Lessons learned for your CAD-designs

Skip to first unread message

Marcus Wolschon

Jan 29, 2011, 3:38:06 PM1/29/11
to thingiverse
I'm currently working on the gimble for my new steadycam.

What I do is to document the shortcommings of each version and how I
solved them.
(at least I try to.)

I'd really like to see more of this to spread the knowledge and enable
more complicated things.

lessons learned so far:

* on a heated bed the fist layer melts a bit and widens. This reduces
the diameter of my vertical holes.
countermeassure: an 0.25mm tapper

* for holes <6mm diameter+0.5mm in diameter is fine for screws.
* +1mm for things that should fit without much friction.

* Holes for 608 bearings should have +1mm in diameter
* they also absolutely should have a 1mm tapper to guide the bearing
into position. (else it will jump out or get stuck at odd angles when
you try to force it into a snug fitting hole)
* have a smaller hole inside the back-wall of the hole, so the inner
part of the bearing has no friction with the back walll and rotates

* supplying a a bill of materials, assembly-instructions and
technical drawing of your thing with the important meassurements
displayed is a very nice thing for *finished* designs. It helps people
build your thing.
* Also don't forget a photo of the real, assembled thing once you are
finished yourself.

* a dumb-bell makes an excellent blunt force to insert a 608 bearing.

* rounded or at least tappered edges make any object much more hand-
friendly. A 45° tapper in the vertical is absolutely no problem.

* 27mm is about the maximum you should do vor non-teardrop vertical,
round holes. The overhang works well but you already need the machine
properly calibrated. Machines doing less well will already have
problems with this contruct.

* ALWAYS run the STL your CAD-program generates through netfabb or
something similar to
** align it to the base-plate for printing (I've seen a fair share of
non-aligned STL files and am sick of finding out about the issue after
lengthy Skeinforge-runs only to go back, correct them and wait for
skeinforge again) and
** fix any mesh-errors. (even high priced parametric, volumentric CAD-
systems do sometimes create minor issues in the STL.)

Marcus Wolschon

Jan 30, 2011, 5:57:44 AM1/30/11
to thingiverse
Another lesson:

608 bearings don't react well to being beaten out of a too tight
fitting socket.
They look more sturdy then they are:

Marcus Wolschon

Jan 31, 2011, 2:09:58 AM1/31/11
to thingiverse, Open Manufacturing

Hello everyone,

I finished the description of
(Only actual test-shots made with the steadycam missing.)

This is how I like my designs to be documented.
(Not restricted to thingiverse.)
* constructions history (to learn from the mistakes of others)
* parts list
* print-time and how much material it uses
* wiki-page with my notes on the project, it's status and links
* assembly instructions
* photos of the assembly
* photo of the thing
* constructions drawing of the thing and it's parts including
important measurements and text-annotations
* pdf with movable 3d-views of the thing and it's parts, qr-code,
license, links and construction-drawings of thing and parts
Kind of a role model against all these things with no description at
where you need to run software just to find out if the size of the
thing and if it's metric or imperial.

What do you think about this style of documenting a complex thing?
Sadly my license does not allow me to do the animated explosion-views
for assembly in 3d with links to the parts list.
(I played with that in a 30 day evaluation and have seen it in action
in the BfB-assembly-manuals. Coooool. ;) )

Reply all
Reply to author
0 new messages