Closed mryall-mawson closed 10 months ago
We have confirmed the cause of this issue was an incorrect footprint for the LTC4313, which has been present since the 2.0 board design. The PCB uses a pitch 0.65mm footprint when the part uses 0.5mm.
Some boards are fine because the difference is small and the pads managed to connect to the right place. We are going to try to fix the remaining boards with some hot-air rework and additional solder.
The proper fix will be to change the footprint for U9 to the DFN-8 3x3mm footprint with 0.5mm pitch, but remaking the boards for flight is not possible now.
JD has fixed and tested all the v2.2 boards now. PCB design still needs fixing.
Fixed in PCB design
Looks like a problem with the I2C buffer we're using, LTC4313-1. Watching the I2C lines with an oscilloscope on the board side shows no signal when the Beaglebone is transmitting messages.