The screen is the last thing that happens
By the time the ring reaches site, most of the work is already buried. A circular plinth four metres across, about 300 mm thick, poured and cured with eight anchor bolts cast into it. A steel base frame bolted on top. The two grey boxes the ring stands on are not decoration: power distribution in the left one with its vent mesh facing outward, control and receiving gear in the right one.
Estate managers usually ask why the base is so big next to the screen. It is not padding. It is counterweight, and I will come back to that.
A ring has no interchangeable cabinets
On a sphere the cabinets come off moulds by latitude band, so a whole screen might need six shapes and every cabinet on a band is identical. A ring runs the other way. The face splits into concentric rings of wedge cabinets, narrow at the inner edge and wide at the outer, so each ring is its own shape and nothing crosses between them.
Two consequences land at the factory rather than on site. Spares get ordered per ring and labelled that way, because a mixed bag of ring spares is a bag of parts that will not fit. And cabinet numbers have to be fixed on the drawing before production, since once the face is on you cannot count rows from the outside.
Centre line before anything heavy moves
The instrument in the foreground is a total station, and on day one it is the most important tool on the job. We mark the centre of the plinth, transfer that point up onto the frame, then check both verticals against it. Target is a millimetre on the centre point and one per mille on plumb.
The reason is plain arithmetic. At three metres, one degree out of plumb puts the top 52 mm off centre. Twenty millimetres nobody notices. Fifty-two reads as an oval, and once the panels are on there is no arguing with it.
Lifting is where a ring gets bent
A ring is a thin wall with a hole punched through it. Pick it up from a single point and the load runs through one side of the hoops, the frame sags into an egg, and panels that fitted on the floor stop fitting. Two or four points, spread across the diameter, slings taken up to a spreader beam so the horizontal component does not squeeze the ring inward while you are lifting it.
Before the crane releases, measure across the ring both ways. Those two diameters should agree within 2 mm. If they do not, set it back down and straighten it on the ground. The frame will not recover on its own, and every cabinet you force on afterwards carries that error for the life of the screen.

The hole is the hardest part
A solid disc is straightforward. Take the middle out and three new problems appear. Rain lands on the inner top edge and runs down the inside face. Cables have nowhere to hide, because the back of the screen is visible from the far side of the plaza. And there is no wall behind the thing, so every driver, every switch, every terminal has to live inside the ring or inside the base.
So: routing stays inside the frame depth, every pass-through gets a grommet so a sharp edge cannot saw through insulation in the wind, and the lowest point of the inner edge gets a drain hole. Otherwise water sits in the bottom of the ring through a wet season and finds the connectors on its own.
Do the wind sum once, then stop guessing
A three metre ring with a one metre hole has roughly 6.3 m² of frontal area. In a 12 beaufort gust, about 0.7 kN/m², that is four to five kilonewtons pushing sideways, centred about 1.5 m above the plinth. Call it 7 kN·m of overturning moment on a base 1.5 m wide.
Which is why the concrete looks oversized. It has to be. We size the plinth with a factor of two against uplift on the windward anchors and specify chemical anchors rather than expansion bolts, because an expansion bolt under reversing load in a 300 mm slab works loose in a couple of seasons, and then the screen is leaning and nobody can say when it started.
Sixty cable ends, sixty paper tags
Look at the top arc of the ring. Every cable hanging out of the frame has a paper tag taped to it. That is not housekeeping. A screen this size has somewhere between thirty and sixty cable ends, and with the sun behind you and a black frame in front, they are all identical.
The tags carry numbers from the drawing. When a quadrant comes up dark on first power-up, the difference between a twenty minute fix and a two day one is whether the man on the platform can read a number and walk straight to the right receiving card.
Three checks, and the crane stays until they pass
Geometry and plumb while the frame is still empty, because that is the cheapest moment and there is nothing to take down. Then a crosshatch pattern at the distance people will actually stand from it, not with your nose against the surface. Then a low grey field at dusk, since a step between concentric rings that nobody can see at noon reads as a patchwork at seven in the evening.
That last one is calibration, not hardware, but it is far cheaper to run while the platform is still hired and the crane has not driven off.
What we would change on the next one
The lower face panels should stay in their crates until the base bolts are torqued down. It is tempting to fit them early, because working at two metres off the ground beats working at four metres on a lift, and it does save a couple of hours.
It also adds weight to the lift at the exact moment you want the ring as light and as stiff as possible, and it puts finished panels in the air for twenty minutes with nothing between them and the ground. Lighter lift, nothing to scratch, and the two hours come back somewhere less expensive.

