We have experienced a series of problems with ADP5073. I hope this could help someone with similar issues.
The design is same as the application circuit in the datasheet and the design tool, with 12V input and -2V output. Parts are selected from the parts the design tool suggested. However there are two errors in the design: SYNC/FREQ pin and EN pin are float on the PCB.
To start, I grounded the SYNC/FREQ pin by connecting it to a ground node (not exposed pad, which is also connected to ground) and extended the EN pin outwards as such that it's large enough to probe with a tip and made sure there were no net errors before assembly.
The circuit is stable before EN is applied, with current draw less than the power supply can display (10 mA). However, when EN is applied (2.1V) the device draws maximum available current and forces the power supply to lower the voltage, e.g. 12V input drops to around 6V when the current is limited around 300 mA, and to 0.9V when it's below 100 mA. AVIN and PVIN are both tied to 12V. Datasheet says when AVIN drops below 2.55V the regulator turns off, but even at that levels the device continues to draw current.
Thinking that the PCB must have some sort of error, original or due to modification, I assembled the part on a QFN/DIP adapter card with no other components on, and set up the same circuit externally. I used two different parts that are procured at different times for this.
The adapter had no exposed pad connection, so initially I left it float. In this case both circuits gave an output around zero volts. When tied to ground however, finally a negative voltage appeared on the outputs, but with one giving a stable -0.25V and the other -1V, with the same external circuit.
Other unexpected observations:
- The circuits behaved differently when EN is applied at the same time with Vin vs. when EN is applied after Vin. However, this behavior was not consistent and differed with time, that is in some trials EN was needed with Vin, but not most of the time.
- VREG voltage differed between circuits and between measurements, like 1.3V and 3.4V.
- The current draw when enabled was usually around 20 mA, however in some cases it jumped to 60 mA when VREG is probed for voltage measurement and stayed at that level.
Then the real problem occured: I changed the inductor in the external circuit, the result was about the same, and after a few other trials the initial short circuit problem showed up. Now one of the two parts sinks the current, the other gives around zero voltage output with 50 mA draw, with the same external circuit.
I know there are too many uncontrolled variables here, so this is not a specific question on this case but obviously devices are getting damaged somehow and there's not much information on the internal structure to figure out what might have gone wrong. I'll get an evaluation module but I don't think it will be much helpful on this.
Although the station is safe my first suspect was ESD, but we didn't have any error like this with similar setups including ADP5053, and EN pulldown resistance seems intact.