The FC carrier boards were built back in June 2010, but the SPS section was just completely dead in the water - it wouldn't output any voltage at all. We sent the boards to Screaming Circuits and they X-rayed the giant LGA LT4619 micromodules. As we had feared, all of the solder got sucked into the vias-in-pads and there were no connects everywhere. Next time: don't just fill the vias with solder paste, actually bake them to get rid of the flux which takes up most of the volume of the paste. SC went in and hand tented the pads and the micromodule pads, and then reworked that chip.
Finally, we had output voltages we could use. A few real problems then surfaced:
- U1000: SGND and GND weren't connected... anywhere. Added zero ohm resistor.
- R1003 (47k) was dropping way too much voltage: 11.37 V to 10.28 V. For some reason, the MAX5902 has a DRAIN input resistance << 650 Kohm they spec.
- Reduce R1003 to 1K ohm and re-attached the over voltage protection
- U1012 has active low enables that are, of course, pulled high.
- Took pins 1 and 19 and tie them to ground.
- D+ and D- may be switched; see breakout board mixup
- TODO: verify D+ and D-.
- USB2 doesn't work.
- TODO: This is a software issue - turn on USB2 in the MCP5200 peripheral registers.
- Still not tested
- CAN
- Ethernet