GobiSerial and GobiNet kernel drivers reconnect

Are you buying through a distributor here in North America? It will be easier for the distributor FAE teams to open a ticket with us for further support. Jon

[quote=“curry-te, post:16, topic:15813”]
AT+CGREG?

I would rather not have another person (who can’t help) in the middle of solving my problems. I have tried getting my buyer involved in my problems but they don’t have the experience. There is no reason why this Sierra forum shouldn’t help me.

I need to know why I would get an ERROR from the AT!SCACT=1,1. Is there a way I can debug this problem? I can perform all the other AT commands. Do I have to have a good RSRP level while performing the AT!SCACT=1,1. If I have low power levels, I should at least get an “OK” but then my eth1 IP address would not be created.

Tim, What is your company name again? We need the distributor to open a ticket and you can still work directly with me on this issue. For Sprint and Verizon, it is usually AT!SCACT=1,3 as legacy CDMA carriers typically use what is called the Class 3/Context 3 APN. What does AT+CGDCONT? show? AT+COPS?

Jon

Our Company name is Evaporcool Solutions. I called Future Electronics and left a message to open a ticket. Please don’t let there submission delay your responses.

I have been using AT!SCACT=1,1 for a long time because Sprint said it was the right pid. I tried 1,3 but I still get ERROR.

Can you check my Gobi versions. They are different version but same date between Serial and Net below.

I typed:

modinfo GobiSerial

filename: /lib/modules/4.9.11-02316-g91a2d54-dirty/kernel/drivers/usb/serial/GobiSerial.ko

license: Dual BSD/GPL

description: GobiSerial

author: Qualcomm Innovation Center

version: 2018-08-24/SWI_2.34

srcversion: 4591907522522961516424B

alias: usb:v05C6p920Cddcdscdpiciscipin

depends: usbserial

vermagic: 4.9.11-02312-ga264544324c3-dirty SMP preempt mod_unload modversions ARMv7 p2v8

…

…

…

parm: debug:Debug enabled or not (int)

parm: flow_control:flow control enabled or not (int)

parm: ignore_gps_start_error:allow port open to success even when GPS control message failed (int)

parm: delay_open_gps_port:Delay Open GPS Port, after device ready (int)

parm: iusb3_zlp_enable:0 = Disable , 1 (default) ZLP on USB3 in QDL mode (int)

Is the version of my GobiSerial correct?

I typed:

modinfo GobiNet

filename: /lib/modules/4.9.11-02316-g91a2d54-dirty/GobiNet.ko

license: Dual BSD/GPL

description: GobiNet

author: Qualcomm Innovation Center

version: 2018-08-24/SWI_2.53

srcversion: 9B510198CC64EBC84E177F9

alias: usb:v1199p9110ddcdscdpiciscipin

alias: usb:v1199p68A2ddcdscdpiciscipin

depends: usbnet

vermagic: 4.9.11-02312-ga264544324c3-dirty SMP preempt mod_unload modversions ARMv7 p2v8

parm: debug:Debuging enabled or not (int)

parm: qos_debug:QoS Debuging enabled or not (int)

parm: interruptible:Listen for and return on user interrupt (int)

parm: txQueueLength:Number of IP packets which may be queued up for transmit (int)

parm: iTEEnable:-1 : Ignore TE flow Control, 0 : TE Flow Control disabled, 1 : TE Flow Control enabled (int)

parm: iRAWIPEnable:RAWIP enabled or not (int)

parm: iQMAPEnable:-1: Auto, 0 : QMAP disabled, 1 : QMAP enabled (int)

parm: iMaxQMUXSupported:-1: Auto, Max QMUX instance support (int)

parm: iIPAlias:0 = virtual adapter , 1 (default) = IP alias (int)

I submitted the AT commands below:

AT+CGDCONT?

+CGDCONT: 1,“IP”,“cc.m2m.ispsn”,“0.0.0.0”,0,0,0,0

OK

AT+COPS?

+COPS: 0

OK

OK on !SCACT=1,1 and COPS 0 means “unknown network” and is usually an indication of no cell services. What does AT!GSTATUS? show? If the RSRP level is ~-100 dBm, then you are far-cell and usually will have a hard time registering and starting the LTE PDP data session. This is likely a coverage issue, assuming that the module has been provisioned OK with Sprint. Jon

Jon,

I received the Sierra dev kit yesterday and found some interesting results. The dev kit comes with large paddle antennas. I have been using the Taoglas antennas in the past. The kit with paddles gives me -100 dB which allows me to successfully call the AT!SCACT=1,1. I placed the Taoglas antennas on the kit and my signal strength went down to -117.0 dB and I get ERROR after submitting the AT!SCACT=1,1. I think the Taoglas antennas have been my problem all along. I ordered a Yagi directional antenna and I will let you know those signal strengths. Not having a good signal is very frustrating.

Good to know and Qualcomm typically defines -95 to -120 dBm RSRP as far-cell and thus unlikely to establish an LTE PDP session and transfer IP packets. Taoglas is one of our preferred partners for antennas and let me know if you need to reach their U.S. VP for support. I can email him with your issue. Regards, Jon

I have the FAE at Future Electronics looking into the bad Taoglas performance. We think it might be because we are getting the antenna with a different termination connector for the tiny EM7455 EM4 connector. This EM7455 connector is very hard to manage and should be changed to a larger one.

I see and good to know you are working with the FAE. Jon

I am back testing with the larger/better antenna. I can now get a consistent virtual eth1 address but it is the wrong IP address. I should get 10.x.x.x. but I get 24.x.x.x. Sprint has no idea why and the IP address goes away after about a minute. When I call the AT!SCACT=1,1, I get a different IP address. I think it has something to do with the Sprint network but I need better proof from the EM7445 output. Is there any data we can extract from the EM7455 to prove that it is a Sprint issue and not the FW issue?

I am still not sure which kernel modules should be blacklisted. I have about 10 blacklisted but some folks only blacklist two. Is the blacklist documented for debian 4.9.11?

If the IP address is changing, and the session is disconnecting, and the RSRP is good, then this is likely a network issue. The blacklisting is show in the application developer guide. Linux QMI SDK - Application Developer’s Guide.pdf (893.4 KB)