The DRC correctly identifies missing connections when pads don’t connect to the zone, but it doesn’t show me where they are.
For example, it highlights the overall zone (GND) and some random connected pad on C304, but the problem is below, with the two ground pads just below the dialog box. If I highlight the zone in the DRC, the whole board is illuminated.
This isn’t helpful. Shouldn’t the disconnected pads be zoomed to? In this case, it wasn’t hard to find, but sometimes it’s very difficult to figure out where the real problem is.
This isn’t helpful. … In this case, it wasn’t hard to find, but sometimes it’s very difficult to figure out where the real problem is.
You are right, the zoom-factor+position for the “unconnected items” drc-error is not the best.
Shouldn’t the disconnected pads be zoomed to?
This sounds trivial, but:
you have two separated GND-islands (the big overall GND-zone and the little unconnected island in the middle of the picture).
every island already has some pads connected
so every island alone seems to be ok
how should the kicad drc-algorithm decide which of these islands is the unconnected part? Judge by already filled area? Judge by already connected pads?
And it gets more difficult if you imagine a case with multiple isolated floating gnd-zone islands.
My idea would be to zoom full-screen onto the remaining unrouted ratsnest-line which somewere must exist on the board.
In this case there are two things that are unconnected: the zone that contains C304, and the island that’s not connected. All it has to do is zoom to the island that’s not connected when we click, instead of the whole zone (since the other item is the C304 zone). I don’t see why that’s complicated, or why it matters which is the “main” one. It should be handled like having broken traces.
It also doesn’t show ratsnest lines. Pads shouldn’t be considered connected unless they’re connected to the zone, AND that zone isn’t an “island”. That shouldn’t be too different than the algorithm that determines unconnected islands.
