We have had many instances of the AD7147 hanging up and no longer responding to SPI commands. Most times the chip can be power cycled and normal function returns. It is not possible to get normal function back by re-initialising and/or sending the software reset command. The micro is remote to the AD7147 and communicates along a 1m loom. I am deeply suspicous that the AD7147 is receiving corrupt SPI commands with bit errors that put it into this strange mode. Indeed I have proven that if you write to locations 0x044 and 0x043, the chip will hang up.
Has anyone seen this type of thing before? Any clues on how to get it going again without doing a power cycle?