Over-the-air with x310 UE gNB connection issue

[devprge]

Hi,
I am trying replicate the Over-the-air Setup using two Ettus USRP X310, but I am currently having trouble getting the UE and gNB to connect with each other. I am following this guide
https://docs.srsran.com/projects/project/en/latest/tutorials/source/srsUE/source/index.html
In my set up I have two computers each computer are connected to their own X310. One computer has the UE on it while the other computer has the gNB and 5gCore on it. I do not have a device to provide an external clock for the two X310, so I am using the internal clock. I was able to bring up the gNB, but when I try to bring up the UE it constantly prints out RF status and does not establish connection with the gNB. I tried using the configuration in the USRP X300/X310 Troubleshooting section but got the same message. I tried changing the tx/rx gains and the srate. I also ran the srsran_performance script to see if that will help.
I attached my configuration files and log. Do you have any suggestion what parameters I can try?

Computer Spec:
CPU: Intel Xeon w5-2465x
Memory: 32GiB
OS: Ubuntu 22.04.3 LTS
OS Type: 64-bit

gnb.yml.txt
ue_rf.conf.txt
gnb.log
ue.log

[pgawlowicz]

please check this thread: #859

[devprge]

I tested the configuration files shared in that discussion and I am still unable to get the UE and the gNB to connect.

Here is the logs and configurations I lasted used:
gnb.log
ue.log
gnb.yml.txt
ue_rf.conf.txt

First question I have is the transmit frequency for the UE and gNB suppose to be the same? I connected the gNB Ettus TX output to a spectrum analyzer and I can see peak frequency is center around 1842.5MHz. Then I connect the UE Ettus TX to spectrum analyzer and I also see it outputting a signal that is around 1842.5MHz. Is it normal that both the UE and the gNB are transmitting in the same frequency? The log file said it is using FDD, so I think the Ettus 1 Tx frequency should match the Ettus 2 Rx frequency and the Ettus 1 Rx frequency will match the Ettus 2 Tx frequency.

Second question is what can I try to fix overflow issue on the UE? When I look at the UE log it seem to be able to detect the gNB signal, but the UE throws an overflow error right after.

[devprge]

I used the configuration file in #859. I changed the clock to use the internal clock and the tx and rx gain and still was getting the cell overflow error from the UE. It looks like it is tuning the rx channel then the cell search fail. This is lines 51-107 in the ue.log.

2024-11-27T15:55:06.460147 [PHY-SA ] [I] [    0] Cell Search: Going to IDLE
2024-11-27T15:55:06.460149 [PHY-SA ] [I] [    0] Tuning Rx channel 0 to 1843.20 MHz
2024-11-27T15:55:06.460222 [RF     ] [I] Mapping RF channel 0 (device=0, channel=0) to logical carrier 0 on f_rx=1843.2 MHz
2024-11-27T15:55:06.461110 [PHY-SA ] [D] [    0] run_stack_tti: from recv
2024-11-27T15:55:06.461111 [PHY-SA ] [D] [    0] run_stack_tti: calling stack tti=0, tti_jump=1
2024-11-27T15:55:06.461111 [PHY-SA ] [D] [    0] run_stack_tti: stack called
2024-11-27T15:55:06.461111 [PHY-SA ] [D] [    0] SYNC:  received 30720 samples from radio
2024-11-27T15:55:06.461112 [PHY-SA ] [D] [    0] SYNC:  state=IDLE, tti=0
2024-11-27T15:55:06.461112 [PHY-SA ] [D] [    0] Discarding samples and sending tx_end
2024-11-27T15:55:06.461198 [MAC    ] [D] [    0] Running MAC tti=0
2024-11-27T15:55:06.461199 [MAC    ] [D] [    0] Update Bj: lcid=0, Bj=0
2024-11-27T15:55:06.461199 [MAC    ] [D] [    0] Update Bj: lcid=1, Bj=0
2024-11-27T15:55:06.461200 [MAC    ] [D] [    0] RA:    IDLE:    rar_window_start=-1, rar_window_length=-1
2024-11-27T15:55:06.461200 [MAC-NR ] [D] [    0] Running MAC tti=0
2024-11-27T15:55:06.461201 [MAC-NR ] [D] [    0] nof_lcids_with_data=0, nof_lcgs_with_data=0, last_non_zero_lcg=-1 [lcg0=0, ] [lcid0=0, ]
2024-11-27T15:55:06.461202 [STCK   ] [D] tti_tprof: duration=0 msec
2024-11-27T15:55:06.462375 [PHY-SA ] [D] [    0] run_stack_tti: from recv
2024-11-27T15:55:06.462376 [PHY-SA ] [D] [    0] run_stack_tti: calling stack tti=0, tti_jump=1
2024-11-27T15:55:06.462376 [PHY-SA ] [D] [    0] run_stack_tti: stack called
2024-11-27T15:55:06.462376 [PHY-SA ] [D] [    0] SYNC:  received 30720 samples from radio
2024-11-27T15:55:06.462377 [PHY-SA ] [D] [    0] SYNC:  state=IDLE, tti=0
2024-11-27T15:55:06.462377 [PHY-SA ] [D] [    0] Discarding samples and sending tx_end
2024-11-27T15:55:06.462471 [MAC    ] [D] [    0] Running MAC tti=0
2024-11-27T15:55:06.462471 [MAC    ] [D] [    0] Update Bj: lcid=0, Bj=0
2024-11-27T15:55:06.462472 [MAC    ] [D] [    0] Update Bj: lcid=1, Bj=0
2024-11-27T15:55:06.462472 [MAC    ] [D] [    0] RA:    IDLE:    rar_window_start=-1, rar_window_length=-1
2024-11-27T15:55:06.462473 [MAC-NR ] [D] [    0] Running MAC tti=0
2024-11-27T15:55:06.462473 [MAC-NR ] [D] [    0] nof_lcids_with_data=0, nof_lcgs_with_data=0, last_non_zero_lcg=-1 [lcg0=0, ] [lcid0=0, ]
2024-11-27T15:55:06.462474 [STCK   ] [D] tti_tprof: duration=0 msec
2024-11-27T15:55:06.463207 [PHY-SA ] [I] [    0] Cell search: Setting SSB configuration srate=30.72 MHz; c-freq=1843.200 MHz; ss-freq=1838.550 MHz; scs=15kHz; pattern=A; duplex=fdd;
2024-11-27T15:55:06.463415 [PHY-SA ] [D] [    0] run_stack_tti: from recv
2024-11-27T15:55:06.463415 [PHY-SA ] [D] [    0] run_stack_tti: calling stack tti=0, tti_jump=1
2024-11-27T15:55:06.463415 [PHY-SA ] [D] [    0] run_stack_tti: stack called
2024-11-27T15:55:06.463415 [PHY-SA ] [D] [    0] SYNC:  received 30720 samples from radio
2024-11-27T15:55:06.463416 [PHY-SA ] [D] [    0] SYNC:  state=IDLE, tti=0
2024-11-27T15:55:06.463416 [PHY-SA ] [D] [    0] Discarding samples and sending tx_end
2024-11-27T15:55:06.463424 [MAC    ] [D] [    0] Running MAC tti=0
2024-11-27T15:55:06.463425 [MAC    ] [D] [    0] Update Bj: lcid=0, Bj=0
2024-11-27T15:55:06.463425 [MAC    ] [D] [    0] Update Bj: lcid=1, Bj=0
2024-11-27T15:55:06.463425 [MAC    ] [D] [    0] RA:    IDLE:    rar_window_start=-1, rar_window_length=-1
2024-11-27T15:55:06.463425 [MAC-NR ] [D] [    0] Running MAC tti=0
2024-11-27T15:55:06.463426 [MAC-NR ] [D] [    0] nof_lcids_with_data=0, nof_lcgs_with_data=0, last_non_zero_lcg=-1 [lcg0=0, ] [lcid0=0, ]
2024-11-27T15:55:06.463426 [STCK   ] [D] tti_tprof: duration=0 msec
2024-11-27T15:55:06.463451 [PHY-SA ] [I] [    0] Cell Search: Running Cell search state
2024-11-27T15:55:06.464443 [PHY-SA ] [D] [    0] run_stack_tti: from recv
2024-11-27T15:55:06.464443 [PHY-SA ] [D] [    0] run_stack_tti: calling stack tti=0, tti_jump=1
2024-11-27T15:55:06.464444 [PHY-SA ] [D] [    0] run_stack_tti: stack called
2024-11-27T15:55:06.464444 [PHY-SA ] [D] [    0] SYNC:  received 30720 samples from radio
2024-11-27T15:55:06.464445 [PHY-SA ] [D] [    0] SYNC:  state=IDLE, tti=0
2024-11-27T15:55:06.464484 [MAC    ] [D] [    0] Running MAC tti=0
2024-11-27T15:55:06.464485 [MAC    ] [D] [    0] Update Bj: lcid=0, Bj=0
2024-11-27T15:55:06.464485 [MAC    ] [D] [    0] Update Bj: lcid=1, Bj=0
2024-11-27T15:55:06.464485 [MAC    ] [D] [    0] RA:    IDLE:    rar_window_start=-1, rar_window_length=-1
2024-11-27T15:55:06.464485 [MAC-NR ] [D] [    0] Running MAC tti=0
2024-11-27T15:55:06.464486 [MAC-NR ] [D] [    0] nof_lcids_with_data=0, nof_lcgs_with_data=0, last_non_zero_lcg=-1 [lcg0=0, ] [lcid0=0, ]
2024-11-27T15:55:06.464487 [STCK   ] [D] tti_tprof: duration=0 msec
2024-11-27T15:55:06.465586 [RF     ] [I] Overflow

Here is my config and log:
gnb.log
gnb.yml.txt
ue_rf.conf.txt
ue.log

[pgawlowicz]

try coreset0_index: 12 in gnb config

[devprge]

I was able to fix cell selection failureand over flow, by updating the ue_rf.conf ssb_nr_arfcn to match the gNB dl_ssb_arfcn. Now the UE is transmitting. I now seeing CRC failed in the ue.log. Do you know what might be cause the CRC failures?

This the my config and log right now:
ue.log
ue_rf.conf.txt
gnb.log
gnb.yml.txt

I see metrics on the gNB but it rarely happens

[pgawlowicz]

Could you try to apply this patch #826 (comment) to srsUE, recompile and try again?

[devprge]

Do you know if the incompatible prach_cfg.freq_offset might be causing the CRC failed in the UE?

2024-11-27T22:49:25.245715 [PHY-SA ] [E] [ 0] prach_cfg.freq_offset=1 is not compatible with LTE

ue_rf.conf.txt
ue.log
gnb.yml.txt

[pgawlowicz]

you need coreset0_index: 6 for 10MHz cell

[devprge]

I updated coreset0_index to 6 and change the gain settings and RRC is still connecting then releasing right after.
gnb.log
gnb.yml.txt
ue.log
ue_rf.conf.txt

[devprge]

Looks like I am getting closer. It successfully establish connection but it still release RRC in like 1 second after establishing connection.

gnb.log
gnb.yml.txt
ue.log
ue_rf.conf.txt

[pgawlowicz]

Did you apply this patch #826 (comment) to srsUE?

[devprge]

I applied the patch on the srsUE, but it still behave the same. I get Recieved RRC Release right after recieving RRC Connect.

I am going to try to change the clock source from internal to gpsdo or external to see if I can get it to work with those clock setting. I was not able to get external it would always throw
Error: couldn't set sync source: RuntimeError: Reference Clock PLL failed to lock to external source.
srsRAN ERROR: Unable to create radio session.

I was able to bring up the gNB and the UE with setting the clock, sync, clock_source, and time_source to gpsdo. Is there something I am missing from my config if I switch to using gpsdo for my clock and sync setting?

gnb.yml.txt
ue_rf.conf.txt

gnb.log
ue.log

[pgawlowicz]

could you use external clock in gnb, but leave the sync empty? For srsUE: time_source=internal,clock_source=external

[devprge]

I gave up on the two x310 and got two b210. The two b210 work using the config files in the guide.
I was able to proceed forward in the guide. Everything seem to be working except the TCP iperf3 test. The UE keeps releasing RRC after a few second running iperf3 using TCP. Do you know what is casing that?

[pgawlowicz]

could you share your gnb and srsue logs?

[devprge]

Here is my config and log file for the two b210 setup. I added two set of logs

gnb.yml.txt
ue_rf.conf.txt

Log 1
gnb1.log
ue1.log

Log 2
gnb2.log
ue2.log

[pgawlowicz]

the configs look good.
What are the PCs in terms of performance? (CPU, ram)?

[devprge]

My computer spec is:
CPU: Intel Xeon w5-2465x
Memory: 32GiB
OS: Ubuntu 22.04.3 LTS

I am using internal clock on both. Does the over the air tutorial works with using internal clock on both devices?

[pgawlowicz]

Do you have a possibility to check with an external clock?