How an 83-year-old inventor beat the high cost of 3D printing | Hacker News Reader

How an 83-year-old inventor beat the high cost of 3D printing

techland.time.com

125 pointstechnologizer36 comments

Great work, and if the spools continue to be $50 I definitely would want one. However I don't think the spools are $50 because they cost $50 to make, they probably cost $1 to make. It's the razors/blades or printer/ink cartridge business model. I'm sure all the 3D printing companies are making fat profits off the spools of plastics, and this will push those prices down a bit.

ABS pellets are $1000-2000 per ton ($0.45 - $0.91 per lb).

I'm surprised nobody is selling cheap spools in the $5/lb range.

EDIT:

Acutally, they do.

http://www.alibaba.com/product-gs/542698056/Best_Quality_ABS...

Speaking from experience, Chinese plastic filament is almost always out-of-spec enough to mess up prints on today's 3D printers.

Interesting ... I've been playing around with the idea of having a hopper (mounted above the 3D printer) that fed pellets to the print head and used the same heater to extrude the filament right as it was being joined to the work-piece.

The two stage process is convenient for existing printers, but wouldn't it be nicer if new printers took pellets as their raw material?

Based on the numbers for cost of pellets vs cost of filament, it seems reasonable for someone to simply integrate this and make a printer that takes pellets instead of filament.

The device shown doesn't look overly complicated, and I would guess/hope the "crucial core" parts cost even less than the competition's requirement since that was for the whole machine.

You would require a Bowden extruder to build a pellet-filament engine. There's no way a Mendel I2 or similar models could handle the inertial load on the Y and Z axis.

And the video shows rather slow extrusion speed. My printer tip velocity is about 2x faster.

Dont get me wrong: reduction of filament costs by a factor of 10 is AMAZING.

My guess is that it'd be a bit of a pain to switch between filament colors like this.

Maybe when multi-head extruders become very common.

Not really. Just run maybe 50g of new material. You'll end up with blended colors, and then solid color of new material.

If you are really deadset on perfect color, you can take it apart and detail clean it.

HOWEVER, this is also a way to make new colors not purchasable: you can mix quantities of each color in a bucket, and then put the mixture in the hopper. Out comes new color. This is how the injection molded plastics companies do their products. In the drawings/contract, exact mixture is specified.

This is where the Dum Dums Myster Flavor comes from!

http://mentalfloss.com/article/30823/what-mystery-flavor-dum...

What a fun fact, thanks for sharing! I found a video that shows the whole process: http://www.youtube.com/watch?v=xUpGTvPrCG8

I believe it's the same case with Airheads mystery flavor

For a number of reasons this isn't as useful as you might hope, many have been mentioned (the mass on the extruder head being perhaps the most killer)

That said it seems like it is fast enough that you could connect the filament coming out of it into a typical filament extruder head and use the slack available as a signal to turn on the filament maker or not.

That would mean you would always have "fresh" filament to work with, and in moist environments that can improve repeat ability since PLA especially doesn't "age" well in humid environments.

I actually have access to such a printer - but it's massive and can't really print objects smaller than a few inches. We've tried scaling the extruder down, but it seems that pellet/flake extruders just don't have the control and homogeneity required for high-resolution printing at small scales.

Here's some media coverage of our device: http://www.economist.com/news/science-and-technology/2156557...

Contests are a great way to attract a lot of interest to neglected problems.

So if you can make it at home so cheaply, why can't it be mass produced in a factory even more cheaply? It doesn't seem like shipping pellets would be much less expensive than shipping filament.

It's because the demand for plastic pellets is radically larger than for plastic filament.

Since the filament side of things has been caught in a sea of change - where sudden demand swamps supply - companies making the filament are able to start and hold prices from a high point. This kind of event is common, and it usually just takes time + competition to bring prices down.

I would like to see this technology built into existing 3d printers. I would also like to see 3d scanners built in. I envision a single device that will allow you to put in a model, have it scanned, then replicated using cheap plastic material. You could even used recycled plastic with a macro shredder.

The guys at makibox have been tinkering with this idea. I think it is on the back burner until they ship their printers. (per request of printer pre-orderers)

http://makibox.com/details/product/ramen_kit

THis is awesome - the only modification which I think would make it better would be to have the filament loop back to a spool that is attached to and turned by the same motor that is pushing the material out.

In the video - he shows the filament being spooled across the room by a device with a separate power source. Make it a singular closed little process.

This home-made filament dramatically improves the economics of 3D printing. For instance, producing 392 chess pieces in a particular color requires one kilogram (2.2 pounds) of plastic. Buy one spool of mass-produced filament, and that will cost you about $50. Buy a kilogram of pellets and make your own filament, and the cost goes down to $10. Buy 25 kilograms of pellets in bulk, and you can print the chess pieces for just $5.

-- Seems to meet the "10x better" criterion.

Nice work.

Maybe. What is the price of 25 kg of filament in bulk?

I know almost nothing about producing filament, but if I had to guess, I would say its price is high because of low demand. If demand goes up, competition will drive the price down.

It's partly low demand (relatively speaking) and partly because the machines to make it have historically been on the expensive side (i've heard something like half a million or so). Then there's the time that it takes to run a batch of filament and all the other normal business issues there. It'll definitely come down but it'll likely only approach the range that has been hit with lyman's machine. The pellets have high demand and extremely high quantity (injection moulding). That being said, even when commercial costs come down to this realm they'll still likely have more consistent and higher quality filaments (longer runs tend to even out issues in the plastic).

I'm already thinking of ways to automate this machine to "print money".

Run a fan at the end of the filament guide. Then, attach a mouse recovered wheel encoder to monitor amount of filament extruded. Then make a simple geared wheel that turns the rate equal to extrusion (with some given slack). You can even stagger it the similar way a yarn winder weaves on the coil, to allow even fill on the spool.

After the circuit monitoring this process indicates the right amount of filament on spool, it kills the system and alerts operator to finish spool prep.

I could probably make a tidy profit if I made 20 machines and undercut filament prices of everyone on the market now. And even charge a bit extra for customized filament color (mixing beadstock would be easy).

What would likely help that even more is to do it in some larger spools (say 20kg spools) so that you can do larger runs and then chop it up later. There is a possible market here to do that, but I don't know what kind of margins you could expect; I've got no idea what the power the filament extruders use. I'll know when/if I finally get my filabot from the kickstarter, it's a very very similar design to lyman's but came a bit later. The main thing I'm after for it is that it'll come with a grinder to let me take things like plastic bottles and try to print with PET plastics and other strange materials.

Yep, you can get 9800 chess pieces for a total price of $49000. (392 per kg x 25kg x $5). Still a significant investment :)

minus the 980 errors :)

$5 per 392 pieces, not $5 each. So that's 9800 pieces for $125.

oops, you're right.

You should check your math there.

They talk about the filament ending up $5/kg. So the cost is $5/kg * 25 kg = $125.

392 pieces/kg * 25 kg = 9800 pieces.

So you end up with about 306 (9800/32 pieces) chess sets for $125. That's less than a dollar a set. I don't know what you'd need with that many sets but it really does change things.

Seems to me this method is patented. How are they going to sell these kits?

Reference: http://en.wikipedia.org/wiki/Plastics_extrusion

Care to quote any section of that page that mentions patents?

"Patent", "Patented", "Pat.", "USPTO", "Proprietary", or "Protected" do not appear in the body of the article.

Also, there is this: "The first thermoplastic extrusion was in 1935 by Paul Troester in Germany." Looks like a decent prior-art claim.

This one for example: http://www.google.com/patents/US4118163

I know it won't be a problem when people use this at home, but Lulzbot is going to sell assembled versions. I hope they don't get in trouble for that.

We deal with your type fairly regularly over at hackaday and other discussion sites.

The Lyman extruder documents (including BOM and assembly and 3 view drawings) are published for free. The parts are easy to obtain.

How are you (the possible patent holder) going to stop individuals from sourcing the parts and building these machines?

The reprap machines are also, Im sure, violating multiple patents in the US. Go ahead and try to stop the international audience from furthering the science and art. I'll just use TOR and download the assembly drawings from evil foreign servers.

"You wouldn't download a car, would you?"

A whole class of patent holders are about to get napstered. Once you can simulate just about any electronic board with a RasPi for $25 and make just about any plastic part on your 3d printer, what good will a patent even be?

I'm making popcorn.

Damn straight I would.

I also do casting, and natural PLA has a nice burnoff for it. I also have access to a metal lathe and a few other pieces of heavy equipment. I can make. out of junk and scrap, pretty much anything I want. My next work will be a CNC table using a mounted dremel (printed bracket, of course).

And some of people who know what I do already have asked "can I make them guns for them?". Yeah. I could do nice guns. Dont really care too much about 'vs human' and 'vs animal' ballistsics.

I'm also working on a cheap 3d scanner to feed fodder into my system :) Im sure I breaking copyrights AND patents, and maybe a few trademarks too.

I was not talking about individuals but companies like Lulzbot who are going to sell them pre-assembled.

Reply on news.ycombinator.com