STM32WB55RG possible hardware failure - chip gets very hot and cannot find target

[nickdzi]

Hi everyone,

​

EDIT: This is on the NUCLEO-WB55RG development board

Was using the board last week without any problems. Started using it this morning and the board boots up, but it cannot connect to the computer to dump code. The chip under the removable shield (I'm not sure whether its the M0 or the M4 chip, I think its the M4) is getting very hot.

​

All the components/soldering/traces look fine so I'm leaning towards internal failure of the chip under the shield. Is there any way I can verify whether this board is broken for good?

​

The LED boot sequence is different (I have two boards), the working board boots as follows:

1. Plug USB in via computer - board powers on
2. LED 6 turns red ON
3. LED 5 turns ON
4. Program starts and all is fine

However, the suspicious board boots like so:

1. Plug USB in via computer - board powers on
2. LED 6 turns red ON
3. LEDs 4 & 5 flash on then off x2
4. LED 6 turns red OFF and green ON
5. LED 6 turns green OFF and red ON
6. LEDs 4 & 5 turns ON
7. MCU under shield becomes very hot and cannot connect via ST-Link

Any input is appreciated, happy to send the board to place of purchase to get it assessed, but if there's something that can be done in house before that then I'm happy to give it a try.

[Harvey White]

Any idea of what LED 4 and 5 flashing mean?

Likely an "I can't download"

Assuming that the two boards are identical, you have a number of very generic possibilities:

1) a pin is shorted to ground that shouldn't be, and an I/O section is dumping power into ground

2) same possibility for a short to VCC

3) VCC is not 3.3 volts but it much higher, please check regulators.

4) VCC is lower because of a short to ground, check regulator and see if the regulator gets hot

I've never had a chip get hot without a short to ground (on an active pin pulled high, check voltages!) or a pin pulled low (same check voltages)

Look for solder splashes, mis-wiring (assuming a different revision of boards), and so on.

Assuming the same software, it looks like hardware from here.

[nickdzi]

Thanks for the response, I did consider a shorting pin, I'll have a look around the voltages you mentioned and see if that helps. Thanks again

[Harvey White]

I had a pin that was going betwen 0 and 1.4 volts on a 3.3 volt system I don't thing that anything got hot, but it was an active pullup pin being set as a 1 and an active low pin being set to a zero, and the common point being the same wire This is common in TTL wiring where two lines are shorted (or designed to be) together. One active pullup and one active pulldown on the same lne. This can be explained any number of ways, but from same board to same board, the software revision (and setup) is most likely.. Please note that this did NOT overheat the chips on an STM32, allhough it did on an XMEGA series processor (different hardware!)

[Harvey White]

This suggests a bad connection to it, or something else odd is happening. I'd think (and expect) factory boards to have undergone testing before shipping. You've disconnected everything, powering it up only from the USB. Have you checked the jumpers between the working and non-working boards to encertain that they are the same? That can get you.

I've designed my own boards, and not loading has always been a hardware (or jumper) fault.

As a silly test, make sure that all the chips are in identical positions. How that might have made it through test is another matter, but it would be an item on a checklist. As I mentioned, check the jumpers as well between boards.

[Harvey White]

I should add that if you are powering this externally, do you have the proper voltage and polarity or are you powering this strictly from a USB connection?

[nickdzi]

After doing some testing with a multimeter I have discovered two things:

When the JP2 jumper connecting 3.3V to VDD is not bridging then there is no strange LED sequence and the chip does not get hot... when the jumper is connected (and power is provided via USB/ST-Link) then nearly all pins become shorted to VDD??

Is the chip shot?

Would be helpful to hear from an ST member, if none see this I will have to open a support ticket.

[Harvey White]

I guess the only thing to ask is if the chip is running on 3.3 volts o not. If it is, and you draw too much power, then that suggests a bad chip (since a shorted bypass capacitor would get hot and not the chip). It seems to me that you're on the track to a suspect chip, but why and how I'd have difficult telling. Have you measured the 3.3 volts on the chip (check at the bypass capacitors for an easy look)? I assume that the 3.3. volt regulator is also getting hot, since the power draw ought to be excessive. Have you checked the USB power with any of the cheap USB meters?

It certainly sounds as if the chip is drawing too much power. These chips are not designed to get that hot during normal use.

The VCC is correct, is it not? It's one of the first things I'd check, along with power draw.

[nickdzi]

When JP2 is disconnected (bridges the 3.3V from the regulator to the chip) there is 3.3V on the regulator output and 5V on the ST-Link 5V input has 5V. When the jumper is connected the output of the regulator is 2.7V with the input as 3.2V.

[Harvey White]

Suggests that the regulators are being asked to supply too much current, and that's likely pulling down the USB 5.0 vols. Don't try running this board on anything but a current limited supply, it's probably bad enough already.

A trick might be to run it off a power supply limited to about 50 ma. Put it through on the 9-12 volt inputs, but start off at zero. Note, as you bring up the voltage to the proper level, where the supply starts to current limit and what voltage is on the processor. Unfortunately, you may not be able to disconnect things from the processor, so you may not be able to see if an external source is pulling down the processor. Since the processor is able to run at voltages lower than 3.3 volts, you might get something informative.