Closed mrnuke closed 8 years ago
I have addressed this in https://github.com/KiCad/Housings_SSOP.pretty/pull/12
In future please feel free to make pull requests yourself to fix these issues :)
These are manufacturing examples with git commit 2933ce3
This specific example was manufactured at oshpark, with default kicad settings. oshpark's reference rendering is included below.
The mask clearance looks quite big. The footprints in the library has not the mask clearance defined so they use global settings. Maybe you have set it too large? I guess the setting should be zero and the manufacturer modifies mask according to his technology. I've used the the clearance set to zero and my boards looks fine (it may depend of manufacturer).
Mask clearance was set to 0.2 mm (8 mil). For oshpark, the tightest is 0.15mm (6 mil), and that sort of figure seems to be common among commodity manufacturers.
I see two ways to go about this.
In my opinion, assuming a clearance of 0.15mm would work, a difference of 0.05mm should not cause such a dramatic outcome on a human-visible feature.
@mrnuke can you post a screenshot of the silkscreen gerber layer you sent to OSHPark? Comparing the screenshot of the KiCad render with the boards you received, it looks to me like the silkscreen should have been on the final board.
The middle part (SSOP-16_4.4x5.2) for example, the top line does not look like it should intersect with the solder mask at all - it looks like it has been completely removed, either by OSHPark or by KiCad when it was plotted...
Looks like they added their mask clearance to yours. Won't they accept just zero mask clearance on gerbers?
I uploaded the kicad_pcb file [1] directly. If you zoom in enough, there shouldn't be an overlap with soldermask (there's a clearance of 0.025mm/1mil).
[1] https://gist.github.com/mrnuke/10e3027b3196ba24202a1db71d90fb26
The small line that is supposed to be the pin 1 indicator is too close to pad 1. Once soldermask clearance is taken into account, the silkscreen line would end up falling on bare FR4. This happens with 0.15 mm pad mask clearance.
I've tested Oshpark and seedstudio, and with both manufacturers, the pin1 indicator line is removed, and there is no way to identify the correct orientation of the footprint.