Grrr! Wasted hours...

[colorado.rob]

I just spent many hours chasing down a problem trying to get current on an STM32L4 chip down to 3uA in stop mode.  Once I initialized the RTC or UART, I could never get the current consumption below 250uA.  I thought at first it was a board/GPIO issue, but the RTC uses no GPIOs.  Even initializing just this one IP block caused the problem.

I decided to try this on a Nucleo64 dev board I had lying around and the same code just worked. 

I finally convinced myself that I could have damaged the MCU during the initial PCB power-on where I had a wiring mistake that took one of the power domains well below its 3V3 nominal value.  The weird thing is the rest of the chip worked flawlessly.  It was just this odd Stop mode current issue.  The MCU even drew the expected current in standby or shutdown mode.

I replaced the MCU and the code I was working on for two days just worked. 

[Mr. Scram]

This is why I generally order two or more of something. I love being able to do a sanity check with a second part. It typically shows me the problem isn't in the part, but in the designer.

[coppercone2]

or the programmer
or the compiler
or the cable
or the physical location (i had a mcu that kept randomly resetting when it was in a particular location, the thing was so bad I had to put a notebook ontop of the esd mat and it was fine, the things on isolated brass stand offs!)
or the angle (cracked socket)

[LapTop006]

Last week for me being convinced I'd killed a board due to overvoltage (the old HP supply it was hooked up to makes it *very* easy to add an extra few volts), then the next day while using that board as a reference noticed I'd simply failed to solder the data pins on the UART and that was why the board seemed dead.

[dmills]

I spent a good week chasing 'faulty relays' on a batch of production board at work.
Could NOT for the life of me reproduce on my bench, but the test department were having all kinds of issues...

The problem?

The test departments signal source had a honking great loudspeaker in **JUST** the wrong place, and they were resting the DUT on top of the signal generator causing the speaker magnet to hold the relay in! Oh how we laughed.

Regards, Dan. 

[Zucca]

well below its 3V3 nominal value.

So you can damage the MCU if you apply 1V instead of 3V3? I always thought the components damages are due to overvoltage not undervoltage situations. Wired...
Maybe there were others MCU power rails on, while one was down.... 

[xaxaxa]

looks like esd damage to me