Huge power consumption on RTC ofLPCXpresso54S018M EVB.

[jcxz]

I use are LPCXpresso54S018M EVB. I need the RTC in the LPC54018 to work after the main power supply is removed. I initialized and enabled the RTC in my firmware. I powered the VBATT line (J10:11) with a coin cell battery CR2032. While the main power is present, the VBATT line consumption is == 0. It's Ok. However, after removing the main power supply, the VBATT line consumption increases up to ~41 uA! According to the datasheet, consumption should be less than 1 uA. Why? How can I achieve current consumption of 1 µA or less?

[Harry_Zhang]

Hi @jcxz 

The ~41 µA is most likely leakage through the RESETN circuitry , not RTC consumption. When main power is removed, VBAT > VDD; If VBAT> VDD, the external reset pin must be floating to prevent high VBAT leakage.

The LPCXpresso54S018M-EVK connects RESETN to its pushbutton/debug circuitry, so the unmodified EVK does not satisfy that measurement condition.

To reduce the current:

Isolate the MCU’s RESETN pin from the EVK reset network.

BR

Harry

[jcxz]

I read about this requirement in the datasheet (If VBAT> VDD, the external reset pin must be floating to prevent high VBAT leakage).
But I don't understand how to achieve this?
I'm not talking about a EVB schematic. EVB - only for prototyping. But in our working device's circuit, MCU RESETN will be connected direct to the RST output of the external WDT-chip (TPS3823-33DBVR).
Can you recommend: How to correct the connection MCU.RESETN <-> WDT.RST connection's schematic so, that the VBAT current consumption does not exceed 1 µA?
I am attaching a schematic of the WDT <-> MCU.RESETN interconnection in my device:

In this picture, the wire named "Reset" is connected to the RESETN pin of the LPC54005JBD100.

[Harry_Zhang]

Hi @jcxz 

"Floating" here means that RESETN should have no external DC path during battery-only operation (VBAT present, VDD removed). The MCU's internal reset pull-up does not need to be disabled; the datasheet already accounts for that condition.

In the proposed watchdog circuit, RESETN remains connected to the unpowered TPS3823 output as well as the external RC network. Therefore, the datasheet condition of a floating RESETN pin is not met. This is a plausible explanation for the observed ~41 µA current, although the exact leakage path has not yet been experimentally verified.

To meet the datasheet condition, we recommend inserting a normally-open isolation switch between the watchdog reset output and the MCU RESETN pin. All external reset components (pull-up resistor, capacitor, watchdog output, etc.) should remain on the watchdog side of the switch, while the MCU side should not contain any external pull-up, pull-down, capacitor, or other DC connection.

The selected switch should explicitly support a low-leakage, high-impedance state when its supply voltage is removed.

BR

Harry