kicad-developers team mailing list archive
Mailing list archive
Re: Question about odd pad shapes
On Thu, Jul 05, 2012 at 01:13:40PM -0700, Roger Marin wrote:
> I'm one of those who silently read the Kicad developer discussions.
> Please forgive my jumping in here...
The list is public and feedback is welcome (and useful)
> Yes, if needed, you *can* create a filled copper zone in a module.
Technically yes, since you can use trapezoidal pads to tile whatever area you
want (since they can degenerate to axis aligned triangles, I think it could be
proved that such a decomposition is always possible).
But drawing a polygon in the module and saying 'I want that filled with copper'
no, that's not possible. It *could* be done with either a new kind of module
primitive or really enhanced pad support (look in gerber aperture macros, the
primitive 4 allows you 4000 vertices in a *pad* :D); there is also the
conventional gerber fill which is already used for regular zones.
> circumvent some of the more frustrating design enforcement features
> that are a part of the Kicad design philosophy.
Uhmm... part is design philosophy, part is missing features:P Look at gencad,
in that old format you can't even *express* trapezoidal pads... (in kicad we
don't have padstacks, so I guess the score is even)
> Attached is a "thrown together" .MOD that started life as a .BRD,
> was stuffed through BRD2EMP, imported into Pcbnew's module editor,
> then saved as a .MOD . Please note that text cleanup or pin
> renumbering was not done in the module editor phase.
Going *too* crazy with strange pads can give headaches with board testers.
Also it seems that your converter raster fills an area which is a big no for
gerber these days:P
As always it depends on your fabricator software/engineer (I've seen boards
rejected because they used AM and their importer was flaky and didn't recognize
> BTW: Today, after creating the test EMP, I found that the module
> editor from the latest main bzr release throws a biuParse() exception
> while loading the EMP. No such problem exists with the stable release
Is the 'loading' kicad nano enabled? I filed a bug with EMP handling a while
ago, maybe they're related (it was a really weird scaling issue, but who know).