Confusion about understanding PPK observations

[pd@learner]

I have custom board  with nrf 52832. I am using SDK 15.2.0  and softdevice s132 V6.1.0  to transmit advertising packets periodically. I have written simple application that reads battery status and transmitted it in manufacturing data  of advertising packet 

By referring online Power profile   I got average current 20 microAmp for following configuration.

Tx Power = 0 dBm

Tx Payload = 27 bytes

BLE Interval = 1000 msec

DCDC = No.

sensor = lis2dh12

Power Profiler Kit nRF6707 is used for average current calculation. PPk setup as follows

SW2 = External

SW4 = Reg

COM = DK

 I have captured 2 images with different average current for above configuration I am little bit confused which value i will consider for calculation.

a)

b)

kindly remarks on this. These remarks will be valuable for me.

[Sigurd]

Hi,

The uploaded images are very small, so it's not possible to see the details. Could you upload some larger images?

[pd@learner]

a)

b)

[Sigurd]

pd@learner said:

In image there is value below 04:20 which is different from value below 04:21. What does it indicate?

Sigurd said:

In picture a) you seem to have a small peak right before 04:21,

 Could you measure how high the peak goes ?

What Low frequency clock source are you using for the Softdevice ? (What is the value of NRF_SDH_CLOCK_LF_SRC in sdk_config.h)

[pd@learner]

NRF_CLOCK_LF_SRC_XTAL clock source used for softdevice.  one modification I have made in 

nrf_sdh_enable_request functionas follows.

nrf_clock_lf_cfg_t const clock_lf_cfg =
{

.source = NRF_CLOCK_LF_SRC_RC,
.rc_ctiv = 16,
.rc_temp_ctiv = 2, // 0
.accuracy = NRF_CLOCK_LF_ACCURACY_20_PPM //0
};

[Sigurd]

Hi,

pd@learner said:

NRF_CLOCK_LF_SRC_XTAL clock source used for softdevice.

 Then if you see this peak at a regular interval, it could simply be the LF RC clock calibration you are seeing.

 

pd@learner said:

.accuracy = NRF_CLOCK_LF_ACCURACY_20_PPM //0

You should set this to 1 (i.e. NRF_CLOCK_LF_ACCURACY_500_PPM) when using NRF_CLOCK_LF_SRC_RC.

[pd@learner]

I have not seen that at regular interval.  But in this thread https://devzone.nordicsemi.com/f/nordic-q-a/15363/change-clock-source-nrf52832-app-beacon-example  I have found that xtal_accuracy is ignored for RC. Also in this thread https://devzone.nordicsemi.com/f/nordic-q-a/54286/what-are-the-three-source-differences-of-low-frequency-clock-precision-comparison-power-consumption-comparison/219814#219814  I have found that the LFRC clock can not be more accurate than 250 PPM. so which value I will use? Does it depend on hardware?

[Sigurd]

pd@learner said:

I have found that xtal_accuracy is ignored for RC.

This behavior was changed starting from s132_nrf52_5.0.0-2.alpha. From release notes:

"The RC oscillator accuracy can now be set to any of the defined NRF_CLOCK_LF_ACCURACY values, and there is no default anymore. In other words, the nrf_clock_lf_cfg_t::accuracy
parameter now has the same functionality when used with the RCOSC clock source as with the XTAL clock source"

pd@learner said:

so which value I will use? Does it depend on hardware?

Due to errata 146, I recommend using 500 ppm when using the LF RC oscillator.

[pd@learner]

I am not using external 32.768 KHz crystal on custom board. I want to know that following settings will be correct or not?

NRF_CLOCK_LF_SRC_XTAL clock source used for softdevice.  

nrf_clock_lf_cfg_t const clock_lf_cfg =
{

.source = NRF_CLOCK_LF_SRC_RC,
.rc_ctiv = 16,
.rc_temp_ctiv = 2, // 0
.accuracy = NRF_CLOCK_LF_ACCURACY_500_PPM //0
};

[Sigurd]

pd@learner said:

I want to know that following settings will be correct or not?

Yes, the settings looks correct.