Rostock Mini build, Part 1
I have had my prusa i3 3D printer for over a year now. I really enjoy messing around with it and have wanted to build another one basically since I got the first one. Well, since I got my shapeoko mill, I can now cut out a lot of the parts I would normally have had to buy somewhere else or sent away for cutting. So, now that I had the capability (and a little bit of spare income), I decided to finally start building my next printer.
The second printer
For my second printer I wanted to try something different. I had a cartesian printer already, and delta style printers looked very cool. So I decided to build a Rostock Mini 3D printer. The rostock design is actually depreciated, but I had several of the components already printed and figured I would give it a shot anyways.
This printer only has a 6×6 build area, but so few of the things I actually print are that large, so this shouldn’t be a huge limitation. Delta style printers boast higher positioning rates, meaning potentially faster printing speeds. The bed on a delta is also stationary, so you can affix and level it easier. The printer I am building will use 1.75mm filament, a remote placed bowden extruder, and an all metal hot end. All three of these things are different for me compared to my other 3D printer. I know enough about programming and how to read the source of the controller software to be able to tweak the right settings, so I should have fewer problems getting it up and running compared to my first time.
The specific hot end I purchased is the Hexagon hotend from makerfarm.com. It looks fairly well made, and all my parts from makerfarm already, so this was just convenient. I primarily print in PLA plastic, but with this all metal hot end, I should be able to try out some more exotic materials as well. I plan to use a geared bowden extruder, mostly because I do not think my motors are torquey enough to handle direct drive.
Printing/Cutting the Parts
All of the parts I am using I have printed/am printing myself on my Prusa i3. I also came up with a couple of the design changes myself.
I recently cut out the top and bottom plates out of ¼” clear acrylic. A couple of issues with the bit slipping during cutting were remedied by drilling out the holes manually with a cordless drill afterwards.
I also cut out nema17 motor gaskets out of ⅛” thick cork sheet. These should help stop vibration and heat transfer from the motors to the plastic brackets. Eventually, I plan to add these gaskets to both my shapeoko and other 3D printer too. Also, I cut out a cork insulator for the heated bed. This is ¼” thick cork sheet, and has a bunch of holes cut into it so that all the fasteners do interfere.
I also printed 3 spool drums for my linear motion. For this printer, I am going to use high tensile braided fishing line as the drive. Normally, people use belts and toothed pulleys to do this, but I liked how this idea looked and wanted to try it out. The brand of fishing line is called Spectra line, and its tensile strength is something like 70 lbs, which is way higher than this machine will work at. The pulleys were designed by another reprap user, who uses them on his own delta style derivative (richrap, he uses them on his 3DR printer). I also printed V groove pulleys to use as idlers for the line.
I designed a spool holder to hang my spool off the back of my extrusion brace. I cut these out on my CNC using .177″ thick ABS plastic. Extra slots are there for the option to add a cross plate later if needed. More likely it will just be supported by three bolts on each side.
Assembling the parts
Assembly was pretty straightforward, though I did run into a few hiccups.
The smooth rods are from MisumiUSA and the bearings are from Amazon. The rod and bearings work very well together: very little friction and they glide nicely. The rods fit so tightly in the bottom and top sections already, that it will likely not need added nuts to hold it together.
It was quite a task assembling the joints into the carriages and effector. I used a soldering iron to heat the M3 nuts and press them in flush to their traps, but this caused its own problems later. The basic idea is that the joints rotate on the ends of 12mm long M3 bolts, that are tightened into the inset nuts so that the bolts do not rotate with the joints. Getting the bolts into the nuts and joints from the outside accessible points is easy, but when you have to get the ones that you cannot reach with a screw driver, it is a different story. You can only turn an allen wrench about ¾ of a turn before you have to remove it and start over again. The problem with the inset nuts came up during this part. When I had melted the nuts into the traps, it must have displaced some plastic into the threads of the nuts, so that when I tried to screw in the bolts, the melted plastic became a thread locker and often stripped the delicate allen key slots on the bolts heads. After several hours of trying to force these in, I had cracked the carriages or effector in some places, but I attempted to heal these cracks with some heat from a soldering iron. On the surface at least, these cracks have been welded closed. Only time will tell if they are completely garbage.
Finally, all 12 joints and their nuts and bolts were connected, lubed, and rotating freely for the most part. I used PTFE grease on all the joints to try to get them to rotate more freely and I imagine the plastic will wear in a little with working to improve movement.
I then attached the carriages to the linear bearings on each tower. Each carriage is attached by two zip ties per bearing. This does not really seem like the best solution, but again, the carriages were already printed, so for now I will use them. Some other designs have carriages that wrap around the bearings and then clamp down with nuts and bolts. If these carriages become problematic, I may switch to these.
Next up I attached each of the three stepper motors for motion control. 4 M3x12mm bolts and a cork gasket for each motor.
I started to attach the printed pulleys to the motors, but was only able to attach one before I ran out of time. Threading the hole with the spectra line required a sewing needle. The hole was so small that a needle was the easiest way to string it through. Each pulley also uses three bolts to fasten to motor shaft.
What’s left to do
There are several more parts I need to print, and some more things I need to assemble before this machine is done. I did not receive all my electrical components before the holiday weekend, so I have to wait to assemble all those associated parts. I am still waiting on 4 stepper motors, a hobbed bolt (filament driver), a RUMBA control board, a hot end and a heated bed. (UPDATE: These are all accounted for as of 5/29/14).
For my design, I decided to add a section of aluminum extrusion that I had laying around as a sort of backbone support for the printer. The extrusion will span the top and bottom plates, and have adjustable feet as to not make the machine un level. I will mount a spool holder (CNC’d), the extruder (CNC’d mount plate, printed extruder) and the controller board (CNC’d adaptor) to this extrusion. This should allow me easy access to my controller, convenient placement of my spool, and easy access to my extruder. Additionally, it should add some rigidity to the structure, as rostocks are supposed to suffer from lack of rigidity. I originally designed printed parts to closely hold the extrusion to the top and bottom frame pieces. Once I printed them and test fit them, I found that the effector had lost about an inch of travel in that direction. With this design already having a smaller build area, I felt there was no reason to lose any more. The extrusion mount has been redesigned, but I ran out of time to print them out. I may change them to be cut out on my CNC and then assembled.
Wiring and calibrating may end up being their own part by themselves. I want the wiring to look clean for this printer, and will take my time in order to accomplish that. Like I mentioned earlier, I have no experience with delta style printers, but I think a lot of my other reprap knowledge will transfer over well.