3D-printing with a granulat extruder
I’ve been following Precious Plastic for about a year now and I have to say, it’s a wonderful project! One, I’d like to contribute to at some point. Currently, I’m in my last year of industrial design studies and I’ll start soon my Bachelors Thesis on designing a 3D-printable monocoque frame for an electric scooter.
I’m planning to actually print & test the frame next spring/summer. As my primary tool I will use my Hangprinter, that I’ve been developing the past 1.5 years. The Hangprinter is a frameless printer-concept by Torbjorn Ludvigsen. The printer consists of a base unit with all the electronics and motion mechanics, and three anchor points on ground level for dyneema lines, which are used to position the print head. It is a quite unique concept, but foremost it is easily scalable to (theoretically) any size needed by only extending the lines.
My printer has currently a 2.5 m tall tetrahedron frame, a heated print surface and the e3d-online SuperVolcano filament extruder attached to it. I did this summer two big print projects on it. One was a giant high heel for a finnish shoe designer and for the other project we printed the shell parts of a prototype robot. What I learned, was that the current FDM-technology is far too slow for a printer this big and of course that in long term it is not viable to print anything else than waste plastic.
This is also the reason, I’ve been interested in Precious Plastic. You guys have come up with nice machinery for shredding and extruding plastics. If you want to print with waste plastic, the first step is to shred it. My initial thought was to make my own filament, but after second thoughts I realized, that it is just a unnecessary step. So, I began to study commercial pellet extruders and their open source counterparts. Finally, I designed my own extruder and I’m now in short time getting it ready for testing.
Tech specs are as followed:
Effective screw length: 225mm
Heating: 3 x 180W /230V
First, I’m going to test commercially available recycled HDPE and PP pellets in order to rule out issues caused by raw material impurities and quality, but eventually my plan is to test shredded flakes.
I know that both HDPE and PP are going to be really tough to print due to their heavy warping properties. That is something, that can possibly be addressed only by a controlled heat chamber. But before that is going to be on my todo list, my idea is to mix chopped carbon fibers into the polymer. Fiber-enhanced polymers usually print nicely, since the warping is mostly eliminated. I’m hoping this will be the case also with HDPE and PP.
The printer frame is going to need an update to be able to handle the stress caused by the pellet extruder. I already have a new set of BLDC-motors with rotary encoders, that will do the heavy lifting. Each of them is rated with 1.8kW and there will be 4 of them in total. Also new Dyneema line is on the way. A single line will be able to lift around 140kg. The whole system runs on a Duet 2 Wifi controller and ODrive BDLC-driverboards.
So, what do think? Do you have any questions, feedback, etc.? Do you think this setup could be beneficial for Precious Plastic?
After switching to 4:1 reduction the extrusion speed went through the roof 😀
I was able to extrude around 80 g in 6,5 minutes, which equals in almost 750 g/h. And I’m pretty confident, that’s even not the maximum yet. Had to stop, since the hopper section got too hot and the pellets got stuck in it.
That will be no issue after I get the actively cooled heat sinks delivered. I’m still aiming for 2 kg/h and I think that could be feasible!
Ok, time again for an update!
Lately, I’ve been working on the extruder again. Mostly doing cable management and designing a cooling system for the feed section. I had an issue of clogging the feeding tube with prematurely molten pellets. As a solution I added three heat sinks with active cooling onto the barrel. Still to be tested, but I think it will work.
I wanted also to have the thermocouple sensor boards as close as possible to the heater bands, so there would be only minimal interference in the signal lines. Now, the data and the power wires gets terminated in a single junction box. Ideally, there would be only one single cable sleeve clipped to the pellet feeding hose later, but for now I’ll go with this solution.
Next, I will focus on an anti-ooze mechanism for the nozzle to prevent leaking during moves across the print. With a screw extruder you won’t be able to retract filament, and so you will have to come up with a valve-like solution. More on that in a later post.
Please let me know if you would be interested in beta testing. The beta extruder will be available later this year for a reasonable price.
Short update again:
So, I managed to get the barrel cooling working. There’s still room for improvement, but at least now the temperature stays quite stable. Later, I might have to switch to beefier fans and a bigger heat sink.
Currently, I have a 5 mm nozzle installed. Including the die swell, the extrusion width is 8 mm 😀
Last couple of days we’ve been doing testing with HDPE. With virgin pellets for now, because we’re still in lack of a shredder. Hope, that will change soon. But, in any case, the throughput with moderate speed is currently 1.55 kg/h. We already have a more powerful stepper, which will replace the standard Nema 23. That will double the torque and then we will pass our target of 2 kg/h by… I don’t know yet, but it will be much 🙂
Super happy with this and it gets better by every day!
Well, actually, that’s a good question! I don’t really have a method of monitoring the motor power output. I can set the stepper driver amperage and the control software gives me information about set and actual speed, but I think for real measurements I would have to hook a torque meter on the motor shaft?
The die swell seems to be quite essential. For now the nozzle diameter is 1.75 mm, but I was thinking about a nozzle size of 5-6 mm in order to get the carbon fibers nice and easily through the orifice.
One thing, I have to figure out somehow is the e-steps/mm value, which is usually determined by the difference 100 mm filament extrusion and the actual measurement. But… I’m not extruding filament :/ any ideas?
Thank you for your feedback!
Yes, indeed the use of HDPE/PP will be a challenge in every way, but as I mentioned, I’m going to mix long (6-25 mm), chopped, carbon fibers into the polymer in order to gain maximal strength and rigidity. I’ve been printing quite a lot with CF-enhanced filaments like CF-Nylon and CF-PET/G and you can achieve pretty amazing stuff with them, although the fibers in those filaments are quite short. Usually they have to go through a nozzle of 0.4/0.6mm.
The nozzle size I’m aiming at with my pellet extruder would be 5-6mm. This enables me to use much longer fibers, maybe even a continuous fiber tow. The polymer acts as a binder material to keep the fibers together. For sure, this process won’t give you parts as strong and stiff as the cross laminated carbon fiber mats, but I’ve read some research that stated that the strength could be as high as 45% of a traditional carbon fiber part. The material testing of the printed CF-HDPE/PP will be also part of my thesis.
For now, let’s say it is a hunch, that there could be something in this approach. In 5-6 months I can deliver facts.
As for the size of the stepper motor and the reducer, I used the Dyze Design Pulsar extruder as benchmark. The stepper I’m using is Nema 23 with a 47:1 reducer. We’ll see, if that is going to generate enough torque. According to the torque curve, it should be able to output 40Nm at low speed. The screw profile is a standard 3-stage compression profile, channel depth starting at the 4 mm feed section and going up to 1.5 mm at the metering section. The L/D ratio is around 14:1.
The mass of the extruder and its inertia will be an issue, but I regard it to be solvable by a stronger/stiffer frame. The BLDC I’m going to use have enough power and speed. I’m not sure though, if I manage to come up with an improved frame on schedule. My backup plan is to bolt the extruder to the gantry of our school’s CNC router.
I’ll keep you updated 🙂
Thank you for posting this. I had not seen the hangprinter materials. Sounds like you have and exciting project for your thesis.
A few questions and comments:
How did you size the stepper motor/gearbox for your extruder? What does the leadscrew profile look like?
Doing a scooter frame design with HDPE or PP sounds challenging given the low structural properties. I wonder if you will end up with mostly solid cross sections.
The hangprinter design appears to be sensitive to printhead inertia. Putting your extruded on the print head will increase both the mass and the moment of inertia. The printhead that @3dseed https://davehakkens.nl/community/forums/topic/3d-printer-for-pet-particles-no-filament-ever/ is using in his design appears to be a lot smaller. Perhaps for larger prints, develop a larger or flexible hopper that can feed a smaller printhead.
Please post updates, this is an interesting approach.
This is a cool project. Following.. 🙂
Looks like a great start. Do you have a way of monitoring how hard the motor is working?
What nozzle size do you plan on using? I noticed that when I create a bead of hdpe with my injection machine, it tends to expand quite a bit.
Actually, your back-up plan of bolting it to a CNC machine sounds like a great trial run for optimizing the print parameters.
Good luck, great work
I’m really happy to tell, that I was able to fix few remaining issues with the extruder and got today to extrude first little bits of LDPE. In January I will tidy up the cable mess around the extruder and fit couple of cooling fans to it. And of course push it to the limits XD
The overall throughput is something I need to figure out first thing. My goal is 2 kg/h. The speed, the first extrusion came out, is making me hopeful, that the goal isn’t too far fetched.
@alinna, great stuff , thanks for sharing ! beside the @3dseed commercial grade variant, there 2 more printers in progress here : https://github.com/plastic-hub/noah. @timberstar rolled out good brain work for larger frames, perhaps that may help in some ways.
judging by the pictures (vanplestik) and my own experience building smaller extruders, 0.75Kw 3p at 1:30 delivered til now good results which will be utilized in the ‘noah’ project for the larger printer.
there is also a discord channel for the PP related task force : https://discord.gg/AtM6zsS – bump there if you need any help 🙂
This looks so promising! Good work so far!
I was disappointed I could not see the videos…. Not sure why as I don’t speak the language that the errors were in… :\
Nevertheless, I am looking forward to seeing more!
You must be logged in to reply to this topic.