STM32U535: high current consumption on VBat pin then the battery is discharged

[hdmi87]

In my device based on STM32U535, I use the internal RTC with LSE with external 32.768KHz crystal as the clock source. The VBat pin is powered with an external rechargeable battery. The charging is done externally, and the internal VBat charger is not activated (VBE bit of PWR->BDCR2 register is not set).

When the voltage on VBat pin is in the working range (above approx.1.65 V, which is the minimal VBat voltage according to the datasheet), the current consumption in VBat circuit is about 0.7..0.8 uA. But when the battery is discharged below this threshold, the current consumption suddenly increases up to 8 microampers, and then is slowly decreasing while the battery is being discharged and the voltage is decreasing. And if I power up the MCU again after that, the battery domain is reset and the time is lost.

Why does the high VBat current occur? Is it described somewhere in the datasheet? Could it be caused of some misconfiguration of the battery domain?

[AScha.3]

>Is it described somewhere in the datasheet?

You see the limits for "normal" working cpu in ds ?  If you dont respect this, anything could happen, at too high or low voltage. 

>Why does the high VBat current occur?

I assume, here the brownout/reset happens -> battery domain is reset .

And you should be glad, it stopping/reset at its minimum working voltage .

Otherwise it would/could do anything at voltages, that are too low for correct working cmos gates

and if raising the voltage then, RTC time..etc could be at any random value - THIS would be bad. Right ? 

[hdmi87]

You are right that the Vbat voltage is outside of the normal working range, but it's hard to agree that anything could happen in this condition (like in C, if you read value of a variable before its initialization, you may get demons flying out of your nose).

If it is so, who can guarantee that the current is limited by 10 microampers and not, say, 10 milliampers? If there is no limit of this current at all, in theory too high discharging current can damage the rechargeable battery. I think it should be some value that the current does not exceed, or at least it should be specified in the datasheet that the current should be limited externally (e.g. by a series resistor).

Also, if the battery domain was not previously initialized and a low voltage (1 - 1.5 volts) is applied to VBat pin, the current is less than 100 nA. Why doesn't the battery domain enter the same state after brown out reset?