![](/static/253f0d9b/assets/icons/icon-96x96.png)
![](https://aussie.zone/api/v3/image_proxy?url=https%3A%2F%2Flemmy.world%2Fpictrs%2Fimage%2Fe4083c7c-83c3-4258-8370-708e95e309b3.png)
They definitely don’t know what they’re doing. They featured this one, which is a death trap. It has a disclaimer that it might not be safe above 120V, but it’s absolutely unsafe and a code violation in the US, where we use 120V (and are very litigious). The disclaimer says they’re trying to get it approved which implies they believe it could be and that the design is sound, but fundamentally it cannot meet code in the US for mains voltage use.
Even if the design were sound, there are material requirements, and having seen the quality of prints some people find acceptable, there’s no chance allowing random people online to print their own boxes is safe.
I think they basically run the contests and feature things based on “ooh this is neat” and “this will excite people to use 3d printers”. It’s a marketing thing, and I guess I accept it because I have low expectations of even pretty-good businesses. But if it’s illegal…someone should probably let them know.
The best I have is to be careful to minimize dependencies, and minimize when I change the number of faces an object has, but of course that’s unavoidable sometimes. I don’t buy it that all CAD tools have the same problem or that this is how real professional CAD designers would work, though.
To minimize dependencies for example, instead of drawing the sketch for pad 2 directly on a face of pad 1, I might draw it on the base plane and transform the sketch to line up with pad 1’s face. The main consequence is that I need to manually move pad 2’s sketch if I change the size/position of pad 1. It’s a tradeoff, because I’m giving up some of the benefits of parametric CAD in exchange for easier fix-up.
I agree, mapping a datum plane to a face should have the same topo naming issue as just drawing on the face, so I don’t know why the guide would suggest that. The comment below about mapping datum planes to a simplified skeleton is interesting though.
The good news is that the next release (which sounds imminent) apparently improves it quite a bit.