LMD: A new, less wasteful metal 3D printing technique | Hacker News Reader

LMD: A new, less wasteful metal 3D printing technique

core77.com

144 pointssurprisetalk29 comments

I'ts neat, but it's not that new and it has pretty limited use cases due to the poor surface quality, which albeit has improved from earlier wire-feed systems.

Norsk Titanium has been doing wire-feed 3D printing for many years, and is even delivering FAA certified parts made using a similar wire-feed deposition system and plasma arc as the power source. Looks like the laser gives you a higher resolution surface finish, but all those parts are going to still need machining on most of their critical surfaces for real world use anyway. You cannot tolerate a surface finish like that on a fatigue critical part. You also can't do dye penetrant or mag particle inspection on a surface like that without getting all kinds of spurious indications. Once you have the part clamped and indexed in the CNC to machine the bores and mounting features you might as well skim the whole thing.

https://www.linkedin.com/posts/norsk-titanium-components-as_...

https://www.theverge.com/2017/4/11/15256008/3d-printed-titan...

The carrot is that you get near-forging strength levels without having to buy a very expensive, very very long-lead closed-die titanium forging, or having buy an pretty expensive rectangular forged rectangular block of titanium and machining 90% of it away.

BTW that 3D printed lobed exhaust noise supressor is cute and all but that thing would fall to bits in hours if installed on a real jet aircraft.

One of the interesting use-cases in the video is machining the part while it's being printed. This lets you machine inside surfaces and do a few things that would usually be tricky to do. Thinking of inside-machined almost closed manifolds. I think inserts and multi-material also open up some interesting design possibilities.

It's neat, but there aren't answers to many important questions:

- How fast is this process? All videos are sped up.

- Is the resulting material close to isotropic, or is the layer to layer bond weaker than the longitudinal bond. That's the usual weak point of 3D printing. Heating up the previously laid down area with lasers is a good idea, because it gets you away from trying to weld a hot thing to a cold thing. That never gets a really good bond.

- How much laser power does this take? They say "small" lasers, but don't give the power level. Probably over 100 watts each on stainless steel.

The interviewer doesn't seem to know enough about metalworking to answer the right questions.

A modest sized machine for this would be useful. Would have liked to have had one in the TechShop days.

An even better question is what are the metal's granular properties like? Most metal prints are fairly isotropic, but their strength compared to a machined sample is significantly worse.

You can probably do annealing / hardening / tempering of the printed part to improve its mechanical characteristics.

the integration kit specs say 1000w 9 x 450nm diode lasers.

I can't tell if that's 9x1000w diode lasers, or 9*111W diode lasers.

Given the 2kw-5kw average power need they quote, i'm guessing it's the latter.

I highly doubt it's 9kW, that would require some crazy cooling.

The lasers are at the other end of the fiber optics feeding the print head.

It takes about a 1KW laser to cut stainless steel sheet. A 100W laser is about right for plastics and wood. I suspect each laser is about 1KW, but they are not all turned on at the same time.

That's an expensive system.

Sure, but then i also see: "Cooling:Water-cooled deposition head. Chiller lncluded"

(spelling error in original :P)

the deposition head is where the lasers are.

In general, fully automated systems do not require persistent human labor, and thus only the cost of the machine/consumables/space/power constrain fabrication capacity.

Generally these systems are likely slower than mature subtractive CNC Mills.

"Is the resulting material close to isotropic"

Some processes hit above 98% density, but for hobby level machines it is rarely above 90% (3 cubic inches of 316L a day on a 120v 1kW max outlet.) The oxide inclusions may be an issue in some materials, but it depends on your use-case and process.

Best of luck =3

Well if the machine is slow enough it may not make enough products throughout its lifetime to pay back for its investment cost.

A key difference with additive manufacturing is it allows internal geometry/manifolds that are difficult or impossible using traditional fabrication methods.

Metal is always expensive, but hollow parts are not as weak as one would expect. =3

It makes zero sense to do this for metal but then have to remachine the result for precision or finish. Their "reckoning that it's cheaper" seems highly inconsistent with the cost of most billets and the speed of most VMC/HMC/etc

They do offer an integration kit. https://meltio3d.com/metal-3d-printers/meltio-engine-integra...

But the print times listed here are totally unimpressive. Take the turbine blade.

6 hours 16 minutes for that is nutso.

Here's one being machined: https://www.youtube.com/shorts/6OFvkDfIvUw

This is common speed. See https://www.youtube.com/shorts/oCqTEWknC40 for another example.

The end result of the printed turbine blade is worse than a roughing bit. Which, as you can see here:

https://www.youtube.com/shorts/4qjHWeznTKc

Can do it much faster.

If you want mirror finish, it's a bit slower. Still not 6 hours.

See https://www.youtube.com/watch?v=r5yCRXcIGSs

That's the best case - the other models timings are even worse.

No commercial machine shop should could afford these timings - they'd go out of business instantly.

Someone making their own stuff (IE in-house machine shop) maybe.

Also, the automation is hugely lacking.

The integration kit/etc looks like it would block any useful automatic loading/unloading.

This might be a useful technology someday, but as long as billets are cheap and machines are fast, this would have to be very cheap and very fast to be useful.

But it's not - it's 230k for the m600, and 150k for the m450.

You can get a fine 5 axis VMC for that pricing.

I suppose that re-machining only makes sense if you need to machine only few surfaces, e.g. the flanges of a complex pipe / manifold. I suppose that complex geometries with thin walls is where metal 3D-printing mostly shines, because such things are rather hard to make in any other way.

Also I think that if printing a whole blade + machining does not make sense (but can be a demo of sorts), printing only a small chipped / worn area onto an existing blade + machining it may make sense, if the bond is strong enough.

If this process can be used on aerospace grade aluminum lithium alloys, they can print isogrids onto any surface without throwing away 50% of the material.

This is neat, but the technology itself isn't "new": the basic idea has been around for decades, and I've seen videos from companies demonstrating this at trade shows for a while as well. This is also how Firefly Aerospace 3d-prints their rockets.

Yep, iv seen them printing that rocket and talking about how they had to add extra processing steps to account for thermal warping initially.

RepRappers originally put their early filament onto empty welding wire spools.

Now we are loading welding wire spools onto 3D printers.

Everything has come full circle.

There are plenty of use cases for this type of manufacturing.While the surface quality is rough, it will print directly onto high finished sections and then create impossible to machine or cast geometries. Unfortunately parts are likely to be heavy as well as rough, so its not going to be much use in airospace.....at least for flying parts, but for wild wierd and wonderfull jigs, to then build production parts on, it could be the cats meow.

Printing with Metal Paste Deposition:

https://news.ycombinator.com/item?id=42908858

Why does the metal move from the wire to the work? Surface tension? Gravity? Ionic attraction?

And speaking of ionic attraction, how feasible would an ionic deposition version of this be?

Like electroplating but selective.

(downvote this comment, unrelated to article)

What a horrible website. Menu-banner floating near the top, bisecting the text, and the undismissable cookie banner taking up almost 100 pixels. It was a genuine chore to read and scroll.

You should use a browser that has a "reader mode".

Seems like a more costly and complicated version of Wire Arc Additive Manufacturing (WAAM). What's the benefit of using lasers vs. simple electrical current?

Precision, I think. Much easier to confine the heat to the material you want with a laser.

According to this: https://www.mdpi.com/2411-5134/8/2/52 you get better heat control / precision at the cost of being slower. Apparently people are also working on a process combining arc and laser.

One thing I notice with physical stuff is that hundreds of tiny individual advances (each one seemingly inconsequential) seem to really add up over time. Like a 2025 FDM printer is "fundamentally the same" as a reprap from 2014 but I get much more utility from the 2025 printer.

There are hobby metal printers that may eventually be applicable to regular consumer markets.

The reason "open public development" stalled was people were tired of subsidizing cloner companies with engineering projects, and disillusioned by the complete lack of community loyalty to the original authors. i.e. people proved they also didn't want to help pay the original sunk cost just like cloners, and would still get pissy when asking the hapless for support.

This is why some have our own metal printers... and the public gets 20 year old glue dispensers on a flimsy CNC platform.

Try building your own, it is actually not as hard as people assume. lol =3

Most of the current metal printers have a wacky vendor-specific supply chain, for feedstock and also post-processing. It's not really worth it for me to buy the equipment when I need an ongoing contract for it, and it will severely drop in value / usefulness once the vendor stops supporting it.

If I build my own it's probably gonna look like a mig stinger gaffer taped to a second hand kuka ;)

I have a CNC taig and a manual mill and lathe. Like most home shops I'm stuck at 90's tech level or below.

A lot of innovation has happened in the CNC space since the 90s but not much of it seems to have trickled down :/

Nice, we've had a CNC taig mini-mill for small <1mm mill stuff around for ages, and linuxCNC is actually surprisingly reliable once setup.

Our machines internal CAM release candidate is slowly getting more features, and the slicer part likely will end up FOSS once it meets its goal.

You could also check out this great group (kits were around $8k if I recall):

https://iro3d.com/

Or build metal + ceramic composites:

https://www.rapidia.com/

YMMV, and remember safety... Ti powders can fail in dramatic ways... =3

Reply on news.ycombinator.com