We have a custom design running off of a primary battery pack. The plan was to run the DW1000 until the batteries were completely depleted, then go out and replace the batteries. There is no low voltage shutoff in the design, the boards will continue to run until the pack can no longer supply adequate voltages.
What we have found is that as the units continue to operate with battery voltages that bring the main 3.3V rail down below 2.8V, we have a high occurrence of finding damaged DW1000s.
When units are repowered with supplies that provide 3.3V on the main rail, SPI reads from the DW1000s read either a 00s or all FFs. Further investigation shows a leakage path inside the DW1000 from 3.3V to VDDLDO / VDDLDO2 (which our design feeds with 1.8V). This leakage puts ~3.2V on VDDLDO / VDDLDO2 even when 1.8V is not connected. All other pins and signal levels appear as expected.
Is there any history of damaged DW1000s from low input voltages?