We use the LMZ13608TZE/NOPB part to create 5V rail on the board with external 24V serving as an input voltage. During design phase we applied our derating criteria for temperature and maximum load so that under normal operating conditions there is a lot of operational margin and the module is not stressed at all. Out of nearly 700 units produced so far we had 28 failures related to this LMZ module. While this may not seem like a high failure rate it does stand out especially when you consider the possible failure modes. We’ve identified three so far – lack of output voltage, short to ground and one that concerns us most – either slow or immediate rise of output voltage beyond the regulation point and up to the input voltage level. I don’t have to tell you how destructive presence of 24V on 5V rail can be.
Vast majority of these failures occurred during our test process, mostly during first power-up but there were few that happened when the test was well underway.
Based on the datasheet this module employs multiple protection schemes including output overcurrent and overvoltage. Yet they don’t seem to be active when the failure happens.
At this point we’re looking for any tips and pointers as to the reason for the failures. We were wondering if there is any existing field data from similar cases that we can relate to our experience. We are concerned that the same problem may also show up as a latent failure on deployed units in the field which of course is the least desired scenario.
Looking forward to your response.
Derek
