Power Consumption in Low Power Mode nearly 100% higher than expected

[ARich.6]

Hello,

I've tested a few weeks now without solution, hopefully somebody can help me with this topic.

I try to reach the Low Power mode (LP Run or LP Sleep), but when I enter it, the consumption is still at 150µA (~270µA in debug mode).

I even tested with a simplified initialization without any peripherals except a EXTI to wake up from interrupt.

Do I miss something essentially? Why don't I come down to <10µA?

Code example:

int main()
{
...
	while(1)
	{
		LowPower_EnterLowPower();
		if(wakeup)
		{
			wakeup = false;
			LowPower_ExitLowPower();
			// just some dummy alive testing:
			HAL_GPIO_TogglePin(DEBUG_PIN_OUT_GPIO_Port, DEBUG_PIN_OUT_Pin);
			HAL_Delay(3000);
			HAL_GPIO_TogglePin(DEBUG_PIN_OUT_GPIO_Port, DEBUG_PIN_OUT_Pin);
		}
	}
}
 
void EXTI_Btn()
{
	wakeup = true;
	__HAL_GPIO_EXTI_CLEAR_FLAG(GPIO_PIN_7)
}
 
void LowPower_ExitLowPower()
{
	SystemClock_InitConfig();
	HAL_ResumeTick();
}
 
void LowPower_EnterLowPower()
{
	__HAL_RCC_GPIOC_CLK_DISABLE();
	__HAL_RCC_GPIOB_CLK_DISABLE();
	__HAL_RCC_GPIOA_CLK_DISABLE();
	__HAL_RCC_GPIOD_CLK_DISABLE();
	__HAL_RCC_GPIOH_CLK_DISABLE();
	HAL_SuspendTick();
	__HAL_PWR_CLEAR_FLAG(PWR_FLAG_WU);
	HAL_PWR_EnterSTOPMode(PWR_LOWPOWERREGULATOR_ON, PWR_STOPENTRY_WFI);
}

Used MCU STM32L05, Clock config HSI16 at 4MHz, 32.768kHz LSE for LP; measured with STM32CubeMonitor-Power and X-Nucleo-LPMO1A, verified with Multimeter and R&S Power Supply; (In run mode/wake up: ~1,5mA during button pressed, 1,2mA during HAL_DELAY)

Cube IDE 1.6.1, FW_L0 V1.12.0

[ARich.6]

Additional information:

• Problem remains in Release build configuration (with Power Reset, no debugger connected).
• No big difference and power levels between 3V - 3,3V
• Power consumption only best-case at 150µA, sometimes higher up to 380µA under same test condition, same code, power reset, LP Monitor 3V-3,3V, etc. (seems like a general offset, because also consumption is higher in same value in run mode ~1,66mA instead of ~1,5mA)

[Uwe Bonnes]

Test on another chip to be sure no non-fatal damage to the chip has happened.

[ARich.6]

Thank you Uwe Bonnes for the tip. Unfortunately other chips behave similar. I tested now with 6, one of them really seem to a defect (doesn't go to sleep mode at all).

[ARich.6]

Hello again,

I have to admit I've done some mistakes. Maybe the summary is helpful for others. But still there must be some improvement regarding the specs:

It's not enough to unplug the Debugger USB cable. All cables (espacially VDD Pin1) must be unplugged, otherwise the debugger drains ~200µA.

For my new test the lowest level I reached 110µA (no Debugger, only supply).

However these 110µA still seemed a bit high for LP. So I really checked that there is no HAL initialization at all, expect the EXTI pin and all other pins as analog inputs. Now it went down to 30µA. Datasheet describes <10µA for LP modes though.

I was hoping that the entry of LP mode would stop some peripherals automatically. But it seems that I have to make a complete Uninit (entering LP) and Init again (leaving LP). Let's see how consumption drops/rises if I use LPTIM or RTC to wake up.

Still some tips, why i don't reach <10µA with the current setting?