Hacker News Reader

Comment by eternityforest | Hacker News Reader

Doing CAD design is really interesting. A lot of stuff is just 2.5D, extrusions of 2D sketches sitting on other 2D sketches.

Then you accidentally make something truly 3D by intersecting things and realize you have no idea what you're looking at, couldn't imagine it if you closed your eyes, couldn't replicate it if you had a picture of the result and didn't know the 2D inputs that made it... and then you realize there are probably people out there who can see that entire design in their head.

To me it's like unicycling on a tightrope or skateboarding or realistic oil painting or playing piano well. I have no real concept or reference point for what that experience must be like.

View on HN

Interesting. With a very strong reference for the experience of skateboarding, reading your sentence made me think about how hard it is to explain! I suppose when it’s going well, it feels like body/mind flow, when it’s going poorly it feels like physics :)

That kind of thing is super amazing! I've secretly suspected for a while that people who are able to skateboard or do similar things must live in 4 or 5 dimensions.

I can't figure out any kind of way someone could do that with anything resembling what I understand "thought" to be like, so I assume you must be able to process entire sequences and their alternate possibilities simultaneously?

Just drawing a 3D slanted plane to match the front of my printer made my head hurt: https://www.printables.com/model/526981-ender-3-s1-quad-z-br...

Coming from 3d sculpting to CAD was...eye opening. It really makes you realize how much of design constrained by the CAD tools, which are mostly constrained by manufacturability.

If true 3d printing* ever gets cheap, it'll be interesting to see how much form will be able follow function, rather than manufacturing cost.

* true 3d, as in overhang are allowed. Something like a cheap FDM is more 2.5d, since overhangs aren't allowed.

You can print with supports to print any shape.

Yea, but this requires soluble supports, without joined parts/assembly.

Overhangs up to 45 degrees work fine, and usually steeper. Bridges also work, most of the time, although I am having trouble with them at the moment.

They're not quite unconstrained, and the layer adhesion strength makes some geometries not strong, but they're much more than 2.5D.

Well, there's SLS which is tolerant of overhangs and other exotic features. Formlabs has a machine in the 5 digit price range, and I wouldn't be surprised to see competitors come out with a 4-digit one in the near future.

You could probably build a 4-axis FDM machine for less than that too, which might be better for average users(No powder mess)

The axis count isn't the limiting factor. 4 axis will still require soluble supports to achieve what SLS can (captured parts).

With powder bed printers overhangs and detached captured parts aren't a problem. Try making your own from an old inkjet printer.

Where could someone interested in doing this learn more?

CAD is hard! I think about this a lot as I'm doing a makerspace at an elementary school and adults think it needs a 3D printer, but what are kids going to print on it?

I think for me, the prerequisites for mastering CAD were 1) the practice I got visualizing 3D shapes so I could translate them into unambiguous mechanical drawings on paper (I swear, I'm not that old but my college was behind and we were the last class to actually do mechanical drawings with a pencil), and 2) having a procedural thought process from coding so that I could sequence the CAD operations to get where I want.

Try Blockscad3d.com

It's a block language where you program the object rather than manipulate shapes. It works well for my brain. Ymmv

Grasshopper for rhino is awesome too. It's actually really relevant for this topic. It's the only syste I know that integrates so well into a precise cad design system and also allows you to do complex stuff easily. I love it. Drawing a stellarator precisely is probably not that hard on it... Will see if I can throw something out quickly today.

For me CAD is pretty much the opposite, I rely heavily on the apps, I spent about a week trying to learn to draw before deciding I didn't really want to spend months or years on it.

I'm not a mechanical engineer, I mostly only do incidental CAD and hobby level work, so it's not really essential to have the deep understanding of space that real MEs need.

I often don't know what sequence of operations I'll need until I actually open the app. Generally it's more of an "Oh I need a mounting hole, lets look around on the screen and see where one could go" thing, a lot of the thinking is in the app rather than in the mind.

Of course you can't make nicely parametric things without a lot more thought so I will often wind up having to redo things that aren't one offs...

Reply on news.ycombinator.com