from the waiting-for-the-high-speed-3-D-copier-machine dept.
MIT's new desktop 3D printer technology increases speeds up to 10x
[A] team of engineers at MIT have shown off a 3D printer capable of creating builds up to 10 times the speed of their consumer counterparts. According to the team, objects that take around an hour to print on conventional systems are done in a matter of minutes.
The system is built around FDM — the same technology used in most desktop 3D printers, which deposits melted plastic layers to build up a structure. MIT made some key tweaks to the print head in order to speed things up, including a screw mechanism that feeds filaments through at higher speeds by getting a tighter grip on the plastic than the traditional pinch wheel model.
The print head also added a laser near the new mechanism capable of melting the plastic much more quickly. Those pieces were coupled with a speedier gantry capable of moving the print head at a higher speed in line with the upgraded print speed.
(Score: 2) by Freeman on Thursday November 30 2017, @06:16PM (1 child)
Pick Two
Joshua 1:9 "Be strong and of a good courage; be not afraid, neither be thou dismayed: for the Lord thy God is with thee"
(Score: 2, Touché) by khallow on Thursday November 30 2017, @06:32PM
(Score: 0) by Anonymous Coward on Thursday November 30 2017, @06:30PM (6 children)
Sounds like they've simply borrowed ideas from injection molding, is there anything really new and interesting they thought of?
(Score: 1, Touché) by Anonymous Coward on Thursday November 30 2017, @06:45PM
:-D
(Score: 2) by crafoo on Thursday November 30 2017, @08:35PM
It's MIT, so no.
(Score: 0) by Anonymous Coward on Thursday November 30 2017, @08:40PM (3 children)
Which is itself a valuable invention.
(Score: 0) by Anonymous Coward on Thursday November 30 2017, @09:04PM (2 children)
Applying existing methods for very similar processes to another process isn't really an invention, it's common sense. Injection molding machines have had screw feeds since their inception. Lasers are new (I think at least, I've been out of field for a few years), but are they capable of doing high-volume work or are they hobbyist-scale only? Usually melting/heating the media in the head too fast is a cause of trouble anyway - feed interruptions, burnt material leading to contamination or clogging, etc. On the other hand, it might keep the screw from heating up as much as they tend to, which would prevent some problems... I think I need to finally get around to building a 3D printer.
(Score: 0) by Anonymous Coward on Thursday November 30 2017, @10:57PM
Make sure you add lasers. And,while at it, sharks.
(Score: 3, Informative) by Zinho on Thursday November 30 2017, @11:00PM
This isn't the kind of screw you're thinking of; they're using filament that has a threaded profile on the diameter, then spinning a barrel nut around it to drive the filament. Their logic was that the traditional drive wheels didn't have enough engagement with the filament, so they figured out a way to grab more filament at once. $DIETY only knows where they found threaded filament...
"Space Exploration is not endless circles in low earth orbit." -Buzz Aldrin
(Score: 3, Informative) by Zinho on Thursday November 30 2017, @10:56PM (3 children)
The laser is the part I have questions about in this rig; neither filament feed rate nor linear travel present a serious limit to how fast you can lay down material. The real bottleneck these days seems to be heating capacity on the print head. If you're feeding Xmm of filament per second and the filament has a cross section of Ymm^2 then you have to supply enough heat to take X*Ymm^3 of plastic up to its melting point, melt, and then reach the desired deposition temperature. The heater is usually run with a Pulse Width Modulated (PWM) supply so that when you want to stop feeding for a bit (say, move to another part of the print w/o bridging the two and leaving strings between) you can quickly shut off the supplied heat.
So, are they running PWM on the laser? Do they have some other method for controlling the filament temperature (i.e. analog control of the laser intensity)? Are they heating the filament directly, or are they heating the print nozzle and letting that transfer the heat to the filament? if it's direct filament heating, how are they keeping the laser aperture from getting fouled?
I've found the arxiv.org version of the paper, [arxiv.org] no need to give any money to Elsevier. This answers most of my questions:
* laser shines into a quartz heating chamber, no leak path for melted filament
* laser is a pre-heater only, takes room-temp filament to just under the melting point
* melting is done with a conventional heating block
* laser is a 50W near-infrared model, 808 nm; note that this is more power than my entire rig uses (Prusa clone)
* laser power is adjusted to match feed rate, no further detail given on laser control
Of course, the paper raises a bunch of other questions, too:
* the screw drive depends on having filament with a 4-40 UTS thread profile on its OD; where are they buying threaded filament???
* what is the market for this? Their servos are 400W AC units that cost >$500 each, [anaheimautomation.com] that alone puts them far outside the hobbyist market. I doubt we'll be seeing this on a Prusa system anytime soon; they're probably looking at breaking into the industrial market with those kinds of part costs.
* they mention that the laser heating effectiveness depends on things like how transparent the filament's base plastic and dyes are to infrared light; this means that they'll need to re-calibrate the system for each filament used. No question here, just a big "buyer beware" sign.
All in all interesting; I'll give them props on the impressive build quality they're delivering; the chair model they show in figure 4A has some impressive overhangs that don't appear to have support material under them. If this catches on it'll be a while before the supply chain catches up to supplying enough threaded filament to make this even an option. I'd love to think that people buying lots of servo motors would perhaps bring down the cost, but I'm not holding my breath there, either. This has the feel of being a proof-of-concept build where they bought the most expensive components they could afford with their grant money and then took it out for a joy ride to see how fast it would go. There's a lot of room for value engineering in this design, which they'll need to do before going commercial.
"Space Exploration is not endless circles in low earth orbit." -Buzz Aldrin
(Score: 3, Informative) by Virindi on Friday December 01 2017, @12:44AM (1 child)
The chair model is clearly printed on its side. You can see in the other photos that their ooze control is hardly great; this is an obvious consequence of having such a long melt zone. This is a well known speed vs. ooze tradeoff...for instance, the E3d v6 vs. volcano nozzle systems. In theory you could make the melt zone as long as you want, but that would just mean more melted material available for oozing.
Several of us have long considered a filament preheater, so it is nice to see someone actually make one. Knowing MIT though, it will be patented.
(Score: 2) by Zinho on Friday December 01 2017, @05:35PM
good catch, the grainy pictures tricked me on that. Informative mod for you, thanks!
"Space Exploration is not endless circles in low earth orbit." -Buzz Aldrin
(Score: 0) by Anonymous Coward on Saturday December 02 2017, @11:13AM
> filament with a 4-40 UTS thread profile on its OD
Since the plastic filament is soft, it could be an ideal candidate for a rolled-in thread, done in the print head, just before laser pre-heating. Thread rolling is how threaded bolts and all-thread rod (steel) are mass produced.
Another choice would be to cut the threads with a die, but this would generate chips that are a nuisance to get rid of.
Either way, once a threading process is started, it keeps going. It takes some effort to start running a die down a rod (metal or otherwise), but after that the die is self-feeding.
(Score: 0) by Anonymous Coward on Friday December 01 2017, @02:31PM (1 child)
"The only problem is, the video shows nothing of the sort. What’s more, when pushed for details, the creators of the video are now claiming the same thing."
https://hackaday.com/2017/11/29/peer-review-in-the-age-of-viral-video/ [hackaday.com]
(Score: 1) by ElizabethGreene on Friday December 01 2017, @04:13PM
OP is about the MIT FastFFF extrusion system, a high volume small orifice extrusion system. Hackaday's peer review was of the UMich magic acceleration and look-ahead algorithm.
The latter felt like hokum, the former an actual advance in the state of the art.