It seems to me that there’s a fine connection between the two pads, so why is the ratsnest airwire still there?
I found a previous post from August 2016 that said this was a known bug, although that seemed related to copper pours/polygons, and this board is just in the beginning stages of being routed, wires only. Of course, I’m new to KiCad, so this may be something I’m missing?
Here’s the screen shot and board file, if you’d like to try to reproduce the bug(?):
If I were to guess, you didn’t stop on Q1, pin2 and made a track start/end there, but just assumed KiCAD would interpret copper tracks of the same net going over/touching/intersecting with a pad of the same net would make them connected for the DRC… this alas was a wrong assumption.
Can you do me a favor (if that is a wrong assumption by me) and take another screenshot of the same area with Tracks in outline mode active?
[Edit]
There is some red around that rectangular pin1 of J10… maybe there the end track is not sitting right on top of that TH pad?
Personally I route with pads in outline mode, no silkscreen and no val/ref fields (just have my small personal ones active on Eco1.User, small text in bright green having text value %R instead of **REF), to better see what I’m doing. Having show clearance for tracks at all times also helps to spot problems with sections.
@jwatte, In case you didn’t catch it . . . when routing the trace that passes over Pin 2 of Q1, you must slow down long enough to click on the pad itself. Otherwise, KiCAD doesn’t know that you want an electrical connection between the trace and the pad. Yeah, they ARE touching each other, but KiCAD isn’t smart enough to know that an electrical connection will naturally happen.
I’ll let you decide whether this behavior is a significant bug, an annoying quirk, or a bit of charming eccentricity. For now, it’s a fact of life.
In open gl mode sometimes ratsnets stay there even when the connection is already made. (The unconnected count in the lower part of the window also looks as if the connection is not made. But drc does not detect that as unconnected.)
If this is the problem in your case, restarting pcbnew resolves the problem.
Thanks for the answers – the cause was, in fact, that the pad didn’t coincide with the center of a vertex of a trace.
The reason I didn’t assume that was the problem was that I had routed it by stopping at that pad. However, I typically route with a default width, and then go through and fatten up the traces that are going to carry significant current, and somehow, in the process of applying E (change trace width) to these traces, it got un-stuck from the pad. No idea how that happened, but it did.
