Shockburst; power consumption in deep sleep

[Svein]

Hope someone can bring some clarification to me.

I am studying the principle of Shock-burst and a part of that study is to go through every step in the examples of “esb_low_power_ptx_pca10056” and “esb_low_power_ prx_pca10056”. (SDK17.0.2)The code is running on a custom board built around nRF52840. 

Regarding “esb_low_power_ptx_pca10056” I can’t fully get the understanding of:
- Function “recover_state()”. What is the purpose and what it is doing?
- When calling function “system_off()” I measure an average power consumption of 4,2 micro Ampere. But when we run our BLE app on the same HW, the power consumption in deep sleep is 0,5 micro-Ampere. Power consumption measured used PPK2. Why is power consumption in deep sleep 10 times higher when running Shockburst.

Kind Regards

Svein

[Kenneth]

Hi,

You can find the intention behing the recover_state() in the documentation for the ESB example:
https://infocenter.nordicsemi.com/topic/sdk_nrf5_v17.1.0/esb_examples_low_power_ptx_prx.html

I don't have a good explaination directly why you don't achieve sub uA when going to system OFF, it's either some current though a GPIO and/or something related to the RAM retention I can see is done before calling NRF_POWER->SYSTEMOFF = 0x1.

Maybe you are able (for test) to move  NRF_POWER->SYSTEMOFF = 0x1; straight after gpio_init() for comparison?

Kenneth

[Svein]

Hi Kenneth,

Thank you very much for the quick and clarifying answer. Regarding power consumption, you were right that it was a GPIO that drew power. Had just completely forgotten that our design had a GPIO that behaved the opposite of what all the other GPIOs do :-O  (I am testing SB on a custom board designed for BLE 5.x)

BR,  Svein