when you print normally (0.4mm nozzle, 0.2 layer height) for every 100mm movement you use only 3mm of filament, there is no way you can stretch fibers 30x. So only way this can sort of work if you print 1.75 filament with 2mm wall width and 1.2mm layer height, so the cross section is the same. Looking forward for part 2.
I use the markforged Mark 2 printer at work and they have a cool workaround around your problems. The fiber strand encased in plastic is very thin under 1 mm. They use 2 nozzles with 2 filaments feeded into it: one for bulk printing and the second for fiber reinforcement. The fiber nozzle has a cutter that severs the fiber when it has to move to a new area. Take their approach. Make thinner filament. Use 2 nozzles. Print only on certain levels with fiber.
That's great, but you're going to need a filament cutter right after the nozzle to prevent the nozzle from moving and pulling what's already printed(For having a continuous thread of carbon fiber). You can probably activate it with the retraction
There is a reason the Mark Two uses two separate spools for fiber and plastic. The plastic likes to expand and shrink throughout the printing process, but the fiber is completely unaffected at those temperatures. Instead, the print head of the Mark Two injects the filament into the extruder at the melt head. The only way to make a carbon infused filament work on a classic extruder is to use short-strand fibers. Same with fiberglass, which sounds fun to experiment with but, really, don't. The side effect of printing with your medium and fiber on two seperate reels is that you can also select what layers get fiber and which don't, optimizing strength vs cost. You also have to consider how to cut the strands; the closer to the extrusion head, the shorter your strands can be. In short, there is a reason the Mark Two is so expensive, and their stranglehold on the patent is only part of it. Good luck!
I can see a concept like this genuinely working well, but likely with a pretty custom setup. For example we can still get very accurate prints from a 1mm nozzle. So if you could use 1mm diameter filament and a 1mm nozzle with a custom hotend setup designed to push very close to the nozzle, it could probably go pretty smoothly. Of course you can't have any breaks in printing, but that could also be solved with a cutter right below the nozzle. Still lots of issues there, but hey, big props for trying and it would be awesome to see it really work.
Brave try at something most people consider impossible! Without sophisticated machinery to make your filament you are "stuck" at running the strand down the middle. Since one needs to wet the strands to get strength of the composite I would use way less carbon fiber to give the filament more chance to fuse together with the fiber. And as stated below a smaller nozzle (0.8 to 1.2 mm) but increased layer width and thick, but no more than 80% of the nozzle diameter, might force the plastic into the fiber and give a bit more pressure on the build plate to get some adhesion.
I’ve been wanting to try this for over a year now , glad I ran into this video! I’ve seen a company from France named Zero3D /Zero D at Formnext Germany and he had a filament making machine built exactly for this. He had filaments with continuous glass fiber aswell. He said he was able to print with them on a sovol sv06 and the prints were looking promising. Not pretty but strong. This would be really interesting to try in a system with a filament cutter guillotine and experiment different metals for the blade to see how it gets cut. If something like that works it might enable next level strong parts on desktop machines, even if that has like an extra purge line added manually to do the cutting part. He had less fiber laminated inside which i think might help getting better desktop results. Dosing the material is a challenge here having the fiber volume inside and most companies went for a separate feed for the continous fiber that gets laminated in the nozzle to overcome that.
Just an idea that popped to my mind when seeing the video's begining : with axial fibers, you need to ensure that the layer lines have the same cross section as the polymer in the filament, elseway your printer will stretch the fibers as it wil try to lengthen the filament by printing it at a different cross section to keep volumetric flow constant (the polymer needs it). Why have the fibers purely axial and not coiled around a main thin filament like a spring, which would then be encapsulated in aouter shell (like with your pulltruded attempts) ? as such the fibers would be able touncoil themselves into staright sections to allow for filament lenghthening. It's an idea that would need a lot of work in order to have a reliable fiber coiling machine, but could definitely work. I am thinking about how to make continuous fiber filaments for a long time and coiling around a mandrel filament might be a good trick
Start with a high quality mechanical pencil and use it to dispense the carbon fiber strands only then pass it through a fine brass tube while being mixed with a UV activated resin/polymer once the wet strands exit the brass nozzle a UV laser is directed towards the deposition contact to instantly solidify the wet strands (BTW brass is self lubricating and easy to solder and available in small diameters look for EDM drills that is why I'm suggesting it) good luck my friend.
I'm working on the same project lately for my master's thesis. We had some similar ideas. Try pulling the roving through a stock hotend (0,6-0,8nozzle) ;) I'm about to make my first video on yt to share what I have found so far.
I wonder if you can make a nozzle that has an inner, separated channel to deposit carbon fibers directly into the print I made a mockup in onshape of a more ideal nozzle (would have to be metal 3d printed) called "carbon fiber depositing nozzle", its public so it should show up if you look it up the idea is that the melted filament and already printed fibers would pull the fibers through the nozzle as the printer moves
Just took a look at your nozzle, that's a fascinating design! It's definitely close to the limit of what a metal printer could make, but if it could be made, that might be the best solution to continuous CF printing!
Try vase mode with a 1mm nozzle. At 0.8 layer hight. There is nothing on your test printer to cut the fibre filament on retraction moves. But very good effort. Will look forward to your next video attempts. Altnatively try a thinner core of carbon fibre if you want to continue using the 0.4 nozzle. Additionally I think it's a very interesting concept the current issue with carbon fibre filaments if the size of the chopped fibre is too fine to actually provide much more strength. Using your idea. If you were able to say add 2-3mm strands in as a core due the the nozzle melting, stretching and squashing the plastic you may end up with a very usable filament without the need for a cutter and retain 0.4mm nozzles. I imagine the end result would be somewhat akin to fibre glass mixture.
I've had this idea for a while but with a 0.5mm filament, that way we can retain the print quality by using a .5mm nozzle and only merge a few fibers inside the filament, making it possible for the printer to break it with travel moves 😇
First of all: CONGRATS getting your first video out! Biting the billet and uploading before what may be considered “success” was the right call. So the results…I can’t believe you got filament that looks that good out of the “rigs” you used. There are a few factors I’m aware of that were working against you: PET is not easy to print; it likes to be around 280C and the bed at 80C, and you’ll have struggles with bed adhesion without a PEI build plate. I don’t think that an ender can print straight PET well at all - I suggest investing in a roll of PET(no G) to use as a control. Now about the extrusion rig…you can get lucky, but it needs to be temperature controlled and high wattage. PETs temperature window for melting without crystallization is not large. PET also needs to be heated up very quickly, then cooled off quickly, the goal there is to make it amorphous. Considering you were PULLING the extrusion through with your hands, I think we can assume it didn’t come out of the nozzle at 260C-300C. I know it’s probably not in the budget yet, but a secondhand 3d printer seems like a great way to get some of the parts you need for cheap, like a PID controller, motors, thermocouple, power supply, and maybe even the heat block will be enough. Okay so after you have tuned and successful print settings for PET and your filament is amorphous, the name of the game is now Consistent Extrusion Diameter. Think about it; if the variance is 0.5mm, the printer’s flow rate is off by 30% EVERYWHERE. What happens if you try to print a benchy at 70% or 130% flow rate? You don’t get a boat. So to get the extrusion diameter tolerance better is simple: consistent speed. This is where that motor from that secondhand 3d printer can help. Lastly; I’m not an expert on filament manufacturing, I just have gathered a lot of knowledge surrounding it. If you come across conflicting information I wouldn’t be surprised, and if you have conflicting experiences, find out why before dismissing any of it. Again, great work and im excited to see this channel grow!
I made one of those diy pet pulling machines, but modded, i made a 2 stage melt zone, one big long heatblock lower temp +-210C with a big hole so you get a perfect U shape then trough a modded nozzle to 1.7+-, at 250-260C this 2 stage made me able to increase spooling rate and gives way better results since you don't stretch thé pet, or overheat it with 1 stage. Such a design works perfect with your idea. I would go for thinner Carbon fiber stain, this wil give to much back pressure in thé nozzle, and increasing heat with pet above 270 to decrease pressure does not work, it cristallizes, ( visibel white), and it's just overal bad print Quality, no work around that, so keep thé temperaturen at 260-265, also i tryed using a pretty modded Ender 3 for this, bowden setups Will always fail with this. If i where you i would take out some variables out of thé game, like a bowden setup. If it works out i think it would work. Great first video🎉 Also btw, PET is very hydroscopic, so you need to keep it dry. I also saw a command about the strains not able to stretch, which makes sense, so you could try extrusion testing normal pet that you pulled with the same heat and strip width so you know how luch plastic there wil be and hiw much CF, wich makes you able to calculate the wall width, it will not be 2mm by 1.2mm like he/she mensioned.
for better adhesion between plastic and fiber, you could try dipping them in a solvent that disolves the plastic, then even between the fibers will have a little bit of plastic. Con: you will need to bake out the solvent before printing. Also, i wonder if something like heating your filiment could help. like, oven, or run over a flame quickly.
I have made a LOT of PET bottle filament. Like 20+kg of it. What you're trying to do here is VERY interesting and VERY possible with a few tweaks. In fact I have thought about this but with glass fibres instead. When done correctly the PET tape pulled from the bottle will form a tube as it is pulled through the heater and nozzle and the trick is to get it just hot enough to soften and shrink without actually melting. Getting the volume correct is key. Too much PET tape pulled though and it jams and snaps. Too little and it doesn't form a proper tube. Considering how little filament is actually laid down when printing I think less carbon fibers would be better and definitely individual fibers instead of a tow or tape. But you're definitely getting close. The main problem I see with this is the slicer and the logic it uses to create the line path. Ideally it would be one single line wihtout ANY retracts or travel moves. Almost like the entire print is in vase mode but wihtout it printing only the outer wall. I suspect an IDEX or multi toolhead printer would be best for this with one toolhead providing non-CF filament for supports and areas where the CF is not needed/wanted and a second for just the PET-CF filament. Proper Printing did a great series on trying to print resin with a cartesian 3D printer and I think laying down a bead of resin that has a line of CF embedded in it would be a game changer. Between what you're doing here and what he did there I think the two of you could REALLY have something big. I just wish I could give this video more than one like.
Ich would guess that you could do it with a setup that looks a little bit like a wire-edm Maschine. One Side pulls the fibers through a molten platic bath. The other side keeps a little resistance so the fiber is always straight. I hope you know what I mean 😂 Best regards from Germany 👍
Haha, great intro. Pity about the audio quality for that bit. It looked like you were printing your first layer too quickly. I'd also use a brim, just so it has an easier time sticking down before it makes any sharp turns. Printing with a big nozzle isn't easy. Every time a new line begins, you need to cut the fibre. Otherwise, there's a practical limit to fibre length. This is doubly-true for retractions; you do not want to drag a long strand back into the nozzle! I subscribed, so you're obligated to continue ;)
I think you can get away with much fewer strands also dont make it a continuous strand , make overlapping strands so they can slide past each other . then try and make much smaller diameter filament, larger diameter nozzle , vase mode. Higher temperature. Problem is that printing this way only makes it stronger parallel to layer lines. Very cool concept though , I could still see applications for it.
instead of making something that cuts the fiber while lifting/retracting the nozzle, maybe you could try to make everything work in vase mode so it does one continous move.
I you misundretood how the printer uses the filament.the filament gets "streched" by a factor of "alot" which carbon fiber can't do .that is a fundimental problem you have to overcome before everything else in my opinion. cool video 👍
I wonder if this more like instead of printing the whole part like this, have some designed channels where this filament could create strength. But the shape is made by other plastic. Also I think you need to consider the carbon fiber bend radius, in your part. I don't think it can bend to the radius of your printed part. Start big with big curves and go down from there.
I basically had the same idea yesterday evening, I've been searching for parts and stls since. I think printing in vase mode is a must. How many fibres does this tow have? I think 3k should be good. There are also a lot of community projects on pet-bottle to filament pulltruders (if you want to avoid pulling by hand). The two stage integration of the fibre is a great idea, in my opinion. I think you can also decrease your nozzle diameter a bit (maybe 1.5 mm while keeping the same crosssectional area of the printed line) and decrease the volume flow rate. Carbon fibre conducts heat way better than PET, so you might have to crank the nozzle temp or get a volcano nozzle.
Great try… if i may suggest.. try with half the amount of carbon fiber, better chance to get good print and the carbon will still be there potentially always covered. But you have to get a consistent print without the carbon first. Just a suggestion, i can only imagine the amount of work. Carry on 👍🏼
that's awesome! only real note is that i would highly recommend wearing gloves when working with the fibers-- a lot of those small fibers can get stuck inside you, it's really annoying...
@@BuzzingGooberI'm talking about working with actual carbon fibers, not the chopped up bits in filament. I had to look up what you were talking about, I'm talking from experience working with composites (granted, mostly fiberglass, but the principles are the same) where gloves are mandatory even before you start getting resin into the mix
I thought continious carbon fiber is printed trough a 2 way in nozzle with normal no matter what filament comming from one entrance and a 0.2mm carbon fiber strand (in the case of a 0.4mm nozzle for example) comming from the other strand and beeing combined inside the nozzle ... like those mix nozzles that combine 2 filaments, so you can print any colore you want (as long as it can be mixed by 2 other colors), I think there are some videos on youtube about it
Planning on doing some tensile testing in a future video :) There are some places where the plastic gets between the fibers and others where it doesn’t, I suspect this lack of consistency was from my hotend being pushed to the high end of its limit. Just from playing around with the samples though, the CF ones feel a lot stiffer.
First off, great video! Definitely earned a sub. But -- I only watched the first 30 seconds so far, so maybe the rest of the video will address this -- doesn't continuous carbon fiber filament already exist? It's exorbitantly priced, sure, but exists already and has been used at a consumer level.
That's a great point - I may be wrong about this, but as far as i'm aware, there isn't any continuous CF filament that can be printed on a run-of-the-mill printer without modifications. Printers like the Markforged Mark 2 use a spool of carbon fiber in conjunction with a spool of conventional filament, which gets combined in/around the nozzle. That process is a lot more practical for making useful parts, but the process in this video is geared more towards hobbyists, as it just requires a nozzle swap before being able to print a part (albeit a part constrained by the limits of vase mode).
why not use a screw like a normal filament extruder, but also let it push the carbon fiber with the plastic, push not pull. why not a pet plastic tube with a separate fiber running inside. one other option is to sow a 3d shape with the pure carbon fiber, then add resin. well petg with hardened steel nozzle needs about 260C (+30C vs normal nozzle, for cf stuff). diy milk casein plastic with carbon fiber strands might be very nice.
Hey Alex! I just posted a video where I Show my setup for Continuous Fiber Filament making. Feel free to check it out! Regarding your setup: you should carefully select material for the matrix of your filament. (Not a PET from bottles or a gluegun) It should have low enough viscosity at elevated temperature to fuse well with the fibers. Study material properties and choose what best suits your needs) I used PETG and it required 260 deg for good impregnation. Full impregnetion of fibers is essential to achive a filament that will give you good resoults in printng. Good luck! ua-cam.com/users/shortsv8ktu00bYaQ?si=lsMlBJT85p8Yz2l0 Btw. I see in the background you are a fellow RC modeller. If you want to we can get in touch somehow to talk more about the project. Cheers!
So what were the print settings that worked vs the ones that didn't? What helped vs hurt it? That seems like the main discovery of this video but it was skipped over. Also consider a 0.8 mm nozzle so the fiber jams and stretches less
This isnt going to work the way you want it. I think stopping the print and and placing fibers at various levels. Or to create a channel in the design to fill with carbon fiber and epoxy for the extra rigidity.
yes process is possible the vompany you mensionsled doing that for a lot of money but the way you do it is just stupid... at least make filament inteself with some kind of carbon addition and longer nozzle to maintain better temperatures across all of the printed poop you did so maybe then you will have any results... black filaments prints better become of better conductivity and your nozzle is like yours dad pp short and it's surprising that it made something... get at least big vulcano nozzle if not super vulcano one and print slow on 1.75mm and print big thinks not small resolution will be shit... actually the best will be to print on 0.8 and have separate "carbon nozzle" that you for example tie to a neil and melt into the plastic or nozzle within the nozzle where you will cover carbon fiber by fresh hot plastic keep going on maybe one day you will sell yourself 15k machine
![Lp. Сердце Вселенной #60 РОЖДЕНИЕ ЛОЛОЛОШКИ [Финал] • Майнкрафт](http://i.ytimg.com/vi/YoR0pAV9FVQ/mqdefault.jpg)
when you print normally (0.4mm nozzle, 0.2 layer height) for every 100mm movement you use only 3mm of filament, there is no way you can stretch fibers 30x. So only way this can sort of work if you print 1.75 filament with 2mm wall width and 1.2mm layer height, so the cross section is the same. Looking forward for part 2.
Also only Vasemode would work.. cause you cant do movement paths at all..
I use the markforged Mark 2 printer at work and they have a cool workaround around your problems. The fiber strand encased in plastic is very thin under 1 mm. They use 2 nozzles with 2 filaments feeded into it: one for bulk printing and the second for fiber reinforcement. The fiber nozzle has a cutter that severs the fiber when it has to move to a new area. Take their approach. Make thinner filament. Use 2 nozzles. Print only on certain levels with fiber.
That was my thought that you'd want a tool changer setup.
Is this your first video??? Incredible work! It appears the youtube recommendations are working nicely :)
Exactly what i wanted to say!
orca slicer has "vase mode" which prints 1 long continuous line, it might be worth trying
That's great, but you're going to need a filament cutter right after the nozzle to prevent the nozzle from moving and pulling what's already printed(For having a continuous thread of carbon fiber). You can probably activate it with the retraction
good point, guess this would only work for vase mode
There is a reason the Mark Two uses two separate spools for fiber and plastic. The plastic likes to expand and shrink throughout the printing process, but the fiber is completely unaffected at those temperatures. Instead, the print head of the Mark Two injects the filament into the extruder at the melt head. The only way to make a carbon infused filament work on a classic extruder is to use short-strand fibers. Same with fiberglass, which sounds fun to experiment with but, really, don't.
The side effect of printing with your medium and fiber on two seperate reels is that you can also select what layers get fiber and which don't, optimizing strength vs cost. You also have to consider how to cut the strands; the closer to the extrusion head, the shorter your strands can be. In short, there is a reason the Mark Two is so expensive, and their stranglehold on the patent is only part of it.
Good luck!
Incredible progress sir, I'm looking forward to seeing where this project goes!
I suggest to look into importance of proper fiber wetting. Cheers!
I can see a concept like this genuinely working well, but likely with a pretty custom setup. For example we can still get very accurate prints from a 1mm nozzle. So if you could use 1mm diameter filament and a 1mm nozzle with a custom hotend setup designed to push very close to the nozzle, it could probably go pretty smoothly. Of course you can't have any breaks in printing, but that could also be solved with a cutter right below the nozzle. Still lots of issues there, but hey, big props for trying and it would be awesome to see it really work.
Brave try at something most people consider impossible!
Without sophisticated machinery to make your filament you are "stuck" at running the strand down the middle. Since one needs to wet the strands to get strength of the composite I would use way less carbon fiber to give the filament more chance to fuse together with the fiber.
And as stated below a smaller nozzle (0.8 to 1.2 mm) but increased layer width and thick, but no more than 80% of the nozzle diameter, might force the plastic into the fiber and give a bit more pressure on the build plate to get some adhesion.
You need a filament cutter with every retraction
The goal of this project is to make continuous CF work on a stock 3d printer so it's limited to vase mode, at least for now.
I’ve been wanting to try this for over a year now , glad I ran into this video! I’ve seen a company from France named Zero3D /Zero D at Formnext Germany and he had a filament making machine built exactly for this. He had filaments with continuous glass fiber aswell.
He said he was able to print with them on a sovol sv06 and the prints were looking promising. Not pretty but strong.
This would be really interesting to try in a system with a filament cutter guillotine and experiment different metals for the blade to see how it gets cut.
If something like that works it might enable next level strong parts on desktop machines, even if that has like an extra purge line added manually to do the cutting part.
He had less fiber laminated inside which i think might help getting better desktop results.
Dosing the material is a challenge here having the fiber volume inside and most companies went for a separate feed for the continous fiber that gets laminated in the nozzle to overcome that.
Just an idea that popped to my mind when seeing the video's begining : with axial fibers, you need to ensure that the layer lines have the same cross section as the polymer in the filament, elseway your printer will stretch the fibers as it wil try to lengthen the filament by printing it at a different cross section to keep volumetric flow constant (the polymer needs it). Why have the fibers purely axial and not coiled around a main thin filament like a spring, which would then be encapsulated in aouter shell (like with your pulltruded attempts) ? as such the fibers would be able touncoil themselves into staright sections to allow for filament lenghthening. It's an idea that would need a lot of work in order to have a reliable fiber coiling machine, but could definitely work.
I am thinking about how to make continuous fiber filaments for a long time and coiling around a mandrel filament might be a good trick
Start with a high quality mechanical pencil and use it to dispense the carbon fiber strands only then pass it through a fine brass tube while being mixed with a UV activated resin/polymer once the wet strands exit the brass nozzle a UV laser is directed towards the deposition contact to instantly solidify the wet strands (BTW brass is self lubricating and easy to solder and available in small diameters look for EDM drills that is why I'm suggesting it) good luck my friend.
I'm working on the same project lately for my master's thesis. We had some similar ideas. Try pulling the roving through a stock hotend (0,6-0,8nozzle) ;) I'm about to make my first video on yt to share what I have found so far.
please wear gloves
I wonder if you can make a nozzle that has an inner, separated channel to deposit carbon fibers directly into the print
I made a mockup in onshape of a more ideal nozzle (would have to be metal 3d printed) called "carbon fiber depositing nozzle", its public so it should show up if you look it up
the idea is that the melted filament and already printed fibers would pull the fibers through the nozzle as the printer moves
Just took a look at your nozzle, that's a fascinating design! It's definitely close to the limit of what a metal printer could make, but if it could be made, that might be the best solution to continuous CF printing!
Try vase mode with a 1mm nozzle. At 0.8 layer hight. There is nothing on your test printer to cut the fibre filament on retraction moves.
But very good effort. Will look forward to your next video attempts.
Altnatively try a thinner core of carbon fibre if you want to continue using the 0.4 nozzle.
Additionally I think it's a very interesting concept the current issue with carbon fibre filaments if the size of the chopped fibre is too fine to actually provide much more strength. Using your idea. If you were able to say add 2-3mm strands in as a core due the the nozzle melting, stretching and squashing the plastic you may end up with a very usable filament without the need for a cutter and retain 0.4mm nozzles. I imagine the end result would be somewhat akin to fibre glass mixture.
I've had this idea for a while but with a 0.5mm filament, that way we can retain the print quality by using a .5mm nozzle and only merge a few fibers inside the filament, making it possible for the printer to break it with travel moves 😇
Great effort. Eagerly awaiting the next iteration!
First of all: CONGRATS getting your first video out! Biting the billet and uploading before what may be considered “success” was the right call.
So the results…I can’t believe you got filament that looks that good out of the “rigs” you used. There are a few factors I’m aware of that were working against you:
PET is not easy to print; it likes to be around 280C and the bed at 80C, and you’ll have struggles with bed adhesion without a PEI build plate. I don’t think that an ender can print straight PET well at all - I suggest investing in a roll of PET(no G) to use as a control.
Now about the extrusion rig…you can get lucky, but it needs to be temperature controlled and high wattage. PETs temperature window for melting without crystallization is not large. PET also needs to be heated up very quickly, then cooled off quickly, the goal there is to make it amorphous. Considering you were PULLING the extrusion through with your hands, I think we can assume it didn’t come out of the nozzle at 260C-300C.
I know it’s probably not in the budget yet, but a secondhand 3d printer seems like a great way to get some of the parts you need for cheap, like a PID controller, motors, thermocouple, power supply, and maybe even the heat block will be enough.
Okay so after you have tuned and successful print settings for PET and your filament is amorphous, the name of the game is now Consistent Extrusion Diameter. Think about it; if the variance is 0.5mm, the printer’s flow rate is off by 30% EVERYWHERE. What happens if you try to print a benchy at 70% or 130% flow rate? You don’t get a boat. So to get the extrusion diameter tolerance better is simple: consistent speed. This is where that motor from that secondhand 3d printer can help.
Lastly; I’m not an expert on filament manufacturing, I just have gathered a lot of knowledge surrounding it. If you come across conflicting information I wouldn’t be surprised, and if you have conflicting experiences, find out why before dismissing any of it.
Again, great work and im excited to see this channel grow!
I see what you were doing with the zoom in at the end, very cool idea!
I made one of those diy pet pulling machines, but modded, i made a 2 stage melt zone, one big long heatblock lower temp +-210C with a big hole so you get a perfect U shape then trough a modded nozzle to 1.7+-, at 250-260C this 2 stage made me able to increase spooling rate and gives way better results since you don't stretch thé pet, or overheat it with 1 stage. Such a design works perfect with your idea. I would go for thinner Carbon fiber stain, this wil give to much back pressure in thé nozzle, and increasing heat with pet above 270 to decrease pressure does not work, it cristallizes, ( visibel white), and it's just overal bad print Quality, no work around that, so keep thé temperaturen at 260-265, also i tryed using a pretty modded Ender 3 for this, bowden setups Will always fail with this. If i where you i would take out some variables out of thé game, like a bowden setup. If it works out i think it would work. Great first video🎉
Also btw, PET is very hydroscopic, so you need to keep it dry.
I also saw a command about the strains not able to stretch, which makes sense, so you could try extrusion testing normal pet that you pulled with the same heat and strip width so you know how luch plastic there wil be and hiw much CF, wich makes you able to calculate the wall width, it will not be 2mm by 1.2mm like he/she mensioned.
for better adhesion between plastic and fiber, you could try dipping them in a solvent that disolves the plastic, then even between the fibers will have a little bit of plastic.
Con: you will need to bake out the solvent before printing.
Also, i wonder if something like heating your filiment could help. like, oven, or run over a flame quickly.
markforged uses a cutting knife driven by a servomotor(like in rc planes), so they can retract and make the reinforcement
Something else you have to consider is what to do when the filament retracts, or you'll have to stick with vase mode models
I have made a LOT of PET bottle filament. Like 20+kg of it. What you're trying to do here is VERY interesting and VERY possible with a few tweaks. In fact I have thought about this but with glass fibres instead. When done correctly the PET tape pulled from the bottle will form a tube as it is pulled through the heater and nozzle and the trick is to get it just hot enough to soften and shrink without actually melting. Getting the volume correct is key. Too much PET tape pulled though and it jams and snaps. Too little and it doesn't form a proper tube. Considering how little filament is actually laid down when printing I think less carbon fibers would be better and definitely individual fibers instead of a tow or tape. But you're definitely getting close.
The main problem I see with this is the slicer and the logic it uses to create the line path. Ideally it would be one single line wihtout ANY retracts or travel moves. Almost like the entire print is in vase mode but wihtout it printing only the outer wall. I suspect an IDEX or multi toolhead printer would be best for this with one toolhead providing non-CF filament for supports and areas where the CF is not needed/wanted and a second for just the PET-CF filament.
Proper Printing did a great series on trying to print resin with a cartesian 3D printer and I think laying down a bead of resin that has a line of CF embedded in it would be a game changer. Between what you're doing here and what he did there I think the two of you could REALLY have something big. I just wish I could give this video more than one like.
Ich would guess that you could do it with a setup that looks a little bit like a wire-edm Maschine. One Side pulls the fibers through a molten platic bath. The other side keeps a little resistance so the fiber is always straight.
I hope you know what I mean 😂
Best regards from Germany 👍
If you keep trying with one nozzle I'm sure you'll have a very bright future ! Maybe you will sell filament ! Keep pushing !
Haha, great intro. Pity about the audio quality for that bit.
It looked like you were printing your first layer too quickly. I'd also use a brim, just so it has an easier time sticking down before it makes any sharp turns. Printing with a big nozzle isn't easy.
Every time a new line begins, you need to cut the fibre. Otherwise, there's a practical limit to fibre length. This is doubly-true for retractions; you do not want to drag a long strand back into the nozzle!
I subscribed, so you're obligated to continue ;)
I think you can get away with much fewer strands also dont make it a continuous strand , make overlapping strands so they can slide past each other . then try and make much smaller diameter filament, larger diameter nozzle , vase mode. Higher temperature. Problem is that printing this way only makes it stronger parallel to layer lines. Very cool concept though , I could still see applications for it.
instead of making something that cuts the fiber while lifting/retracting the nozzle, maybe you could try to make everything work in vase mode so it does one continous move.
I you misundretood how the printer uses the filament.the filament gets "streched" by a factor of "alot" which carbon fiber can't do .that is a fundimental problem you have to overcome before everything else in my opinion. cool video 👍
I wonder if this more like instead of printing the whole part like this, have some designed channels where this filament could create strength. But the shape is made by other plastic. Also I think you need to consider the carbon fiber bend radius, in your part. I don't think it can bend to the radius of your printed part. Start big with big curves and go down from there.
I basically had the same idea yesterday evening, I've been searching for parts and stls since. I think printing in vase mode is a must.
How many fibres does this tow have?
I think 3k should be good.
There are also a lot of community projects on pet-bottle to filament pulltruders (if you want to avoid pulling by hand).
The two stage integration of the fibre is a great idea, in my opinion.
I think you can also decrease your nozzle diameter a bit (maybe 1.5 mm while keeping the same crosssectional area of the printed line) and decrease the volume flow rate.
Carbon fibre conducts heat way better than PET, so you might have to crank the nozzle temp or get a volcano nozzle.
You might also have better luck with a direct drive extruder. I also saw that PET absorbs a lot of water, so you might want to dry your filament.
what about a dual nozzle system, one for plastic and the second for fiber?
Amazing; great work. Excited to see what progress you make.
How much is that spool of continuous fiber?
Around ~$3 for 30 meters
The second you pulled out a bottle cutter i knew you are an absolute genius
Cool idea, great video!
Thank you for answering the question, that I thought, I was the only one asking.
This can definitely be automated with some stepper motors and a loom the system
Great try… if i may suggest.. try with half the amount of carbon fiber, better chance to get good print and the carbon will still be there potentially always covered. But you have to get a consistent print without the carbon first. Just a suggestion, i can only imagine the amount of work. Carry on 👍🏼
that's awesome! only real note is that i would highly recommend wearing gloves when working with the fibers-- a lot of those small fibers can get stuck inside you, it's really annoying...
We get it. You watched Nathan's videos on carbon fiber. Bravo. You're an expert now.
@@BuzzingGooberI'm talking about working with actual carbon fibers, not the chopped up bits in filament. I had to look up what you were talking about, I'm talking from experience working with composites (granted, mostly fiberglass, but the principles are the same) where gloves are mandatory even before you start getting resin into the mix
I thought continious carbon fiber is printed trough a 2 way in nozzle with normal no matter what filament comming from one entrance and a 0.2mm carbon fiber strand (in the case of a 0.4mm nozzle for example) comming from the other strand and beeing combined inside the nozzle
... like those mix nozzles that combine 2 filaments, so you can print any colore you want (as long as it can be mixed by 2 other colors), I think there are some videos on youtube about it
Using a hot glue gun as a filiment extruder is genius!
Cool! Did you test the tensile strength? Also, does the plastic get in between the fibers, or is it just a shell? Is this a problem?
Planning on doing some tensile testing in a future video :) There are some places where the plastic gets between the fibers and others where it doesn’t, I suspect this lack of consistency was from my hotend being pushed to the high end of its limit. Just from playing around with the samples though, the CF ones feel a lot stiffer.
Great experiment! Good video
Does your printer have direct extruder mod? If not - buy/make it direct drive, using something like BMG extruder
Starting is the hardest part, couple more tries, and you got money printing machine
The problem is that layer adhesion is going to be the weakest point. Unless you can make slicer modifications in away that make up for that.
Amazing work. Superscribed!
First off, great video! Definitely earned a sub. But -- I only watched the first 30 seconds so far, so maybe the rest of the video will address this -- doesn't continuous carbon fiber filament already exist? It's exorbitantly priced, sure, but exists already and has been used at a consumer level.
That's a great point - I may be wrong about this, but as far as i'm aware, there isn't any continuous CF filament that can be printed on a run-of-the-mill printer without modifications. Printers like the Markforged Mark 2 use a spool of carbon fiber in conjunction with a spool of conventional filament, which gets combined in/around the nozzle. That process is a lot more practical for making useful parts, but the process in this video is geared more towards hobbyists, as it just requires a nozzle swap before being able to print a part (albeit a part constrained by the limits of vase mode).
This is awesome thanks for sharing
You need a cutter for retraction
BTW. CF is basically new asbestos.
I though of the exact same concept last year when i was making pet filament but i got lazy
You should first wind the carbon fiber into the plastic then pulling through the glue gun
Haha 13 seconds in and I'm already subbed admittedly I subbed as soon as I saw this video title but, that joke let me know it was a good decision.
Hey man! Great vid- I’d wear gloves tho
why not use a screw like a normal filament extruder, but also let it push the carbon fiber with the plastic, push not pull. why not a pet plastic tube with a separate fiber running inside. one other option is to sow a 3d shape with the pure carbon fiber, then add resin. well petg with hardened steel nozzle needs about 260C (+30C vs normal nozzle, for cf stuff). diy milk casein plastic with carbon fiber strands might be very nice.
PET is a very strong engineering filament. You’re not going to print that on an unmodified ender 3
IMO embedding the fibers in the plastic is a very limiting approach. The fibers and the plastic need to be able to move at different speeds.
Cool!
Excellent!
0:12 😂😂😂😂😂
Hey Alex! I just posted a video where I Show my setup for Continuous Fiber Filament making. Feel free to check it out! Regarding your setup: you should carefully select material for the matrix of your filament. (Not a PET from bottles or a gluegun) It should have low enough viscosity at elevated temperature to fuse well with the fibers. Study material properties and choose what best suits your needs) I used PETG and it required 260 deg for good impregnation. Full impregnetion of fibers is essential to achive a filament that will give you good resoults in printng. Good luck! ua-cam.com/users/shortsv8ktu00bYaQ?si=lsMlBJT85p8Yz2l0
Btw. I see in the background you are a fellow RC modeller. If you want to we can get in touch somehow to talk more about the project. Cheers!
👌🤏
So what were the print settings that worked vs the ones that didn't? What helped vs hurt it? That seems like the main discovery of this video but it was skipped over. Also consider a 0.8 mm nozzle so the fiber jams and stretches less
This isnt going to work the way you want it. I think stopping the print and and placing fibers at various levels. Or to create a channel in the design to fill with carbon fiber and epoxy for the extra rigidity.
yes process is possible the vompany you mensionsled doing that for a lot of money but the way you do it is just stupid... at least make filament inteself with some kind of carbon addition and longer nozzle to maintain better temperatures across all of the printed poop you did so maybe then you will have any results... black filaments prints better become of better conductivity and your nozzle is like yours dad pp short and it's surprising that it made something... get at least big vulcano nozzle if not super vulcano one and print slow on 1.75mm and print big thinks not small resolution will be shit... actually the best will be to print on 0.8 and have separate "carbon nozzle" that you for example tie to a neil and melt into the plastic or nozzle within the nozzle where you will cover carbon fiber by fresh hot plastic
keep going on maybe one day you will sell yourself 15k machine
hopefully some filament company will pick up on this and create CCFPLA / CCFPETG