NRF54L15 BLE power consumption spikes

[Maya.]

Hi,

I am testing the BLE beacon sample on the NRF54L15 PDK, and it was successful as I was able to detect the beacon using the NRF Connect scanner. However, when monitoring the power consumption of this sample, I noticed additional power spikes beyond the expected advertising spikes.

They appear to happen at intervals of around 4 seconds. These spikes are seen when config_bt setting is enabled

The advertising intrevel is set to 10seconds.

using sdk version v2.8.0
toolchain version v2.8.0

Could you help me understand why these spikes are occurring and suggest ways to eliminate them?
thanks in advance for your help

[Håkon Alseth]

Hi,

 

Looks like there is some CPU activity that is occurring. Could possibly be due to the logging being enabled? Try setting CONFIG_SERIAL=n, to see if the logger thread is the reason why there's spikes.

 

Kind regards,

Håkon

[Maya.]

Hi Håkon,

I have already disabled the logs before checking the power consumption. The image provided above is with the logs disabled.
Below is my current prj.conf file, in case it helps..

CONFIG_BT=y
CONFIG_LOG=n
CONFIG_BT_DEVICE_NAME="Test beacon"

I just tested by also setting CONFIG_SERIAL=n in the .conf file, but I'm still seeing the spikes

[Håkon Alseth]

Hi,

 

I see the same spikes in my testing.

The peaks are occurring due to this errata: https://docs.nordicsemi.com/bundle/errata_nRF54L15_EngB/page/ERR/nRF54L15/EngineeringB/latest/anomaly_L15_30.html

 

When the system is running on HFINT (ie. high frequency internal RC oscillator), the GRTC needs to be calibrated in an interval. This will happen every CONFIG_MPSL_CALIBRATION_PERIOD, which is default to 4000 ms.

 

Kind regards,

Håkon

[secondbest]

Hello,
I have similar problem. https://devzone.nordicsemi.com/f/nordic-q-a/116897/nrf54l15-dk-sleep-current-problem

I have brand new DK not PDK and I am seeing same problem. How can I solve this problem?

Thanks.

[Håkon Alseth]

Hi,

 

Please see my former answer. This is an expected current peak due to the workaround in the below linked erratum.

 

Kind regards,

Håkon

[secondbest]

Okay. Got it. Actually, I did not see this type of peaks at Nordic's Youtube demo and I think there is a way to remove it. I feel there is no way. I have one more problem. If I apply 3V supply voltage, I am getting strange sleep current. If you have time, please check my topic. Thanks.

[Emil Lenngren]

Can you give a code sample (or high level idea) how the workaround should be done in case the MPSL is not used? The GRTC peripheral doesn't seem to have any calibration functionality according to the data sheet?

Is the workaround needed if the GRTC runs on the LFXO?

[Håkon Alseth]

Hi,

 

Emil Lenngren said:

Is the workaround needed if the GRTC runs on the LFXO?

Yes, it is needed in cases where GRTC is used and HFCLK is sourced from HFINT.

Emil Lenngren said:

Can you give a code sample (or high level idea) how the workaround should be done in case the MPSL is not used? The GRTC peripheral doesn't seem to have any calibration functionality according to the data sheet?

this is currently being worked on and evaluated, and a non-MPSL implementation is scheduled. Unfortunately, I do not have any more detailed information to share at this point in time.

 

Kind regards,

Håkon