AT+COPS=0 behavior

[Vincent44]

Hello Nordic Team,


I have devices running nRF9160, modem firmware 1.3.6, tied with MNVO SIM card that provide access to multiple operators.


Providing an example seems to be best option to illustrate my question.

Let's say there are 3 PLMN available for a given device :
- AAA, with good signal quality
- BBB, with medium signal quality
- CCC, with bad signal quality 


In this case, I want to force my device to connect to CCC, because eventhough it has bad signal quality, I know it's the only one that provide PSM.
To do so, use AT+COPS=CCC.

After some time, my device disconects from the network, for one of the following reasons: apply FOTA, firmware bug generating watchdog reset, modem config changed to enable GPS, battery swap ...

Then, it needs to re-connect to the network. 
Let's say that I use AT+COPS=0. What is the modem's behavior ?
- It connects back to CCC, even if signal quality is not the best avaible, but it's considered as "prefered" ?
  If so, is this "prefered" state cleared by a XFACTORYRESET ?
- It perform a "global" PLMN scan, and automatically selects the one providing best coverage ? In the example AAA.


My final goal is to implement a logic that "suggest" a PLMN to a device, but never forces it.
Forcing using (a) specific(s) PLMN , using AT+COPS=XXX or %PALL is not ideal because if the device is moved to a location where the the said PLMN(s) is not available will result in the device no longer sending data.


Thanks for you help.

Vincent

[Achim Kraus]

If this is mainly about the PSM support:

Did you consider to switch to the nRF9151 and the "proprietary PSM" option for that?

AFAIK, there is a HPPLMN function on the SIM, but if the user HPPLMN list is implemented, depends on the SIM card itself.

There is also a "forbidden list", but you may need to populate that again and again after modem restart, also depending on the SIM card.

If you're interested in an idea, how such a implementation of those lists may look like, see modem_sim.c . 

[Vincent44]

Hello,

Thanks for your help.

Yes, switching to nRF9151 is the next step, but I already have several thousands of devices deployed.

For context, my issue is mainly for Spain, where my devices can connect to 3 different MNO. Two of them provide PSM and the last one doesn't. But since network search is AFAIK based on signal strength, I end up with ~25% of my devices connected to the MNO without PSM. For these device, battery life will be divided by 2, which is a shame when you know that simply choosing the right MNO can fix this.

"Forbiden list" is kind of what I need, but it's too scrit for my usage. In case the device is located at a place where only the MNO without PSM is available, I want it to connect (better less battery life than no data at all).

[Achim Kraus]

LTE-M? Then it may get much more critical in the future. You will need to ensure, that your SIM really have LTE-M subscription, not that no PSM is caused by fallback to LTE on old base-stations. And so the luck of LTE-M subscription may turn in a disaster, when those base-stations get their update.

There was never really a proof, that this is possible at all. But 2 years ago there have been rumors that in the UK that cause quite a lot devices to go offline for a couple of weeks/months. 

[Vincent44]

Yes, I am talking about LTE-M.

I know that my algorithm needs to take into acount several things into account, such as the device can move, the antennas might get an update, one MNO may stop operating, ...

That's the reason of my original question: I need to fully understand what's happening on some modem commands in order to adjust my algorithm.