cancel
Showing results for 
Search instead for 
Did you mean: 

Unused pins connected to ground. Could they cause short?

VVest.1
Associate II

I just had my first STM32 pcb done and I've been following the reference designs and checked many times everything that there is no shorts etc. However when I power the circuitry 3.3V side draws >200mA and the STM32L4S5ZIT6 does heat up when looked through thermal camera. I did connect all unused pins directly to ground. Could this be the culprint that immediately when powered it is like shorted? It makes no sense that these pins would be output high default when power on. I'm not even able to program the chip. And yes the chip is not 180 degrees positioned wrong.

AN4555 reference design says for some reason PC3 pin (29) to connect to 1u capacitor. I used this pin as ADC to sense battery voltage. I'm not finding any explanation why this pin was connected to capacitor. Could this be the culprint?

12 REPLIES 12
Uwe Bonnes
Principal III

Connecting unused pins to ground should not cause a short, as long as no code initializes the GPIO to output high. So test with keeping the CPU under reset if the high current still happens. If yes, tripple check for shorts on the pins. Otherwise check if bootloader is active, Maybe it puts some pins high. From some own observation, I thing that a single grounded pin with active high current tops perhaps at 50 mA (for F1, L4 may be stronger.

Otherwise PC3 is somehow related to VLCD. The datasheet does not explain about that. That needs to be reviewed @Khouloud OTHMAN or other ST employee

VVest.1
Associate II

Hi, thanks for the answer. Boot0 is tied to ground through 10k resistor.

One other thing is VDDUSB and VDDIO2. I have tied them the same VDD than the other MCU power pins. I'm not sure should these be treated differently? I'm not using usb connectivity on this MCU.. I have checked with microscope all the pins and certainly no shorts between the pins..

I've tried the ST-Link mode "Under reset" and reset mode "Hardware reset" it still wont recognize the mcu..

I wonder if you have placed test points on your board. Check them first. Then look for short circuits. If you have a copy of the PCB, check if there is any short circuit in the same track. Unless the board is designed correctly, there is no reason for extra heat.

VVest.1
Associate II

At least it is not complete short circuit as it limits the current to 225mA and from the 3.3V supply measured 129kohm resistance.

VVest.1
Associate II

I have isolated everything else from the board except i2c pins which go to external ADC chip, 32kHz, 24MHz oscillators and decoupling capacitors. MCU is the only one to heat up. It draws 225mA. Could it be faulty from factory? Should I redesign the pcb and leave those unused pins floating and test again? I'm out of ideas. I'd really appreciate your help.

Dear @VVest.1​ 

If the costs aren't really a matter of limitation, it worth replacing the microcontroller and check if the dismantled PCB with other elements assembled, performs its job correctly. Then if dismantled board on which only the microcontroller is dismantled does its job nicely, then check the assembled board. At the time of soldering, check the orientation of UC on board.

VVest.1
Associate II

Thank you all for your answers. Just a sanity check that the orientation is correct? I'm 99% sure it is but anyhow. In the picture top left corner pin 1 and from there ccw numbering goes. Thank you in advance.0693W000008yd07QAA.jpg

Uwe Bonnes
Principal III

That is right. With the text visible normal, Pin 1 is left bottom

Uwe Bonnes
Principal III

If you can, have a close look with a thermal imager. Maybe that give you a clue.