mdadm marks hdd faulty even though it's in pristine health?

Surely it's just that the drive really is faulty.

You are seeking a complex answer in timeout/spinup, when the reality is

[97148.490321] print_req_error: I/O error, dev sdc, sector 225357808

Controller could not read or write a specific sector to the drive. Cache would normally accept a write while the spin up proceeds.

Normally this is only ever seen on genuinely faulty drives, irrespective of what smartctl says.

Does swapping the drive for a different make a difference?