Ben Jackson wrote: > So I've been thinking about how to do the soldermask DRC. To recap, > I want to include a soldermask registration tolerance in the DRC parameters, > and warn if the mask is anywhere too close (risking an overlap with the pad) > or too far (risking exposing a nearby trace or adjacent plane). > > One particular problem is that a copper pad's clearance (from other > copper) is a uniform distance all the way around the pad, producing a > rounded rectangle. On the other hand, the hole in the soldermask for > the pad is simply a rectangle that's bigger in X and Y by the mask amount. > So if a pad has a 10mil clearance and a 10mil mask, there will be an arc > of exposed copper even with ideal mask placement. This is most obvious > when there is a flood of copper (with a polygon) around the pad. You > can turn on the soldermask and zoom in to see. The maximum allowable > mask would be the clearance/sqrt(2) (via a bit of trig). So a 10mil > clearance would allow a max ~7mil mask. If you had a 3mil registration > tolerance for the mask, it would only allow a mask of either 3 or 4mil. > If it had a 4mil max error there would be no possible mask (given a 10mil > clearance) that would be guaranteed not to possibly hit the pad or expose > the adjacent copper. > > Do people think the mask *should* have square corners, or should they > be rounded to conform to the clerance, and leave more leeway? >
It seems like the rounded mask would be better. -Dan _______________________________________________ geda-user mailing list [email protected] http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
