What's new
What's new

Desktop metal printing gets cheaper

TonyStark

Aluminum
Joined
Dec 23, 2017
Though this was interesting - BASF released a material for 3d printing metal parts on practically any desktop printer. They partnered with Ultimaker so essentially for $5k you can have the same setup as the metal printers from Markforged or Desktop Metal ($200k systems). Certainly seems like a “disruption” to both of those manufacturers.

BASF to commercialise its Ultrafuse 316L metal filament with industry partners

The only part missing is the sintering furnace, and BASF says they will be offering this as a service with industry providers. Does any know what would be required to have a small scale sintering furnace in house?
 
Presumably localized heat and fuming would be the biggest issue, with power supply and space on the minor side. Not sure of the baking/sintering process of this new compound, or whether any pressure is needed along with the thermal requirements. Time for some reading...
 
let me know when a part cost of material and labor and machine costs and maintenance costs is less than $10./lb total part cost per 1000 parts per year average
.
just saying add up all the costs or "total cost to manufacture" per part average of 1000's of parts per year. its like tooling cost when cnc machining somebody wanting to save 5 minutes machining but spend over $100. more on tooling per part
.
total cost to manufacture per part long term average data needed. obviously if any machine starts needing repairs, calibration, upgrades and or high tooling costs that effects things too. that is some machines are very expensive to maintain. or parts made end up being very expensive
.
the cool factor ends quick when you get the bill
 
let me know when a part cost of material and labor and machine costs and maintenance costs is less than $10./lb total part cost per 1000 parts per year average
.
just saying add up all the costs or "total cost to manufacture" per part average of 1000's of parts per year. its like tooling cost when cnc machining somebody wanting to save 5 minutes machining but spend over $100. more on tooling per part
.
total cost to manufacture per part long term average data needed. obviously if any machine starts needing repairs, calibration, upgrades and or high tooling costs that effects things too. that is some machines are very expensive to maintain. or parts made end up being very expensive
.
the cool factor ends quick when you get the bill

You forgot to mention how much money you make on overtime.
 
let me know when a part cost of material and labor and machine costs and maintenance costs is less than $10./lb total part cost per 1000 parts per year average
.
just saying add up all the costs or "total cost to manufacture" per part average of 1000's of parts per year. its like tooling cost when cnc machining somebody wanting to save 5 minutes machining but spend over $100. more on tooling per part
.
total cost to manufacture per part long term average data needed. obviously if any machine starts needing repairs, calibration, upgrades and or high tooling costs that effects things too. that is some machines are very expensive to maintain. or parts made end up being very expensive
.
the cool factor ends quick when you get the bill

I completely agree, I don’t really see this as replacing machining anytime soon. Basic things like precision bearing bores are not possible with additive (yet). To me this replaces casting and reduces machining time, or allow the creation of parts not possible with subtractive technologies.

Where the math becomes interesting is when you have a lot of these printers running in tandem to increase throughout dramatically. If a printer cost $250k and take 2 hours to print a part, that doesn’t make a lot of sense. But if you can buy 50 printers at $5k/piece, that time per part drops to just a few min per part. Plus you get a lot of extra redundancy in your process (ie if one machine goes down the whole process doesn’t stop).
 
Hi TonyStark:
You wrote:"But if you can buy 50 printers at $5k/piece, that time per part drops to just a few min per part"

But if you calculate the energy budget for those 50 machines you are still way way behind on that measure compared to making your parts some other way.
That's the elephant in the room that almost no-one talks about.
Cheers

Marcus
Implant Mechanix • Design & Innovation > HOME
Vancouver Wire EDM -- Wire EDM Machining


 
unfortunately 3d printing doesn't scale well. with FDM, you get a time penalty when the volume you're printing goes up. if you double the size of a part, you cube the volume, all of which is being extruded thru a tiny nozzle.

3D printing is kind of like a microwave - a complementary tool really good at a few things, but not about to revolutionize kitchens anytime soon.
 
Hi TonyStark:
You wrote:"But if you can buy 50 printers at $5k/piece, that time per part drops to just a few min per part"

But if you calculate the energy budget for those 50 machines you are still way way behind on that measure compared to making your parts some other way.
That's the elephant in the room that almost no-one talks about.
Cheers

Marcus
Implant Mechanix • Design & Innovation > HOME
Vancouver Wire EDM -- Wire EDM Machining



My FDM printer(similar to ones that can use the BASF material) use about 200w. So 10kw for 50. That’s about the same as an average size VMC. We aren’t talking about high powered laser sls machines with heated / atmosphere controlled build chambers.
 
....
The only part missing is the sintering furnace, and BASF says they will be offering this as a service with industry providers. Does any know what would be required to have a small scale sintering furnace in house?

Most sinter furnaces are not small. It does not scale nicely and cycles are certainly not fast.
There are lab sized units but expensive to operate even if piggybacked off the bigger plant atmosphere and vacuum systems.
Sintering is a very fussy process with a ton of control variables. I've seen installing new HVAC in the room make a total mess of things.
Bob
 
3D printing will be useful for the things that machining centers are not good at. Like things that are difficult to hold, things that would have to be cast and then machined and can be done in one shot.

I was talking to a guy about T shirts, and he was saying that basically no one actually silkscreens anymore. They do more computer generated stuff. It is a much slower technology but so much more flexible than silkscreen they make more money and I think it takes less talented workforce. Similar to offset press vs laser printing, 30 years ago everything 3 color was run on a printing press, which is blindingly fast compared to color laser, and frequently better quality, but man, who wants to print 200 sheets when you can do it right on your desktop.
 
there are some things that are not easy to machine but really easy to print, like globoid worm gears. Do you need a gearbox that's 11.5:1? Try to source that out. BTW, hypocycloid reducers have been patented only few years back, which is exciting because I'd not expect in this day and age to see new mechanical mechanisms and there are quiet few more with more complex mathematical shapes ... easy to print, nightmare to machine even on CNC.

As technology becomes available, we'll see more interesting mechanical shapes becoming common since the tech to make them is becoming cheaper as well.

One dynamics about parts and cost is the inventory. Just-in-time concept had dramatic effect on business in the past. I can see that some plastic parts can be custom made here in US on as needed basis, especially for parts that are not a commodity, you don't need to stock a whole lot and keep money tied up in inventory. That's a plus.

In terms of availability of sintering ovens. There is a huge community of metal casters who melt everything from aluminium to cast iron and even steel. Lots of tech too, from plain oil burners to electrics and induction. This is all hobby level casting from scrap parts. I don't know what's needed for sintering ovens, but my guess is, it's probably not too far out of reach.
 
3D printing is great for prototypes, one-offs, and research (the fictionalized recreation of a dinosaur vocal chamber in a movie mirrors what some scientists have done) but it's very poorly suited to production.

I read a story not too long ago where a U.S. Navy facility 3D printed prototype titanium helmet ducts to add rebreather capability to the standard deep diving suit. Apparently helium has gotten so expensive they feel the need to recycle the breathing gas mix rather than just vent it as in the standard open circuit rig.

IMO if you occasionally need to do 3D prints in metal you would probably be better off creating a plastic pattern that can be sent to someone who does lost wax casting. As I understand it jewelers use plastic patterns all the time, customizing and adding detail before imbedding/burnout/casting.
 
IMO if you occasionally need to do 3D prints in metal you would probably be better off creating a plastic pattern that can be sent to someone who does lost wax casting. As I understand it jewelers use plastic patterns all the time, customizing and adding detail before imbedding/burnout/casting.

there are filaments designed for lost wax (filament) already, the filaments that are soluble in water and apparently easy to burn out as well. Regular filaments can also be burned out with care.

but the same could be said about plastic casting. If you need to make 20+ plastic parts, it's could be faster to mold one in silicone and cast with liquid plastic mixes.

The problem becomes again, casting need special design considerations in mind, i.e. how the metal will flow, how it would shrink, system of sprues and chillers. This is especially true for bigger parts and precision parts. Unless you have casting experience, you'll run into a whole lot of surprises. This is also true for casting plastic parts as well. You need to figure out how to feed and vent the mold so no air is trapped inside. Again, experience matters here.
 
The company I work for makes tooling for the foundry industry. A couple years ago we got a job from Siemens asking for a 2 or 3 up run of a small casting. It was a simple armature with a bore or something along those lines that was getting sent for machining afterwards anyways.

We wound up sending the job to be metal 3D printed at the additive manufacturing lab at the college I was still attending at the time. I spoke with the technicians that worked on the job and learned about the issues they ran into with stresses in the parts and dimensional issues that resulted from it. They wound up having to add ribs and other features to improve rigidity during solidification of the part(much like practices used in the foundry industry), while it was being printed. This lead to post print cleanup beyond the expected surface finish touch-ups.

As far as one off goes it's not bad but it's also not great. a 3D wax printer for investment casting use is way better for one off small parts, in both surface finish and part stresses/metallurgical control and dimensional accuracy. For larger one off parts, a 3d printed sand mold is often the way to go from machines like an ExOne or voxeljet printer. Better control over solidifaction stresses, more flexibility and control over metallurgy and the cost is extremely reasonable. We'll do one up castings from 3D printed molds with a turn around time of 3-4 weeks, reasonable surface finish on the castings and the price per mold is very competitive.

3D printed metal will have its place, but it's not replacing machining anytime soon, if ever, and I don't see it fully replacing production foundries anytime soon either.
 
there are filaments designed for lost wax (filament) already, the filaments that are soluble in water and apparently easy to burn out as well. Regular filaments can also be burned out with care.

but the same could be said about plastic casting. If you need to make 20+ plastic parts, it's could be faster to mold one in silicone and cast with liquid plastic mixes.

The problem becomes again, casting need special design considerations in mind, i.e. how the metal will flow, how it would shrink, system of sprues and chillers. This is especially true for bigger parts and precision parts. Unless you have casting experience, you'll run into a whole lot of surprises. This is also true for casting plastic parts as well. You need to figure out how to feed and vent the mold so no air is trapped inside. Again, experience matters here.

I'm not arguing against what you said, it just reminded me of something.

An engineer I worked with some time ago designed a forming die for an inhouse compnent we needed for a prototype. Easier and faster than machine form solid. I was the person who did the die shoe(s) work, assembled, then actually stamped the few pieces we needed. Anywho, one look at the die and I said "that isn't going to work" ... it did work, and damn near flawlessly. :eek:
Moral of the story, sometimes things work you don't expect to. ;)
 
unfortunately 3d printing doesn't scale well. with FDM, you get a time penalty when the volume you're printing goes up. if you double the size of a part, you cube the volume, all of which is being extruded thru a tiny nozzle.

3D printing is kind of like a microwave - a complementary tool really good at a few things, but not about to revolutionize kitchens anytime soon.

What about UVM/DLP/MSLA type of printing? My cost in time is the same for 1 piece as it would be for 100 pieces (granted that 100pcs would have to fit on the bed).
 
What about UVM/DLP/MSLA type of printing? My cost in time is the same for 1 piece as it would be for 100 pieces (granted that 100pcs would have to fit on the bed).

While that is true that doesn't account for post-processing time for support removal, cleaning, etc. That's a lot more time than most people think. You are also dealing with a very long print time and increased risk of print failure. It's pretty common to come back to your machine with the nozzle dragging around a nice big birds nest of melted filament. Markforged has a much lower failure rate - mostly from scanning the bed, compentation and auto leveling - but you are going to pay for that. We evaluated their printer as a potential source for some of our parts but with the long print times the amortization of the machine just couldn't make sense.

The furnace will need to be pressurized otherwise you are going to end up with a very porous part.

3D printing is getting much better - just a few years ago it was good for making Christmas Tree ornaments - but it's still got a ways to go.
 
What about UVM/DLP/MSLA type of printing? My cost in time is the same for 1 piece as it would be for 100 pieces (granted that 100pcs would have to fit on the bed).

since you're developing a 2d sheet vs a 1 dimensional line at a time, you get a speed boost for increasing part count using those techniques. However, DLP etc are more limited in a number of ways: high-maintenance goo, leaching problems, comparative structural weakness...
 
While that is true that doesn't account for post-processing time for support removal, cleaning, etc. That's a lot more time than most people think. You are also dealing with a very long print time and increased risk of print failure. It's pretty common to come back to your machine with the nozzle dragging around a nice big birds nest of melted filament. Markforged has a much lower failure rate - mostly from scanning the bed, compentation and auto leveling - but you are going to pay for that. We evaluated their printer as a potential source for some of our parts but with the long print times the amortization of the machine just couldn't make sense.

The furnace will need to be pressurized otherwise you are going to end up with a very porous part.

3D printing is getting much better - just a few years ago it was good for making Christmas Tree ornaments - but it's still got a ways to go.

Post-processing time would be the same, no? It wouldn't matter if you clean 1 part at a time 100 times or 100 parts in row, when all is said and done you still need to clean 100 individual parts. While I agree that there would be more "relative risk" with printing more parts at a single time, this would just come down to someone having confidence in their capabilities and their machine.

Also, the methods I mentioned previously (UV Masking, DLP, etc.) have nothing to do with a filament nozzle as they are resin-based, so don't worry about that...
 








 
Back
Top