has anyone got the i.MX53 internal temperature sensor to work as described in AN4380?
Initially I could not access the PHY registers (the acknowledge bit was never set) but I found this was
because the SATA module did not have a clock (we haven't connected an external clock).
Blowing the fuse at 180C in fusebank 4 to set SATA_ALT_CLK_REF to 10 (internal USB Phy) fixed that
(I couldn't find another way to do it in software without blowing the fuse)
So I can now access the registers and implement the algorithm described in the application note.
The readings seem fairly stable but I get (at ambient temperature) m1=520 m2=683
Which, according to the formula in the AN is -106.6 C
Something is clearly wrong...