As I wrote, in certain conditions (usually at nights and on weekends) power consumption on 4G drops to 110-130 mA, and then further lowers to 30-60 mA when the device is idle and the display turns off, which doesn’t really differ much from 2G power consumption. And it can stay so for hours. Then it suddenly goes back to 200-300 mA and remains so even when display is turned off. And it can take hours before it again drops to low levels, eating battery like crazy. There are no active processes with any increased CPU usage, so it must be the radio part eating that energy.
Anyway, it’s really difficult to tell what’s causing it. Maybe even without the actual cell/band switching taking place, in certain conditions (maybe depending on momentary signal strength or quality, or current mobile network “load” - quantity of subscribers momentarily using that specific cell/band vs. its capacity) a lot of “talking” between the phone and the network is taking place and prevents the phone (at least its baseband/modem part) from idling and reducing power consumption. Such explanation would actually correspond with the times at which I usually see lower power consumption (mostly at nights), i.e. periods of time when there’s a much lower mobile network load (less active subscribers, less traffic), less interferences, etc.