On one project, a customer reported unexplained attach and ping instability in a dense LTE-M deployment. A competitor claimed they had fixed the issue using “CGI-based resolution”, triggering speculation that we’d missed something obvious.
We reviewed everything: RSRQ looked fine, but control channel decoding was unstable. SIB scheduling from neighbouring cells was overlapping, creating interference at just the wrong moment. A classic case where on-paper RF looked good, but the device still struggled.
Later, we confirmed the truth:
CGI reporting is defined in 3GPP TS 36.331,
But not supported in most CAT-M (Rel-13) devices because they lack measurement gaps and don’t decode SIBs from neighbouring cells.
The real issue? Likely control channel sensitivity and firmware-level decoding performance, something far less flashy than a config flag.
Field testing matters. Because sometimes, the fix isn’t a feature, it’s a better stack.
LinkedIn: 