Currently for rev2, and it's supporting boards there is a bit of a chunky BOM, and while some items are shared between the connector boards, the PHYs and the main board, it could be improved.
The following things could vastly simplify the BOM at the cost of ease of assembly:
Those are the first two things that come to mind, there are also some DFM optimizations that would need to be made to make assembly easier, that ties into the BOM consolidation, but there are some other things as well:
The following other misc items are also tied in but don't fall exactly in line with DFM, if we want to have a case, it would be cheaper to go for a standard extruded aluminum shell, but this limits our ability to do module stacking like planned, meaning you need to pull the device out of the case to swap things like PHYs or connectors, which is a big downside, but on the upside it would mean you don't need custom milled cases for each part.
We could also possibly squish the board a big to make it less long, compress the power supply area a bit to save board space.
Also possibly rotating the design a bit to have a more standard flow-through so the USB and power connectors are on the bottom rather than side, we would also need to rotate the FPGA anyway if we keep using the ECP5 so we can use the high-speed IO on the SCSI PHY connector.
Thoughts/additions @dragonmux, @esden?
Currently for rev2, and it's supporting boards there is a bit of a chunky BOM, and while some items are shared between the connector boards, the PHYs and the main board, it could be improved.
The following things could vastly simplify the BOM at the cost of ease of assembly:
0201or0402where possibleThose are the first two things that come to mind, there are also some DFM optimizations that would need to be made to make assembly easier, that ties into the BOM consolidation, but there are some other things as well:
The following other misc items are also tied in but don't fall exactly in line with DFM, if we want to have a case, it would be cheaper to go for a standard extruded aluminum shell, but this limits our ability to do module stacking like planned, meaning you need to pull the device out of the case to swap things like PHYs or connectors, which is a big downside, but on the upside it would mean you don't need custom milled cases for each part.
We could also possibly squish the board a big to make it less long, compress the power supply area a bit to save board space.
Also possibly rotating the design a bit to have a more standard flow-through so the USB and power connectors are on the bottom rather than side, we would also need to rotate the FPGA anyway if we keep using the ECP5 so we can use the high-speed IO on the SCSI PHY connector.
Thoughts/additions @dragonmux, @esden?