A strange problem: some overvoltage problem (?) obviously interferes with the cmu; Setting external xtal and watchdog works fine - for most cards based on EFM32LG380F256. However some cards/processors are DOA. This problem concerns as much as 5% of a production run.
Next, some test may have generated (minor) overvoltage condition, and further 5% will die; these will not boot when power is applied.
When re-programming the chip, it will start first time with the JLink connected, but not if power is re-applied. Watchdog also gets disabled, and will not be active again. This WD problem is probably a sign of the clock unit failure.
I am at loss how to resolve this - maybe I need an external check of hardware and initialization?
And maybe Erase Chip is not enough prior to re-programming?
Appreciate hints and/or some paid help here!
Do you mind clarify a little more in detail about the overvoltage, do you mean you have applied over-voltage signal on some GPIO of the device? If this is the case what voltage you have applied, on what pin?
My views are my own and do not necessarily represent the views of Silicon Labs
The overvoltage is only an assumption as we have never managed to simulate such a condition in the lab. Field failures are relatively few ( but higher than the last generation product based on 80C32).
No specific IO pins are involved as this is a 'closed' system ( LED display in an enclosure. The enclosure is earthed through a 1Mohm||10nF RC to 0V)
The assumption of some over voltage is based on the fact that cards may go through programming & test first, then to fail when installed in products. It is a suprisingly stable precentage failing.
Some of the failing cards may be fixed just by reprogramming, but most will not start at subsequent power up.
After reprogramming, CMU may be affected in a way we do not understand: The processor will not start unless the J-Link pod is connected, plus watchdog is disabled...
I think you might need to create an internal ticket with our tech support team here - https://www.silabs.com/support
We will definitely need more information about the nature of failure. What is the level of overvoltage? For how long is it exposed to that voltage? We might need to see your schematic and learn more about your hardware. This is best done through a support ticket. Please create one following the link above.