MSP430F2619S-HT: MSP430 Flash Data Retention and life questions

Peng,
1) I would not expect operation at 125c to cause a flash bit to be corrupted.
2) Radiation definitely could cause errant behavior. However, the radiation would likely need to be significant to cause this type of error. Soft errors from neutrons can cause memory and flip flops to change state. However, this is fairly low probability, and primarily at higher altitudes.
I will need to check with others on potential impact of high magnetic fields. Could you indicate magnitude of magnetic field exposure?
3) I expect that you did have some type of corruption to the firmware. Most likely this would be a systematic problem, with firmware writing to wrong location or something similar.
I would suggest that if this occurs again, to dump the existing firmware to evaluate what has changed. This could help point out the source of the error.
Does the firmware do write and/or erase operations to flash as part of its normal operation?

Regards,
Wade

Hi Wade,

Thanks for reply.

The mcu was put into oil well which was about 125°c and thousands psi environment, it was operating normally till 11 days, the magnitude of magnetic field is unavailable this time. We also tested dozens of device which could operate at 200°c for months.
We tried to recreate the same failure mode and intentionally corrupted two bits of firmware in HEX file and could observe the same behavior (unresponsive through TX&RX) except tiny current consumption difference.
With the normal operation, mcu samples data and saves to external flash. The data is saved in internal ram for seconds before write to external flash.
We are trying to identify the failure mode is related to hardware or firmware, isolated or systematical.

Any thoughts?

Thanks,

Peng

Just to be pedantic, I believe that neutrons are not the issue, but high energy particles and gamma with high LET.

(And yes, I do it for a living.)

Thanks for the additional information.
Also, I agree that your issue would not likely be due to neutron SER as Keith pointed out. This is especially true since you are not operating at high altitude. Additionally, FLASH is pretty robust to radiation.

I am uncertain of effects of high PSI on a plastic package (unless you are using KGD version?)
Presumably your room test power cycle test that failed was performed at atmospheric pressure?

Based on your description, you do not modify the device FLASH during operation, and only when programming firmware. Can you confirm this statement?

I am going to request some insight from our QA engineer. He may be able to provide some additional insight into possible causes and effects of high PSI and/or magnetic fields.

Regards,
Wade

Sorry for the confusion. The die we used has a package protection and it is not exposed in high pressure environments.
We did power cycle test at room temperature and atmospheric pressure, and you are right, we do not do flash operation during our normal operations.

Thanks,
Peng