6 ms·

Gotta admit the combo of 3D printed parts, high pressure and hot water makes me a little nervous but that looks like a very cool device

wwykeham 3y ago
There's an important extra ingredient that you didn't specify: FEA. He did a proper finite element analysis, which is the only way that I would be willing to be in the same room as a 3D printed pressure vessel. Temperature was interesting though. I assume it partially because the hot water was mostly in contact with metal rather than plastic. I don't know what he printed it from but ABS's melting point[1] is only just above water boiling point. We've got a UV cured 3D printer at work (projet) which has some much better temperature properties. [1] Technically "glass transition temperature" if you want be all materials-science about it
- beAbU 3y ago
  I'm quite certain they conducted FEA on the Titan DSV's hull. And look how that turned out.
  - speed_spread 3y ago
    They then also proceeded to _repeatedly_ exceed the design parameters (depth) that FEA had been performed with. You can't beat stupid.
  - plagiarist 3y ago
    I, too, thought of the Titan when I saw that photo of all the leaks under only espresso PSI.
- criddell 3y ago
  Modeling with FEA is great but I wonder if he also did a fatigue analysis. It might not blow up the first time, but will it blow up three years later after the parts have been pressurized and thermally cycled 1000 times?
  - MarkMarine 3y ago
    It’ll probably fail at some point, but the failure mode is probably a crack and a leak, especially on the threaded connections. That said, making this out of steel would be much safer. Seems like the person that wrote this is pretty uniquely frugal, and doesn’t value their time, so I’m going to wager that will give them a happy chance to redesign the part. They could have done some contract design work in the time they spent on this and bought a nice espresso machine plus had some money to stick in the bank, but I suppose that isn’t the point.
- mastax 3y ago
  Did the FEA account for the weakness from the layer lines or did it pretend it was an injection molded part (I'm assuming the latter).

wwykeham 3y ago

There's an important extra ingredient that you didn't specify: FEA. He did a proper finite element analysis, which is the only way that I would be willing to be in the same room as a 3D printed pressure vessel. Temperature was interesting though. I assume it partially because the hot water was mostly in contact with metal rather than plastic. I don't know what he printed it from but ABS's melting point[1] is only just above water boiling point. We've got a UV cured 3D printer at work (projet) which has some much better temperature properties. [1] Technically "glass transition temperature" if you want be all materials-science about it

beAbU 3y ago
I'm quite certain they conducted FEA on the Titan DSV's hull. And look how that turned out.
- speed_spread 3y ago
  They then also proceeded to _repeatedly_ exceed the design parameters (depth) that FEA had been performed with. You can't beat stupid.
- plagiarist 3y ago
  I, too, thought of the Titan when I saw that photo of all the leaks under only espresso PSI.
criddell 3y ago
Modeling with FEA is great but I wonder if he also did a fatigue analysis. It might not blow up the first time, but will it blow up three years later after the parts have been pressurized and thermally cycled 1000 times?
- MarkMarine 3y ago
  It’ll probably fail at some point, but the failure mode is probably a crack and a leak, especially on the threaded connections. That said, making this out of steel would be much safer. Seems like the person that wrote this is pretty uniquely frugal, and doesn’t value their time, so I’m going to wager that will give them a happy chance to redesign the part. They could have done some contract design work in the time they spent on this and bought a nice espresso machine plus had some money to stick in the bank, but I suppose that isn’t the point.
mastax 3y ago
Did the FEA account for the weakness from the layer lines or did it pretend it was an injection molded part (I'm assuming the latter).