NUCLEO-U575ZI-Q excessive current consumption

[Dadigno]

I'm using the NUCLEO-U575ZI-Q board with the PWR_ModeSelection example, and I'm measuring the current consumption with an ammeter through the JP5 (IDD) connector.

The current I measure in the different low power modes is always around 4.52 mA, which does not match the values stated in the datasheet.

I'm probably making a configuration mistake. To use the example, I followed the README included in the SDK.
As toolchain, I'm using CMake and VSCode as the IDE.

Where am I going wrong?

[Andrew Neil]

To measure really low power modes, you need to ensure that all sources of current leakage are disconnected - in particular, the ST-Link.

Unfortunately, the NUCLEO-U575ZI-Q  doesn't have the break-off ST-Link section - which is the easiest way to do that.

So you will have to thoroughly study both the User Manual and the schematics, and make all the required disconnections.

Note that just the User Manual may not be sufficient; eg, see:

Excess Current Consumption Running B-L072Z-LRWAN1 From Battery??

The User Manual can be found on the 'Documentation' tab of the board's Product Page:

https://www.st.com/en/evaluation-tools/nucleo-u575zi-q.html#documentation

The Schematics should be available on the 'CAD Resources' tab:

https://www.st.com/en/evaluation-tools/nucleo-u575zi-q.html#cad-resources

#NucleoLowPower #LowPower

[Dadigno]

@Andrew Neil 

I tested the PWR_EnterStopMode firmware, but the current consumption didn’t go below 500 µA.
So I modified the example code by disabling the SysTick,

  /* Switch OFF LED1 */
  LED_Off();

  SysTick->CTRL &= ~SysTick_CTRL_ENABLE_Msk;

  /* Enter STOP 0 mode */
  EnterSTOP0Mode();

Then I changed the EnterSTOP0Mode function to call LL_PWR_SetPowerMode(LL_PWR_STOP2_MODE) since STOP2 mode is the one I’m interested in.

void EnterSTOP0Mode(void)
{
  /** Request to enter STOP 0 mode
    * Following procedure describe in STM32U5xx Reference Manual
    * See PWR part, section Low-power modes, STOP 0 mode
    */
  /* Set STOP 0 mode when CPU enters deepsleep */
  LL_PWR_SetPowerMode(LL_PWR_STOP2_MODE);

  /* Set SLEEPDEEP bit of Cortex System Control Register */
  LL_LPM_EnableDeepSleep();

  /* Request Wait For Interrupt */
  __WFI();
}

Now the current drops to around 10 µA, which is still too high compared to the ~3.9 µA stated in the datasheet.

Configuration:

• SRAM2 page1 & page2 and ICACHE retained
• RTC disabled
• Power supply via JP5 using STLINK-PWR 3.3 V
  
  

I think it’s the same issue I mentioned in my other post, where I’m using a custom board with an STM32U595 MCU

https://community.st.com/t5/stm32-mcus-products/unexpected-current-consumption-in-stop2-mode-on-stm32u595/m-p/847317#M287272 

[Andrew Neil]

Yes, you will need to disable SysTick - otherwise it will keep waking the CPU!

---

@Dadigno wrote:

Now the current drops to around 10 µA, which is still too high compared to the ~3.9 µA stated in the datasheet. 

---

So you're chasing ~ 6 µA.

At this level, you have to be really careful about all possible sources of leakage - capacitors can become significant!

Have you carefully checked the schematics?

Are you sure that the ST-Link is completely disconnected?

[Dadigno]

I’m currently checking every possible connection.

However, I’m wondering why design a dev board for the STM32U5xx, a mcu specifically intended for low-power applications, and then make it so difficult to properly disconnect the ST-LINK to measure true low-power consumption? -.-

[Andrew Neil]

---

@Dadigno wrote:

I’m wondering why design a dev board for the STM32U5xx, a mcu specifically intended for low-power applications, and then make it so difficult to properly disconnect the ST-LINK to measure true low-power consumption? -.-

---

Indeed.

I made a similar comment in the thread I linked earlier.

[Dadigno]

Update: I haven’t modified the board yet, but with the following configuration I managed to reach an average current consumption of 6.2 µA, compared to the 3.9 µA stated in the datasheet:

• SRAM2 page1 & page2 and ICACHE retained
• RTC disabled
• Power supply via JP5 using STLINK-PWR 3.3 V
• USB connector CN1 not plugged in (~3 µA difference)  <- Thats makes the difference
• Perform a complete power cycle (~200 µA difference) <- I think related to this 
• STOP2 power mode
• SysTick suspended

Regarding the connections, from what I can tell by looking at the schematic, it seems that only the SWD pins (SWDIO, SWCLK, SWO) and VCP connect the U575 to the ST-LINK through 47 Ω resistors: R3, R10, R11, R12, R13. But in theory they shouldn't cause any issues...

[Andrew Neil]

You didn't mention setting the GPIOs ?

---

@Dadigno wrote:

But in theory they shouldn't cause any issues...

---

Yes, they will!

The ST-Link STM32 is not powered, so you will be leaking some power back into it.

In most cases, that's not noticeable - but, when you're worrying about ~ 2 µA, they become significant !

There will also be a connection to the ST-Link to measure V_Target ...

[Dadigno]

---

@Andrew Neil wrote:

You didn't mention setting the GPIOs ?

---

GPIOs are set to analog  ( also SWD )

---

There will also be a connection to the ST-Link to measure V_Target ..

---

When JP5 is removed and VDD_MCU is used as power ( UM2861, 7.4.1 VDD power supply input ) then Vtarget (VDD) is no connected to the stlink 

[Dadigno]

I tried to remove every possible connection I found between the STM32U5 and the ST-Link, so I removed: SB32, SB34, SB42, SB44, SB46, SB45, SB6, R11, and R12 (SWCLK, SWDIO). Unfortunately, this made no difference in the current measurements, which remain around ~6 µA on average.