Hi everyone!
I designed the board shown in the attached picture. It’s built around an STM32H745 microcontroller, along with crystal oscillators, a DAC, an EEPROM, and an NE555 that generates a reset signal when a button is pressed long enough.
After assembling the board, I noticed that it draws an unusually high current (over 800 mA). The power supply goes into compliance and reduces the voltage, while the microcontroller itself becomes very hot to the touch. I noticed that the current rises with time to the compliance and over.
To troubleshoot, I removed all external components (oscillators, DAC, EEPROM, transistors, and NE555), leaving only the microcontroller, but the excessive current consumption persists. I’ve checked carefully and I’m quite confident there are no shorts on the board.
When I connect via ST-Link using STM32CubeProgrammer, the tool successfully detects and identifies the MCU. However, within a few seconds, the connection drops because the supply current rises above the compliance limit, causing the voltage to collapse.
For comparison, I have a Nucleo board with the same MCU family, and it doesn’t exhibit this behavior: the current consumption is much lower, and the device remains cool.
Does this increasing current consumption point to a short circuit I may have overlooked, a faulty microcontroller, or possibly a layout issue?
According to the reference manual RM0399 (page 274-276) I chose the bypass configuration (config 6 in figure), which is the one that I want (I provide external 3.3V). I checked the schematic and it seems that this configuration is respected. Did I get something wrong?
Thank you in advance for your help
"Bypass" mode requires that you generate the correct voltage for VCAP somewhere else on the board. You cannot tie VCAP to 3.3 V. It will damage the chip.
You probably want LDO mode, which is the most basic. In LDO mode, VCAP should have 2x 2.2uF caps bypassed to GND.
> When I connect via ST-Link using STM32CubeProgrammer, the tool successfully detects and identifies the MCU
It's rather impressive for the hardware to still remain responsive under these conditions. Protection diodes are putting in overtime.
