4
\$\begingroup\$

I have (had?) an STM32F429ZI. I was working with the ADC on PC0 and due to a hardware malfunction, 12V ran through the pin when the maximum voltage an ADC pin supports is 3V. Since then, whenever I connect the STM to my computer via the USB in CN1, LD5-Overcurrent constantly glows. In addition, the IC close to LD5, U4, heats a lot. The part number of U4 is ST890C-GK035.

So, I'm really not sure if I just killed the MCU or if the board is in some kind of secure state?

I'm using Atollic TRUEStudio to work with it, but I can't upload any program and neither use the debugger.

Overcurrent LD5 glows. Close up.

\$\endgroup\$
11
  • 8
    \$\begingroup\$ It is probably dead. \$\endgroup\$ Commented Aug 19, 2021 at 15:47
  • 1
    \$\begingroup\$ Sounds dead to me as well. \$\endgroup\$ Commented Aug 19, 2021 at 16:17
  • 4
    \$\begingroup\$ R. I. P. ………..;) \$\endgroup\$ Commented Aug 19, 2021 at 16:47
  • 2
    \$\begingroup\$ what do you mean by STM? \$\endgroup\$
    – jsotola
    Commented Aug 19, 2021 at 16:57
  • 1
    \$\begingroup\$ @jsotola STMicroelectronics \$\endgroup\$ Commented Aug 19, 2021 at 18:50

4 Answers 4

6
\$\begingroup\$

Sounds dead to me. U4 probably developed a short circuit. Some boards do have polyfuses which get hot in case of a short circuit and won't allow the board to turn on again until after they cool down. However, U4 is not a polyfuse and the fact that U4 gets very hot is a dead giveaway that U4 is dead.


Update from comments: apparently U4 is a switch which is part of the power supply circuit. In this case it is also possible that something downstream of U4 has a short circuit. That component is probably also getting hot. That could be the main CPU U8 in which case the board is useless and you should get a new one.


If you are lucky, other parts of the board may still work, including the actual STM32 chip U8. The busted U4 is stopping them from getting any power because it's a short circuit. You'll certainly have to remove U4 to get anything to work again. If U4 has something to do with the USB connection then you might not be able to use the board anyway. If U4 has something to do with the power supply then removing it could cause further damage. No way to predict that without seeing the schematic or at least seeing what kind of chip U4 is.

If you want to get actual work done it is probably more productive to get another board. It seems they are not very expensive. If you value your money more than your time, you may remove or replace U4 just to see what still works.

According to the manual, the top part of the PCB is intended to snap off (hence the cut line) and you can reconnect it with a few wires. So if the main part is working you could still power and program it using the top part from another board. If you value your money more than your time. Note that the ADC unit in U8 is likely to be dead regardless.

\$\endgroup\$
2
  • \$\begingroup\$ U4 is "1.2 A current limited high side power switch with thermal shutdown".. st.com/en/switches-and-multiplexers/st890.html \$\endgroup\$
    – Ben Voigt
    Commented Aug 19, 2021 at 18:22
  • 3
    \$\begingroup\$ There's a very good chance that U4 survived (its maximum current rating is much larger than anything else on the board) and it's getting hot because it's passing the current limit, while the thermal design to keep it cool was designed for the actual current the board was supposed to draw. The fault is somewhere downstream, almost certainly the microcontroller which was hit with twice its permitted voltage. \$\endgroup\$
    – Ben Voigt
    Commented Aug 19, 2021 at 18:26
15
\$\begingroup\$

12V connected to 3.3V tolerant MCU pin means the MCU is now permanently damaged and possibly other components as well. If the USB port on your PC still works and is not damaged, you can consider yourself lucky.

Edit: I double checked and PC0 is a five-volt tolerant (FT) IO pin. It has an absolute maximum voltage of VDD+4.0V, which means that if the standard 3.3V supply voltage was present, the absolute maximum for the pin was 7.3V. The FT pins do not have internal clamping diodes to VDD supply voltage.

\$\endgroup\$
4
  • \$\begingroup\$ Can IO clamp diodes do anything to prevent permanent damage or at least limit the damage to that single faulty IO and keep the others safe? \$\endgroup\$ Commented Aug 19, 2021 at 17:38
  • 2
    \$\begingroup\$ @TirdadSadriNejad Your comment made me double-check and the PC0 pin in question is a 5V tolerant pin which does not have a clamping diode to VDD. \$\endgroup\$
    – Justme
    Commented Aug 19, 2021 at 17:55
  • 1
    \$\begingroup\$ I once smoked my computer by connecting 24V to the 5V USB rail. Literally. There was smoke pouring out of my keyboard. That was an expensive lesson. OP got really, really lucky. \$\endgroup\$
    – CurtisHx
    Commented Aug 20, 2021 at 12:32
  • 1
    \$\begingroup\$ @TirdadSadriNejad: They used to be somewhat more capable in that regard than they are nowadays. Sometime around 1995, a coworker working with a grounded PIC-based board that had a 1K throughhole resistor in series with a pot input accidentally connected the input to mains. The resistor was obliterated (a couple of upsidedown-L-shaped pieces of metal were all that remained) but after the resistor was replaced the board worked fine. I would not expect newer chips to be nearly so rugged. \$\endgroup\$
    – supercat
    Commented Aug 20, 2021 at 19:35
3
\$\begingroup\$

12V ran through the pin when the maximum voltage an ADC pin supports is 3V

Probably, it also depends on how much current you sent through it. The absolute maximum ratings have been violated so the proper functioning of the chip is not guaranteed.

If you can't read the flash back from the chip then its probably dead. You could buy a new STM32F and unsolder the old one and put the new one on (or find someone with good soldering skills to do it for you).

\$\endgroup\$
2
\$\begingroup\$

I have quite some experience of accidentally connecting higher voltages to 3V3 ADC inputs on various MCUs. They all tend to fry when that happens, no matter brand or type. The MCU overall could still work in some "limp home" mode, but the protection diodes etc on the pins are likely fried, and likely the ADC peripheral as well.

The MCU U8 is very likely fried. Depending on what "heats a lot" means, U4 could be broken or just drawing lots of current. Some 40-50 degrees isn't hot for a power MOSFET. A dumb idea is to press a finger against the IC to check - you may burn yourself. An old hobbyist trick is to shut off supply, then apply a tiny droplet of water on top of the IC. If it goes up in fumes when you turn the power on, then the heat is some >100 dgr C and the part is likely toast. Note that some parts have thermal shut-down as high as 85 dgr C though.

I'd switch out the MCU U8 first and see if that helps, if not then replace U4 as well, along with any TVS etc diodes nearby. However, replacing a LQFP144 without damaging the PCB pads requires quite some soldering skills and a desoldering gun. If you don't have the skills and equipment, just forget about it and buy a new board.

\$\endgroup\$

Not the answer you're looking for? Browse other questions tagged or ask your own question.