Most writing about rural 5G is about spectrum and beamforming. I want to write about the boring layer underneath it, which is power. In a lot of eastern Montana and northern Arizona, the site you are looking at is not plugged into anything. It generates its own electricity, stores it, and spends it carefully, the same way a satellite does.
I got curious about how the pieces fit, so I walked one of these sites and then tried to reconstruct it as a 3D model back at my desk. Here is what I found and what the reconstruction got right and wrong.
The load is smaller than you think
A rural sector site is not a data center. Roughly, what pulls power:
• Radio units for three sectors, plus baseband, in the 300 to 800 W band depending on traffic and how many bands are lit
• Microwave links for backhaul, since fibre never reached this location, maybe 100 W
• A small heater or ventilation blower that runs on a thermostat, which is the surprise item
• Overnight lighting and monitoring electronics, negligible
That matters because everything else in the design falls out of it. Panels, batteries, and the physical layout of the ground are all downstream of a number in the low hundreds of watts, averaged over a day.
The array is sized for December, not July
The array was two rows of ground-mounted panels on fixed tilt frames, oriented to the south. The naive question is "how much sun does this place get," and the wrong answer is the annual average. Designers size against the worst month with the shortest sun path, because a site that runs out of battery in a December week is an outage.
Fixed tilt, no trackers. Trackers add moving parts and a service call on a road that is not plowed. Every decision at these sites is the same trade: efficiency versus the number of times a truck has to come.
Where the batteries live, and why they are shaded
Under the raised edge of the array, on a long concrete pad, sit two battery cabinets and one telecom cabinet. Being under the panels is not an accident of tight land. The panels cast shade for most of the afternoon, and battery chemistry is happy in the shade. Cable enters the cabinets from the bottom through sealed ports so water cannot run in, and the runs leave through a shallow trench that is visible as a pale line in the grass.
I kept waiting for a generator, and there was one, small, behind a ridge, clearly a last resort rather than a plan.
Reconstructing the site as geometry
This is the part I actually wanted to solve. Photos gave me objects, not relationships. I could not tell, from the ground, whether the array row would put shade on the cabinets at 3 p.m., or how much of the winter sun path the monopole's own structure cut across the panels.
I am not a 3D modeler and I have no desire to become one. What I did instead: with Shapezo, you select a region by dragging a box on a map, and it generates a 3D model of that area from it, terrain, roads, fences, buildings, and the utility poles. I boxed about a square kilometre around the site.
What the model gave me:
• Real elevation, so I could check the horizon the panels see
• The road and its grade, so I could see where a truck actually turns
• The monopole position relative to the array rows
What it did not give me: cabinet dimensions, panel tilt, anything behind the fence finer than a metre or two. Fine for terrain questions. Useless for engineering questions. I want to be honest about that boundary, because I read a lot of posts that pretend a generated model is a survey.
What I took from the exercise
The site is a small, self-contained power system with an antenna on top. It works because the geometry is dumb and conservative: fixed tilt, south-facing, shaded batteries, oversized array, short cable runs, and a road that gets maintained.
Rural connectivity does not fail because of modulation schemes. It fails because a panel got iced, or a road got washboarded, or a battery aged out. That is the layer I did not know existed, and now I cannot stop looking at it.
Top comments (0)