Part Number: TPS3823
Hi Team,
A customer is reporting an issue where a single unit of TPS3823-33DBVR exhibits a watchdog timeout period that is shorter than the minimum specification outlined in the datasheet.
Below are the detailed observations, test conditions, and troubleshooting steps completed so far:
Problem Description:
- Device: TPS3823-33DBVR (Mass-produced project).
- Issue: The datasheet specifies a minimum watchdog timeout period T_tout of 0.9 seconds. However, on the failing board, the actual measured timeout period is approximately 710 ms.
- Failure Rate: This is an isolated/single-unit failure. This is the second instance observed over several months. Other boards from the same batch, tested under identical conditions, pass normally and meet the datasheet specification.
Test Conditions & Troubleshooting:
- Test Method: After the board powers up normally, the watchdog input (WDI) signal is manually shorted to GND to stop toggling. The time from the last WDI transition/short until the RESET signal triggers is then measured.
- VDD Behavior: During the entire test window, VDD remains stable and within the normal operating range. The issue is not triggered by a VDD dip or UVLO event.
- WDI Pin State: The customer confirmed this behavior is not related to WDI floating or entering a high-impedance state. Furthermore, a floating WDI would activate the internal watchdog timer, which would typically prolong or delay the reset, whereas the current issue is an early reset trigger (710 ms vs 0.9 s).
Given that this has been isolated to a single component (1 pcs abnormal), the customer rules out a schematic/layout design flaw. They would like to investigate this as a component-level abnormality.
Could you please help analyze the potential root causes for a shorter-than-specified watchdog timeout period on a single unit?
Thanks and best regards,
Tony Liu
