6 ms·

3D printing titanium: Learning to learn from success

73 points by pencerw 11y ago

pencerw 11y ago
Hey all, just posting an update to some older work that I did (and that got posted here by someone else) a few months ago. tl;dr: It's totally possible to print titanium bicycle parts, but the business model needs a bit more work still. For reference, this is the thread from my last blog post on the subject: https://news.ycombinator.com/item?id=9216474 https://news.ycombinator.com/item?id=9216474 Cheers, Spencer Wright
- Gravityloss 11y ago
  Why is it so expensive, 500 per part? You said that you can lower the cost 25-40% by doubling layer thickness. From that I infer that the printing time is a large part of the total cost?
  - pencerw 11y ago
    I go into the pricing in a bit more detail in the older post (http://pencerw.com/feed/2015/3/15/3d-printing-titanium-and-the-bin-of-broken-dreams http://pencerw.com/feed/2015/3/15/3d-printing-titanium-and-t...) but the gist is that 1) the powder is expensive, and 2) the machines are really expensive, and 3) the people who know how to run them are rare and highly prized. Essentially, reducing printing time (by doubling layer thickness) means that you don't need to amortize as much of the cost of the machine during the build.
- Gravityloss 11y ago
  Also, why threads in the part and not a ~~bolt~~ nut insert? To save weight?
  - pencerw 11y ago
    Do you mean "why threads and not a nut insert?" If so, yeah - for sure, that would be easier. Partly I wanted the challenge, though. There are a lot of applications where nut inserts aren't optimal, and in general they'd be heavier too. In the coming months I'm going to focus on a part that gets glued to a post, which obviously will skirt the issue altogether.
- nitrogen 11y ago
  I look forward to seeing how you apply gongkai to manufacturing. Knowledge sharing has facilitated the explosive growth in web entrepreneurship; surely something similar is possible in other industries, like 3D printing metal.
  - pencerw 11y ago
    Yeah, that's something I've spent a bunch of time thinking about. I've gotten a lot of emails from people around the industry saying they'd like to be able to share more process data, but I'm not sure what the best approach is. If anyone out there has any thoughts, I'd LOVE to collaborate.
    - nitrogen 11y ago
      I am not at all connected with the manufacturing industry, but I wonder if a group like ASME could help.
- Obi_Juan_Kenobi 11y ago
  My layman impression of the part is that it looks like a traditional piece. Is there anything about it that couldn't efficiently be done with traditional machining? If not, is the next step to leverage the additive process to produce a unique design? Meshes, hallows, etc.?
  - taneq 11y ago
    This was my thought - it's an awesome read and 3D printing definitely has its place, but its strength is generally in one-off or very-short-run custom items rather than building mass produced parts.
    - pencerw 11y ago
      Those are definitely good use cases, yup. There are a few things I can do to increase the value of this part (or similar ones), and in general the high end cycling market tends to require a lot of 1) lightweight structures, which additive is good for, and 2) customization. It's also worth thinking about this in context: Even without offering any customization, the cost to buy my part in batches of ~40 actually works out pretty well (caveat: this is a decidedly high end bike part). With forged, machined, and composite parts, that's an extremely low purchase quantity; in most cases, you'd never come close to amortizing the tooling & setup time. That said, I'm definitely interested in exploring other products, and new business models for how to sell them. This part is just the first step in that process :)
  - pencerw 11y ago
    Totally fair question. This part would be really hard to manufacture traditionally - there's a hollow in the middle of the part that couldn't be machined, so you'd have to machine it in two parts and then weld it. That said, this design is very much NOT optimal, and there's a lot I can do to tune it for the process. I'm working on that more in the coming months; first I needed to 1) prove that the process could produce a functional part, and 2) get access to more advanced design software (topology optimization + lattice structure generation). Stay tuned!
chrisbennet 11y ago
The page I get is "about.blank"
- dang 11y ago
  The URL was mangled but I think we fixed it.
  - chrisbennet 11y ago
    Works now. Great reading!
jlebrech 11y ago
does that mean you could 3d print joining parts and connect them all with carbon tubes? post bike kits to people for them to assemble.
- pencerw 11y ago
  Yes! Once I get the seatpost (this part: http://pencerw.com/feed?tag=seatpost http://pencerw.com/feed?tag=seatpost) done, I want to start working on a full bike frame, made to order and assembled individually.
  - jlebrech 11y ago
    I think connecting (carbon) tubes could be a way to go.
    - pencerw 11y ago
      Yeah, for sure. Carbon fiber tubing is great when you need a long structural element - no need to print that, in my opinion.
- PhasmaFelis 11y ago
  I can't find it right now, but at some point someone had made a kit of bike-frame joints, with the tubes to be supplied by the builder from locally available materials. PVC pipe and hand-whittled wood with bark still on were demonstrated. It was really cool.
- xavieralexandre 11y ago
  VRZ 1. a tack bike frame with 3d printed lugs https://vimeo.com/34293503 https://vimeo.com/34293503
  - pencerw 11y ago
    Yeah! That one's pretty nice, though personally I probably would have skipped the gold coating on the lugs.
jlebrech 11y ago
Would it be possible to 3d print a light mesh and finish off the process with anodization? that way the part would be 3d print more rapidly.
- pencerw 11y ago
  Anodization is definitely an option - you can do some pretty wild colors, too. And a mesh is in the works :)
  - jlebrech 11y ago
    how much faster is 3d printing a mesh rather than a solid, you could 3d print a lot more, then chuck it in a bath to gain thickness and strength.
    - pencerw 11y ago
      Build time scales pretty closely with the volume of material you're printing, but meshes often require less material to achieve the same mechanical properties. As far as post processing to improve mechanical properties goes: I'm not aware of anything like that, unfortunately. Titanium has a pretty high melting point and a very good strength/weight ratio, so I'm not sure what you could add to improve it.
PhasmaFelis 11y ago
I wonder if it's possible to 3D-print titanium mail (aka "chain mail") at a competitive price? Titanium is great stuff for high-end mail, strong, lightweight, shiny, and corrosion-free, but it's even more insanely labor-intensive than ordinary aluminum or steel mail, especially if the links are welded or riveted for strength; prices appear to start at well over $2000 for the equivalent of a T-shirt. I think you'd want to print it in a sort of collapsed/"wadded-up" form, to reduced the print size and to insure you can get enough separation between the links that they don't bind together. You'd need some pretty fancy algorithms to translate the client's measurements into a wad of printed mail that will be the right shape/size when spread out, but once you got that down you could just plug in the numbers and watch it go. If you could get the price down to something decent, I think you'd get a lot of business from SCAdians/fantasy fans/etc.
- JabavuAdams 11y ago
  I've thought a bit about automated mail-manufacturing. Additive manufacturing is the new hotness, but seems to bring only problems w.r.t mail, not benefits. Wire is easy to bend and spot-welding or riveting it although perhaps intimidating seems like it would be easier to get right than all the problems additive manufacturing introduces. Drop my a line (email in my profile) if you want to talk more about mail manufacturing. It's on my project bucket list.
- jackcarter 11y ago
  Here's a video about 3D-printing a dress, which uses the "wadding-up" solution you mention: https://www.youtube.com/watch?v=wdRswasftfI https://www.youtube.com/watch?v=wdRswasftfI
- pencerw 11y ago
  As others on this thread hinted at, I'm not sure that would be economical in titanium. But Nervous System (designers of that dress - website is http://n-e-r-v-o-u-s.com/ http://n-e-r-v-o-u-s.com/) are awesome at that stuff when it comes to polymers!
- VLM 11y ago
  Supermarket butchers have mail gloves. I worked in a store as a kid and watched coworkers put on mail gloves then cut meat. There's probably more supermarket butchers than SCA members and the wear and tear must be tremendous compared to SCA meetings. If the labor cost is lower that would be a win, but also the weight of the gloves must be pretty high and tiring over time (which is why illegally they mostly spent their time on the shelf instead of on hands, although they would occasionally wear them). So if 3d printing could be lighter...
JabavuAdams 11y ago
Thanks for this -- what a treasure trove of information! I love being alive in 2015.
wanderfowl 11y ago
Knowing nothing about 3D printing, this was among the more awesome things I've read this week. Thanks so much for the detailed write-up, particularly focusing on the unknown-unknowns.
- pencerw 11y ago
  :) Thanks!

73 points by pencerw 11y ago

Hey all, just posting an update to some older work that I did (and that got posted here by someone else) a few months ago. tl;dr: It's totally possible to print titanium bicycle parts, but the business model needs a bit more work still. For reference, this is the thread from my last blog post on the subject: https://news.ycombinator.com/item?id=9216474 https://news.ycombinator.com/item?id=9216474 Cheers, Spencer Wright

Gravityloss 11y ago
Why is it so expensive, 500 per part? You said that you can lower the cost 25-40% by doubling layer thickness. From that I infer that the printing time is a large part of the total cost?
- pencerw 11y ago
  I go into the pricing in a bit more detail in the older post (http://pencerw.com/feed/2015/3/15/3d-printing-titanium-and-the-bin-of-broken-dreams http://pencerw.com/feed/2015/3/15/3d-printing-titanium-and-t...) but the gist is that 1) the powder is expensive, and 2) the machines are really expensive, and 3) the people who know how to run them are rare and highly prized. Essentially, reducing printing time (by doubling layer thickness) means that you don't need to amortize as much of the cost of the machine during the build.
Gravityloss 11y ago
Also, why threads in the part and not a ~~bolt~~ nut insert? To save weight?
- pencerw 11y ago
  Do you mean "why threads and not a nut insert?" If so, yeah - for sure, that would be easier. Partly I wanted the challenge, though. There are a lot of applications where nut inserts aren't optimal, and in general they'd be heavier too. In the coming months I'm going to focus on a part that gets glued to a post, which obviously will skirt the issue altogether.
nitrogen 11y ago
I look forward to seeing how you apply gongkai to manufacturing. Knowledge sharing has facilitated the explosive growth in web entrepreneurship; surely something similar is possible in other industries, like 3D printing metal.
- pencerw 11y ago
  Yeah, that's something I've spent a bunch of time thinking about. I've gotten a lot of emails from people around the industry saying they'd like to be able to share more process data, but I'm not sure what the best approach is. If anyone out there has any thoughts, I'd LOVE to collaborate.
  - nitrogen 11y ago
    I am not at all connected with the manufacturing industry, but I wonder if a group like ASME could help.
Obi_Juan_Kenobi 11y ago
My layman impression of the part is that it looks like a traditional piece. Is there anything about it that couldn't efficiently be done with traditional machining? If not, is the next step to leverage the additive process to produce a unique design? Meshes, hallows, etc.?
- taneq 11y ago
  This was my thought - it's an awesome read and 3D printing definitely has its place, but its strength is generally in one-off or very-short-run custom items rather than building mass produced parts.
  - pencerw 11y ago
    Those are definitely good use cases, yup. There are a few things I can do to increase the value of this part (or similar ones), and in general the high end cycling market tends to require a lot of 1) lightweight structures, which additive is good for, and 2) customization. It's also worth thinking about this in context: Even without offering any customization, the cost to buy my part in batches of ~40 actually works out pretty well (caveat: this is a decidedly high end bike part). With forged, machined, and composite parts, that's an extremely low purchase quantity; in most cases, you'd never come close to amortizing the tooling & setup time. That said, I'm definitely interested in exploring other products, and new business models for how to sell them. This part is just the first step in that process :)
- pencerw 11y ago
  Totally fair question. This part would be really hard to manufacture traditionally - there's a hollow in the middle of the part that couldn't be machined, so you'd have to machine it in two parts and then weld it. That said, this design is very much NOT optimal, and there's a lot I can do to tune it for the process. I'm working on that more in the coming months; first I needed to 1) prove that the process could produce a functional part, and 2) get access to more advanced design software (topology optimization + lattice structure generation). Stay tuned!

chrisbennet 11y ago

The page I get is "about.blank"

dang 11y ago
The URL was mangled but I think we fixed it.
- chrisbennet 11y ago
  Works now. Great reading!

jlebrech 11y ago

does that mean you could 3d print joining parts and connect them all with carbon tubes? post bike kits to people for them to assemble.

pencerw 11y ago
Yes! Once I get the seatpost (this part: http://pencerw.com/feed?tag=seatpost http://pencerw.com/feed?tag=seatpost) done, I want to start working on a full bike frame, made to order and assembled individually.
- jlebrech 11y ago
  I think connecting (carbon) tubes could be a way to go.
  - pencerw 11y ago
    Yeah, for sure. Carbon fiber tubing is great when you need a long structural element - no need to print that, in my opinion.
PhasmaFelis 11y ago
I can't find it right now, but at some point someone had made a kit of bike-frame joints, with the tubes to be supplied by the builder from locally available materials. PVC pipe and hand-whittled wood with bark still on were demonstrated. It was really cool.
xavieralexandre 11y ago
VRZ 1. a tack bike frame with 3d printed lugs https://vimeo.com/34293503 https://vimeo.com/34293503
- pencerw 11y ago
  Yeah! That one's pretty nice, though personally I probably would have skipped the gold coating on the lugs.

jlebrech 11y ago

Would it be possible to 3d print a light mesh and finish off the process with anodization? that way the part would be 3d print more rapidly.

pencerw 11y ago
Anodization is definitely an option - you can do some pretty wild colors, too. And a mesh is in the works :)
- jlebrech 11y ago
  how much faster is 3d printing a mesh rather than a solid, you could 3d print a lot more, then chuck it in a bath to gain thickness and strength.
  - pencerw 11y ago
    Build time scales pretty closely with the volume of material you're printing, but meshes often require less material to achieve the same mechanical properties. As far as post processing to improve mechanical properties goes: I'm not aware of anything like that, unfortunately. Titanium has a pretty high melting point and a very good strength/weight ratio, so I'm not sure what you could add to improve it.

PhasmaFelis 11y ago

I wonder if it's possible to 3D-print titanium mail (aka "chain mail") at a competitive price? Titanium is great stuff for high-end mail, strong, lightweight, shiny, and corrosion-free, but it's even more insanely labor-intensive than ordinary aluminum or steel mail, especially if the links are welded or riveted for strength; prices appear to start at well over $2000 for the equivalent of a T-shirt. I think you'd want to print it in a sort of collapsed/"wadded-up" form, to reduced the print size and to insure you can get enough separation between the links that they don't bind together. You'd need some pretty fancy algorithms to translate the client's measurements into a wad of printed mail that will be the right shape/size when spread out, but once you got that down you could just plug in the numbers and watch it go. If you could get the price down to something decent, I think you'd get a lot of business from SCAdians/fantasy fans/etc.

JabavuAdams 11y ago
I've thought a bit about automated mail-manufacturing. Additive manufacturing is the new hotness, but seems to bring only problems w.r.t mail, not benefits. Wire is easy to bend and spot-welding or riveting it although perhaps intimidating seems like it would be easier to get right than all the problems additive manufacturing introduces. Drop my a line (email in my profile) if you want to talk more about mail manufacturing. It's on my project bucket list.
jackcarter 11y ago
Here's a video about 3D-printing a dress, which uses the "wadding-up" solution you mention: https://www.youtube.com/watch?v=wdRswasftfI https://www.youtube.com/watch?v=wdRswasftfI
pencerw 11y ago
As others on this thread hinted at, I'm not sure that would be economical in titanium. But Nervous System (designers of that dress - website is http://n-e-r-v-o-u-s.com/ http://n-e-r-v-o-u-s.com/) are awesome at that stuff when it comes to polymers!
VLM 11y ago
Supermarket butchers have mail gloves. I worked in a store as a kid and watched coworkers put on mail gloves then cut meat. There's probably more supermarket butchers than SCA members and the wear and tear must be tremendous compared to SCA meetings. If the labor cost is lower that would be a win, but also the weight of the gloves must be pretty high and tiring over time (which is why illegally they mostly spent their time on the shelf instead of on hands, although they would occasionally wear them). So if 3d printing could be lighter...

JabavuAdams 11y ago

Thanks for this -- what a treasure trove of information! I love being alive in 2015.

wanderfowl 11y ago

Knowing nothing about 3D printing, this was among the more awesome things I've read this week. Thanks so much for the detailed write-up, particularly focusing on the unknown-unknowns.

pencerw 11y ago
:) Thanks!