sizes for the divan bases we see dumped or freecycled, or your kit could specify a bed from IKEA with a note that exact size is flexible.
♦ Tinderwood construction is a problem, but a kit that specified lining with snipped-out sheet steel from old computers and filing cabinets would ignite less.
Just a thought. I have put a shorter version of this message on one of LAOS laser printer pages and am not sure if there are other projects to suggest it to.
i am building something similar in size, and i am going for supported rails to ensure no sag.
i have received a sample set of the above rails and they are fantastic! (i use UK suppliers as i am in europe).
Links to this post: