STM32F0x1 Sleep Mode Conundrum

Is the SLEEPONEXIT bit set by chance?

I am not setting it, but I'll check its state the next time I'm in the room with the board and report back here. I assume it should be zero -- yes?

I have had success with the 152 part using ST's library function:

PWR_EnterSTOPMode( PWR_Regulator_LowPower, PWR_SLEEPEntry_WFI );

The 152 is an M3 part, your 303 is an M4, so there may be differences on the 051, which is an M0. Is there an equivalent library function that you could use as a reference for the 051? Ultimately it comes down to calling the WFI instruction within the library function, but it may include some chip specific code prior to the WFI instruction itself.

If it helps the low power specific parts that use the above reference library function can be found in the source file STM32L_low_power_tick_management.c of the demo documented on the following page:

Regards, Richard.

• formatting link
  Designed for microcontrollers. More than 107000 downloads in 2013.

• formatting link
  IoT, Trace, Certification, FAT FS, TCP/IP, Training, and more...

demo.html

I found their STM32F0xx library, and looked at the function you mention

-- all it does is to clear or set SLEEPDEEP depending on whether you want sleep or stop, then call your choice of WFI or WFE.

So, no dice. But at least I know that I'm doing what ST intends.

How odd. It seems to be connected with having the system tick clock enabled. I had inadvertently left in some code that runs that.

My power consumption is still about two orders of magnitude above what I want, but I think that's a board problem, not the processor.

And for some strange reason I need two in a row. And not WFE followed by a NOP -- it wants two WFE instructions in a row.

Ah, the ways of the ARM are mysterious.

Tim Wescott wrote: ...

Any floating lines? They draw much unwanted current.

I found some, and figured out the magic finger-ring combination to go really to sleep. I'm down to 250uA at this point, but I'd like to get lower still.

errata says something about excess current for analog pins on certain packages

"Additional current consumption in the range of tens of ?A per pin can be observed if VDDA is higher than VDDIO"

Already noted, not an issue. I am using the ADCs, but VDDA is tied do VDDIO. The application is such that the measurement process goes 1: everything stops, 2: the ADCs get sampled.

Come on, tell us what is needed!

-- Uwe Bonnes snipped-for-privacy@elektron.ikp.physik.tu-darmstadt.de

Institut fuer Kernphysik Schlossgartenstrasse 9 64289 Darmstadt

--------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------

It's in another post -- I had the systick clock running, which apparently preempts trying to turn off the core.

And for some bizzare reason I need to call WFE twice in a row. I don't know why -- but it's working now.