5 ms·

Well, the issue with 3D printing as I see it is 3D modeling. Anyone can fire-up a word processor and produce interesting looking documents and flyers. A lot l

Well, the issue with 3D printing as I see it is 3D modeling. Anyone can fire-up a word processor and produce interesting looking documents and flyers. A lot less people can fire-up a paint program and produce interesting images. A lot less people than that can fire-up a 2D CAD program and produce 2D mechanical or architectural drawings. And yet a lot less people than that can run a 3D modeling program and product much beyond trivial trinkets. Beyond "Look! I printed this tiny Statue of Liberty from a file I got online" and "Check out the letter A I made!" lies real mechanical design. This is where mechanical things start to become useful. At that level you now need far better understanding of how to design multi-component mechanical assemblies, materials, etc. Anyone can do a little pyramid in Sketchup. Not everyone can design a full robotic hand with differential drive and force compliance. I imagine most folks buy 3D printers, print a few things they can get online and that's the end of it. A very few are inspired to learn some more and do some basic modeling. I can only see a very, very few go all out, learn advanced modeling and mechanical design and start to extract value out of their 3D printers.

maturz 10y ago
It's pretty easy to create some useful things with https://www.tinkercad.com/ https://www.tinkercad.com/ I had no trouble using it with no previous 3d modeling experience.
badsock 10y ago
This is one area where I think VR will produce a solid win. IMO the steepest parts of the modelling learning curve are the result of the complexity required to manipulate a 3D world using 2D interfaces. With VR, both the input and output interfaces have the same dimensionality as what you're building - I think that'll have a big impact on how easy it is to get into, and consequently how much content is created.
- ffwacom 10y ago
  The interface to cad programs is trivial to learn, mechanical design isn't. It's like saying the hardest part about programming is using the IDE.
  - grogenaut 10y ago
    To ellaborate, when you're using cad you have to give thought to how the part will be produced as well. There are a lot of shapes that 3d can't build, there are minimum angles, all sorts of restrictions. And there's the nitty gritty of actually fitting real shapes together in the real world, having the right flange sizes for holding screws, nuts, etc. And then you get into wear, balance, etc.
  - badsock 10y ago
    You're right. When I wrote the above, what I had in mind was sculpture (what I've mostly see consumer 3D printers making these days - representational things like D&D set pieces, or cosplay props). The interface of programs like Maya, for making very complex 3D shapes, I wouldn't call trivial. But in hindsight the context of the discussion is more about industrial design of functional consumer goods, which I completely agree is a highly skilled trade, and while learning the CAD program is still a lot of work, that's one of the least of the requirements.
- semi-extrinsic 10y ago
  I don't think so. I do a bit of AutoCAD at work, mainly for pressurized vessels (i.e. pretty simple stuff, just has to follow the regulations). 95% of the time, you're working with a 2D sketch in either the x, y or z plane and extruding from that. The other 4.9% of the time, you make some other plane based on those extrusions and do 2D sketches in that plane. 0.1% is reserved for 3D splines.
- rebootthesystem 10y ago
  You are confusing UI with mechanical design. It's like confusing learning Python with producing solid software products. OK, let me illustrate the point with a question: You have two parts that attach to each other. The parts are joined using a set of six 5 mm bolts with washers and locking nuts. The hole pattern is circular and has a radius of 3 inches. What are the minimum diameter and dimensional tolerances you should specify in order to guarantee mating across multiple parts and across thermal variations? OK, same two pars. Now they mate using no screws. They use a one-time-only locking tab and catch system along the inner periphery. What's the geometry and what are the tolerances? Interestingly enough, none of the above talks about how to use the UI of some 3D program to solve the problems.
mbrock 10y ago
For random DIY stuff you don't always need advanced modeling. If I had a 3D printer right now I'd want to make a circular gasket with an outer diameter of 78 mm. That's the simplest modeling imaginable but it would be useful for me. It's like how a power saw is useful even if you don't know advanced furniture construction...
- sslalready 10y ago
  OpenSCAD is great for these things. A gasket would be the difference() between two cylinder()s for example. Check this out: https://www.google.de/amp/s/cubehero.com/2013/11/19/know-only-10-things-to-be-dangerous-in-openscad/amp/?client=safari https://www.google.de/amp/s/cubehero.com/2013/11/19/know-onl...
  - grogenaut 10y ago
    It's even easier in sketchup, inventor, or solidworks. And I don't have to feel like I'm wrestling with an HP calculator when I'm using it.
    - ascorbic 10y ago
      Except Sketchup can't handle exotic shapes like the "circle".
      - grogenaut 10y ago
        Yeah that's why I take advantage of the student discount that taking a class at the local technical college offers and get full solid works for $170/year.
        jjeaff 10y ago
        But then of course, that license is only valid while you are a student. So after the class, you might as well have just pirated it since either way, you are breaking the license.
phaed 10y ago
I got up and running with Solidworks in about an hour after watching a couple of Youtube videos. It's point and click, it's not rocket science.
- rebootthesystem 10y ago
  That's pretty funny. I've been using Solidworks for somewhere around 12 to 14 years. I also run Siemens NX. Clarification: Not saying it's impossible to learn. All I am saying is that a very, very, very, very small percentage of 3D printer buyers develop the chops. Don't confuse an intuitive UI with mechanical design. It's the difference between the intuitive UI of a word processor and writing a novel. The same could be said of, say, Python. Easy to learn from a few videos on YouTube. There's a vast difference between that and actually producing good, efficient, fast, sophisticated and accurate software. I've been programming, designing hardware and electronics for over twenty years. There is no way anyone can obtain the equivalent chops and experience from a bunch of videos on YouTube. Maybe they can learn to use Altium Designer, Solidworks, a compiler or two and Xilinx's FPGA toolset. That does not mean they can actually design products at all. That takes years or learning and dedication and is hard. Learning to drive Solidworks from a few YouTube videos does not make that person a mechanical engineer or a mechanical designer beyond the basics. Example, unrelated to 3D printing: Design a differential, cable-driven wrist actuator with 3 degrees of freedom. Use a combination of 3D printing, machined parts and off-the-shelf hardware. Use SW to run load analysis on the parts and assemblies. Calculate allowable dimensional tolerances. Produce mechanical drawings with annotations for class of fit of various components as required. Now make the entire assembly (all components) parametrically driven by a reference sketch. Automatically produce a new set of prints and revisions if any dimension is changed. In other words, if you change one gear all other components in the actuator redefine themselves based on the new gear and so does the assembly. If the new gear is large enough to accommodate four mounting holes rather than two, it should implement those changes automatically. I still hold that creating anything truly useful for 3D printing requires background that most people don't have and are not interested obtaining through hard, self-directed work.
  - WillPostForFood 10y ago
RandyRanderson 10y ago
I completely agree. To extend this, even if you can 3d model something there is no reliable way to convert that into something printable. Some of the basics aren't even implemented, such as: * You have to model explicitly the model walls. That is you can't just say "using X material, print this model with whatever minimum thickness is necessary". * Even if the model visually looks correct, you may have to remove hidden surfaces or non-continuous (or connected) surfaces. * Even if an online service like shapeways verifies a model it may not print correctly. And it's not always clear how to fix it. * Solid models are extremely expensive and so you have to create a shell. Shells have several issues like print orientation, holes for the excess material to drain out, thickness of solid, overhangs, etc edges. * Typically shapeways, etc either don't support certain formats or don't support them well. This means you have to convert your model into whatever format they support. This can introduce more errors. * Automatically re-sizing - strangely enough if you can print a say 10cm tall model on a printer you can't necessarily print a 15cm version, even if it fits in the print volume. Given all these things, even if you have a decent model, there is a large subsequent step in turning that model into something that can print on a specific printer and material. It's likely that printer hardware makers have to create some software that takes popular native model formats like .stl, .blender, etc and convert it reliably into a printable model for that specific printer/material.
- svarrall 10y ago
  Check out RenderFab[0], this is exactly what they have solved. [0] www.renderfab.com
  - RandyRanderson 10y ago
    I hope they succeed!
- rebootthesystem 10y ago
  Correct. My point was to say that 3D printers end-up collecting dust because what's needed to make them useful is beyond the realm of what most people --not all-- are willing to devote time, effort and money to learn. Without learning these things the 3D printers become useless very quickly.
campbellmorgan 10y ago
I think this is the key point. I have years of experience modelling for 3d animation and have been trying to print a complex object for the last few weeks. Modelling an object is one thing, but modelling a visually interesting object that doesn't require support structures everywhere is a real skill that takes some mastering
- WillPostForFood 10y ago
  Why not just use support structures?
  - TeMPOraL 10y ago
    Because if your object is tricky enough, you'll have to spend a lot of time removing them, risking damage to the object. Dissolvable supports are a thing, but I haven't seen many people with double-headed extruders in their printers.
ht_th 10y ago
> Anyone can fire-up a word processor and produce interesting looking documents and flyers. Actually, even this is harder than it seems. Having worked as a teacher, I saw many teacher create their own materials. They created handouts, pamphlets, worksheets, exams, short instructional materials, and some even whole text books. Compared to profession stuff from publishers, these self-made artifacts were often awful. The artifacts worked, of course, but typesetting, layout, illustrations, printing, and editing were almost always sub-par. The most common example of this was trying to fit a worksheet or exam on one sheet A4, printed on both sides. As a result, my colleagues and I chose to use smaller fonts, smaller margins, decreased image size, less whitespace, and so on. Over-all, we ignored the intended readers (also known as our students), and focused on secondary matters such as ease of distribution. On the other hand, where these instructional artifacts did shine was with respect to their practical applicability and usefulness because they were custom made (by the teacher) to fit a very specific situation (their classroom). Which is why readily available computers, software, and printers are such a powerful and awesome tool for a teacher.
Something1234 10y ago
SketchUp was amazing when it came to doing really simple 3d stuff. I wish it would work on Linux, and was still free.

maturz 10y ago

It's pretty easy to create some useful things with https://www.tinkercad.com/ https://www.tinkercad.com/ I had no trouble using it with no previous 3d modeling experience.

badsock 10y ago

This is one area where I think VR will produce a solid win. IMO the steepest parts of the modelling learning curve are the result of the complexity required to manipulate a 3D world using 2D interfaces. With VR, both the input and output interfaces have the same dimensionality as what you're building - I think that'll have a big impact on how easy it is to get into, and consequently how much content is created.

ffwacom 10y ago
The interface to cad programs is trivial to learn, mechanical design isn't. It's like saying the hardest part about programming is using the IDE.
- grogenaut 10y ago
  To ellaborate, when you're using cad you have to give thought to how the part will be produced as well. There are a lot of shapes that 3d can't build, there are minimum angles, all sorts of restrictions. And there's the nitty gritty of actually fitting real shapes together in the real world, having the right flange sizes for holding screws, nuts, etc. And then you get into wear, balance, etc.
- badsock 10y ago
  You're right. When I wrote the above, what I had in mind was sculpture (what I've mostly see consumer 3D printers making these days - representational things like D&D set pieces, or cosplay props). The interface of programs like Maya, for making very complex 3D shapes, I wouldn't call trivial. But in hindsight the context of the discussion is more about industrial design of functional consumer goods, which I completely agree is a highly skilled trade, and while learning the CAD program is still a lot of work, that's one of the least of the requirements.
semi-extrinsic 10y ago
I don't think so. I do a bit of AutoCAD at work, mainly for pressurized vessels (i.e. pretty simple stuff, just has to follow the regulations). 95% of the time, you're working with a 2D sketch in either the x, y or z plane and extruding from that. The other 4.9% of the time, you make some other plane based on those extrusions and do 2D sketches in that plane. 0.1% is reserved for 3D splines.
rebootthesystem 10y ago
You are confusing UI with mechanical design. It's like confusing learning Python with producing solid software products. OK, let me illustrate the point with a question: You have two parts that attach to each other. The parts are joined using a set of six 5 mm bolts with washers and locking nuts. The hole pattern is circular and has a radius of 3 inches. What are the minimum diameter and dimensional tolerances you should specify in order to guarantee mating across multiple parts and across thermal variations? OK, same two pars. Now they mate using no screws. They use a one-time-only locking tab and catch system along the inner periphery. What's the geometry and what are the tolerances? Interestingly enough, none of the above talks about how to use the UI of some 3D program to solve the problems.

mbrock 10y ago

For random DIY stuff you don't always need advanced modeling. If I had a 3D printer right now I'd want to make a circular gasket with an outer diameter of 78 mm. That's the simplest modeling imaginable but it would be useful for me. It's like how a power saw is useful even if you don't know advanced furniture construction...

sslalready 10y ago
OpenSCAD is great for these things. A gasket would be the difference() between two cylinder()s for example. Check this out: https://www.google.de/amp/s/cubehero.com/2013/11/19/know-only-10-things-to-be-dangerous-in-openscad/amp/?client=safari https://www.google.de/amp/s/cubehero.com/2013/11/19/know-onl...
- grogenaut 10y ago
  It's even easier in sketchup, inventor, or solidworks. And I don't have to feel like I'm wrestling with an HP calculator when I'm using it.
  - ascorbic 10y ago
    Except Sketchup can't handle exotic shapes like the "circle".
    - grogenaut 10y ago
      Yeah that's why I take advantage of the student discount that taking a class at the local technical college offers and get full solid works for $170/year.
      - jjeaff 10y ago
        But then of course, that license is only valid while you are a student. So after the class, you might as well have just pirated it since either way, you are breaking the license.

phaed 10y ago

I got up and running with Solidworks in about an hour after watching a couple of Youtube videos. It's point and click, it's not rocket science.

rebootthesystem 10y ago
That's pretty funny. I've been using Solidworks for somewhere around 12 to 14 years. I also run Siemens NX. Clarification: Not saying it's impossible to learn. All I am saying is that a very, very, very, very small percentage of 3D printer buyers develop the chops. Don't confuse an intuitive UI with mechanical design. It's the difference between the intuitive UI of a word processor and writing a novel. The same could be said of, say, Python. Easy to learn from a few videos on YouTube. There's a vast difference between that and actually producing good, efficient, fast, sophisticated and accurate software. I've been programming, designing hardware and electronics for over twenty years. There is no way anyone can obtain the equivalent chops and experience from a bunch of videos on YouTube. Maybe they can learn to use Altium Designer, Solidworks, a compiler or two and Xilinx's FPGA toolset. That does not mean they can actually design products at all. That takes years or learning and dedication and is hard. Learning to drive Solidworks from a few YouTube videos does not make that person a mechanical engineer or a mechanical designer beyond the basics. Example, unrelated to 3D printing: Design a differential, cable-driven wrist actuator with 3 degrees of freedom. Use a combination of 3D printing, machined parts and off-the-shelf hardware. Use SW to run load analysis on the parts and assemblies. Calculate allowable dimensional tolerances. Produce mechanical drawings with annotations for class of fit of various components as required. Now make the entire assembly (all components) parametrically driven by a reference sketch. Automatically produce a new set of prints and revisions if any dimension is changed. In other words, if you change one gear all other components in the actuator redefine themselves based on the new gear and so does the assembly. If the new gear is large enough to accommodate four mounting holes rather than two, it should implement those changes automatically. I still hold that creating anything truly useful for 3D printing requires background that most people don't have and are not interested obtaining through hard, self-directed work.
- WillPostForFood 10y ago

RandyRanderson 10y ago

I completely agree. To extend this, even if you can 3d model something there is no reliable way to convert that into something printable. Some of the basics aren't even implemented, such as: * You have to model explicitly the model walls. That is you can't just say "using X material, print this model with whatever minimum thickness is necessary". * Even if the model visually looks correct, you may have to remove hidden surfaces or non-continuous (or connected) surfaces. * Even if an online service like shapeways verifies a model it may not print correctly. And it's not always clear how to fix it. * Solid models are extremely expensive and so you have to create a shell. Shells have several issues like print orientation, holes for the excess material to drain out, thickness of solid, overhangs, etc edges. * Typically shapeways, etc either don't support certain formats or don't support them well. This means you have to convert your model into whatever format they support. This can introduce more errors. * Automatically re-sizing - strangely enough if you can print a say 10cm tall model on a printer you can't necessarily print a 15cm version, even if it fits in the print volume. Given all these things, even if you have a decent model, there is a large subsequent step in turning that model into something that can print on a specific printer and material. It's likely that printer hardware makers have to create some software that takes popular native model formats like .stl, .blender, etc and convert it reliably into a printable model for that specific printer/material.

svarrall 10y ago
Check out RenderFab[0], this is exactly what they have solved. [0] www.renderfab.com
- RandyRanderson 10y ago
  I hope they succeed!
rebootthesystem 10y ago
Correct. My point was to say that 3D printers end-up collecting dust because what's needed to make them useful is beyond the realm of what most people --not all-- are willing to devote time, effort and money to learn. Without learning these things the 3D printers become useless very quickly.

campbellmorgan 10y ago

I think this is the key point. I have years of experience modelling for 3d animation and have been trying to print a complex object for the last few weeks. Modelling an object is one thing, but modelling a visually interesting object that doesn't require support structures everywhere is a real skill that takes some mastering

WillPostForFood 10y ago
Why not just use support structures?
- TeMPOraL 10y ago
  Because if your object is tricky enough, you'll have to spend a lot of time removing them, risking damage to the object. Dissolvable supports are a thing, but I haven't seen many people with double-headed extruders in their printers.

ht_th 10y ago

> Anyone can fire-up a word processor and produce interesting looking documents and flyers. Actually, even this is harder than it seems. Having worked as a teacher, I saw many teacher create their own materials. They created handouts, pamphlets, worksheets, exams, short instructional materials, and some even whole text books. Compared to profession stuff from publishers, these self-made artifacts were often awful. The artifacts worked, of course, but typesetting, layout, illustrations, printing, and editing were almost always sub-par. The most common example of this was trying to fit a worksheet or exam on one sheet A4, printed on both sides. As a result, my colleagues and I chose to use smaller fonts, smaller margins, decreased image size, less whitespace, and so on. Over-all, we ignored the intended readers (also known as our students), and focused on secondary matters such as ease of distribution. On the other hand, where these instructional artifacts did shine was with respect to their practical applicability and usefulness because they were custom made (by the teacher) to fit a very specific situation (their classroom). Which is why readily available computers, software, and printers are such a powerful and awesome tool for a teacher.

Something1234 10y ago

SketchUp was amazing when it came to doing really simple 3d stuff. I wish it would work on Linux, and was still free.