IMXRT1176 SNVS LP RTC CoinCell Battery consumes very quickly

[burhanhagi]

I have a circuit in below:

burhanhagi_0-1739880021261.png

Below is also the schematic part of IMXRT1176 SNVS blog.

burhanhagi_1-1739880083842.png

I am inserting a CoinCell battery for RTC running when power is OFF. 

I would think that Coincell battery will keep supplying SNVS LP RTC at least for 6-8 months while Power is completely OFF. However, after we collected a product, then insert a CoinCell Battery and then put that product into the shelve; our Coincell battery consumes all its energy and exhausts within 4-6 days long. After that RTC gets zero, not keeping its datetime anymore.

So, in my design with STM32F7 microcontroller, it draws nearly 800nA current form Coin cell Battery while power is gone out. So it consumes all its energy within 8 -10 months long. But with IMXRT1176 product, it is much far away from that. 

I measured the current from CoinCell with IMXRT1176. It draws roughly 35-40uA current from Cell when Power is OFF. As you know, this is so much current.

So what is my wrong? How do I set this current to a range less than 1-2uA ? Do I have to add some code to my below BSP part? Waits your helps. Thanks

burhanhagi_2-1739880724708.png

[Gavin_Jia]

Hi @burhanhagi ,

Thanks for your interest in NXP MIMXRT series! 

When measuring VDD_SNVS_IN the current consumption should be typically around these values (depending on the temperature). 

Gavin_Jia_0-1740024369573.png

If the consumption is similar to these specifications, it's likely that the issue lies elsewhere on the custom board. At what point did you measure the data?

Also, we can rule out software effects by testing if we are in SNVS mode. 

You can ensure SNVS mode by lookin into the PMIC_ON_REQ pad. Before entering SNVS mode, or upon exit, this pin should be HIGH. During SNVS mode, it will be LOW. For more information about this, look into Section "25.3.1 Reset and Power Scheme" and "25.3.2 Finite State Machine (FSM)" of the RM.

After eliminating the influence of software, we can check the hardware circuit together.

Best regards,
Gavin

[burhanhagi]

Hi @Gavin_Jia,

I noticed that I had not put imxRT1170 into SNVS mode so far. 
I overviewed the example dcdc switch power mode in SDK. So, seemingly I need to add some codes into my software. It may later be meaningful to examine the Hardware consumption from Coin Cell. 

However, according to your table in your post,  I got a little bit unwilling whether working and implementing it into my software or not. Because table shows min current 3.8uA consumption while only RTC is active&running and in SNVS mode. 

Please I need confirmation that when I implement every required steps for SNVS mode, Will 3.8uA be min reachable target value? Or Is it possible to get a lower value such as 1uA or even smaller?

Thanks in advance

[Gavin_Jia]

One other thing, you might want to check out this AN: https://www.nxp.com/docs/en/application-note/AN13104.pdf

A customer has effectively reduced the current by integrating section 4.7 method in their project: 
4.7 SNVS_XX pin leakage

[burhanhagi]

@Gavin_Jia , I performed tests on eval board. 

The current dropped till 3.4 - 3.5uA levels when measured over R419 resistor according to the schematics SPF-32171_C3.pdf while IMXRT1170 switches into SNVS mode and RTC is active and running.

I added the line in below in AN13104.pdf to see if getting lower of current possible. However, it does not affect the current anyway. 

IOMUXC_SNVS_GPR->GPR37 |= IOMUXC_SNVS_GPR_GPR37_SNVS_TAMPER_PUE_MASK;

So, if anything more improvements can't be made to drop current to a level as low as 1uA, I will close the ticket.