RRC mode / Are there debugging options to find out why RRC mode changed to connected

Is there a way to find you the reason for RRC Connect?

I'm using a simple software developed based on NCS AWS IoT sample. I've implemented the SW to test power consumption in different modem idle modes. The SW updates data to AWS periodically and is idle between update periods. For some reason the SW operates differently using different networks and I don't know why.

I had trouble with PSM earlier as I mentioned in this ticket: https://devzone.nordicsemi.com/f/nordic-q-a/80120/asset-tracker-v2-power-consumption-with-thingy-91

Now I know better and I know how to set appropriate PSM parameters that are accepted by network. The thing that I don't understand is why my device constantly transitions between RRC idle and RRC connected randomly when the device is in DNA network. The same RRC mode toggling is not present in Elisa or Telia networks, in those networks the idle modes work as expected.

In DNA network the device very seldom stays in RRC idle for 60 seconds which is the shortest possible active timer value granted by the network. As a result the modem usually won't go to PSM during the 20 minutes that is the keepalive time for AWS IoT MQTT.

If the modem goes to PSM in DNA network, it stays in PSM until the next periodic data update. Thus I'm guessing the RRC toggling is network (not AWS) originated, but I would like to know if there is something that I can do to prevent the toggling.

I've attached a log file with 2min data update interval. In the beginning of the log there are only application writes. After a reboot there are library debug logs with the following flags:

# debugging
CONFIG_LTE_LC_MODEM_SLEEP_NOTIFICATIONS=y
CONFIG_AWS_IOT_LOG_LEVEL_DBG=y
CONFIG_LTE_LINK_CONTROL_LOG_LEVEL_DBG=y
CONFIG_NET_LOG=y
CONFIG_MQTT_LOG_LEVEL_DBG=y
CONFIG_DATE_TIME_THREAD_STACK_SIZE=2048
CONFIG_DATE_TIME_LOG_LEVEL_DBG=y
  • Hi IIKKA,

    I am waiting for the reply from some LTE RRC protocol expert, in the mean time, I want to confirm something that may related, are you still testing with the following configuration?

    - Thingy:91 (1.4.0 & MFW 1.3.0)
    - Asset_tracker_v2 (low_power_conf + GPS disabled) under NCS1.6.0

    Are you able to verify with MFW1.3.1+NCS1.9.1? Just share another recording with DNA network if it sitll have same issue.

    Best regards,

    Charlie

  • Hi Charlie,

    Thank you for the reply. Are you referring to the ticket https://devzone.nordicsemi.com/f/nordic-q-a/80120/asset-tracker-v2-power-consumption-with-thingy-91 or would you like me to build Asset_tracker_v2 (NCS 1.9.1) with modem traces to test this issue?

    My tests at the moment are with a more purpose built FW. I have Event Manager based implementation like the Asset Tracker v2, but the FW only implements AWS shadow communication. With the test FW I can change PSM/eDRX parameters and the FW reports network response back to shadow. In addition I can turn on/off uart logs using shadow. I have done the implementation to test different configurations while the HW is connected to Power Profiler Kit II in source mode.

    The test FW is built using NCS 1.9.0 and I'm using MFW 1.3.1. I have nRF9160DK with B1 version of nRF9160 SIP and Thingy:91 v1.4.0 and Circuitdojo nRF9160 Feather V4. These two both have nRF9160 SIP B0 version.

    I don't see any functional difference between the devices in this test. All three devices have the same issue, in Telia network RRC idle is ok, but in DNA network RRC constantly changes to connected. 

    I am not using Asset_tracker_v2 from NCS 1.6.0 for testing anymore. I have implemented a FW that expands the asset_tracker_v2 functionality and uses also the nRF52 chip on Thingy:91. This FW is built using NCS 1.7.1. With this asset_tracker_v2 modification I can see similar issue with DNA network as I see with the simple test FW.

    Br,

    Ilkka

  • Hi IIkka,

    Thanks for the detailed explanation. I think I have collected enough information, I will let you know if there are any further findings.

    Best regards,

    Charlie

  • Hi Ilkka,

    I have an internal discussion with our modem developer. 

    After analyzing the recording you shared, we found there is Paging always in the log before PSM can be entered. Probably the reason is that the IP address in DNA SIM is public and there is a lot of polling coming in. The Network pages the modem and it just responds.

    The possible solution from the device side which I know is to set up IPv6 to avoid these Pagings from NW. As the source of the issue is from the network side, I suggest you contact DNA to get more information. Just turn back to us if you need any assistance.

    Best regards,

    Charlie

  • Thank you for the reply. I will share the information with DNA. Maybe they can change the visibility of the IP address.

    Br,

    Ilkka

Related