Keepout Detection (DRC) Dodgy

[davidperrenoud]

From bitsybof...@gmail.com on July 24, 2013 22:10:09

See the attached images (and fzz), you will see

The grid size is set to 0.1mm
The keepout is set to 0.3048mm (12mil)
The DRC passed
The bottom center pin, and the trace running above it, appear based on the grid to be about 0.25mm apart (~10mil), less than the keepout and the keepout, by eye, should have failed them.

Attachment: keepout2.png keepout3.png keepout1.png transistor-test-jig-mini.fzz

Original issue: http://code.google.com/p/fritzing/issues/detail?id=2678

[KjellMorgenstern]

DRC does not complain, although the distance seems 0.3mm, and I have set the keepout to 0.31 mm
keepout.fzz.zip

[KjellMorgenstern]

In the above keepout.fzz.zip, DRC complains when keepout is set to 0.3375mm but does not complain for 0.3370mm .From the image, I think it should already complain at 0.3mm

[KjellMorgenstern]

Note: This error means about 1 dot at 800 DPI. To fix it, the DRC either needs to calculate with vector graphics, or use a much higher DPI (eg. 2000 dpi)

[failiz]

Could it be that increasing the DPI helps also to fix #3902 and #3639? Not sure if it is computationally expensive...

[KjellMorgenstern]

I don't think they are related. The DRC works on raster graphics, and such effects are inevitable (of course increasing the resolution would mitigate this. Not sure if there is any machining that precise that it is a problem already right now)

However, parts placement like in the other issues is a vector thing, and i hope it can be solved by just being more careful with numerics.