Saturday, September 17, 2011
Boot-strappable Open Laser Cutter: Part 2
I thought I'd post a quick update to my Open 3D-Printable Boot-strappable laser cutter) project. While some of the finishing touches or aesthetic details (such as side panels, or a top) are still incomplete, the machine moves fantastically under EMC2, and I've had much of the design built for some time. The main issue that I'm encountering right now is in aligning the laser beam to be accurate over the course of the 1m x 1m travel area, which I'm finding is not easy (but more on that later). Here's a tour of the printed parts, so far:
The system uses linear rails and pillow block bearings for linear travel on each axis. Upon each of the pillow blocks mounts a carriage -- in the case of the bottom axis, these carriages contain (1) two shaft holders, for the shafts on the top axis, (2) a pinch mechanism to hold onto a timing belt, and (3) either a mirror holder/idler mount (on the left side), or a NEMA 17 motor mount (on the right side). This is the mirror mount version, with the idler for the top axis also visible.
The timing belt goes through a pinch loop to keep it taught and transfer its motion to the pillow block. This pinch loop has an open center, so one can use large loops of timing belt without having to cut the belt and break the loop (I think this is a handy feature, incase you plan on making a larger machine, or using the belts for something else later).
The carriage for the NEMA 17 stepper motor that drives the top axis sits on the right hand side of the lower axis. I'm using the NEMA 17 motors from Adafruit ( [link] ) since they're relatively inexpensive at only $14.
The top carriage slides on the two shafts held by the lower axis, and also contains a mount for the Z-mirror and focusing lens. These optics are relatively inexpensive, and are from Light Object's build-your-own bundle ( [link] ) that a lot of folks on buildlog.net are using to build their own laser cutters.
The NEMA 17 stepper motor brackets for the bottom axis are much as they were in the last post, and have worked out pretty well so far. They mount to the aluminum extrusion with standard M5 bolts, and are reversible -- so they can face either left, or right. These are the same Adafruit steppers as used above, and there are two used on the bottom axis (one to drive each side).
The idler pulley brackets for the bottom axis are much as they were in the first iteration, only I've added a little more support to the design so they can take a bit more stress. They mount to the standard 20mm Misumi aluminum extrusion using the same standard M5 bolts that everything else uses. I've also added in a nylon bushing between the timing pulley and the M3 bolt that it's spinning on, which makes it glide fantastically better. All in all, I estimate that each idler mount with the hardware (minus the pulley) is only a dollar or two -- so very inxpensive. (Thanks to everyone who posted suggestions in the comments of the last post!)
The static mirror mount holds a mirror that deflects the beam at a right angle, from the CO2 laser, down the bottom axis. It then deflects off the mirror held on the bottom axis carriage toward the Z mirror on the top axis carriage, and onto the work piece.
This piece went through a bunch of iterations, largely because it's about 80mm high and needed to be rigid enough to not jiggle while the machine was moving. The hitch to this is that there are a bunch of bolts for mounting the mirror (top) or mounting the whole thing to the aluminum extrusion (bottom), so it can't easily be made solid. In the end this design works out pretty well, and is fairly rigid.
The laser mount was a fun and challenging part to design. The laser is about 2" in diameter, and also needs adjustment screws with another half inch or so of travel on either side. In my design, the laser also sits up 3 or 4 inches off the height of the aluminum extrusion. This is to allow a lot of depth to the cut chamber, and potentially enable some SLS experiments with a broader spectrum of materials than the Open Selective Laser Sintering project was able to use.
In the end, while I had designed some single piece holders, they were too large to print on anything but a RepRap, and so not too useful to the Makerbot folks. They also used a fantastic amount of plastic, so I decided to rethink the design. In the end I settled on using these 3D printable mounts, and mounting the laser tube itself into a piece of ABS pipe. While I was definitely into a "print everything you can" mindset, in this case I decided that a quick trip to the hardware store and $10 was more than worth the 10 hours of printing custom holders. I was also a little more comfortable with tapping the ABS pipe, and believing in its rigidity to hold the CO2 laser at exactly the right level. Even being off by a few millimeters could mean the beam hitting something it shouldn't (like a printed mount), and causing a catastrophic (and undoubtedly spectacular) failure.
This mount I'm actually very happy with, and it fits the CO2 laser wonderfully. Rather than use the CO2 laser for testing, for safety reasons I'm using a laser pointer that I've placed into a long ABS tube for alignment.
The CO2 laser tube tends to arrive in a fantastically large package to keep it safe (here, with a 3D printer and ikea desk for comparison ;) ).
The tube itself is about 70cm long, and about 2 inches in diameter.
The laser mount tube has a bunch of large nylon bolts placed at 120° angles to help adjust the laser path.
Here's the complete optical path, with both axes near their home positions. The total travel is quite large (about 1 meter by 1 meter).
For the controller, I'm using Bart Dring's Laser Interface Driver board. Honestly, I'd considered making my own board, but for just under $150 with the Pololu stepper controllers and connectors for something that's predesigned, it's really not that bad. I figured that the parts and a run of boards would cost me about the same to make, and this has already had the bugs worked out and has run quite well with EMC2 out of the box. I have had nothing but positive experiences using it, and it's even worked out well with the 2 steppers for the bottom axis plugged into a single driver. My only complaint is that it's a bit tricky to calibrate the voltages in on the stepper drivers, but that's more an issue with the spot Pololu's chosen to place the test point, and really you only have to do it once.
With the lights off, it's much easier to see the test beam in this picture. I'm having some issues aligning the optical path, and so I'm at a bit of an impasse. The beam seems to be at the same height at either end of the bottom axis, but in the middle it appears to move up a few millimeters (which corresponds to the entire carriage moving down a few millimeters). This isn't great, appears to get amplified as it moves down the +1meter on the final optical path before being sent to the workpiece. By the time the beam hits the Z mirror when the top carriage is at the far end (all the way to the right), it can wander quite a bit -- say up to 5mm or more. The beam width of the CO2 laser should be about 6-7mm by the time it reaches that distance, and the aperture to the Z mirror is only about 9mm wide -- so any divergence from being true really isn't okay.
As near as I can tell, this dip is coming from a bit of a sag with the rods. Placing a little bit of upward pressure on the bottom of the pillow bearings tends to level the optical path out, and so the rods appear to be sagging -- likely from a combination of their own weight, and the weight of the top axis. The rods are fairly large in diameter -- 12mm, or 1/2 inch -- so I didn't expect any sag, but I've also never constructed a machine that has a meter of travel before. It's not clear to me whether there's a quick hack I could do to help support the axis (perhaps sticking another piece of aluminum extrusion under the pillow bearings, along with some tiny wheels for it to travel on and keep the distance constant) -- or whether there's a better and inexpensive solution, like moving towards something like Bart's Makerslide v-rail based aluminum extrusion. I've noticed that the lasersaur folks seem to be using aluminum extrusion too, and are also dealing with similar distances to my design (I think their travel is about 2x4 feet, which is about the same as this design at 3x3 feet). That being said, I've noticed they've been in beta for quite some time, and haven't seen a video of their system cutting at full travel, so I'm not sure if they're also having issues with sag in their aluminum extrusion system at those large distances. Doing a couple of quick sag tests with some spare extrusion, it looks like it definitely sags less than the rods -- so it definitely has the potential to be a replacement if it comes to that. I'd just feel safer if there was an existance proof out there first for such large travels.
That's my quick update! Because this is an open project, I've posted the Google Sketchup and STL files to Thingiverse (available here: [link]), though there are still a few bugs left with alignment and such. Still, there are a bunch of interesting design elements that might be useful for other projects.
Thanks for reading, and please feel free to send along your comments or suggestions!
[Laser Cutter part 1] [OpenSLS Part 8] [http://cogsci.mcmaster.ca/~peter/]