The least photogenic week of the build

Nothing on this ball lights up. It is a 2 m shell on castors, some sixty hours in: cage welded, jig work signed off, roughly two hundred ribbon cables already hanging off it. The modules arrive next week. Nearly every photograph of an LED sphere is taken three weeks after this one, when it is lit and the content looks expensive — which is why buyers end up asking about pixel pitch and never about the thing in this picture.

Fair enough, you cannot see the frame once the modules go on. But the frame and the loom decide most of what you will complain about later: the seam that opens, the cabinet that runs hot, the module nobody can reach without pulling half the ball apart.

The cage is welded from cut profiles, not rolled from a sheet

What you are looking at is 24 vertical ribs and a set of latitude rings, welded and then checked against a jig. The ribs are not aluminium tube bent once and trimmed to fit. Every band of a sphere has a different radius, so a rib running from the top pole down to the equator changes curvature the whole way. You can roll a cylinder in one pass. You cannot roll a sphere in one pass.

That is also why a 2 m shell ends up as 24 cabinets instead of two big curved panels. The cage takes the load: ribs in compression, handing the pull up to the hoist point, so the modules are only ever asked to be modules.

The black goes on before the welding is finished

Every rib is coated before it goes into the jig, not after the cage is complete. A welded joint leaves a heat-affected zone either side of it, and that is exactly where rust starts if the coating is applied as one pass at the end of assembly. On a unit headed for a coastal site, the difference shows up in year two as pale brown weeping lines under a seam — the sort of thing that never gets fixed, because fixing it means opening the shell.

One more reason the frame in this photo is worth looking at: the finish is a spec you can ask about, and almost nobody does.

Every cable is cut before the modules exist

Every ribbon in the photograph is already the right length for the module it will feed. That is not tidiness. With the modules on, the gap behind a cabinet on a 2 m sphere is a few millimetres — nothing gets cut, stripped or re-routed in there afterwards. Loom it first, or plan on opening the shell again once it is built.

The white tags on the ends are positions, not part numbers. Cabling a sphere is closer to wiring a building than to plugging in a monitor: each cable reaches exactly one socket, and the map has to exist on paper before the first one is cut. Get that map wrong and the ball still works — it just stops being serviceable, because the only panel you can reach is the one feeding the wrong module.

One hook carries the whole ball

Look at the top of the frame: one attachment, one chain. A 2 m shell like this runs around 160 kg before a single module goes on, and all of it hangs from that one point when the ball is lifted into an atrium. The frame does not feel that as weight so much as bending — the crown ribs pull, the equator sags, the bottom keeps trying to go down.

So the hoist is not welded to the thin section of a rib. There is a reinforced ring at the crown and a centre post behind it spreading the pull across the upper cage. From 4 m up, that turns into a steel tower and the shell hangs off the tower instead. If a supplier shows you a sphere hanging from a single welded eye bolt on a 6 m ball, that is not engineering, that is optimism.

The plastic at the bottom is not packaging

The wrap around the base is there for the next three weeks, not for the container. Once the lower modules go on, people walk around this ball all day with screwdrivers, module frames and trolley handles in their hands. The bottom of a sphere is the part nobody looks at until it is scratched, and then it is the first thing every visitor sees from the floor.

It doubles as an honest label. A wrapped sphere is an unfinished sphere — on our floor that is how you tell at a glance which units are still open.

The other two balls are a timetable, not a background

On the left, another cage still hanging on its chain, further behind in the process. On the right, a finished black shell that has had its modules and its first power-on. Three units, three stages, one afternoon. None of that is parallel: cage, loom, jig fit, modules, ageing, crate. Each stage waits for the one before it, and skipping ahead is how a finished ball gets opened up again.

A workshop running several orders at once looks like chaos in a photograph. It is actually a queue, and the position of your unit in it is the only honest answer to "when will it ship".

Three checks before a module goes on

First, the cage against the jig: diagonals measured both ways, a shim under any rib that reads long. Second, the hoist: the shell goes up off the floor and the diameter is measured again — if a ball goes oval under its own weight, you want that news now rather than in the atrium. Third, the loom: every cable tagged, every tag checked against the map, continuity tested on the runs that will end up behind the hardest modules to reach.

A sphere that passes all three gets its modules. One that fails goes back on the jig, which costs a day. The alternative costs a service visit and a flight.

Three questions that separate two quotations

When two sphere quotes differ by 30%, the gap is usually sitting in this photograph rather than in the pixel pitch. Ask what the ribs are made of and how thick they are. Ask what the hoist point is rated at, and whether that rating covers the shell alone or the shell plus modules plus rigging. Ask whether the cable loom is pre-cut and pre-tagged in the factory or made up on site.

A supplier who answers all three without forwarding your email to somebody else builds the frame in-house. The one who goes quiet on the second question is buying shells in and putting their badge on the crate.

It looks better in three weeks

In three weeks this ball will be lit, the content will be looping, and nothing in this photograph will be visible. That is how it should be — a screen is supposed to hide its insides. It just means the week with nothing to look at is the week worth asking questions about.

A 2 m LED sphere frame on the factory floor, welded cage with roughly two hundred ribbon cables hanging in the grid before any module is fitted
Same unit, full frame: cage on castors, loom finished and tagged, base still wrapped. The finished black shell on the right is two stages ahead.