Make Your Own Filament

According to [Alex] it is easy to make your own rolls of 3D printing filament, even though existing off-the-shelf solutions don’t work very well. His explanation for this is economics. He built a filament extruder using a high torque induction motor and gearbox that was locally sourced. He argues that shipping heavy gear around would make a similar extruder commercially unattractive. He sunk about $600 into the device but estimates that a company would need to charge at least $1,500 or more for the same thing. That may seem steep but as [Alex] points out, a 1 kg roll of filament really only has about 750 grams for filament and plastic pellets cost $2 to $3 per kilogram.

There are other costs, of course, like the electricity required to heat and move the plastic. Still, the system appears to use about $1 of electricity for every 10 kg of filament. You can see the process in the video below.

If you think about it, the mechanism isn’t too different from a 3D printer. You heat plastic, force it through a nozzle, and it cools off. The big differences are you are not moving around and have to manage the pellets using a screw feed. It turns out the screw and associated components make up a large part of the machine’s cost.

The other key component is a 1 HP motor. A typical motor will run at 1800 RPM, so you also need a gearbox to slow things down. You’ll also need drive electronics, heaters, and temperature control. If you pay retail for everything, you are going to have trouble matching the $600 price tag. However, the motors and quite a bit of it can be found used or salvaged. A lot of the details are in the second post. The details of the part of the machine that winds the new filament are in yet another post.

If you don’t want to spend quite as much, you can also make a smaller version that can produce about 2 kg per hour as opposed to the 5 kg per hour that the big machine makes. The little sister uses an eBike motor and the whole thing should come in for well under $500. There are several other posts linked from the original ones, including notes on the water bath required, measuring filament thickness, and even selling filament for profit.

We’ve seen a lot of takes on making filament. One even claims to cost about $100.

52 thoughts on “Make Your Own Filament

  1. Note if you do a Youtube search for “filament extruder” there are ( actual ) hundreds of projects on there, most being succesful.

    Really wish there was a core Open-Source project aiming at taking the best of all that’s around and packing it into an Open-Source design, would probably get a lot of traction.

    Years ago Reprap would have taken on the challenge, but unfortunately that community has gotten diluted by all the China-$100-machine users, pretty sad.

    1. Also if you just want to buy one of these, check out RobotDigg’s youtube channel/shop, they have a pretty good one for sale, lots of different configurations too :

    2. Add to that the cheap and easy availability of tons of filament in better-than-back-then tolerances, packed dry, etc that makes DIY extruding not worth the cost of admission except for really big players.

      I probably go through 5-10 kg per year, and if the stuff comes out at half the price, that’s saving $10/kg, so… The machine does just barely pay itself off, but then there’s all the space it takes up in my cramped basement workshop.

      I’ve talked with a bunch of people who’ve made their own, and they all agree that the real trick is in keeping the tensions and cooling rates consistent on the post-extruded filament is the key. I’m surprised that we don’t see any closed-loop systems with (a couple?) thickness gauges and some slow PID. How do the big factories do it?

      But yeah. Reprap has been successful — now everyone can have a decent printer for not that much money! Hooray, we won. Just not in the way we thought we were going to…

          1. What’s not likely ?? The designs for the printers are not Open-Source, who cares if on top of that they also use Open-Source firmwares… that’s not the point.

          1. For a hardware design to be Open-Source, the source files for it must be available under an Open-Source license. Is this the case of the Ender3 ?

          2. > You’re claiming anything that’s cheap is closed source in your post.

            I’m not. It’s the overwhelming majority though. Ender is a rare exception. And I wasn’t talking about Ender, I was talking in general.

          1. That’s interresting, thanks. There’s a good reason why I presumed this wasn’t the case / asked if it was the case or not, though: it’s extremely rare for Chinese machines. It’s the exception confirming the rule really.

        1. This “open source” does not translate well to hardware.
          Sure the “Ender 3” is “open source hardware”.
          It’s just badly documented, just like lots of software.

          Get some high resolutions photo’s form the ‘net. Then buy yourself a handfull of extruded Aluminimum tubes, some wheels, rods, etc, and screw all the parts together.

          You can clone lots of rero’s from github / gitlab / etc for free, but real hardware costs real money.
          Hand assembly is also more elaborate than running make on a cloned repo.

          And this is sort of what those Chinese kits are. They are a bag of parts, for about the same (or less?) than the cost of the parts themself. They just need some skill and time to put together.

    3. “Years ago Reprap would have taken on the challenge, but unfortunately that community has gotten diluted by all the China-$100-machine users”

      What?!? How does that work?

      How does “dilution” prevent a community from developing something? If there are just as many people interested in developing things as there were before then there is nothing stopping them from collaborating just like they used to and doing more development. The fact there are a lot of people factory producing stuff based on their previous work and many more consuming it shouldn’t be any hindrance at all.

      Remember all that “wealth without money” stuff from the heyday of RepRap? We were all supposed to be printing all our every day objects. I don’t know if we can ever reach all the way to that sort of 3d printed utopia (and have a lot of doubts) but we are still moving in that general direction.

      I think that if anything the people who would be part of the RepRap project today, the ones who started it and the younger hackers and makers who would have had they been born later are just working on the next aspect of the problem. They are working on the designs. They aren’t trying to re-invent the 3d printer. They are designing the stuff that we will print with our printers. And that is as it should be because today that is what we need most.

      Remember when all we ever printed were shot glasses, replacement printer parts and maybe a whistle? Then there was a deluge of decorative figurines. It seems to me that the development of practical, functional objects has been slow but it is accelerating. It is accelerating slowly but it is happening. Likewise there seems to be more development in 3d-printing oriented engineering techniques. For example look at what we have seen here on HaD recently involving 3d printed springs.

      It seems to me that RepRap is alive and well, not suffering at all from “dilution”. It’s just that today’s “RepRappers” are doing what is needed today, thing development more than printer development and they might not even realize that they even are “RepRappers”. And yet they are advancing the RepRap project’s goals.

      No doubt, as we eventually see big, game changing advances (such as practical home metal printing I hope!) we will see temporary uptakes in more traditional RepRap development to refine them. But in the times in between the action is in the things, not the printers.

      1. The Reprap community has moved from actually developping things, to trying to provide help to the onslaught of people with $100 printers who can’t make them work, and who flood the community. It’s time consuming, it’s everywhere, and it’s directly corelated to the death of actual research within the project.
        You are making claims about the Reprap community that make it very much sound like you aren’t an active part in it, you have very odd optics on the project, very abstract, not really matching the reality of what’s going on today…

        1. Nope, you’re getting to the right result for the wrong reasons. Helping newbs is not what killed it.

          “Good enough” is

          I have a scratch built behemoth “rep-strap” failure in my garage. In my craft room is a MPSM, that’s good enough I don’t ever really “need” the broken one, hence it remains abandoned.

          Same reason 90% of rep-rap people won’t work on filament, the $15/kg with good tolerance is “good enough” that unless you plan to run a business selling Filament or run a print farm, you don’t care enough to work on it.

          Additionally, most of the best researchers, engineers, and makers in 3d printing have been hired by for profit companies.

          I’m sure dilution effected those who remained, but it’s much less than the fact that so few remained.

          Most of the people were there because they wanted a 3D printer that makes great prints, and now we have that.

          1. This is wrong. Reprap is still making progress. It’s also branching out to other manufacturing techniques. That’s just *much* slowed down, in big part thanks to the chinese clones, in several ways ( support, lack of revenue for devs, closed designs not contributing to the ecosystem, etc… ).
            If you think Reprap is the way it is now because of “good enough”, you haven’t been very involved recently.

        2. $100 printers are so incredibly good now, all that’s left is the fancy stuff like optical end stops and position feedback, and they’re probably going to have that soon enough.

          The slicers, however, still have a ton of room to grow. Nonplanar layers will be revolutionary.

          1. « $100 printers are so incredibly good now, » that can only come out of the mouth of somebody who doesn’t have to help newcomers with these machines. The 3D printing community is flooded with users having all sorts of issues with these machines, most of which are quality-related, ie the machines are utter shit. Some are better than others, but no, they are definitely not *incredibly good*, most are still lowest-possible-quality crap that lives only a few months at best, and only after quite a bit of hacking to get it to work in the first place.

          2. Also, no, they are most definitely not going to get optical endstops or position feedback, for the extremely good reason that neither of these is needed. You can make a perfectly fine printer without these, so you’re not going to see them become common, and even less so on the cheapest machines…

        1. Yes, it’s also making progress on individual pieces of the machines ( electronics, hotends, etc ). It’s also branching out into other fabrication methods ( was the first SLA printer before they exploded, was Open-Source, lemon-curry ), just last week I’ve seen work on mixing 3D printing and CNC milling on a single part. Modern Reprap electronics ( Smoothieware ) have the unique/new ability to mix these techniques in a single firmware while maintaining standards compliance. There’s a lot going on, just because one is not informed doesn’t mean the information isn’t there.

  2. “1 kg roll of filament really only has about 750 grams for filament”
    This is exactly how rumors are created, one person writes something questionable on a simple website, then another website with some more leverage copies the content. Although I assume that the author here also finds the info questionable and therefore distances himself from it by mentioning “but as [Alex] points out”). But the harm has already been done… the seed of rumor has been planted. Perhap now within a few moments this goes viral and suddenly the general public start to believe that ALL spools of filament are sold as scams… And all because gross and nett weight seem to be confusing. Therefore, I like to see some reference here, because every spool of 1kg I bought was always perfectly 1kg of filament according to the seller(s) who always mentioned the nett weight explicitely as nett weight and not as total weight. Perhaps I should have checked it? Perhaps I should also check my bags of sugar, bottles of milk, slices of bread?

    But printing your own filament is more then metling pellets, you need to cool it in a controlled way, wind it in a controlled way otherwise you have many spools of low cost but fiddely filament eventualy resulting in dubious quality 3D prints. But other then that this is a very interesting project, looking very nice too.
    I always thought it would be nice to build one. But then I realized, do I really need one, if I only use 2 or 3 spools per year?

    1. I did do a quick check and the filament I buy is 1kg net weight. But I don’t know for sure that no one sells 1kg gross weight spools, so I am not ready to say Alex’s statement was wrong. I did notice that gross weight seemed all over the place with some claiming 1.1 kg and some as high as 1.4 kg. I wonder if there is really that much difference in the weight of the spools?

      Just as an example, here is a “1 kg spool” (in the title) but in the specs it does say 0.9kg weight along with the note (minor deviation): Now, granted that isn’t 750 grams, but… Certainly if some 1kg net spools are 1.4kg gross and you used that spool to get a 1kg gross spool, you’d have 600g of plastic, so he may well have spools with that kind of net weight.

      1. Every spool of filament I have purchased has had the stated amount of filament on it. To advertise it as 1kg including the weight of the spool would be false or deceptive advertising.

  3. It seems inefficient to have to go through the filament stage at all. We need a printer with a hopper on the top: chuck granules/discarded prints in the top, get new print out at the bottom!

    1. For a pretty neat pellet printer, have a look at this one.
      “The Part Daddy 3D Printer from SeeMeCNC”
      It is a little larger than my Deltaprintr!

      I was talking to a fellow who produces his own filament as he teaches 3D printing.
      He grinds the old prints up and feeds them into his extruder. It is mounted vertically so the filament comes out straight down. It cools then loops past a line scan camera that controls the reel winder back up the top. That sounds like a better way than running horizontally.

      1. Yah. I think this might be the same thing you are thinking. If I were to build a pellet extruder for making filament I would consider building it vertically so that I could eventually integrate it into a large format printer. It would be kind of cool if you could use that same extruder both for printing large objects and by parking it, swapping out the nozzle and setting a winder on the bed make filament for a smaller printer.

      2. You nailed it. Maybe the a grinder option for recycling makes rolling your own filament more palatable vs. buying more precise off the shelf filament. I have tons of prints that could be ground up, early iterations etc. Someone needs to invent a part shredder like they use to junk whole cars but smaller scale.

  4. This guy is quite the hypocrite. He boasts about how cheap filament will be per kg when you make it yourself, and he sells his filament on his website for $35~$50/kg.

    1. No, he is not. He is selling his time. Not everyone will make their own filament spools, by not having the skills, tools, or willingness to do so.

      He have his own costs besides the filament. He have to maintain his site, monitor payment, mail shipments, buy more materials. Things almost all of us don’t have to, and he is not required to do all that for free. And he is not required to charge for his work the cost plus a penny or two. He can charge as much as his customers are willing to pay.

      I don’t care if a guy sells me for $35 a filament that took him $0.50 to make but it’s comparable to a filament I buy for at least $50. I don’t care if a guy sells me for $150 an Arduino that replaces the mainboard of my washer machine that sells for $250.

      Let them earn their money! I am saving money, and it’s all that concerns me.

      Otherwise computer programmers will have to charge close to zero, because all they do is convert caffeine to code, and spend just a little on power.

      1. Yeah, it’s frightening seeing reactions like that, some people have zero idea how business works. They somehow think if somebody is making a profit, they have to be abusing a customer. How about earning a living ???

    2. How does that make him a hypocrite? If he said it was somehow morally wrong to sell filament at greater than cost then yes, he would be a hypocrite. I don’t see where he has said that.

      You can say he is selling his time but even that isn’t 100% accurate. Who sits there and things “What is the bare minimum amount of money for which I would be willing to spend the next hour of my life producing this item for the benefit of some stranger”? Prices are set by the market. It’s supply and demand. Currently that is the price range for which the market is willing to buy a roll of filament. Why would he sell it for less? Is he a charity?

      Maybe if you want filament for a good cause, education, PPEs for nurses, something for the poor, etc.. and you ask nicely he will give you a better deal. Or maybe not. Maybe he only has the time and resources to produce for sale just enough to monetarily support his own hobby by selling it at market prices.

      But he did share the information of how you can get your filament cheaper, cutting him out of the deal if you want to. That’s pretty cool. Don’t be a jerk.

    3. I don’t think they’re a hypocrite as it depends on volume. But they sure have been obnoxious about promoting this to /r/3dprinting and related subs over the last few months.

    4. I tried selling for less but prospective buyers thought filament is of lower quality (probably that’s why it’s cheap?) so many were skeptical of buying it. Pricing it along the normal range made more convincing.

    5. i worked for a small package delivery service once upon a time.
      a thing kept happening:
      drivers would leave and start their own business, fail then wind up back as drivers.


      because they figured if they were getting paid a certain amount to do each parcel delivery, why not cut out the company.. then proceed to charge their per parcel pay rates and have no overhead for anything else resulting in making less money, while having less customer volume to enable them to cluster locations and have sane routes which also drove up cost, while also having to try to manage EVERYTHING not just delivery/pickup which also added inefficiency to a solo op and drove up functional costs. the list goes on and on.

      i feel that while the specifics are different, the concepts apply directly.

      it may seem reasonable in a web delivery amazon style world to think, heyo i can just crunk out this machine and make product cheaper even accounting for power and such.

      but to actually operate you are either doing a lot of prework investment producing ahead of time and bulking stock or being on demand, and little things like small volume shipping itself eats a margin. further if you are preworking to build stock for bulk how are you protecting the plastic? now you are adding a vacuumpackager perhaps? maybe boxes.
      what about those spools its going on?

      how about machine part wearout? where are you storing your bulk production? are you going through amazon and taking losses through them or doing your own storefront and eating costs there as well if you want to handle cc payments?

      bottom line is until you actually do it, theres probably a whole book worth of things to consider you might not see yet.

      that said odds are you might find if you are local or contact that individual you might have more flexibility as well, which you won’t generally with the amazon likes.

      i suppose a tldr would be something something economy of scale is more than just producing more is cheaper.

  5. I’ve had a Filastruder for a couple years and I love it. Sometimes it’s hard to source polycarbonate, but when I can get pellets, the Filastruder makes excellent filament with super high tolerances. I print in a lot of composites (carbon chop polycarbonate and carbon chop nylon, primarily – for stiffness and disappearing layer lines), so I don’t always use the Filastruder. I have made ABS and PETG with it, and they work as well as commercial filaments in my E3D V6 Gold.

    Just a suggestion if anyone wants a commercial option with warranty and service/support.

      1. Yes, it’s not the thing I built. If you read my guide, I used this image because for explaining the extruder parts and how it all fits together. My extruder is a bit messy and not this clean, so I used the cleanest example I could find.

    1. Yeh, i had…
      ‘looks like a generic Chinese piece of kit with those Louvre coverings/ build/feel about it’.
      Sorry, sounds lame, but that’s what popped into my head.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.