FiPy LTE not connecting after Upgrade Firmware

[tuftec]

I was having issues when trying to run the v18.1 code while attaching to Vodafone Aus. I upgraded to the latest dev version and a number of little issues just disappeared. I am now getting reliable connections every 10mins or so but every 2 - 4 hours my system crashes unexpectedly. No o/s errors reported. Just fall s back to the repl. No idea why yet.

[Paul Thornton]

@steve-williams a change in how the firmware handles things could cause more power usage. That said I'm not aware of anything specific that would have caused this.

I'd suggest giving the latest rc firmware for the fipy a test. We've had a bunch of lte/stability improvements in the last few rc builds that could help.

[Steve Williams]

@paul-thornton Thanks for that advice, I will try that. However I am trying to deliver a finished prototype for team review, so I would prefer to use release version software really. Downgrade is perhaps an option at this stage - previous version worked fine for a long time!

[akimmels]

I am wondering if I have a similar problem.

Setup is FiPy on Pyscan module.
FW version 1.18.2 and 39529 on the modem.
When lte.attach() is called, and then quickly repeating AT+CFUN?, the modem first reports
CFUN: 1. Shortly after it reports +SYSSTART and then CFUN: 0.
As per Sequans AT manual, +SYSSTART is associated with modem reset/startup.
Could this suggest that the modem reboots due to lte.attach() call?

Below snap from terminal:

[Jordan Reyes]

My thought is that the Fipy Firmware is not the issue. but the Sequans firmware 39529 modem update was the issue. No problem running any Fipy Firmware old or new. Issues arise when using modem firmware 39529

[Steve Williams]

@steve-williams I have been trying a number of different options and combinations of GPY and Sequans firmware and I have not found a solution to the problems I am experiencing. While I strongly suspected the Sequans firmware, downgrading to 33080, then up to 38638 and 39529 showed no real difference to the problem I am having getting LTE to attach properly.

With each Sequans version I tried Pycom GPY Release F/W GPy-1.18.1.r3, r4, r7 and also GPy-1.18.2. I also tried release candidates GPy-1.20.0.rc2, rc3, rc4 and rc7 - no joy with any of them or any combination.

I had no difficulty with any of the firmware installs and my code doesn't seem to mind which version/s are running. But I still can't get LTE to attach. But I least I have gone through previous versions which did work OK previously.

From here I can only try and prove that either the Sequans module itself (processor or radios) or perhaps the Expansion Bd. 3.0 is the cause. More testing to do now I have eliminated some possibles.

Anyone else having trouble or lucked onto a solution?

[akimmels]

@akimmels I realized I should probably clarify this a bit.

What this suggest to me is that when lte.attach()is called, the modem seems to reset itself, without lte.attach() returning or giving an error.
So you think the modem is still trying to attach, but in fact it is reset....

This can be seen from the CFUN returning 0 indicating that both RX adn TX of the modem are disabled. It can also be seen from the current the module is draining, which is much lower in the reset state than when the RX and TX are active.

I flashed NB-IoT firmware (instead of LTE-M) into the modem, and was then able to attach and connect to a network and test connectivity. So this proves that the hardware is OK....

[Xykon]

I have discussed this with Sequans but before we can do any type of troubleshooting we need more details.

• The network operator being used

• The API commands used to setup the connection

• The LTE firmware version currently installed (print(lte.send_at_command('ATI1')))

• The MicroPython firmware currently used (os.uname())

• The output of print(lte.send_at_command('AT+CEREG?')) after trying to attach.

• The output of print(lte.send_at_command('AT+SQNMONI=9')) after trying to attach (might return ERROR)

@akimmels This is normal behaviour under certain circumstances. The lte.isattached() command should automatically re-enable the radio in this case. It is not advised to enable/disable the radio manually as this will cause the MicroPython task monitoring the LTE modem to fail.

[Xykon]

I have also uploaded a new firmware package for CAT-M1: https://software.pycom.io/downloads/sequans/CATM1-41065.zip

The password is the same as for other Sequans firmware downloads.

This is a pre-release that has not been fully tested yet. It should be used only for users who have issues with the latest official release.

[Steve Williams]

@xykon said in FiPy LTE not connecting after Upgrade Firmware:

It is not advised to enable/disable the radio manually as this will cause the MicroPython task monitoring the LTE modem to fail.

When you say don't manually enable/disable the radio, what commands that would be used to do this? I do use init/deinit in my code (to reduce power use between LTE use), is that OK?

I will try the new firmware you advised of, and also get the information you request, this is a vexing problem and I am keen to resolve it. As I posted earlier, I have tried a number of different Sequans and GPY firmwares to try and identify the problem but had no success yet.

[Steve Williams]

@xykon Just installed the 41065 firmware on my GPY, running 1.20.0.rc7. Sadly, my let.attach problem is NOT fixed..

[Steve Williams]

@xykon I am in Australi and use the Telstra service. The lte configuration I use is:

I have followed you instructions and taken snapshots of the command/result for each, Please note that the correct command to send AT commands is lte.send_at_cmd(instruction). as follows:

After that I uploaded the new seqans firmware OK, but it hasn't fixed the issue..

Hope the above helps to resolve the problems being encountered..

[Xykon]

@steve-williams Can you please try with:

from network import LTE
lte = LTE(carrier="telstra")
lte.attach()

Please do not send any commands starting with "AT!" to the modem unless expressly advised by Pycom. The command you are using has been discontinued in recent firmwares.

[Steve Williams]

@xykon OK, I wasn't aware of that.. the code I am using has that instruction included and has done for months, and worked fine until recently? Where was the information on this provided -( so I can look for any other useful information)?

[Steve Williams]

@xykon I tried that set of instructions - still doesn't attach.. :-(

[Jordan Reyes]

@xykon do any of these updates location specific? Wouldn't think so but wanted to double check I am in the USA usually verizon. Tried the upgrade here and it still did not work.

[jason_gladen]

@xykon I have also tried the "41065" firmware from your recent post

I still get No attach.
I'm in California this that an issue?.
It was working before I tried the firmware update. (but was not very stable)
I have two FiPy now that don't attach...
Can I roll back? or is there some extra setup I need to do to make it work?

Ran this:

from network import LTE
import utime, os

print('instantiate the LTE object')
lte = LTE()

utime.sleep(4)
print(lte.send_at_cmd('ATI1'))
utime.sleep(5)
print(os.uname())
utime.sleep(4)
print('attemting attach')
lte.attach()
utime.sleep(10)
print(lte.send_at_cmd('AT+CEREG?'))
utime.sleep(30)
print(lte.send_at_cmd('AT+CEREG?'))
utime.sleep(60)
print(lte.send_at_cmd('AT+CEREG?'))
utime.sleep(10)
print(lte.send_at_cmd('AT+SQNMONI=9'))

Got this:

rst:0x1 (POWERON_RESET),boot:0x13 (SPI_FAST_FLASH_BOOT)
configsip: 0, SPIWP:0xee
clk_drv:0x00,q_drv:0x00,d_drv:0x00,cs0_drv:0x00,hd_drv:0x00,wp_drv:0x00
mode:DIO, clock div:1
load:0x3fff8028,len:8
load:0x3fff8030,len:2064
load:0x4009fa00,len:0
load:0x4009fa00,len:19208
entry 0x400a05f8
Initializing filesystem as FatFS!
Pycom MicroPython 1.20.0.rc7 [v1.9.4-2833cf5] on 2019-02-08; FiPy with ESP32
Type "help()" for more information.
>>> Running ***************\test5.py

>>>
>>>
instantiate the LTE object

UE5.0.0.0d
LR5.1.1.0-41065

OK

(sysname='FiPy', nodename='FiPy', release='1.20.0.rc7', version='v1.9.4-2833cf5 on 2019-02-08', machine='FiPy with ESP32', lorawan='1.0.2', sigfox='1.0.1')
attemting attach

+SYSSTART


+CEREG: 2,0

OK


+CEREG: 2,0

OK


ERROR

>
Pycom MicroPython 1.20.0.rc7 [v1.9.4-2833cf5] on 2019-02-08; FiPy with ESP32
Type "help()" for more information.
>>>

[Xykon]

@jason_gladen If you are using Verizon, you should use lte = LTE(carrier='verizon') although the latest firmware should also detect this based on the SIM card.

I see that the modem is restarting after attach, which should be because the modem is detecting a Verizon SIM card and restarts in "Verizon" mode. After that you need to run lte.isattached() at least once because we need to tell the modem to re-enable the radio after the restart.

Can you please also check if the carrier was set correctly: print(lte.send_at_cmd('AT+SQNCTM?'))

We have a large customer in the US who has been using the 41065 firmware successfully with Verizon for two weeks, so we know the firmware is working with Verizon.

@Steve-Williams Could you please provide the output of print(lte.send_at_cmd('AT+CEREG?')) again after trying to attach with the new code I provided? You may also need to run lte.isattached() several times after the attach command as it's possible the modem restarted after running the attach() command.
Can you please also confirm the carrier was set correctly: print(lte.send_at_cmd('AT+SQNCTM?'))

I'm also reviewing this case with Sequans and will let you know once I have more feedback.

[jason_gladen]

@xykon said in FiPy LTE not connecting after Upgrade Firmware:

lte.isattached()
lte = LTE(carrier='verizon')

from network import LTE
import utime
import sys

utime.sleep(3)
print( "{0}: lte=LTE()".format(utime.ticks_ms()) )
lte = LTE()

try:
  utime.sleep(3)
  print( "{0}: lte.attach()".format(utime.ticks_ms()) )
  lte.attach()

except Exception as e:
  sys.print_exception(e)

utime.sleep(3)
print( "{0}: lte.isattached()".format(utime.ticks_ms()) )
ret = lte.isattached()
print( "{0}: Ret = {1}".format(utime.ticks_ms(),ret) )


utime.sleep(10)
print( "{0}: lte.isattached()".format(utime.ticks_ms()) )
ret = lte.isattached()
print( "{0}: Ret = {1}".format(utime.ticks_ms(),ret) )


utime.sleep(10)
print( "{0}: lte.isattached()".format(utime.ticks_ms()) )
ret = lte.isattached()
print( "{0}: Ret = {1}".format(utime.ticks_ms(),ret) )


utime.sleep(10)
print( "{0}: lte.send_at_cmd('AT+SQNCTM?')".format(utime.ticks_ms()) )
ret = lte.send_at_cmd('AT+SQNCTM?')
print( "{0}: Ret = {1}".format(utime.ticks_ms(),ret) )


utime.sleep(10)
print( "{0}: lte.isattached()".format(utime.ticks_ms()) )
ret = lte.isattached()
print( "{0}: Ret = {1}".format(utime.ticks_ms(),ret) )

utime.sleep(10)
print( "{0}: lte.send_at_cmd('AT+CEREG?')".format(utime.ticks_ms()) )
ret = lte.send_at_cmd('AT+CEREG?')
print( "{0}: Ret = {1}".format(utime.ticks_ms(),ret) )

>>>
6185: lte=LTE(carrier='verizon')
24403: lte.attach()
Traceback (most recent call last):
  File "<stdin>", line 13, in <module>
OSError: the requested operation failed
32385: lte.isattached()
32635: Ret = False
42638: lte.isattached()
42818: Ret = False
52821: lte.isattached()
53001: Ret = False
63004: lte.send_at_cmd('AT+SQNCTM?')
63037: Ret =
+SQNCTM: standard

OK

73040: lte.isattached()
73221: Ret = False
83224: lte.send_at_cmd('AT+CEREG?')
83257: Ret =
+CEREG: 2,0

OK

>

Again with LTE(), this worked before I upgraded

18250: lte=LTE()
23112: lte.attach()
26724: lte.isattached()
26850: Ret = False
36853: lte.isattached()
37452: Ret = False
47455: lte.isattached()
47635: Ret = False
57638: lte.send_at_cmd('AT+SQNCTM?')
57671: Ret =
+SQNCTM: standard

OK

67674: lte.isattached()
67855: Ret = False
77858: lte.send_at_cmd('AT+CEREG?')
77891: Ret =
+CEREG: 2,0

OK

>

[Steve Williams]

@xykon Hi, so I put together this little routine, just so i can test various GPY and SEQUANS firmware combinations fairly quickly. I plan to test both CATM1 and NB modem firmware.

# Testing LTE connection
# Start with importing core modules for features, interfaces etc.
import uos
from network import LTE
from utime import sleep

print('Instantiate the LTE object to Telstra (AUS)')
lteM = LTE(carrier="telstra")
sleep(5)
print(lteM.send_at_cmd('ATI1'))
sleep(5)
print(uos.uname())
sleep(5)
lteset = lteM.send_at_cmd('AT+SQNCTM?')
print("LTE set to {}".format(lteset))
lteM.attach()
for l in range (0,10):
    lteM.isattached()
    sleep(1)
noltcon = 0
if not lteM.isattached():
    print("Attempting to attach to LTE ", end='')
    while(True):
        noltcon += 1
        if noltcon >= 50:
            # If not attached in 50 sec, flash error then try again
            print(" Failed after 50 attempts")
            break
        elif lteM.isattached():
            print(" Success")
            lteset = lteM.send_at_cmd('AT+SQNCTM?')
            print("LTE attached to {}".format(lteset))
            # Get the current LTE signal quality
            squallteM = lteM.send_at_cmd('AT+CSQ').strip().split()
            rquallteM = int(squallteM[1].split(',')[0])
            # Evaluate the returned signal quality Data
            rssilteM = 113 - ((rquallteM) * 2)
            ltesqual = ("LTE RSSI: -{} dbM".format(rssilteM))
            print(ltesqual)
            break
        print('.', end='')
        sleep(1)
sleep(10)
print(lteM.send_at_cmd('AT+CEREG?'))
sleep(10)
print(lteM.send_at_cmd('AT+CEREG?'))
sleep(10)
print(lteM.send_at_cmd('AT+CEREG?'))
sleep(10)
print(lteM.send_at_cmd('AT+SQNMONI=9'))

and get these results
#################

Running C:\AE Projects\Atom Projects\GPy Projects\JANNE\lib\LTE_Tests.py

Instantiate the LTE object to Telstra (AUS)

UE5.0.0.0d
LR5.1.1.0-41065

OK

(sysname='GPy', nodename='GPy', release='1.20.0.rc7', version='v1.9.4-2833cf5 on 2019-02-08', machine='GPy with ESP32')
LTE set to
+SQNCTM: telstra

OK

Attempting to attach to LTE ................................................. Failed after 50 attempts

+CEREG: 2,4

OK

+CEREG: 2,4

OK

+CEREG: 2,4

OK

ERROR
##################

Is the code OK, even if the results aren't what I am hoping for?

Question, is the carrier information for 'telstra' now embedded in the SEQUANS firmware? Used to use two AT commands
'AT+CGDCONT=1,"IP","telstra.internet"'
'AT!="RRC::addscanfreq band=28 dl-earfcn=9410"' (Note you advised NOT to use this)
to set that information for the carrier.

As my code stands, I have a common config file holding the information requires for different countries/carriers/locations, so I will have to revisit that - provided i can get LTE to actually attach!

This is becoming a real concern for me as I am due to present a prototype for client review, but a prime requirement is the LTE capability. For me it is very strange that this has just recently stopped working, after trundling along 24/7 without a problem for more than 3 months (other than restarts from occasional code revisions). I haven't had any luck with reverting to previous firmwares either , which is strange?.