Plastic bottle PET filament extrusion for 3d printer: experiments

This will not be a guide to build your own filament extruder, there’s already a very good guide by Ian McMill on instructables, which incidentally I followed to build my own version.

This is a report of my own experience with trying to obtain 3d-printer-worth filament out of plastic PET bottles.

I was enthusiastic of this video of the russian survivalist The Lawyer Egorov (who by the way has a supercool channel on youtube for you hardcore people):

who built a contraption to get a very strong string (pun not intended) out of scrap plastic bottles. This, I thought, would have been great fodder for my next to be PET filament machine.

Fast forward to yesterday night/this morning. I had been sourcing parts off the local shops and eletronics from foreign sellers on ebay, and in the spare time finally built this revision of my machine:

DIY filament extruder

DIY filament extruder, internal electronics
This is the electronics panel, where you can see a SSR, three switches, a PID, and a motor speed controller. The 12V power supply lies on the back.


Extruder motor with fan
This is the motor, taken from the electric window of a honda civic, off ebay, supercheap (I looked all around for it). Most of the metal parts were taken from a local shop. Fan is absolutely needed as the motor gets really hot after a while. You can also see the coupling, with a force breaker plate, and a couple of exagonal keys. I had to shape the coupler square rod myself out of a bigger square rod. Ugh never again.


Extruder tubular assembly
This is the point where the auger bit gets into the tubular assembly, I used several L shaped plates to keep the assembly strictly in place. You can see the push bearing in place, and a closeup of the T junction I used instead of cutting out a window in the metal pipe.


diy extuder nozzle thread
I had experimented with PET already the late evening before this shoot, at 260°C the plastic inside the nozzle became sort of essiccated and very brittle but did not really melt.


extruder SSR
The SSR controlling the heating band circuit


extruder PID
The 220V and 12V lines molex’s, and the PID, a supercheap REX C100 off ebay, chinese replica, SSR version, once and for all this is the pinout: 1-2 AC input to power up the PID, 4-5 the DC output to close the circuit with the SSR, 9-10 the pinout to the K thermocouple (mind the polarity, inverting it will give uncorrect readings)


DIY extruder motor controller
The motor controller, very easy to configure.


extruded PET
This is what was left in the nozzle from the night before, a solidified melted PET which had become grey in color, and was somewhat brittle.


solidified PET
The broken pieces of the stuff in the previous picture


glassy PET
There were also flakes of “glassy” unmelted PET which had become extremely brittle under the high temperature even if it didn’t melt (according to several sources, this is the crystallized form of PET, obtained by removing the water by heating to 80+°C for a certain amount of time, and this is exactly what happened the night before during my tests… this kind of pretreated PET should be the best kind for filament extrusion purposes).


melted PET inside extuder nozzle
Finally some melting was taking place after I raised the temperature to 270°C, the night before I only went up to 260°C (I started off thiking even maybe 190 would have done the trick, silly me!)


brown molten PET
I stuck a piece of PET string inside the pipe to scoop out some of the goo, and this is the result, a brown stuff solidified over the fresh PET.


After I noticed the PET inside the terminal part of the tube was melting, I started the motor, and this is what happened… not much as the pressure inside the tube was low, there wasn’t much material in the column, and since cutting BY HAND the PET string I had was superboring I didn’t really make the small flakes needed for them to be easily transported.

I increased the quantity of PET flakes in the intake, and pressed them down with a screwdriver handle, so the volume of melted mass on the front increased some.

At this point I decided to place the cap with the nozzle drilled in (1.5mm diameter) and in fact some goo started flowing out and quite rapidly solidifying; the quality of this solidified goo appeared “better”, as the surface was very shiny, but it soon stopped flowing out, and when later I unscrewed the cap I realized it was clogged, as a lot of melted material was idling back there and quickly blobbed outside.

extruded PET droplets
The “extruded filament” I got, a couple of little blobs of PET which quickly solidified and became grey from the brownish pulp it was before.

So this is the scenario that presented itself when I uncapped the tube:

molten PET flowout
The molten PET flowed right out of the tube as soon as I uncapped it, the nozzle had probably clogged giving this outcome. The material is soft-ish anyway, and doesn’t pose a big deal in being removed noteven from the threaded insides of the cap.

At this point I decided my adventure with PET had come to an end. Or rather, my experiments did, since I decided since the night before that I wouldn’t be making PET filament for my 3D printer out of plastic bottles: the processing required to this objective it WAY too much. You need to find a way to satisfactorily grind the plastic, after obviously you cleaned the bottles, and after that you should still dry the flakes at 80°+ degrees to avoid water buildup inside which would drasticaly decrease the final product quality. I would totally prefer spending 20€ per roll of cheap black PLA over this.

You also need to take into account the real life scenario, that is, PRINTING with said filament, because that’s the reason you’re doing it, apart from the fact that you’re just a hopeless nerd like me. A printer’s nozzle is SMALL, and nasty things, even if tiny, can clog it; if said nasty things are inside the filament, then no filament filter is going to hep you; they WILL pile up in the nozzle and sooner than later you’ll have to disassemble the nozzle assembly and do nasty things to take it out, like flaming the nozzle to melt the stuff inside and scooping it out with a piece of filament. Don’t trust the cold pull, it didn’t help me when I needed it. Flaming the nozzle instead, it did help me.
Finally, you have to get a decent diameter filament, AND getting that diameter the most constant possible, because if on one side you can change the extrusion rate in your slicing program to take into account the real filament size, on the other there’s nothing you can do to balance a poorly sized filament, and either you will get extruder skipping steps because of a too big filament, or an extruder skipping steps because the filament is too tiny.

In my brightest dreams I supposed I would be producing crystal clear filament out of plastic bottles, heck maybe even multicolored transparent filament since all bottles are slightly coloured.

failed plastic bottle PET filament extrusion experiment
The aftermath of the experiment. You can see the spare PET string I had previously made and originally intended to feed inside the tube.


Final tecnical considerations
By Ian’s own words, with which I agree, the downfall of this method is the feed system: PET can be a great material, yet a DIY extruder like this one does not allow for much leeway, you need to keep the material flowing forward, and for this you need a granulate which can progress easily in the pipe, and at the same time you also need a constant supply in the back.
So, first off, you need to have a ready made supply of flakes, and obviously cannot just improvise by cutting little piece by little piece on the fly like I did (I was desperate to try, I admit it), and even better you need proper sized pellets, not too large but sufficiently thick! Since PET flakes are flat, they easily adjust in the space between the auger bit and the pipe, and they do not progress further very well; also, if you use a string of PET, it will just roll around the auger in one place (where you feed it through) and not progress forward, but instead slowing down and loading heavily the motor by creating a thick wrap on the drill.
You MIGHT melt bottles whole as they are (how?), and use the obtained solidified blob to create smaller pellets, maybe by hammering the glass-like goo. Still, it will take time, and it’s a messy procedure. Unless you can devise a method to do it without much human intervention, it’s going to cost resources anyway, and you may reconsider the whole thing, and maybe just buy pre-made filament, or a commercial extruder plus finding a local supplier of ABS pellets (not your average ebay seller who asks outrageous sums for 3d printing pellets, but a local plastics industry which works with pellets to make for example injection molded items, where you would pay as low as 2€ per kg of granulate).


And now some resources:

  • HDPE filament from milk jugs extrusion test (middle part of a trilogy):



6 thoughts on “Plastic bottle PET filament extrusion for 3d printer: experiments”

  1. Hi
    -projects to be little changed
    -All pipe to be heated
    -To fragmentation material
    -Material to be melted by heat
    -It is poured from above the molten
    thank you

    1. Your message wasn’t quite clear! Anyway rest assured that this kind of design is the standard regarding the extruder build that is also commercially found elsewhere

  2. Hello, I would like to ask about the AWG of the wiring that you used and which part was is used specifically. Thank you!

    1. Honestly, I don’t know/remember and I sold the machine in the meantime, but the heater module is 220V AC so you don’t need a thick wire in the first place, and the motor doesn’t really drain so much that you need to pay attention to that, go with your insticts I’d say, 1mm^2 is enough for the motor I’d say.

Leave a Reply

Your email address will not be published.