IoT case study
A lift company had fitted a GSM module that would not stay connected. The module was not the problem. It was being asked to live on one network, at the bottom of a lift shaft.
The situation
A lift company had installed a GSM module on a lift emergency phone so that anyone trapped could call for help. It tested fine on the day. Then it started dropping out.
That is the difficult kind of fault. The module was not dead, so nothing obvious showed up. It simply worked sometimes and not others, and an engineer sent back to look at it would often find it behaving perfectly. Meanwhile the one thing that connection exists for is an emergency, which is the moment nobody gets to test it first.
The lift company had done nothing wrong. Lift shafts are among the worst places in any building for mobile coverage, and the module had been given a SIM tied to a single network. If that one network happens to be the weak one at the bottom of that particular shaft, the result is exactly what they were seeing.
What we did
Before recommending anything we wanted to know what the signal was actually doing at the module, rather than what a coverage map said about the postcode. Coverage checkers describe the street. They do not describe the inside of a concrete shaft.
The outcome
The module now picks whichever of the major networks is strongest from inside the shaft, instead of being tied to one. If that network degrades, or a mast is taken out of service for repair or an upgrade, it moves to another one and the emergency phone carries on working.
The wider point is about single points of failure in the places least able to carry them. A lift emergency phone is the connection in a building that has to work on the worst day, and it usually sits in the spot with the least signal. Locking it to one network puts those two facts together in the wrong way. The fix was a SIM rather than hardware, which made it quick and cheap, but the reason it worked was doing the testing first.
Questions
A lift shaft is close to the worst place in any building for mobile reception. It is a concrete and steel box, usually in the core of the building and often partly below ground, and it is built to be strong rather than transparent to radio. A signal that is perfectly good in reception can be marginal by the time it reaches the module.
No. It selects the strongest network available to it at that location and uses that one, and it can move to another if the first degrades or disappears. The module does not know or care which network it is on, it just stays connected.
Usually not. If the module is sound then the SIM is the cheaper and quicker thing to change, and that was the case here. We test first so the recommendation is based on what the signal is actually doing rather than on a guess.
On a single-network SIM, the module goes quiet until the work finishes, and nobody necessarily finds out. On a multi-network SIM it moves to another network and the emergency phone keeps working.
It does if the module is 2G only, and many older ones are. EE begins switching off 2G from May 2029, O2 from summer 2029 and Vodafone in spring 2030. If you are already having a module looked at, it is worth confirming at the same time whether it is 4G capable. Our guide to the 2G switch-off dates covers what to check.
It is how somebody trapped in a lift reaches help, and it has to work. That is exactly why the intermittent faults matter more here than they would on a piece of office equipment: the day it is needed is the day nobody can test it first.
Tell us what the device is and where it lives. We will test what the signal is really doing at the device before recommending anything.
