I am still curious to beat up against this snag, and hope to get a Part4
To certain sure the Asic are not involved scratch off the other two due the power pins story at the reply above here,
yeah we must have seen a temperature rise by scratching off the first one which was not
Therefore i think it is a "short" between the copper planes 3V3 and GND
the copper (power) planes are very large and therefore all contribute to dissipate this huge power loss therefore the board still looks cool

under heat detection camera
I think in one of the via(s) like mounting hole or other one(s) a tiny is short exist between the copper planes 3V3 and GND, Maybe it is possible to pick up the flow current by some probe at an scope to see where current is flowing,
other wise 3V3 trace can be cut to see if the short disapears and if there are none 3V3 traces only a 3V3 copper plane cut the pcb into half with a hack saw to catch and trace location of this snag
if the short has gone we know the short has to be at the piece of a half pcb

regards