Yet Another RepRap Style Printer

thingiverse

Hi there, I made another RepRap style printer. Why? Because I wanted a 3D printer, and I wanted to roll my own design just for fun. Now that the printer is up and running, I would not only like to share the design with you, but also discuss some of the design decisions I made. Especially those that turned out to be the wrong decisions, after all the reader should be able to learn from my mistakes. But first some details about the printer. The frame was made using 30mm aluminum extrusions, resulting in a really sturdy, yet relatively light and inexpensive mechanical build. The two extrusions along the x axis are 500mm in length, all other extrusions are 400mm. The footprint of this design is quite large (about 580mm x 420 mm), but space was not an issue, so I am fine with being a bit on the huge side. The printing area is the standard 200mm x 200mm offered by the RepRap heated bed. The z direction also extends 200mm-ish (I have to get the exact amount of travel sometime...). The y axis was designed to be compatible with prusa mendel carriages (see the sources for the version I am using). This approach was quite convenient, but on the other hand designing a custom carriage should not be that hard if you would like to go custom here. I am also using a popular extruder that you can get of the shelf. My hotend is a handcrafted one-of, though. The design is close to the "All steel hot end for Makerbot" (again, see sources). Hotend/Extruder This is the first design decision that I would like to talk about. Buying a complet hotend/extruder from ebay. Do not do that, except when you know that you buy a quality product. It looked good first, but turned out to be terrible! The barrel of the hot end was made from PTFE, which was in contact with the heater block without any additional mechanical support. I do not want to go too much into details when it comes to bad design decisions other people made, but do not buy such a hotend. Period. After a short period of frustration, which ended in the total destruction of said hot end, I decided to make one myself. First I have to say that making your own hot end is fun and quite educating. I had to go through a few iterations, gradually evolving towards a working version. On the other hand, you will need the appropriate tools for this. Without a drill press you might get lost. You should also be aware that printing PLA with a metal hot end is challenging due to the friction inside the barrel. I failed in my attempts so far. ABS works like a charm, as long as I cool the cooling fins of the hot end with a fan (try not to point the fan towards the build platform while printing, this may have side effects with a fan constantly running at max speed). I read about lubricating PLA with some WD40 before pushing it into a metal hot end for some reduced friction, but I have not tried that so far. I am really proud of my custom hot end, so I want to show you some printing action: http://www.youtube.com/watch?v=H8cVUjxpEn0 Ok, back to the topic: Bad design decisions. Z Axis The first thing that I would have made differently today is the design of the z axis. I made a few mistakes there. I had the idea of supporting the z threaded rods with a ball bearing at the bottom, motors held on top. The expected wobble should be minimized with some kind of flexure bearing on the y ends (like in "Wobbless X Mounts for MakerGear Prusa"). Turns out that I designed that flexure bearing far too stiff, resulting in an enourmous amount of wobble. So, what to do if you encounter such a problem? Improvise! In my case the solution was quite easy. I moved the z motor ends a few mm higher. The lower end of the threaded rods is no longer supported by the ball bearing and hangs in mid air. In combination with the flexible shaft coupler, this results in a smoothly moving z axis guided along the smooth rods. The downside is that now the motors have to carry the weight of the whole y axis and the extruder, but so far i have not encountered any problems with that. Z Endstop I am using optical endstops. Although those can deliver a very precise position, my z axis was always a bit off (enough to prevent the first layer to print correctly). After a while I figured out what the problem was. The flag used to trigger the endstop was printed in light-blue ABS. As it seems, this material is not 100% opaque in the infrared used in the photoelectric sensor of the endstop. I immediately got more repeatable results after applying some duct tape on the surface of the flag. Nonetheless, a few days later my endstop position was off again. Seems like the tape is peeling off. I fixed that with more tape for now. More is better, right? In the long run I might replace the flag with one made of aluminum to fix that for good. X Ends As already mentioned, the flexure bearing in the e ends was too stiff to absorb the z wobble, but there were more problems then that. The parts have some tower- like structures designed to hold lm8uu linear bearings. They were designed much too thin, so they just broke when I tried to insert the linear bearings. In the end i removed them completely, drilled a hole all the way through along the axis of the bearings, and just inserted them. They are held by friction only, but that seems to be fine. Initially I planed to use two lm8uu per x end, but with the improvised design only one fits in. Another big problem was the hex hole for the nuts to drive the axis. Due to shrinkage they turned out much smaller than planed, so there was no way of entering two nuts and a spring to account for slipping in the axis. My solution was to cut out excess material until I could hammer one nut into the hex hole. That's it. Slipping is not really an issue, as the z axis tends to move in one direction only during printing. Extruder I am using a wades derivative extuder with a nice idler allowing for quick filament change. Unfortunately this idler easily breaks if you tighten the screws a little too much. I recommend printing one or two of them for repair purposes. If it breaks before you have a spare part, you can try to fix it with acetone. Just don't use too much of it, or the whole part will become rubbery. Other small stuff that can go wrong The feet of the printer are a few mm short to allow for standard screws on the bottom of the y ends. Fixed by using other screws. The belt coming from the y motor end does not have the same z distance to the printing plate as the one coming from the idler side. This is only a few mm and currently ignored. Works anyway. The y motor can collide with the base of the printing plate. Fortunately the heated bed is high enough to slide over the motor. I just cut out a bit of the bed so it won't hit the motor. Electronics I am using a RAMPS 1.4 with Pololus and a RasPi running OctoPrint as a print server. This is a nice combination, but you really have to make sure the firmware on the RAMPS is compatible with OctoPrint. In the testing phase I roasted on Arduino Mega and a Pololu -> make sure to insert the Pololus properly. Do so while the power is off. Also do not add/remove steppers while the electronics are powered on. Double check if everything is connected correctly. Otherwise it might cost you money... Conclusion Designing your own printer is lots of fun and you learn a lot doing so. It is very advisable to read a lot and look at existing designs before starting. I hope I can prevent someone from repeating some of the errors I made. But even if something goes wrong, it does not necessarily mean that you have to start from scratch. Adopt to your current situation, try to improvise. In the end it will turn out that some small wrong decisions can easily be ironed out. Just keep them in the back of your mind in case you plan to go for a version 2.0 sometime.