Part strength, building orientation, and such...

Discussion in 'Guides, Mods, and Upgrades' started by UlrichKliegis, Feb 4, 2016.

  1. UlrichKliegis

    UlrichKliegis Well-Known Member

    Joined:
    Sep 5, 2015
    Messages:
    187
    Likes Received:
    47
    Hi all,
    just stumbled over a very useful remark from Ryan, the creator of the MPCNC. He has designed and makes really heavy duty objects that serve as components in CNC milling machines that allow you to convert not only all kinds of wood but also aluminium to chippings and powder. 3D printed couplers and brackets are essential parts of those machines.
    Here is what he says about the strength of 3D-printed objects:
     
  2. theTroll527

    theTroll527 Well-Known Member

    Joined:
    Dec 24, 2015
    Messages:
    100
    Likes Received:
    17
    That is true about a print in general. You can strengthen things by putting parts together so their z layers run perpendicular to each other.
     
  3. UlrichKliegis

    UlrichKliegis Well-Known Member

    Joined:
    Sep 5, 2015
    Messages:
    187
    Likes Received:
    47
    So that's visible in the construction of the BigBox's printhead. The so called cable bracket is built perpendicular to the motor bracket and hotend carriage. The layers make a compound not unlike plywood but still different.
    Important to keep in mind if one turns an object in the slicer for an alleged better printability.

    Conclusion: Keep the strongest loads parallel to the building plane.
     
  4. Mike Kelly

    Mike Kelly Volunteer

    Joined:
    Mar 7, 2014
    Messages:
    444
    Likes Received:
    89
    Definitely true. The bond strength between layers is the weakest point on a print. You can try adding fillets to move that stress point, but because each layer acts like a shear point you end up pulling apart at the end of the fillet anyways.

    That's why thicker layers are a better idea if you need strong part in all directions. You'll get less layer bonds so less chance of failure. Also the thicker surface area for each strand is greater thus a stronger adherence bond.

    or design your parts to be in multiple pieces making use of the X-Y strength as needed
     
  5. UlrichKliegis

    UlrichKliegis Well-Known Member

    Joined:
    Sep 5, 2015
    Messages:
    187
    Likes Received:
    47
    Or time to start thinking about non-cartesian object generation strategies. Robots exist, I heard... ;)
     
  6. Miasmictruth

    Miasmictruth Well-Known Member

    Joined:
    Sep 4, 2015
    Messages:
    758
    Likes Received:
    112
    I would assume the parts should hold up to compression on the z right? Just not tension.
     
  7. theTroll527

    theTroll527 Well-Known Member

    Joined:
    Dec 24, 2015
    Messages:
    100
    Likes Received:
    17
    Yes that would be true. But you wouldn't want compression on point loads on the z. You do not want to flex the part on the z either. That could delaminate the layers.
     
  8. Miasmictruth

    Miasmictruth Well-Known Member

    Joined:
    Sep 4, 2015
    Messages:
    758
    Likes Received:
    112
    Ahh gottcha I didn't think about it like that. If the flex is very small do you think it would still cause delamination?
     
  9. theTroll527

    theTroll527 Well-Known Member

    Joined:
    Dec 24, 2015
    Messages:
    100
    Likes Received:
    17
    Most likely not. You need enough been that layers curve is enough for the layers to want to slip past each other. For small parts, not very likely for long thin parts much more likely.
     
  10. Sanjay

    Sanjay Administrator
    Staff Member

    Joined:
    Feb 27, 2014
    Messages:
    125
    Likes Received:
    118
    Which material you use makes a huge difference in how strength varies across the axes of printing.

    Nylon is near as dammit just as strong in Z as it is in XY. Copolyesters like Edge are extremely close to this too. PLA can be quite decent, but isn't really comparable to Nylon or Edge. ABS, even when printed in the best of circumstances doesn't do much better than 50% of the strength in Z as in XY.

    Interlayer adhesion is a bit of a black art, and there isn't much good science out there that explains or explores the bonding mechanisms between layers, particularly as those mechanisms seem to be completely different between materials.

    I'm pretty damn sure the reason Nylon is nearly isotropically strong is due to it being so polar and regular, which ends up being able to form hydrogen hydrogen bonds from one chain to another. The reasons for Copolyesters bonding so well is completely unknown to me. We've managed to make huge boosts in how well ABS bonds to itself with the as-yet-unnamed 'Material X' that the hot parts of BigBox print heads are printed in, but honestly the exact mechanics of why this drastic increase has happened eludes me.

    There's much material science yet to be done in extrusion type 3D printing, but there's no reason to settle for crappy interlayer adhesion if you need better strength - even with today's materials.

    Now I want to buy an instron......
     
    mike01hu and Mike Kelly like this.
  11. UlrichKliegis

    UlrichKliegis Well-Known Member

    Joined:
    Sep 5, 2015
    Messages:
    187
    Likes Received:
    47
    Thanks for this know-how-injection. I re-printed the broken PLA part here in PETG and added (subtracted) a matrix of 1.5mm ID cylinders, extending over the whole Z-range from bottom to top of the object, open at both ends.

    This added omnidirectionally pretty coherent perimeter tubes to the (fast honeycomb) 70% infill, and, at least empirically, staying below rupture loads, that added to the strength of the part. 6 tubes in a cross section of 12 x 25 mm^2, three perimeter tracks, according to S3D each 0.35 mm wide, leading to 1mm thick walls around the tubes, do a lot.

    Did you ever try feeding two totally different materials into one output stream using your cyclops for example?
     

Share This Page