Hi,
I’m designing open hardware module based on the Google Coral AI module for Framework 13.
The idea is that you could offload AI workload on this chip thanks to the fast USB 3.0 connection.
What is still missing is the trace matching. Not sure how it’s done in Kicad over components when it differs for each signal - shown in the table in the README.
Another issue that will need to be solved once it’s manufactured and the module is sourced is that the AI accelerator can generate lot of heat.
The AI accelerator is quite old now and I want to just learn how to design expansion module for the Framework 13 laptop.
On top of that try to reverse-engineer the communication and the driver just like Tomeu is doing with the other AI accelerators.