COTS UE can not see 5G network

[peihaoY]

I am following this page(https://docs.srsran.com/projects/project/en/latest/tutorials/source/cotsUE/source/index.html) to establish a connection between UE and gNB, but unlike the tutorial, I am using 5G Docker. I can successfully connect the gNB and srsUE, and they work well. However, I am unable to see the 5G network with my Xiaomi 11 Lite 5G. I have attached my configuration as follows.
UE：
Xiaomi 11 Lite 5G
GNB and 5GCore：
USRP N310, Ubuntu 22.04, UHD 4.5.0
SIM card:
sysmoISIM-SJA5
External clock:
GPSDO
gnb and 5gcore.zip
I have used the SA network, and my two SIM cards do not have services 124 and 125. The PLMNs are 00101 and 99970, band n3 and n78, respectively, and neither of them works. Thank you for your help.

[herlesupreeth]

Try typing *#*#726633#*#* on the dial pad. You should see a hidden menu, there enable Mode SA

[peihaoY]

Actually, I have enabled Mode SA. It cannot see the 5G network.

[herlesupreeth]

hmmm.. have you tried adjusting the Tx and Rx gain values?

[peihaoY]

Yes. I have tried the maximum gain Tx Rx = 65.

[Thanasislt]

Hello, I can confirm this behaviour as well on our testbed.
We deployed Open5GS (SA, v2.6.0) with srsRAN 23.10 (Dec 2023) and a USRP X310 as gnb.
Xiaomi 11 Lite 5G could not connect despite our attempts.

We ended up rooting the device and installing PixelOS. Afterwards, the device connects to the 5G network with no problems.

[brendan-mcauliffe]

Have you forced 5G using an app like shown in the tutorial?

[peihaoY]

Yes. After that, I occasionally see T-mobile 5G, but most of the time there's nothing..

[pgawlowicz]

do you run dockerized version of the Open5GS? could you post your gnb logs?

[peihaoY]

Yes, I used the dockerized version. The gnb log is as follows:
gnb2.log
When the following occurs "Late: 0; Underflow: 2; Overflow: 0;" the log file is as follows:
gnb1.log

[robertfalkenberg]

According to these logs it seems there has been a successful attach and registration, the UE got the IP 10.45.1.3 from the core:

open5gs_5gc  | 02/28 23:49:35.467: [smf] INFO: UE SUPI[imsi-001010123456789] DNN[srsapn] IPv4[10.45.1.3] IPv6[] (../src/smf/npcf-handler.c:528)
open5gs_5gc  | 02/28 23:49:35.467: [upf] INFO: [Added] Number of UPF-Sessions is now 1 (../src/upf/context.c:194)
open5gs_5gc  | 02/28 23:49:35.467: [gtp] INFO: gtp_connect() [127.0.0.4]:2152 (../lib/gtp/path.c:60)
open5gs_5gc  | 02/28 23:49:35.467: [upf] INFO: UE F-SEID[UP:0x1 CP:0x1] APN[srsapn] PDN-Type[1] IPv4[10.45.1.3] IPv6[] (../src/upf/context.c:467)
open5gs_5gc  | 02/28 23:49:35.467: [upf] INFO: UE F-SEID[UP:0x1 CP:0x1] APN[srsapn] PDN-Type[1] IPv4[10.45.1.3] IPv6[] (../src/upf/context.c:467)

Moreover, a data radio bearer (DRB) was also exchanging some data, as you can see in the gNB logs:

...
2024-02-28T23:49:55.819504 [RLC     ] [I] du=1 ue=1 DRB1 UL: RX PDU. pdu_len=66 dc=data p=1 si=full sn=19
2024-02-28T23:49:55.819505 [RLC     ] [I] du=1 ue=1 DRB1 UL: RX SDU. sn=19 sdu_len=63
2024-02-28T23:49:55.819505 [PDCP    ] [I] ue=0 DRB1 UL: RX PDU. type=data pdu_len=63 sn=19 count=19
2024-02-28T23:49:55.819506 [PDCP    ] [I] ue=0 DRB1 UL: RX SDU. count=19
2024-02-28T23:49:55.819508 [GTPU    ] [I] ue=0 UL teid=0x000002: TX PDU. pdu_len=76 teid=0x000002 qfi=QFI=1
...

[peihaoY]

Yes, I think so. I can see some brate on gNB with some overflow. But why my phone always shows "cannot connect to the network"?

Late: 0; Underflow: 6; Overflow: 0;
           -----------------DL-----------------------|------------------UL--------------------
 pci rnti  cqi  ri  mcs  brate   ok  nok  (%)  dl_bs | pusch  mcs  brate   ok  nok  (%)    bsr
   1 4612   10   1   10   1.1k    6    0   0%      0 |  10.3   12    25k    6    0   0%      0
   1 4612    9   1   10    912    5    0   0%      0 |  10.5   12    21k    5    0   0%      0
Late: 0; Underflow: 5; Overflow: 0;
   1 4612   10   1   11    936    5    0   0%      0 |  10.1   12    21k    5    0   0%      0
Late: 74; Underflow: 5; Overflow: 0;

[robertfalkenberg]

Late: 74; Underflow: 5; Overflow: 0;

74 is a lot... you could reduce log level to warning. Maybe this relaxes the stress on the computer a bit.

[pgawlowicz]

hmm, a lot of RF issues. Please try to run this script: https://github.com/srsran/srsRAN_Project/blob/main/scripts/srsran_performance
to set the CPU in the performance mode.

[pgawlowicz]

1. Did you already tried this ?
2. Change PLMN to 90170. The tutorial says:

The PLMN is changed to 90170. Using the OnePlus 8t, we have found that by forcing a roaming scenario, the phone sees the network and attaches more reliably. A roaming scenario is forced by setting different PLMNs in the cell and the ISIM in the phone.

[peihaoY]

1. Neither of these methods works for me; the display remains the same.
2. When I use PLMN = 90170, the phone displays 'forbidden', and there's no output on the 5G core and gNB console. When I use 00101, I can get the logs mentioned above and see some brate and many overflows on the gNB, but the phone displays 'cannot connect to the network.' It shows as in the following picture when automatically connecting, but I still can't access the internet.
   
   I wonder if this warning has anything to do with it?"
   open5gs_5gc | 02/28 23:49:24.579: [gmm] WARNING: Authentication failure(Synch failure) (../src/amf/gmm-sm.c:1366)
open5gs_5gc | 02/28 23:49:35.467: [smf] WARNING: Unknown PCO ID:(0x23) (../src/smf/context.c:3035)
open5gs_5gc | 02/28 23:49:35.467: [smf] WARNING: Unknown PCO ID:(0x24) (../src/smf/context.c:3035)
   Or is there an issue with this mac.pcap?gnb_mac.zip

[robertfalkenberg]

I wouldn't worry too much about this message. I also get those for quite a long time; the UE attaches eventually.

open5gs_5gc  | 03/01 14:33:20.485: [gmm] INFO: [suci-0-001-01-0-0-0-0000051520]    SUCI (../src/amf/gmm-handler.c:880)
open5gs_5gc  | 03/01 14:33:20.565: [gmm] WARNING: Authentication failure(Synch failure) (../src/amf/gmm-sm.c:1366)
open5gs_5gc  | 03/01 14:33:20.846: [gmm] INFO: [imsi-001010000051520] Registration complete (../src/amf/gmm-sm.c:1916)
open5gs_5gc  | 03/01 14:33:20.846: [amf] INFO: [imsi-001010000051520] Configuration update command (../src/amf/nas-path.c:612)

...
open5gs_5gc  | 03/01 14:33:31.042: [upf] INFO: UE F-SEID[UP:0x1 CP:0x1] APN[srsapn] PDN-Type[1] IPv4[10.45.1.121] IPv6[] (../src/upf/context.c:467)
open5gs_5gc  | 03/01 14:33:31.042: [upf] INFO: UE F-SEID[UP:0x1 CP:0x1] APN[srsapn] PDN-Type[1] IPv4[10.45.1.121] IPv6[] (../src/upf/context.c:467)
open5gs_5gc  | 03/01 14:33:31.042: [gtp] INFO: gtp_connect() [10.53.1.2]:2152 (../lib/gtp/path.c:60)
open5gs_5gc  | 03/01 14:33:31.043: [smf] WARNING: Unknown PCO ID:(0x23) (../src/smf/context.c:3035)
open5gs_5gc  | 03/01 14:33:31.043: [smf] WARNING: Unknown PCO ID:(0x24) (../src/smf/context.c:3035)
open5gs_5gc  | 03/01 14:33:31.126: [gtp] INFO: gtp_connect() [10.53.1.1]:2152 (../lib/gtp/path.c:60)

[robertfalkenberg]

Regarding your MAC PCAP, I can see the UE had an active DRB from

41	2024-02-29 19:59:32,208591	10.45.1.3	8.8.8.8	DNS	543			UEId=1     [UL] [AM] DRB:1  [DATA]     SN=0               [63-bytes]   ||   UEId=1     [UL] [AM] DRB:1  [DATA]     SN=1               [78-bytes]   ||   UEId=1     [UL] [AM] DRB:1  [DATA] (P) SN=2               [65-bytes]  (Short BSR LCG ID=0 BS=0)   (PHR PH=51 PCMAX_f_c=51)   (Padding 285 bytes) 

to

375	2024-02-29 20:01:54,129615			RLC-NR	100			UEId=1     [DL] [AM] DRB:1  [CONTROL]  ACK_SN=248     (Padding 63 bytes) 

From 19:59:32 .. 20:01:54 are a few minutes before the UE jumps off.
Could is be this? #100

[peihaoY]

Great! Now it can stay attached without any overflow. They can ping each other. Thanks!

[peihaoY]

I'm having another problem with my oneplus. Why does it disconnect whenever it tries to test the upload speed or transfer more data via iperf? After a while it may connect again automatically, mostly it doesn't. Then the 5g core console comes up with a lot of warnings as shown below. Attached are my log files.
gnb_log+mac+ngap and 5gcore_log.zip

[peihaoY]

Yes, I used an external clock, but the same issue still occurred. My Motorola phone has never had issues with disconnection. So, I'm wondering if it might be an issue with the OnePlus phone? However, you have already tested that the connection with this phone is fine. Do you have any further comments?

[pgawlowicz]

could you install NetMonster app (or similar) and check what is the RSRP and SNR in DL?

[peihaoY]

NetMonster shows Not even a single cell. Another smartphone can see 1 01 5G.SA 3500.

[pgawlowicz]

@peihaoY Any updates on this issue?
If the UE connects to the network and is able to ping, the Netmonster app should show the info about the cell.

[peihaoY]

No update. I cannot use my oneplus. I use Motorola only. Thanks for your consideration!