Standby mode in Cube project. Is SRAM really lost (STM32F042K6)

[Jim Seymour]

Posted on June 14, 2018 at 21:14

I have a project (built using CubeMX) that puts the STM32F042K6 in Standby mode by calling HAL_PWR_EnterSTANDBYMode().

It behaves well and does what I expect.  However, the documentation says 'SRAM and register contents are lost' in this mode.

I only noticed this after I'd written my code - and now I don't understand how it's working.  Does the HAL function somehow take care of this for me?  Or is SRAM really being preserved?  Or am I just getting lucky?

If I read that warning right, I'd need to re-run all the initialization routines when I come out of standby mode to make sure my ADC/UART/GPIO/etc. are set properly.  Is that right?

[waclawek.jan]

Posted on June 14, 2018 at 23:48

Do you debug? If so, then you have DBGMCU_CR.DBG_STANDBY=1 likely set by the debugger, and then you don't really go into STANDBY as the regulator for digital logic remains ON.

The reason for RAM/registers content loss is that in STANDBY the regulator is switched OFF, thus the logic is unpowered - that's the point, to reduce the static consumption and maintain power only for the absolutely necessary circuitry for the clock (RTC, IWDG) to survive and provide wakeup.

JW

[Jim Seymour]

Posted on June 15, 2018 at 00:04

No, I don't do any debugging - though I have 'Debug Serial Wire' enabled in my Cube project.

I'll try it without that setting to see if it makes a difference.

[Jim Seymour]

Posted on June 15, 2018 at 00:17

That didn't seem to make a difference.  I also specifically called HAL_DBGMCU_DisableDBGStandbyMode() which clears the bit in question  - but that also didn't seem to do anything.

My code 'seems' to go into Standby mode when I call 

HAL_PWR_EnterSTANDBYMode()

 - and we see a power drop from ~4 mA down to ~1mA.  But we expected a much bigger drop.