G070CBT6 Strange issue with reset line falling down...

[MKowalski]

Hi, I'm facing a very strange issue with my G070CBT6-based board...

When I start work and power up the board everything works fine for a couple of minutes (let's say 5 minutes ) but later I can observe some disturbances on the reset pin... I connect the oscilloscope and I can see the voltage on the reset pin, after a few minutes of the device operating, starts to fall down slowly and there is some mess on the signal...

Here is an example waveform watched:

Yellow -> 3.3V supply

Blue -> reset line

 

I've tried 3 different pcs of the MCU on the same board and also try soldering bare minimum setup (LDO, MCU, caps, etc.. ) on a totally fresh PCB, and every time the same behavior...

No matter, if the code flashed, is my application code, led blink demo, or totally erased flash.

This is the MCU part of the schematic..  I also tried to pull down the SW_CLK ( there is the BOOT0 also on this pin ) but this ain't help...

When the issue occurs, I'm not able to do anything with the MCU. It is "held under reset" from the SWD perspective. If I leave the board unpowered for some time to "cool down" it starts operating again, but again only for a couple of minutes... The MCU is not getting hot or something... supply is very stable 3.3V...

Any ideas what is going on with this? I have never had similar issues and I doubt that I broke the MCU during soldering...

[SofLit]

Removed the Pull-up on the reset?

The schematics don't show all the connectivity: USART1, IOs, SPI etc .. Is there something connected to the MCU? 

[MKowalski]

Yes, the pull-up is removed, and there is nothing else connected. 

[SofLit]

At this stage I cannot help you further.

You said: "supply is very stable 3.3V.."

Try to monitor your power supply again and find if there is a glitch (try to use a trigger).

Check if the power is the origin of that reset by checking that bit:

[MKowalski]

This is the moment of powering up the board with the reset button pressed and released after about 1s...
As You can see it seems not to be related to the power supply...

[SofLit]

Monitor PWR RST flag..

[MKowalski]

I'm not able to read any register because the board is in hardware reset condition. 

Interesting fact that if I force the reset line to VDD it operates fine...

[waclawek.jan]

This sounds to me like a hardware problem - bad solder joint on VDD/VSS, short from NRST to neighbouring pins/tracks...?

JW

[MKowalski]

Yes, I thought so too, but would it be that repeatable? Same problem with 3 different MCUs and on 3 PCBs?

Tomorrow I'll check in a workshop with more professional tools to make sure it's not my fault, but honestly, I hand-assembled hundreds of similar boards and never had such a problem, so I really doubt that now I messed it up 5 times, one after another, exactly the same way... But who knows? maybe....

[waclawek.jan]

You're right, with multiple pieces of hardware this is unlikely (unless hardware design error).

Have you tried to run some simple software e.g. a plain blinky?

Also, the circuit connected to PD0/PD2, does it account for the 5k1 pulldown mentioned in footnote 3 to pinout table in datasheet?

JW

[MKowalski]

I've assembled one another board and checked each pad of the MCU with a microscope to be 100% sure there are no issues with soldering. I also did not populate the LED so the PC13 does not drain too much current and this particular board is now operating fine for over an hour. So my simple conclusion is that maybe draining too much current from PC13 could break something inside the MCU, causing this issue with previous boards (each of them was first started with the LED on PC13 enabled). Or indeed some design issues ( like too much heat dissipation on GND pin, what I suspect is also possible after reviewing the layout ) make me mess up soldering so many times...

Anyway the board works fine no matter if there is additional pull-up on the reset pin or Vref+ connected to VDD or not. If the issue will not occur again till the end of the day I will consider this topic closed.

BTW Thanks for help @SofLit and @waclawek.jan