Questions about signal quality polling in nRF9160

RE: Questions about signal quality polling in nRF9160

Fredrik — Thu, 09 Jan 2020 11:12:03 GMT

Four SIM cards a lot of confusion later, this finally works. It was indeed an issue with the SIM card being misconfigured as Håkon Alseth suggested.

To anyone having the same issues: Make absolutely sure that the SIM card provided by your carrier supports NB. I recieved conflicting information from customer support.

RE: Questions about signal quality polling in nRF9160

Fredrik — Sat, 14 Dec 2019 11:49:14 GMT

Tried at home, same results unfortunately.

devzone.nordicsemi.com/.../trace_2D00_2019_2D00_12_2D00_14T11_2D00_44_2D00_48.241Z.bin

[00:00:06.709,075] ---> AT%XMODEMTRACE=1,2
[00:00:06.732,971] ---> AT+CEREG=5
[00:00:06.755,523] ---> AT%CESQ=1
[00:00:06.824,432] ---> AT+CFUN=4
[00:00:06.850,219] ---> AT%XSYSTEMMODE=0,1,0,0
[00:00:06.877,227] ---> AT+CFUN=1
[00:00:30.016,876] <--- %CESQ: 60,3,21,3
[00:00:30.032,623] <--- +CEREG: 2,"A0F1","013DD964",9,0,0,"11100000","11100000"
[00:00:35.076,416] <--- +CEREG: 4,"A0F1","013DD964",9,0,15,"11100000","11100000"
[00:00:35.142,730] <--- %CESQ: 255,0,255,0
[00:00:35.155,181] <--- +CEREG: 4,"FFFE","FFFFFFFF",9,0,15,"11100000","11100000"
[00:00:52.689,270] <--- %CESQ: 60,3,24,3
[00:00:52.734,252] <--- +CEREG: 2,"0EDD","02100113",9,0,15,"11100000","11100000"
[00:00:54.002,990] <--- +CEREG: 2,"0EDD","02100113",9,0,14,"11100000","11100000"
[00:00:54.162,994] <--- %CESQ: 255,0,255,0
[00:00:54.171,112] <--- +CEREG: 2,"FFFE","FFFFFFFF",9,0,14,"11100000","11100000"
[00:01:00.727,142] <--- %CESQ: 60,3,19,2
[00:01:00.735,046] <--- +CEREG: 4,"A0F1","013DD964",9,0,14,"11100000","11100000"
[00:01:10.647,857] <--- %CESQ: 60,3,25,3
[00:01:15.767,974] <--- %CESQ: 60,3,20,2
[00:01:20.888,122] <--- %CESQ: 60,3,26,3
[00:01:23.448,181] <--- %CESQ: 60,3,18,2

Telenor agreed to send me a new SIM, just in case the one I got is misconfigured somehow.

Thanks for all your help!

RE: Questions about signal quality polling in nRF9160

Fredrik — Fri, 13 Dec 2019 14:26:55 GMT

I just got off the phone with Telenor who (finally) decided that the subscription I had did not support NB-IoT. My SIM does support NB, however, and we swapped my subscription to allow NB traffic. Telenor is now certain that everything looks good on their side. They also made sure that there is good coverage in my area.

My case is starting to look similar to Terje's case as Telenor asked me to make sure I was using the telenor.iot APN. Simon's answer here indicates that PDP context should be handled by the modem firmware, and I don't know if I should mess with the APN manually.

I'll try connecting at home over the week-end.

RE: Questions about signal quality polling in nRF9160

Håkon Alseth — Fri, 13 Dec 2019 14:14:52 GMT

Hi,

Got confirmation from R&D that the eNB or SIM seems to be the issue.

There is an attach reject with cause value #15 (No suitable cells in tracking area) from the home network for this SIM. This cause value means basically that you could get service from this network, either on some different area or via different access technology.

There were no cells from other home networks tracking areas available so we also made an attempt to another vendor (the one marked with 0EDB) on the log file. That network rejects attach attempt with a cause value that includes the whole network (#14 EPS services not allowed in this PLMN).

Have you tried connecting to NB1 in a different area (at work vs. at home, for instance)?

Kind regards,

Håkon

RE: Questions about signal quality polling in nRF9160

Håkon Alseth — Thu, 12 Dec 2019 15:10:59 GMT

Hi,

Thank you for the trace logs.

I can see that the eNB (the cell tower) releases the connection very quickly. This is very similar behavior with one of my SIM cards that does not support NB1.

I will ask R&D to look at the modem trace as well, in case I have missed something.

Kind regards,

Håkon

RE: Questions about signal quality polling in nRF9160

Fredrik — Thu, 12 Dec 2019 12:56:45 GMT

Sure. Here is the modem trace:

devzone.nordicsemi.com/.../trace_2D00_2019_2D00_12_2D00_12T12_2D00_47_2D00_42.604Z.bin

Pseudo-log of AT commands sent/received:

[00:00:06.754,486] ---> AT%XMODEMTRACE=1,2
[00:00:06.778,350] ---> AT+CEREG=5
[00:00:06.800,872] ---> AT%CESQ=1
[00:00:06.823,181] ---> AT+CFUN=4
[00:00:06.848,907] ---> AT%XSYSTEMMODE=0,1,0,0
[00:00:06.875,885] ---> AT+CFUN=1
[00:00:30.485,015] <--- %CESQ: 54,2,23,3
[00:00:30.500,701] <--- +CEREG: 2,"9FC5","0103A167",9,0,0,"11100000","11100000"
[00:00:38.578,460] <--- +CEREG: 4,"9FC5","0103A167",9,0,15,"11100000","11100000"
[00:00:38.631,622] <--- %CESQ: 255,0,255,0
[00:00:38.644,073] <--- +CEREG: 4,"FFFE","FFFFFFFF",9,0,15,"11100000","11100000"
[00:00:49.760,467] <--- %CESQ: 72,3,22,3
[00:00:49.805,938] <--- +CEREG: 2,"0EDB","020D8126",9,0,15,"11100000","11100000"
[00:00:51.073,577] <--- +CEREG: 2,"0EDB","020D8126",9,0,14,"11100000","11100000"
[00:00:51.231,048] <--- %CESQ: 255,0,255,0
[00:00:51.239,105] <--- +CEREG: 2,"FFFE","FFFFFFFF",9,0,14,"11100000","11100000"
[00:00:56.075,988] <--- %CESQ: 54,2,22,3
[00:00:56.083,831] <--- +CEREG: 4,"9FC5","0103A167",9,0,14,"11100000","11100000"
[00:01:00.846,466] <--- %CESQ: 54,2,27,3
[00:01:49.487,640] <--- %CESQ: 47,2,27,3
[00:01:54.607,757] <--- %CESQ: 54,2,27,3

RE: Questions about signal quality polling in nRF9160

Håkon Alseth — Thu, 12 Dec 2019 12:32:49 GMT

Hi,

My apologies for the delay.

Could you take a modem trace so we can look at what happens?

Here's a howto on enabling modem traces: https://devzone.nordicsemi.com/nordic/nordic-blog/b/blog/posts/how-to-get-modem-trace-using-trace-collector-in-nrf-connect

Kind regards,

Håkon

RE: Questions about signal quality polling in nRF9160

Fredrik — Tue, 03 Dec 2019 11:43:13 GMT

Hi Håkon,

Any update? I edited my previous reply and rejected the "verified answer", but I'm worried you were not notified.

RE: Questions about signal quality polling in nRF9160

Fredrik — Fri, 29 Nov 2019 15:42:07 GMT

It turns out regular consumer SIM cards from Telenor do not support the IoT bands. I ordered a dedicated IoT SIM card from them instead, and NB-IoT works fine with that.

Thanks for your help!

EDIT: NB-IoT does not seem to work with the new SIM either.

I ran the original test again with the new SIM (which I am sure supports NB-IoT) outside in an open area. This is the log:

[00:00:00.431,793] ---> AT+CPSMS=0
[00:00:00.454,589] ---> AT+CEREG=4
[00:00:00.475,494] ---> AT%CESQ=1
[00:00:00.496,215] ---> AT+CFUN=4
[00:00:00.516,937] ---> AT%XSYSTEMMODE=0,1,0,0
[00:00:00.540,435] ---> AT+CFUN=1
[00:00:27.303,375] <--- %CESQ: 94,4,18,2
[00:00:27.311,584] <--- +CEREG: 2,"9FC5","0114CF67",9,,,"11100000","11100000"
[00:00:32.353,790] <--- +CEREG: 4,"9FC5","0114CF67",9,,,"11100000","11100000"
[00:00:32.393,127] <--- %CESQ: 255,0,255,0
[00:00:32.401,489] <--- +CEREG: 4,"FFFE","FFFFFFFF",9,,,"11100000","11100000"
[00:00:42.296,203] <--- %CESQ: 97,4,23,3
[00:00:42.304,412] <--- +CEREG: 2,"0EDB","020D8112",9,,,"11100000","11100000"
[00:00:43.590,759] <--- +CEREG: 2,"0EDB","020D8112",9,,,"11100000","11100000"
[00:00:43.725,555] <--- %CESQ: 255,0,255,0
[00:00:43.733,947] <--- +CEREG: 2,"FFFE","FFFFFFFF",9,,,"11100000","11100000"
[00:00:51.500,823] <--- %CESQ: 94,4,20,2
[00:00:51.509,002] <--- +CEREG: 4,"9FC5","0103A167",9,,,"11100000","11100000"
[00:00:58.853,546] <--- %CESQ: 89,4,20,2
[00:01:03.973,602] <--- %CESQ: 89,4,15,2
[00:01:06.533,630] <--- %CESQ: 89,4,20,2
[00:01:14.213,745] <--- %CESQ: 89,4,10,1
[00:01:16.773,773] <--- %CESQ: 94,4,18,2
[00:01:21.893,829] <--- %CESQ: 88,4,18,2
[00:01:34.694,000] <--- %CESQ: 88,4,13,1
[00:01:39.814,056] <--- %CESQ: 88,4,18,2
[00:01:44.934,112] <--- %CESQ: 93,4,18,2
[00:01:50.054,138] <--- %CESQ: 93,4,12,1
[00:01:52.614,227] <--- %CESQ: 93,4,19,2
[00:01:55.174,224] <--- %CESQ: 88,4,12,1
[00:01:57.734,222] <--- %CESQ: 88,4,17,2
[00:02:00.294,250] <--- %CESQ: 88,4,23,3
[00:02:07.974,334] <--- %CESQ: 88,4,18,2

Around 51.500 seconds I receive modem status 4 ("unknown") with valid tac etc. However, I also receive very strong RSRP values. I can't make sense of this. As usual, LTE-M1 works fine.

RE: Questions about signal quality polling in nRF9160

Fredrik — Mon, 25 Nov 2019 15:10:56 GMT

I called telenor again, and the customer rep said he didn't know but he'd open an internal case and get back to me in a few days. I'll put this on hold until i hear back from them.

Thanks for your help.

RE: Questions about signal quality polling in nRF9160

Håkon Alseth — Mon, 25 Nov 2019 14:42:20 GMT

This is similar to the behavior that I see on my older telenor SIMs (and the iBasis one for that matter), where the network doesn't allow NB1 on the specific SIM:

+CEREG: 2,"9DD1","0145B864",9,0,0,"11100000","11100000"
+CEREG: 2,"FFFE","FFFFFFFF",9

Have you contacted telenor, supplying them with the IMSI (the numbers physically printed on your SIM), and asked them to verify that NB1 is enabled for that specific SIM card? If your SIM is old, it might be that you need a new one.

Kind regards,

Håkon

RE: Questions about signal quality polling in nRF9160

Fredrik — Sat, 23 Nov 2019 13:58:17 GMT

Sure, here is the log:

2019-11-23T13:44:40.183Z INFO Application data folder: /home/fredrik/.config/nrfconnect/pc-nrfconnect-linkmonitor
2019-11-23T13:44:40.306Z DEBUG App pc-nrfconnect-linkmonitor v1.1.1 official
2019-11-23T13:44:40.306Z DEBUG App path: /home/fredrik/.nrfconnect-apps/node_modules/pc-nrfconnect-linkmonitor
2019-11-23T13:44:40.306Z DEBUG nRFConnect 3.3.0 is supported by the app (^3.2.0)
2019-11-23T13:44:40.306Z DEBUG nRFConnect path: /tmp/.mount_nrfconjpLbRj/resources/app.asar
2019-11-23T13:44:40.306Z DEBUG HomeDir: /home/fredrik
2019-11-23T13:44:40.306Z DEBUG TmpDir: /tmp
2019-11-23T13:44:41.916Z INFO Modem port is opened
2019-11-23T13:44:47.696Z DEBUG modem >> AT\x0D\x0A
2019-11-23T13:44:47.767Z DEBUG modem << OK\x0D\x0A
2019-11-23T13:44:50.943Z DEBUG modem >> AT%XSYSTEMMODE=0,1,0,0\x0D\x0A
2019-11-23T13:44:50.963Z DEBUG modem << OK\x0D\x0A
2019-11-23T13:44:55.215Z DEBUG modem >> AT+CEREG=5\x0D\x0A
2019-11-23T13:44:55.231Z DEBUG modem << OK\x0D\x0A
2019-11-23T13:44:59.417Z DEBUG modem >> AT+CFUN=1\x0D\x0A
2019-11-23T13:44:59.465Z DEBUG modem << OK\x0D\x0A
2019-11-23T13:45:22.777Z DEBUG modem << +CEREG: 2,"9FC5","0103A167",9,0,0,"11100000","11100000"\x0D
2019-11-23T13:49:04.906Z DEBUG modem << +CEREG: 2,"FFFE","FFFFFFFF",9,0,0,"11100000","11100000"\x0D
2019-11-23T13:49:13.175Z DEBUG modem << +CEREG: 2,"9FC5","0103A167",9,0,0,"11100000","11100000"\x0D\x0A
2019-11-23T13:52:56.828Z DEBUG modem << +CEREG: 2,"FFFE","FFFFFFFF",9,0,0,"11100000","11100000"\x0D\x0A
2019-11-23T13:53:03.576Z DEBUG modem << +CEREG: 2,"9FC5","0103A167",9,0,0,"11100000","11100000"\x0D\x0A

I used the Telenor SIM for this test.

RE: Questions about signal quality polling in nRF9160

Håkon Alseth — Fri, 22 Nov 2019 15:58:18 GMT

[quote user="fellerts"][/quote]

I tried with a Telenor SIM, and I got the exact same results: valid RSRP with modem status 2 on NB, no issues on LTE-M1. I called Telenor to make sure both LTE-M1 and NB-IoT is included in the subscription, and that there is coverage on both in my area.

Could this be an issue with my DK? It's an early model PCA10090 v0.8.2.

This shouldn't be a problem with the DK.

Could you try programming your kit with "at_client", input the following:

AT%XSYSTEMMODE=0,1,0,0

AT+CEREG=5

AT+CFUN=1

Then wait 10 minutes, and upload the output log here?

Kind regards,

Håkon

RE: Questions about signal quality polling in nRF9160

Fredrik — Fri, 22 Nov 2019 15:23:40 GMT

Hi,

Could this be an issue with my DK? It's an early model PCA10090 v0.8.2.

RE: Questions about signal quality polling in nRF9160

Håkon Alseth — Fri, 22 Nov 2019 13:27:57 GMT

[quote user="fellerts"]In your testing, did you connect to NB-IoT or LTE-M1? As I said, everything works fine in LTE-M1, and I get the same results as you. In NB-IoT mode, however, the modem state never changes. It notifies state == 2 once, then it never changes. I get valid RSRP notifications, though, which is why I'm confused.[/quote]

I tested M1. I have to apologize, I have misunderstood the scenario. I thought the issue was that the RSRP was reported before the ATTACH cam, but it turns out I have not read your questions properly.

You're stating that the ATTACH newer comes. I just tested with iBasis, on NB1, and I see the same.

The other SIM I have is also NB1 capable, so that works just fine on both M1 and NB1, but iBasis does not seem to work on NB1 in this region, unfortunately.

Do you have a Telenor SIM to test with? That should work on NB1 and M1.

Kind regards,

Håkon

RE: Questions about signal quality polling in nRF9160

Fredrik — Fri, 22 Nov 2019 11:26:53 GMT

Håkon,

In your testing, did you connect to NB-IoT or LTE-M1? As I said, everything works fine in LTE-M1, and I get the same results as you. In NB-IoT mode, however, the modem state never changes. It notifies state == 2 once, then it never changes. I get valid RSRP notifications, though, which is why I'm confused.

Thanks for your help.

RE: Questions about signal quality polling in nRF9160

Håkon Alseth — Fri, 22 Nov 2019 09:58:24 GMT

Here's my testing:

I see the same behavior. It seems to be picking up the RSRP in the process of attaching, and reporting that before the ATTACH is fully completed.

By adding timestamps in teraterm, it seems that the RSRP comes around 1 second before the ATTACH is fully complete:

[2019-11-22 10:31:38.293] %CESQ: 30,1,0,0
[2019-11-22 10:31:38.524] %CESQ: 35,1,6,0
[2019-11-22 10:31:39.499] +CEREG: 1

Kind regards,

Håkon

RE: Questions about signal quality polling in nRF9160

Fredrik — Thu, 21 Nov 2019 14:13:30 GMT

Hi Håkon, thanks for getting back to me.

I updated to ncs v1.1.0 and ran the same test again with notifications as you suggested, and I am still a bit confused here.

[00:00:00.464,904] ---> AT+CEREG=1
[00:00:00.492,370] ---> AT%CESQ=1
[00:00:00.537,414] ---> AT+CFUN=4
[00:00:00.564,697] ---> AT%XSYSTEMMODE=0,1,0,0
[00:00:00.594,848] ---> AT+CFUN=1

[00:01:05.232,055] <--- %CESQ: 76,3,26,3
[00:01:05.247,650] <--- +CEREG: 2
[00:01:06.794,067] <--- %CESQ: 255,0,255,0
[00:01:14.438,110] <--- %CESQ: 54,2,19,2
[00:04:56.711,944] <--- %CESQ: 255,0,255,0
[00:05:04.841,613] <--- %CESQ: 53,2,22,3
[00:08:47.977,386] <--- %CESQ: 255,0,255,0
[00:08:55.247,039] <--- %CESQ: 52,2,22,3
[00:12:38.802,825] <--- %CESQ: 255,0,255,0
[00:12:45.652,465] <--- %CESQ: 52,2,23,3

I get the first signal quality update at 01:05 which states that the RSRP is good. Immediately after the signal quality notification, through, I get a modem update with state = 2, which means "Not registered, but UE is currently trying to attach or searching an operator to register to" (from the nRF91 AT Commands Reference Guide). For the next 10 minutes or so, the modem switches between notifying valid and invalid signal quality notifications, but the modem state never changes. Shouldn't the modem state be either 1 or 5 when there is a valid connection?

RE: Questions about signal quality polling in nRF9160

Håkon Alseth — Thu, 21 Nov 2019 13:21:09 GMT

Hi,

[quote user=""]Initially, the modem status reported by the CEREG command is 4 (unknown), and then it flips to 2 (not registered, searching). At the same time I receive valid RSRP values from the CESQ command. The modem status never changes after this. Is this expected? Can I have valid RSRP, <tac> and <ci> on NB-IoT without having modem status equal to 1 or 5? Repeating the same procedure for LTE-M1 always gives initial modem status 4, then 5.[/quote]

Yes, this is expected. Since you are polling the modem via AT commands, it has likely successfully attached somewhere between timestamp 01:02 and 01:12.

If you look at the timestamp, you're receiving the CESQ response before querying CEREG. This indicates that you were already connected when checking the RSRP.

Try running the at_client, and see how that behaves. It will print notifications from the modem as they come in.

PS: I would recommend that you update to ncs tag v1.1.0, as that is now released. It shouldn't be a problem to modem v1.1.0 on tag v1.0.0, but in case you want newer features; like GPS, check that out.

Kind regards,

Håkon