CPSI failing with NO SERVICE,Online

[Timmie]

Hi Folks,

After switching carrier and some other things, I thought I might refresh my One+ Model B.

So I did a git pull to bring everything up to date.. then the usual pio run -t upload -t monitor

Then the console shows:

Code: Select all

...
SD:30514 MB total, 0 MB used
File: /DATA/8.CSV
[CELL] Activating...
CELL:SIM7600E-H
[BUF] 3 samples | 30 bytes | 1/32
[BUF] 9 samples | 90 bytes | 3/32
NO SIM CARD
IMEI:123456789012345
[CELL] Searching...
APN:myapn
[BUF] 15 samples | 150 bytes | 5/32
[BUF] 21 samples | 210 bytes | 7/32
[BUF] 27 samples | 270 bytes | 9/32
[BUF] 33 samples | 330 bytes | 11/32
[BUF] 39 samples | 390 bytes | 13/32
[BUF] 45 samples | 450 bytes | 15/32

+CPSI: NO SERVICE,Online

OK

[CELL] NO SERVICE,Online

I have tested the SIM in a Netgear Nighthawk and it connects to the same APN perfectly, 4G all works great.

As I dug into it a bit I noticed FreematicsNetwork.cpp was refactored recently. We seem to have lost the AT+CNMP and I'm curious if that could be causing the issue ?

I also noted the NO SIM CARD response to AT+CPIN? .. in my case the SIM does not have a PIN, so I keep coming back to:
https://github.com/stanleyhuangyc/Freem ... k.cpp#L321

I'm open to any suggestions and happy to vandalize the code and rebuild. I'm aware that I'm building from master and the risks is poses.

[Timmie]

Just a quick update, I read some of the other similar threads, so tried some of the suggestions:
1. Left APN blank - same outcome.
2. Allowed a bit more time in delays to set things up - same outcome
3. Enabled Wifi - same outcome

I also neglected to mention I'm using firmware v5\teleloger of course.

[Timmie]

Just by way of follow up -
I kept messing with this all weekend, and kept coming back to the AT+CPIN?

After enabling VERBOSE_XBEE I finally saw the error:
+CME ERROR: SIM not inserted

So I removed the sim, opened the case up and reinstalled the sim the right way around, and now its all working again!

So I just wanted to offer 2 thoughts...
1. Surely there's a better way to ensure people put the sim in the right way, draw a picture on the device or something, make the slot a bit easier to "feel" if its inserted correctly, there has to be something.

2. The way the code implements a command-response pattern seems to hide what's going on, I feel like it would be more useful if it surfaced errors better, like "Expected XXX but got YYY" ... as the YYY response would be helpful for people trying to debug these units without needing to switch on VERBOSE_XBEE