OVP for TPS65094x

Hi Antony,

OCP does not inherently trigger a power fault, but will likely cause one. As OCP kicks in, the output current provided by the buck will start to droop and as a result the Vout should start to droop as well, resulting in an PWR FAULT due to undervoltage.

There is a point where OCP will start to trigger and the output will start to droop, but stay within the 8% and not trigger power fault. It's only about 100 mA in width though, so in nearly all cases the OCP will result in undervoltage which results in PWR FAULT.

One important note is that the OCP bit in the OFFONSRC register actually triggers for any PWR FAULT, both undervoltage and overvoltage, so it does not actually necessarily mean an OCP event has happened.

If customer wants to confirm an actual OCP occurred, there is another register they can read:

Hi Antony,

I realized today there may be a fundamental difference in what I am describing as overvoltage protection and what customer may be looking for. I was referencing our PMIC's ability to shutdown in case it starts to regulate high. If they are looking for protection from an external overvoltage source, then they would need separate circuitry. 

The PMIC OVP is intended to shut down the BUCK if it starts regulating outside of the target window. It is difficult to test since the BUCK will do everything in its power to ensure this doesn't happen, especially in DCAP2 architecture.

One option to test on the EVM is to use VCCGI to overpower VNN. Start by shorting VNN and VCCGI with PMIC off. Turn on PMIC by setting CTL2 high (THERMTRIPB), then CTL1 (PMICEN), then CTL5 (SLP_S4B), then CTL4 (SLP_S3B), then CTL3 (SLP_S0B). At this point VNN should be on, VCCGI still off. Set VCCGI to 1.3V by I2C. VCCGI should be able to overpower VNN since OCP for VCCGI > OCP for VNN. At the point that VNN passes the power good threshold for 30 us, the PMIC will shutdown both VNN and VCCGI and reboot with VCCGI off again (default state is OFF until SoC passes I2C VID). It should look like:

The top signal is VNN, the bottom is RSMRSTB. I short a BUCK1 and BUCK2 together and then set BUCK2 to 1.3V and we see BUCK1 power fault. I actually did the testing with a TPS650842EVM-116 but results should be approx the same for a TPS65094x part as well. PMIC tries to restart 10 ms after power fault since I have PMICEN tied high and continues power faulting until the voltage discharges just off screen and boots properly.

Note: they will need to investigate on their board if this test is possible without breaking their system.