on our custom board ( based on MK70FX512VMJ15 and MT47H64M16HR-3 AAT:H TR) we have problems that appear randomly in the field. The device can work up to 20 days without any problems. Then one DDR byte lane can stop working correctly. Thus, it is very similar to "e10521 of Errata KINETIS_K_4N96B".
And we can read all EVEN addresses as 0x00 with byte access. All ODD addresses are still correct.
We have already read a 32-bit value from address 0x400A_E188. This is the current RCR delay value. And it is always like 0x14,0x13,0x12. And it does not change if the K70 DDR controller enters into that abnormal mode. To restore the normal work of the device we just re-initialize the DDR controller. And everything starts to work.
We could see the problem on some devices only. Other devices do not have that issue at all.
We tried to heat the device up to 100°С. And we can't enter the device into that mode in the lab.