NRF52833 high power consumption after sys_poweroff()

[vasimv]

I'm trying to make a BLE beacon with Li-Ion battery and Qi wireless charging circuit. NRF52833 works in high-voltage mode (VDDH connected directly to the battery, internal LDO regulates VDD) There are a lot of problems with different charging ICs, so had to choose BQ25101Y chip which doesn't have battery over-discharge protection. My plan was to put NRF52833 into shutdown mode with sys_poweroff() when battery voltage is lower than 3.0V. But i've encountered weird problem.

When i call sys_poweroff() function in zephyr application at input battery voltage 3.7..4.2V, power consumption (measured with PPK2) is 0.6uA. But if i decrease input voltage below 3.6V - power consumption jumps to around 45uA at 3.5V (25uA at 2.1V). It doesn't change if i change REGOUT1 at all (stays same for internal LDO output at 1.8V or 3.3V).

This behaviour not unacceptable as in real device battery will have quite small capacity (20-50 mAh) and that consumption will over-discharge battery very fast. This is normal in high-voltage mode or i've screwed something? Below is full schematics for my beacon:

[JONATHAN LL]

Hi,


Some comments on the schematic:

You are missing a matching network for the antenna, you only have the radio matching. This is often not the best solution as in most cases some tuning for the antenna is needed for best performance. 


For the DEC3 you seem to have 0.1uF, same as for the VDD decupling, it should be 100pF. Not 100nF.

But if i decrease input voltage below 3.6V - power consumption jumps to around 45uA at 3.5V (25uA at 2.1V).

You have configured the device to operate in High voltage mode, so then you should never go below 2.5V. Its out of spec. 


As for the jump in value at around 3.5V. from 0.6uA to 45uA sounds a little strange. So can you provide a PPK2 file for measurements showing this?


Also do you see this on only one device or on multiple devices? How many have you tested?


Regards,
Jonathan

[vasimv]

JONATHAN LL said:

You are missing a matching network for the antenna, you only have the radio matching. This is often not the best solution as in most cases some tuning for the antenna is needed for best performance. 

Actually, have no idea how to make it, antenna's datasheet shows just 50ohm line (which i don't have space for). Looks like it works quite good as is.

JONATHAN LL said:

For the DEC3 you seem to have 0.1uF, same as for the VDD decupling, it should be 100pF. Not 100nF.

You're right, thanks. Also i've missed connection for VSS (37) to the ground (which doesn't affect much it seems, as MCU has internal connection. But i've soldered it manually anyway). Closest for the size i have right now is 470pF capacitor, will try to solder it.

JONATHAN LL said:

As for the jump in value at around 3.5V. from 0.6uA to 45uA sounds a little strange. So can you provide a PPK2 file for measurements showing this?


Also do you see this on only one device or on multiple devices? How many have you tested?

I've assembled four boards and all of them has the problem. I've removed BQ25101 chip on two to be sure it doesn't affect anything for no effect. Exact voltage threshold varies from 3.3V to 3.6V on different boards. Also, even worse, in active mode (when the beacon sends broadcasts) power consumption jumps to 450-500uA in some voltage range (2.9..3.4V or so). I've tried different REGOUT0 settings, disabled POFCON but nothing changes.

When i limit voltage to 3.6V and connect VDDH and VDD to switch to low-voltage mode, everything works fine, without any consumption jumps) until below 1.4V. Probably something wrong with brown-out stuff in high voltage mode but i don't see any way to change its settings. I'm thinking about switching to BQ25121A as charging+switching regulator chip and use NRF52833 in low-voltage mode only.

Here is screenshots from PPK2 (in logarithmic Y scale), there are two sets - one is in active mode, another after sys_poweroff().

PPK2 screenshots

[JONATHAN LL]

Thanks for all the extra details here. 

Difficult to say what is the issue here, but to continue to investigate I would start removing components and also try and supply the nRF52833 directly bypassing the BQ25101 and BQ51013.  

Does the device connect to any other peripherals that has a different voltage supply or operate at a different level ?  Or could there be any other way that the nRF is connected to different voltages? Like is one of the GPIO's connected to VDDH or something similar? 

Is the SWD pins connected to anything external with a different voltage?  


Schematic looks ok. 

vasimv said:

, everything works fine, without any consumption jumps) until below 1.4V.

that's expected, so nothing strange with the normal voltage mode. min operational voltage is 1.7V.  For High voltage mode it should be at 2.5V where you would see it BOR. 


I would try to cut\disconnect the VDDH trace from everything else and see if supplying only directly to the nRF52 if it would behave the same or not, though I do think that is close to  what you have tested with removing BQ25101.



Regards,
Jonathan