Gpy readiness

[Redferne]

@jmarcelino

I have just built the GPy firmware from the github repo pycom-micropython-sigfox. Added some AT commands to see why our GPy refuses to attach to our internal M1/NB network on band 3, but a network scan using AT+COPS=? comes up empty. Could you please share any Sequans AT commands to help out?

[jmarcelino]

@redferne
AT!="addscanfreq band=3 dl-earfcn=1575"

should force the modem to scan band 3.

(you can find this command in the source already as cmd_Freq, just need to change the parameters and add it to the init sequence GSM_Init[])

Note this is a debug command should not be used in public networks.

[Sam]

@jmarcelino
no traffic observed on random access channels. maybe its not even locking onto the frequency.
any reason why it doesnt support all the bands pycom said they would as part of Gpy? (ie: B8)

how do we get to AT prompt on the Gpy/Fipy so we can run the band lock AT command?

[jmarcelino]

@sam
Sequans has not enabled all the bands in the LTE modem firmware, these will come later (I don't have a date yet)

We are working on arbitrary AT commands from MicroPython. Currently you can add them at the C level / custom firmware as a I suggested in my previous answer. Another way to access the modem (only on the FiPy) is through the LTE_RX, TX, CTS, RTS pins (3.3V levels). You can hold the ESP32 in reset while doing so. Serial rate is 921600bps 8N1 CTS/RTS flow control

[Redferne]

@jmarcelino

I confirmed with our internal network support that the LTE M.1 is on Band 3 and EARFCN 1925. Tried shooting the "addscanfreq..." command and it is accepted but no dice. Modem refuses to register and a AT+COPS=? comes up empty.

AT COMMAND: [AT!="addscanfreq band=3 dl-earfcn=1925"..]AT RESPONSE: [....OK..]
AT COMMAND: [AT+CPIN?..]AT RESPONSE: [..+CPIN: READY....OK..]
AT COMMAND: [AT+CFUN=1.]AT RESPONSE: [..OK..]
AT COMMAND: [AT+COPS=?..]AT RESPONSE: [..+COPS: ,,(0,1,2,3,4),(0,1,2)....OK..]
AT COMMAND: [AT+CEREG?..]AT BAD RESPONSE: [..+CEREG: 2,4....OK..]
C, restarting...
AT COMMAND: [AT+CEREG?..]AT BAD RESPONSE: [..+CEREG: 2,4....OK..]
C, restarting...
AT COMMAND: [AT+CEREG?..]AT BAD RESPONSE: [..+CEREG: 2,4....OK..]
C, restarting...
AT COMMAND: [AT+CEREG?..]AT BAD RESPONSE: [..+CEREG: 2,4....OK..]
C, restarting...
AT COMMAND: [AT+CEREG?..]AT BAD RESPONSE: [..+CEREG: 2,4....OK..]
...

[jmarcelino]

@redferne

Thanks so much for your tests, that looks quite conclusive.
I'll relay it to Sequans for their advice.

[Sam]

@jmarcelino
has these devices been ever tested? looks like we are the guinea pigs testing these for first time on certain bands :(
we find it hard to believe if these were lab tested with correct bands they fail so miserably on live networks.

[kevin]

@jmarcelino
We are also testing modules against our own lab network (NB-IoT in this case) and have had success interoperating with a number of other manufacturers by using their AT command interfaces directly.
We now have the Gpy, but no real way to test it. A basic AT command interface (e.g., via MicroPython as mentioned in another post) would be very useful - when do you expect to have this available? Can you point us to the full AT command reference from Sequans?
Also, we require B8 operation. I did not see this in your list of currently supported bands; what is the reason for this restriction? When will it be supported?

[Electrozest]

@jmarcelino
On the GPy which ESP32 GPIO pins are used for the LTE modem? The LTE_Tx pin is not shown the pinout.
Also should all NBIoT functionality be available from the AT command interface once the LTE modem has new firmware?
Thanks.

[jmarcelino]

Hi @electrozest
You can find GPIO numbers as described in the GPy board configuration:

https://github.com/pycom/pycom-micropython-sigfox/blob/master/esp32/boards/GPY/mpconfigboard.h

#define CONFIG_GSM_TX 5
#define CONFIG_GSM_RX 23
#define CONFIG_GSM_CTS 18
#define CONFIG_GSM_RTS 19

This is relative to the ESP32, so TX is TX from the ESP32 or RX from the modem side, same applies to RX, CTS and RTS.

[kevin]

@jmarcelino
Can you also relay this result to Sequans? We ran a similar test with modified Gpy FW on our Band 8 NB1 lab network. Note the response to "AT+COPS=?" appears to show support only for AT&T, and the CEREG result code is 0 (meaning it is not searching for an operator - maybe due to lack of B8 support?).

[jmarcelino]

@kevin
That's to be expected, you'll need a completely different LTE modem firmware for NB-IoT as well as band 8 support.

[kevin]

@jmarcelino
Thanks. When should we expect that to be available?

[jmarcelino]

@kevin
NB-IoT will be later next month. We're trying to have a test firmware to support all 12 bands as well, but due to lack of certification and calibration procedures that will be for testing only.

[brotherdust]

@jmarcelino I tried your method using my expansion board. I know CTS/RTS is working because it's handshaking correctly. When I take my FiPy off of reset, I can see the AT commands coming from the ESP32. However, it's not responding to anything I type into the terminal. Any thoughts on what could be happening here? Thanks!!

[jmarcelino]

Hi @brotherdust

I think your TX and RX are swapped if you want to communicate with the modem.

Also explains why you see the ESP32 - in this case you want to emulate it so you should not see it.

Please try swapping th RX/TX and keep the ESP32 in reset.

[brotherdust]

@jmarcelino
/me face-palms
There. are. no. words.

Thank you.

[Redferne]

Is there anyway to get to the LTE TX & RTS pin on the GPy?
Are they available on the pads on bottom side of the PCB?
I'm really tired of recompiling each time I think I have some magic AT command to get the Sequans modem to play ball...

[brotherdust]

@redferne Aren't the pins documented? Looks like it to me. If I had a GPy, I would use my expansion board similar to how I did it with the FiPy; just with different pin locations. =)

[brotherdust]

@brotherdust said in Gpy readiness:

@redferne Aren't the pins documented? Looks like it to me. If I had a GPy, I would use my expansion board similar to how I did it with the FiPy; just with different pin locations. =)

Ah, I see now what you are saying. That's a bummer they didn't break it out on any pins. Looks like you'll have to play whack-a-mole for a bit to find those pads!