Anelok: revised case (WIP)

Werner Almesberger werner at almesberger.net
Tue Jul 14 00:52:43 UTC 2015


Ron K. Jeffries wrote:
> Your (clever) case design + electronic design may be useful for more than
> just the anelok password safe.

Thanks ! And welcome to the marvels of open source ! ;-)

By the way, most fun challenges in the case design are still to come.

First, there are lots of sharp edges that should be rounded. This is
a) to make manufacturing easier (many processes dislike zero radii),
b) for visual appeal, c) to make it nicer to touch, and d) to better
distribute mechanical stress. There are (at least) two ways to do
this:

a) design the geometry such that things have sufficiently round edges.
   Nice in theory, a lot of analytical geometry in practice.

b) let FreeCAD fillet all the sharp edges. Like this:

   - here we have the current lanyard hole, with its sharp edge:
     http://downloads.qi-hardware.com/people/werner/anelok/tmp/case-interim/fillet-before.png

   - I select every piece of the edge (selected lines turn green,
     you have to mentally filter out the green PCB in the background):
     http://downloads.qi-hardware.com/people/werner/anelok/tmp/case-interim/fillet-selected.png

   - I tell FreeCAD to work its magic. Et voila:
     http://downloads.qi-hardware.com/people/werner/anelok/tmp/case-interim/fillet-after.png

     (It's now opaque because filleting creates a new object, which
     doesn't inherit the transparency and color settings of the
     original.)

   Looks nice and easy. However, many edges are the result of rather
   complicated sequences of unions and cuts, so I'm not sure how to
   describe those things in terms of the objects I work with in the
   script. And just doing filleting all over the place manually each
   time something gets edited in the model doesn't sound too appealing
   either. For further study.

Second, injection molding is not only about getting the plastic into
the mold but in no small part about getting the mold out of the plastic
afterwards.

For this, the contact surface should be small. This means that walls
should not be parallel to the movement(s) when taking the mold apart.
Instead, they should be slightly angled.

Another issue is that the mold actually must have an escape route.

That may involve cutting the mold into multiple pieces that have to
be removed in a certain order. Not sure what the limits for such a
multi-part mold are, but I'm certain that a crazily complex mold
won't be cheap.

Needless to say, the current design is happy to ignore any such
considerations.

- Werner



More information about the discussion mailing list


interactive