LPN : periodic peaks

[Sameer]

Hello,

I am working on LPN node code from the mesh SDK 3.2.0. When I run the example code, I see a current peak at every 2 sec interval. Please check the image attached. 

I tried changing the poll_interval_ms and other parameters as well but these periodic peaks are always there with 2 sec period. Can anyone please help me understand what these peaks are?? Can I change this interval??

Regards,

Sameer 

[Edvin]

Hello? Where, and in what project (lpn or the friend node) did you try to change poll_interval_ms? And in what function is that poll_interval_ms used? And what did that function call return?

The spikes in your graph look very much like radio spikes, so you are on the right track, but it doesn't look like the polling interval is actually changed.

Best regards,

Edvin

[hwswti_dev]

Hello Edvin, hello Sameer,

I can confirm this behavior and add additional information as I have measured different polling intervals. Please see the measurement below with 30 second polling interval in an established friendship. During each sleep cycle a message for the LPN arrives at the friend node and gets queued up. The polling exchange happens after the 30 seconds of sleep are over and can also be seen in my sniffer. Here is the Power Profiler App Measurement:

However there still is the periodic two second spike in every of my measurements during sleep:

My first guess was it could be a heartbeat message, but that interval seems to be set way higher in health_server.c ( #define ATTENTION_TIMER_INTERVAL 1000000u ). It seems to be some sort of GATT proxy advertising for the nRF Mesh App, because I can see the LPN when I try to reconnect to the network in the App, even if it is asleep. Although I saw none of that advertising in the sniffer.

I have a second issue with the LPN example, which is described here: https://devzone.nordicsemi.com/f/nordic-q-a/42473/mesh-lpn-rx-race-condition/165496#165496 .

So in the test setup with two light switch servers and one LPN, the friendship is terminated with reason NRF_MESH_EVT_FRIENDSHIP_TERMINATED_REASON_NO_REPLY when 5 or more poll exchanges need to happen during polling. Increasing MESH_LPN_POLL_RETRY_COUNT has also not solved the issue, is there any news on this?

Also I want to address another issue. I bought some nRF6707 Power Profiler Kits and was surprised to see that there is no logging function like discussed here devzone.nordicsemi.com/.../exporting-data-from-power-profiler-v2-0-0 since it is mentioned as 'Export measurement data for post-processing' in the Product Brief ( https://www.nordicsemi.com/-/media/Software-and-other-downloads/Product-Briefs/PowerProfilerKit-product-brief.pdf?la=en&hash=81DB9B626015BEB68F135F23A3F8CB9A3804F4BA ). I tried using version 1.1.0, but when I measuring for an hour the last log entry was at 470 seconds, so something seems to not work properly. Is there an update when logging will be implemented in the Power Profiler App?

Best regards,

Florian