Modem 30 minute lockup

[Burt]

This is driving me totally nuts. There are inconsistent pieces of information in documents and code. I am getting a modem reset loop problem. The documentation says the timer does not run when the modem has no power. That tells me I have to leave power on in order to get out of this ridiculous mess that cuts me off from any and all LTE-M and NB-IoT networks. The message from some asset tracker code tells me I can get out by doing a power reset, but that does not work.

How in blazes do I get out of this fricking mess????? This is at least one day down the drain.

How about 5 minutes rather than 30 minutes. How about something that simply restores the modem?

Did I put a bug into the code? I can't imagine how, although I do put in printfs.

Thanks.

Burt S.

[Burt]

Another thing I find a bit strange: in the pdn library, they enable CNEC_ESM events to appear in the log but not CNEC_EMM events. Here is what I get if I enable simple printing of EMM events:

+CEREG: 2,"F460","00F536C9",9
[00:00:05.461,456] <dbg> modem_module: lte_evt_handler: LTE cell changed: Cell ID: 16070345, Tracking area: 62560
[00:00:05.461,608] <inf> app_event_manager: MODEM_EVT_LTE_CELL_UPDATE
+CSCON: 1
[00:00:05.527,862] <dbg> modem_module: lte_evt_handler: RRC mode: Connected
+CNEC_EMM: 6

+CNEC_EMM: 6
+CEREG: 0
[00:00:06.663,574] <dbg> modem_module: lte_evt_handler: LTE cell changed: Cell ID: -1, Tracking area: -1

You see, even if I gray out the debug level messages, I mean, if you imagine they are not present, I would have seen really quickly that an EMM 6 happened. A quick perusal of the AT Commands Reference and following their suggestion to look it up in 3GPP 24.301 would show me quickly that using this SIM, the network is rejecting my IMEI (turns out KORE Wireless support informed me that Twilio only accepts the TAC for old B0 devices; I am using B1).

Burt

[Øyvind]

Hi Burt, I apologize for not being fully switched on, you are right in regards to the EMM events. I was too focused on the modem reset issue that I overlooked these messages in the modem trace. 

I'm not familiar with the Twilio SIM, I think you will need to reach out to your local sales rep for more information on this. The SIM must support LTE-M and NB-IoT, but from what I understand only US based NB-IoT is supported.

[Burt]

Thanks, Oyvind. Yeah, I guess it's unfortunate that I opened the ticket with the "30-minute lockup" title. This was only one issue that arose while trying to work with Twilio SIM cards.

I think I had a few extra questions for you. The one that sticks out in my mind: will AT%XFACTORYRESET=0 wipe out my certificates? In case I want to use it for clearing a reset loop, I want to know how much pain to expect.

Oh, more questions, I saw the NS white paper about the reset loop restriction. I found that they did not define clearly what is meant by modem initialization and modem reset. I usually think of modem initialization in terms of the software (lte_lc.c). Reset refers to power reset, soft reset, WDT reset, and that class of events? Or is there something that can reset the modem that happens independently of a full nRF9160 reset and I won't see in the log? [Maybe dumb questions but because I cannot see when the counter is climbing up to 7 I don't know what got me there.]

Thanks, Oyvind.

Burt

[Achim Kraus]

Sometimes long network searches may be caused by multi-imsi-SIM-cards with auto-switching. You may check that by reading the IMSI (AT+CIMI) frequently, e.g. every minute. If it changes, the modem usually restarts the network search. It's only one cause, but sometimes a nasty one. Even if it doesn't apply in your case, it's good to know it.

[Burt]

Thanks, Achim, yes, this is good to know. If you have examples of multi-imsi-SIM-cards with auto-switching, please let me know that. We use a Twilio Super SIM which is multi-imsi; I guess it is auto-switching, but I don't know when I should expect the imsi to auto-switch. We also use Soracom SIMs; I don't know whether or not those are multi-imsi.