2024-07-11 06:50 AM
I am currently working on bringing up a production board design with the OSD32MP157C-512M-IAA. The board is by design very similar to the Octavo systems osd32mp1-red board and the USB part is very similar to ST STM32MP157-EVAL where VBUS pin in uUSB connector through a bead is connected directly to VBUS_OTG on PMIC and to OTG_VBUS on MPU. The board use I2C, SPI, CAN, GPIO's. If I without any power, plug the board in to my PC the board burn so PMIC VDD3V3_USB (LDO4) and VDD (Buck3) are shorted afterwards. I have found a solution that seems to work. The zero ohm resistor towards OTG_VBUS is changed to 10k and the pull-down is changed from 10k to 100k ohm (47+10k on EVAL board) then the board survive. I belive that the 5V try to lift all internal power
Question is: have anyone seen similar problem first of all on the EVAL board but also on your own design? And have you found other solutions or root cause to the problem?
Solved! Go to Solution.
2024-07-11 07:18 AM
Hi,
The OTG_VBUS of STM32MP1 is somewhat a weak point and could only withstand 6V when both VDD and VDD3V3_USB supplies are present (it is even slightly lower if VDD = 1.71V). If there is no supply, applying VBUS 5V to STM32MP1 could be problematic (usually does not break immediately, but could lead to accelerated aging of the device).
In most cases, to avoid this constrain, the OTG_VBUS of STM32MP1 could be simply left unconnected and forced in SW (either statically if you use only host or device, or dynamically when you use USB Type-C with STUSB1600 as in STM32MP157F-DK2). https://wiki.st.com/stm32mpu/wiki/OTG_device_tree_configuration#DT_configuration_examples
On other side, on STM32MP157F-EV1 board, as far as I remember, we did not got feedback of any board bricked due to this (and we have some customers using similar connection without issues with large product quantities).
Alternatively, VBUS could be feed to a simple GPIO with resistor divider to check presence if needed (this is ok if only 5V and no power delivery negotiation).
Maybe your board has some different overall power scheme or HW which make more stress on OTG_VBUS pin in some situations.
Regards.
2024-07-11 07:18 AM
Hi,
The OTG_VBUS of STM32MP1 is somewhat a weak point and could only withstand 6V when both VDD and VDD3V3_USB supplies are present (it is even slightly lower if VDD = 1.71V). If there is no supply, applying VBUS 5V to STM32MP1 could be problematic (usually does not break immediately, but could lead to accelerated aging of the device).
In most cases, to avoid this constrain, the OTG_VBUS of STM32MP1 could be simply left unconnected and forced in SW (either statically if you use only host or device, or dynamically when you use USB Type-C with STUSB1600 as in STM32MP157F-DK2). https://wiki.st.com/stm32mpu/wiki/OTG_device_tree_configuration#DT_configuration_examples
On other side, on STM32MP157F-EV1 board, as far as I remember, we did not got feedback of any board bricked due to this (and we have some customers using similar connection without issues with large product quantities).
Alternatively, VBUS could be feed to a simple GPIO with resistor divider to check presence if needed (this is ok if only 5V and no power delivery negotiation).
Maybe your board has some different overall power scheme or HW which make more stress on OTG_VBUS pin in some situations.
Regards.