Incredible High Performance Extruder - FuseLAB Formnext 2019

[@JulianMakes]

I would like this extruder as a consumer product, i can see its great potential with it’s high force and high volume accuracy. Great video too

[@MisterkeTube]

I have seen some KZbin footage on an amateur extruder along the same lines, so I hope that attempt and record on KZbin can beat their patent (I hate patents). That said, if you anyway want to push stuff with a screw motion and can accept the extra weight that brings, why not skip the filament entirely and directly extrude pallets?

[@chromaflow9313]

4:16 is when he actually starts talking about the extruder

[@Anyone700]

Actually, I have been working on a very different design of a similar extrusion system. It is comforting to know that it does not work. You cannot call being fundamentally incompatible with flexible filaments, absolutely insane part complexity, and having an encoded servo as a requirement as working. At that point having a filament encoder and a double layer extruder is much more practical. In-fact you can see this in the Dyze typhoon. For the excuse that is justifies itself with power/weight, I would rather see a tool changer or more robust motion system.

[@ezengondolkozom3700]

For that matter: David Leitner on youtube is making a rolling screw extruder. I believe it should have better backlash and dampened frequency curves upon FFT which is what really matters in an extruder for me.

[@dodobarbar]

Yeah I saw that at the show too. Though, beside being something new-ish there isn't much to write home about. It's an expensive way to do what the rest of extruders are doing for much much less money.

[@KasparJohannes]

Especially since it can't even handle any flexible filaments🤔 It's certainly nice engineering though😍

[@patprop74]

you would be surprised how cheaply it can be replicated

[@Inventorsquare]

It is nonetheless an interesting contribution. I'm sure it has inspired others already.

[@Unmannedair]

It would be a significant advantage for driving filament at high speeds which is what I'm using it for. It's a huge waste if your printing at 40mm per sec, but it's invaluable if you're printing at 300 mm per sec. My current goal is 1 meter per second. It's ridiculous, but necessary.

[@Anyone700]

@@Unmannedair Necessary for what? Its cheaper to have many small slower machines running for 0.4-0.8mm nozzles, and cheaper to have higher flow rates at the top end Prusa runs the way they do for a reason. As for the top end, look at the machines the US navy uses for printing boats. 1 m/s is prohibitively difficult for the value

[@patprop74]

That is a very good design, i have no doubt it will work extremely well, and might be the next norm for future extruder designs

[@SusiBiker]

Nicely covered. Looks really impressive. 👍

[@Vector3DP]

Thanks Susi

[@TonyRios]

mechanical engineering is the best I love seeing small mechanisms with little bearings,

[@nebuchadnezzar47]

dat ringing

[@Dave_the_Dave]

If they split it into two sets of those rollers, they could have one ring counter rotate. The net amount of twist would be 0 but there may be still too much twist between them for flex filament.

[@Underp4ntz_Gaming_Channel]

The feeder looks damn cool!

[@Unmannedair]

I already had plans to build something like this from scratch. I would be interested in buying this if they can make it cheaper than I can. Also, flexible filaments are definitely doable.

[@michaelthorsby]

Forget about it, it's patented, lawyer up!!

[@pirobot668beta]

Cool beans. How about hybrid Bowden/direct drive? Bowden tube/extruder; old-school material over time, no pressure advance and no pesky retractions! At the hot-end, small servo-motor (voice-coil?) sits between the Bowden tube fitting and the nozzle. Servo travel moves fitting closer to or farther from nozzle, regulating pressure. For retraction, servo shoves fitting far from nozzle, reducing pressure immediately! While the hot-end is accelerating, the fitting is pulled closer to the nozzle for a bit; this boosts pressure over the steady delivery from the Extruder. Once hot-end is moving at steady speed, servo locks* in position to keep pressure steady. Nearing the end of the segment, servo lets off on the fitting a bit, reducing pressure and returning fitting to 'start' position. *Servo position is under closed-loop control; signals tell servo where to go, it WILL go there and it will use as much power as it takes! Monitoring the drive current gives a direct measure of nozzle pressure! Wouldn't that be nice? Retraction that self-calibrates on the fly... Oh, modulating the control signal with analog data would be a nice way to do lithophanes or other surface patterns.

[@BobbyJ529]

I missed what the benefit was to this design. Was it speed and a more even extrusion? If so, how much faster?

[@ChiralSymmetry]

Two benefits are claimed: (1) First, the filament is pushed/pulled along a longer surface of the filament. This allows for higher extrusion force. That is, the force is spread out, making it less intense at any given point. If all that force were focused at one point on the filament, then you would likely grind filament instead of extruding. (2) The grooves cut into the filament can affect extrusion. In a conventional (BondTech) type of extruder, the groves are (approximately) circular. So if you look at the cross-sectional area of the filament, it goes up and down. That means your extrusion volume goes up and down. Sometimes this causes subtle surface artifacts in the 3d-print. The FuseLAB extruder cuts a helical grove into the filament, and the cross-sectional area is constant, so the extrusion volume is constant. Hope those make sense... There may be other benefits too, but these are the two that I understood from the video. My guess is that the effective "gearing ratio" is very high, much higher than the typical 3:1 of geared extruders. Personally, I like the idea mechanically, but wonder about the added weight. That would come down to actual testing of the machine.

[@riakata]

@@ChiralSymmetry I dunno having rapidly spinning screw threaders running around a soft plastic string seems like a bad idea. With geared extruders they can bite in deeper and with a dual gear system you can even put the peaks/valleys at alternating positions to minimize that effect. I highly doubt this subtle surface artifacts claim. Having very shallow and rapidly spinning drive screws makes the effective area small and the surface velocity high which is also going to cause problems if it gets overloaded it will strip and grid the filament even easier so even if they have a larger thread it roughly nets out to be the same. Not to mention with a screw drive only the first few threads matter so in an overload situation it will just cascade where the first few threads will fail and it will just pull out entirely. Which is dependent on the filament which makes it more difficult than it needs to be. Dirt accumulation will also clog the drive system more and the rapidly spinning parts creates additional hazards, wear on bearings, etc... All in all it will have more contact area by a bit but in reality it probably has worse performance than a bondtech drive. Not to mention you can always swap out the hobb gears to better match your filament with this system that will be tricky to do. It just doesn't seem to make much sense when a simple single drive gear system can produce results the same as a dual drive system minus the surface artificats you can clearly see in the video which already invalidates claim 2. I suspect the rapidly spinning extruder system will introduce vibrations into the system which will ruin surface finish. Consider how beefy their demo machine is and how wavy that surface looks its definitely something to do with the rapidly spinning and speed changing extruder drive system. The inablity to print flexibles and likely weakness with softer non-flexibles like cleaning filaments, harder materials like carbon fiber/metal filled loaded and supports like PVA will make it a non-starter for professional applications.

[@fuselab3d446]

​@@ChiralSymmetry Good analysis! Weight is actually very low, see some of my other replies.

[@Sigmatechnica]

its different and it looks cooler.

[@xxportalxx.]

@@riakata actually you can do out the math, it makes a very subtle surface pattern, but at very small layer heights this is actually visible on large flat surfaces, just watched a vid on it, the gears on a dual gear extruder create much more severe artifacts tho, as when the filament goes through it pries the gears apart, they stop meshing properly, and this creates an aggressive wood grain pattern (with that one you can confirm the cause pretty easily too)

[@ScottKuijlaars]

100% would buy

[@ChiralSymmetry]

Just for reference and the curious: This feed-mechanism is well-known in the thread-rolling industry (large bolts, ball-screws, lead-screws). In thread-rolling, a cylindrical blank is "squeezed" by two or three rollers with a "negative" of the thread shape. In thread-rolling, the goal is to make screw threads. The feeding of the blank is a secondary benefit. In the FuseLAB extruder, the feeding of the filament is the main goal. The helical threading is a secondary "benefit." Here's a video of an industrial thread-roller by LMT Tools: kzbin.info/www/bejne/ipC6ZatofcqWY9k (About 5 min) Random observation: Geometry and math peoples will recognize that a hyperboloid is swept out by the axes of the three rollers.

[@Unmannedair]

You're exactly right. I was already planning on building this for my personal machine. Looks like someone beat me to the draw. I want planning on commercializing though, so good for them. Kinda curious that they tried to patent it though...

[@riakata]

Problem I see is plastic filaments are not good screw making materials. The depth of their "thread" is very shallow and unlike a screw maching machine it has to be able to reverse, stop, change speed rapidly during a print. This is not a good fit for a high rotational inertia system.

[@avejst]

Impressive engineering, Fantastic product Thanks for sharing :-)

[@dougingraham5807]

Although an interesting idea I don't see any advantage and there are a number of disadvantages. The increase in complexity and the inability to handle flex materials kind of makes this a non-product. It is a solution looking for a problem.

[@grantdeisig1360]

Agreed... It seems over-engineered to me.

[@falsedragon33]

Agreed. Patent chasers. It looked to be crawling too.

[@davids.6671]

Looks usefull for Peek and other high strength thermoplastics. Or highly filled stuff.

[@pepe6666]

i think its quite a clever design. considering the bondtech gears are giving prusas a lot of problems with inconsistent extrusions i think this is a welcome addition. personally i still think its the wrong approach tho

[@julienboucaron9780]

Nice stuff. Those hob gear cutters are really interesting. There is a MIT publication for a high speed 3d printer that is using threaded filament like that, probably some patent(s) around too.

[@cowboy124aa3]

Incredible, beautiful design and innovation, speeds look amazingly fast. Only thing i was seeing was wavy ghosting in some of the vase mode prints but still beautiful. Would love to see this go to the public market.

[@dougingraham5807]

The wavy ghosting is perfectly aligned vertically. This means it has nothing to do with the extruder. This is a characteristic seen on well tuned printers. Once you get everything else working well you see this kind of pattern and it is caused by the non-linearity of the micro steps. Switching to 0.9 degree motors will double the number of waves per unit and the amplitude of the waves is often reduced to the point where they are difficult to see.

[@nosiobiad]

The speed shown in the video is nothing special. Cheap chineese printer is able to handle 80-100 with some good stepper motors drivers.

[@riakata]

@@dougingraham5807 The rapidly spinning heavy extruder drive system is going to be a vibration nightmare. Speed changes, 2nd extruder starts/stops, retractions, etc.. are going to be bad times. The level of precision machining that would be required makes it too complex for what is needed as it needs to be perfectly balanced or it will just vibrate the heck out of the machining and likely is. They will need to add damping and mass to the system which will slow it down.

[@butunercouk]

Great engineering 🤙🏼 That design may create an issue with dynamic layer height. 0.05mm layer height for overhangs maybe?

[@wktodd]

I designed a nut like this a decade ago ! Prior art?

[@kevfquinn]

Ingenious. Still trying to visualise how those cylinder teeth work - do they moving freely on their own axis (relying on friction against the filament perhaps?), or are they keyed to the main body rotation?

[@fuselab3d446]

They move freely indeed, and spin around their own axis.

[@kevfquinn]

Oh I get it! For some reason I thought the cylinder gears had spiral threads, but they're a series of rings for teeth. So it's the angle of them that creates the thread on the filament, and it doesn't matter how much they rotate on their own axis. Cool!

[@Zonkobel]

its so easy, that its already genius

[@zachlynn276]

Great solution seems like a no brained but they did a great job using new tech

[@SergioCastanho]

Congratulations on the project. It was excellent. Do you allow us to replicate the project under your supervision?

[@Vector3DP]

It's not my project. You need to talk to @fuselab3d about it

[@SergioCastanho]

@@Vector3DP Tks

[@thehightechredneck9777]

Glad to see some innovation but seems like it would add weight to an already heavy extruder.... I seen something similar on youtube the other day where the heavy things were not on the print head but were on the side of the printer like bowden style but it rotated the bowden tube and drove a feed gear in the print head.... I believe it was said to print flexible around 100mm per second... Love seeing new filament delivery systems... Keep up the good work

[@MrStemkilla]

I wasn't aware that inadiquite exstruder force was an issue.. and even at those speeds im still seeing printing artifacts. I suppose this is meant for large layor heights at fast speeds but i dont see many commercial printing processes that want large layer heights. Everyone wants tinny layers as fast as possible with no defects.

[@slayyou]

bingo

[@marckerger]

Will this system be commercialized sometime?

[@JeffDM]

I could have sworn your narration said the anti-twist device is in the exploded view but I can't figure out where.

[@Vector3DP]

It's the square block part which has two circles on it. Looks kinda like the back end of two bullet cartridges.

[@ameliabuns4058]

that's soooooooooooooooo cool! i mean I don't know if it's good or not but it's very cool xD

[@aronseptianto8142]

how heavy is the assembly, wonder how it'll fare for the benchy race

[@Wellinton-Impressao-3D]

Very good!

[@tvathome562]

looks cool jsut would like to know accuracy for volumetric flow and top extrusion rates, and heat creeep?

[@richard--s]

The extruder looks really fancy, but has some disadvantages: No flexible filaments; it's a complex and heavy extruder which means lower printing speeds or at least slower accelerations (look at the sample objects with very gentle curves. Edges are possible of course, but would need to stop and accelerate that weight); the complexity means a higher potential for failures after some time; etc.

[@riakata]

Yeah I would imagine retractions and speed changes will cause problems with that very rapidly spinning extruder drive system.

[@fuselab3d446]

There seems to be a misconception that this extruder is heavy, it's actually the opposite, this design allows for a very light extruder design. There is a path forward that would allow an extruder below 100 grams. The reason that it can be this light weigh is that we can achieve a better match between needed torque and speed with available motors than what is possible with pinch wheel feeding systems. With pinch wheel feeding mechanisms you need high torque and low RPM, meaning that motor power utilisation is low since power = torque x angular velocity, and the motor is sized for the torque part of the equation, but the speed part is left untapped. That's why the better pinch wheel extruders have gearing, whereas this gearing is inherent in our design (without adding mass). Finally by using a BLDC motor we completely solve the retraction speed issue, and have a near perfect match between motor and system.

[@richard--s]

@@fuselab3d446 OK, that is interesting. Thanks! Good facts are always welcome! So in fact it's not heavy, that's really good!

[@riakata]

@@fuselab3d446 These are not accurate for stepper motors which work best a low speeds and have a falling torque curve with higher rotational speeds and start a maximum torque at zero rpm (They are called stepper motors for a reason they move in steps) so they are well matched to geared extruders that operate at near zero rpm. The rotating bit of the drive system is far more massive it is impossible for it to be lighter than a simple hobbed gear just by volume alone. Made of the same material your system has many more drive components that engage over a larger area. Spinning this at a high speed requires high precision to prevent vibrations which demands more machined components to hold them in place. You do not address how you handle the retraction problem in that you have to stop a much more massive high rpm system and reverse its direction extremely quickly which will cause vibrations.

[@riakata]

@@richard--s They are misrepresenting the system design of stepper motor extruders which have a high torque at low speeds so they are automatically well matched to geared extruders which also move at low rpms in both extrusion and retraction. Stepper motors don't make sense for their ultra-high rotational speed system but high rpms are not something you want unless you really need it as it has complications with balancing, vibrations, stopping/starting, ...

[@wolfthorn1]

I wonder if anyone has tried liquid extrusion. Maybe a spray nozzle. When you use some printers they leave raised ink.

[@Bordpie]

The design as is, is innovative but doesn't bring much new to the table I think. I think there is the potential to make it really good though. They could make the belt drive which turns the extruder work like a Core XY, and take the motor off the moving XY carriges completely. You would just need another belt path above (or below) the ones for the core XY, and loop it around in a similar manner. Remove as much weight from the extruder assembly as possible e.g. make the pulley concentric around the drive section to make it very flat. A disadvantage I can see is that the forces from the extruder belt may interfere with the core XY a bit but that could probably be compensated for in the design. Just an idea.

[@cowboy124aa3]

Only problem i see with that is belt stretch. Over a periodof time it would stretch from the constant motion causing inaccuracies and reduced flow. A smaller bet has less chance of stretching.

[@Unmannedair]

@@cowboy124aa3 the belt stretching would be inconsequential if they put the rotary encoder directly on the extrusion mechanism instead of on the motor that's turning the belt drive.

[@tinkot]

it looks really cool. But functionally it doesn't add anything but complexity, which is definitely a problem for a business when it needs maintenance or replacement

[@michaelthorsby]

Like a few others I cannot see the benefit of this. Like some has pointed out dual gear extrusion, like the BondTech one, seem to be really safe, with very few additional parts needed. It seems to overcomplicate an issue that is already solved. It also looks a bit nightmarish to service compared to normal geared extruders. ....and it cannot even do flexibles?? What???

[@BH4x0r]

I think it is mainly to get good reliable throughput at large nozzle sizes, BondTech etc are pretty good don't get me wrong but having more grip can defo help in certain cases, doubt you'll get any kind of filament grinding this way

[@avongil]

direct copy of a fette thread rolling head!

[@kazolar]

I'm a bit confused what problem are they trying to solve. Extrusion speed? certainly not size. They've over engineered an extruder to be extremely expensive, use a lot of proprietary parts, and less capable than existing solutions on the market. A bmg, or a hemera is capable of feeding a volcano or a supervolcano with plenty of torque to spare. The sample prints appear to have ringing artifacts just on round surfaces, so I can't see how high speed printing will be achieved such a heavy direct drive assembly. Being clever for the sake of clever is not particularly interesting. Good coverage, but if you're asking them questions, and explaining this "innovation" I'd be interested in a more indepth style of coverage -- as in, you covered the what and how, but how about why?

[@riakata]

Yeah they are claiming the head is very light

[@AudreyRobinel]

Nice video, it's indeed an interesting design. I won't buy it, but i'm interested in seeing you or the community testing it. So it's a new way of driving filament, providing high force. But how is that an improvement over dual gear bondtech extruder gears? over the E3D hermea with a dual gear? I mean, those extruders already provide enough force to drive any filament (and are compatible with flexibles...) so what does this one bring to the table? faster extrusion? more precise extrusion distance? I don't feel like those aspects are things that limit current 3D printers; and designs like the hermea and the zesty nimble already provide great torque and speed to the point where the limit would be with the rest of the printer. So, this looks great, but what does it bring that can't be done with other extruders, especially if it means giving up on flexibles (and what about nylons, that are not that rigid?)

[@pepe6666]

much more faster and precise movements. if you up the speed of printing the hot-end is what would slow this machine down. so theyve sped up one part of the system by 10x, just gotta get the other parts of the system up to speed

[@gilb6982]

This is a very big step in 3D printing not really because of the thread rolling system that is a good idea even if not perfect but because there is a brushless motor and an encoder this mean a Close Loop system now install it on all the axis and there you can have a very powerful fast and reliable printer like the big industrial CNC machine don't work with stepper motor because of the precision and speed and reliability

[@Taconiteable]

Thats not true. Stepper motors are more dynamic because of the smaller rotor weight. Stepper motors are perfect for low torque but precise and fast applications. Stepper motors are not suited for high power applications (like milling and turning)

[@marinalobes3778]

OMG that is A. Copie from a Puplic Patent

[@RFi731]

threadless ballscrew mechanism? nice idea! never thought of that for extrusion drives before. but patented huh.

[@air21511]

KISS is ignored here. :D

[@MrBaskins2010]

So they took my price range, burned it alive and danced on its ashes?

[@pavelkanzelsberger6825]

How is that a winner with all that ringing visible in those prints? It's too heavy

[@Vector3DP]

It's lighter, stronger and faster than the new e3d hermea.

[@riakata]

@@Vector3DP But the ringing is just massive and on an easy print too. It is likely heavier, stronger, slower, and much much more expensive than the new e3d hermea.

[@Vector3DP]

I see the artifacts, but that's not ringing. Will need closer investigation.

[@fuselab3d446]

There's no ringing at all. Only thing visible are segmentation artefacts from the STL file. This is the result of the high stiffness of the frame following the theoretical contours very precisely. We've done test tests with pure G2/G3 code and the curved surfaces turn out to be super smooth. We'll post some video's on our website soon.

[@riakata]

This seems like a odd extruder design. It is high mass, high extruder axis inertia, it also cannot handle filled filaments because it will quickly destroy the rapidly spinning extruder drive made of brass. Dual geared extruder drive gears work much better in far less cost, mass, complexity, and won't be eaten alive by abrasive filaments. Soft filaments will be a nightmare to deal with not just flexibles the way the drive works will cause a lot of problems. Looks cool but a filament flow wheel like the moasic pallete has build in and a closed loop extruder stepper would be a far better solution that could track filament jams and motor step loss. The BLDC high speed extruder is functionally threading a soft plastic string which seems very odd.

[@fuselab3d446]

Actually one of the advantages of this system, is that it's really light weight. In it's current form it weighs +-250 grams per extruder, while the one that is currently being finished for the final commercial version, will be around 150 grams per extruder. Further development in the coming years could bring the mass below 100 grams. The brass rollers you mention where chosen for ease of fabrication in the development stage, and since there is only rolling friction, they hold up remarkably well. The final design will have hardened steel rollers, and will handle abrasive filaments perfectly. Finally, this system can't feed flexible filaments (elastomers) but handles soft filament without any problems.

[@riakata]

@@fuselab3d446 What are you including in that mass your entire extruder assembly weighs more than 250g. I'm talking about the entire mechanical head not just the spinning portion. If we just look at the mass of the geared part an e3d gear is like 11 grams. You can't get away with a ultra-light head as you have to hold that rapidly spinning drive system in place without causing excessive vibrations. How does it even handle starts/stops/retractions stopping that much mass is going to cause serious vibrations the drive gear of your system has sigificantly more inertia than a standard geared drive system. The high rotational speeds don't help either as to do a fast retraction you would have to go from high positive rotational printing speed to many multiples of it in reverse vs a gear drive system which is almost always near stationary and has a tiny mass relative to the screw drive system.

[@fuselab3d446]

@@riakata 250 grams is for entire assembly. (1 extruder) Vibrations are no problem at all, at extreme high printing speeds max RPM is still only 600 RPM. Retracts are done within 10ms, so no problems there.

[@rusticagenerica]

I don't even understand how it works.

[@fireblade639]

It can't use múltiple size filament... One head for every size Is expensive... Very.... Is overworking the Solution...

[@Vector3DP]

Extruders never use multiple size filament.

[@fireblade639]

@@Vector3DP so 1.75 AND 2.85 mm aré the same size??? No... And the list Will go up as diferent reinforcements aré added in the future, forcing the material into the nozzle with three rollers with no way to overcome the cleareances Will chew the filament. So very good quality filament Is needed to use those heads, so price increment again, the springy design on others feeders make posible to use low quality filament with no problems, AND the 1.75, 2.85 (sometimes 3mm)

[@TenchoTenchev]

I don't get it with this extruder, It's very complicated, heavy and patented beside of all. Who the hell will want this in the printer? Yeah, it's looking good, but the downsides are too much, and definitely it's only for the show nobody will put this in mass production.

[@odudex]

Impressive to see it working, but, no thanks, I don't want it.