Inside troubleshooting: Finding patterns with 1oT Terminal to avoid disconnections

Tags: IoT
A picture of Muhammad Nabeel Rahim
Written by
Muhammad Nabeel Rahim

Welcome to a new blog series about cellular connectivity and IoT — Inside troubleshooting.

The goal is to share our customer support team's learnings from troubleshooting complex IoT issues worldwide. Each article will conclude with a suggestion to avoid the problem in the first place and better prepare you for your deployment plans.

Let's dive into the first story.

Devices stopped doing data sessions and didn't connect again

One of our customers reported that their devices stopped doing data sessions at random times. Also, they didn't connect to the network again.

Let the troubleshooting begin

After confirming the hardware-specific details, we went straight into 1oT Terminal to analyse the network level logs. We studied the records for several different SIMs that went offline.

We found that each device showed around 1-hour long data sessions just before they went offline. This was a clear pattern to start seeing the underlying cause of the problem.

We suspected that the 1-hour data sessions meant the devices were not completing their restart cycle and were not powering again. As a result, they did not make any connection request attempts to the network.

To confirm the assumption, we performed a few in-house tests.

The test meant that we opened several data sessions, pulled out the battery from the test device without closing the data sessions, and did not turn on the device again.

We learned that the data sessions were closed automatically by the network around the 1-hour mark. The same pattern as we observed with the customer's devices.

All in all, after 1 hour, the network closes a data session if the device is powered off (without properly closing the data session) and doesn't come back online again.

Suggestion to the customer

In light of the findings, we asked the customer to make sure that their devices were completing the restart power cycle gracefully and coming back online to start the network attach process again.

We suggested trying different module-specific AT commands for initiating the power cycle on the module level. Since then, no similar issues have been found.

Example AT commands:

The following example commands are taken from the u-blox AT commands manual. They are relevant to specific u-blox cellular modules. These can be used for gracefully restarting or turning off the modules.


The cellular module will restart (without resetting the SIM card) after detaching from the network and storing the configuration parameters in its Non Volatile Memory (NVM).


The cellular module will restart (resetting the SIM card) after detaching from the network and storing the configuration parameters in its Non Volatile Memory (NVM).


This command turns off the cellular module after storing the current setting in the module’s NVM.

For the MPCI-L2 form factor, this command causes a restart instead of turning off the module. The current configuration is stored in the NVM, and a network detach occurs.

Please note that the above commands might not apply to all u-blox or other devices. Therefore they might not be suitable for every use case. This article is for informational purposes, and 1oT is not responsible for any loss or damage caused by trying these commands. We advise the customer to refer to their device manual for relevant instructions.

Suggestions for the IoT community

If you face similar issues with your device, make sure that the restart cycle completes gracefully, and the devices are turning back on. You also may try using different AT commands for restarting your cellular modules. These AT commands can be found in your device's command manual.

Note: Some networks might temporarily restrict a device's connection with the network due to incorrect network behaviour. This includes restarting and turning off at regular intervals without closing the data session through the correct AT commands.

A sneak peek into the future:

1oT Terminal holds a lot of critical information, such as logs and records of your session history. Yet finding patterns and insights is the human eye's responsibility.

Therefore, we are building an AI-powered predictive system that learns from the data usage and, through interpreting patterns, can detect anomalies before someone notices them.

At the moment, we're testing the AI-powered predictive system internally. But soon, it will be available for testing for the first customers.