Overcurrent on STM32H755ZI-Q Nucleo board

[jdw]

Hello ST Community. I have a STM32H755ZI-Q Nucleo board that I seemed to have blown up the STM32 part. I suspect this because the board is getting an overcurrent condition. But have some questions so I can better understand what may have caused this. The Nucleo board is connected to a base board I built via the CN11 and CN12 connections. I am using a handful of I/Os on the CN11 and CN12 connectors for GPIOs, I2C and SPI, along with connecting ground pins.  In addition have a handful of wires connected to the top of the board via CN7/CN8/CN9/CN10 to another board also for GPIOs, SPI, and I2C and ground. This configuration had been working fine for a few months with no issues. The Nucleo board is configured as it is out of the box with the STLink USB power providing power to the Nucleo board. 

One thing I am doing is that I have the clock configuration set (via STMCubeIDE) to the highest I can get without changing the Nucleo board configuration for SMPS. This configuration allows one to clock the V7 core at 400 Mhz and the M4 core at 200 Mhz. 

All of a sudden, I connected the STLink USB to my USB Hub which is connected to my Mac via USB-C, and within a few seconds the Nucleo board indicated an overcurrent condition. The ST part did not seem overly hot, and I saw no magic smoke. I did find that if I temporarly removed the SMPS 3.3v/1.8v jumper the overcurrent would go away, hence I thought the issue was with the ST part or the SMPS circuit. 

I completely disconnected the Nucleo board from my rig and I changed the one jumper for the 5V config to use external 5V power and with that I was able to power the board enough to have the ST part working again and run my code.  I discovered a few I/Os that I had configured as outputs and using were not working correctly (which made me a bit suspicious that I did something bad electrically to the part) - one GPIO was not changing at all, and another was giving .7v as a high. I don't see a way that I could have shorted these I/Os and blown them up, but anything is possible.

After a while, the Nucleo board stopped working all together, even using an external 5V supply. Provided I did not somehow short any I/Os to a voltage that blew up the part, is there any reason the Nucleo board can't be run at the clock speed I configured and that was the cause, somehow damaging the SMPS external circuit? Has ST tested the board with this clock speed?

I ordered two more boards to I can get my rig going again, but would like to understand what may have damaged the one I have. 

ST, any thoughts?

[AScha.3]

Hi,

first, i am not "ST" - but your question is more general anyway.

So 

>Has ST tested the board with this clock speed?

Sure, this a very basic thing. see ds for typical power/current at all speed settings.

> is there any reason the Nucleo board can't be run at the clock speed I configured and that was the cause

no.

Generally speaking , this kind of SOC / cpu is more critical to wrong settings or handling, because of its internal SMPS : setting this wrong just by software is possible and can damage the chip ( but is up to the user, to do it right or wrong).  More "safe" are the "standard" cpus, with fixed LDO, not easy to kill them.

So i recommend, not to use cpu with smps - except you need it for some reason.

I go this way, because why bringing more risk for failure to design, if not absolutely necessary ?

And if the board did need to much power, you should be alarmed : you doing something wrong, find out and fix your error. But not by giving just more power to the cpu - this only will kill it.

> one GPIO was not changing at all, and another was giving .7v as a high. I don't see a way that I could have shorted these I/Os and blown them up, but anything is possible.

You damaged it; GPIOS very rarely choose spontaneous suicide .

Very common is also damage by static discharge, so its almost sure, if just one time touching some unprotected pins or connections to pins, to damage the chip. If discharge is not killing it instantly, it often shows some higher current consumption of cpu or strange behavior of the pins, so i think, here is the reason for the "bad" pins.

Be careful , to never (!) have any discharge from your body or by connecting something without prior charge equalization .

So read carefully, how to set and use the smps and its settings, and avoid any static discharge, then your new boards will work fine.