Glass roofs, atrium skylights and curtain walls need three different robot classes in 2026. A roof machine climbs shallow slopes on suction and tracks; an atrium unit works tight, often awkward spaces; a curtain wall robot runs vertically on a winch or rail. Buying one type for all three jobs usually fails.
Glass roof: what does a cleaning robot need there?
A factory or warehouse glass roof is the friendliest case, and also the biggest. Runs are long, slopes are shallow, and the glass is usually strong enough to carry a 25 to 30 kg machine. The machine that fits is a tracked unit with a large tank. The Lingfeng S1 class, covering 60 to 100 m2/h on flat panes, is built for exactly this. On a 6,000 m2 roof, that throughput turns a two-day crew job into a single shift.
The risk on roofs is not the glass, it is the edge. Parapets, vents and unguarded drops sit close to the cleaning lanes. A robot that wanders off the glass is a falling object, so tethers, low barriers or a physical stop at the run boundary are not optional.
Atrium skylight: why is this the hardest case?
An atrium punishes machines for being big. The glass sits high above a public floor, the perimeter is often angled or faceted, and the deck space where you would park a robot may be a walkway full of people. Access is the whole problem, not the cleaning.
That is why compact units like the Lingkong K3 get specified here. A narrower footprint gets through service doors and onto tight roof decks. The trade-off is a smaller tank and slower coverage. On an atrium, slow is fine, because the job is small and the real cost is the access arrangement, not the square metres.
Curtain wall: why is it a different machine entirely?
A curtain wall is vertical, and a suction robot designed for a shallow roof simply cannot hold itself against gravity on a vertical pane for long. Curtain wall units use a different approach: cranes, winch lines or fixed rails that carry the weight, with the machine handling the brush and water head. That is a separate product class, a separate budget, and usually a separate vendor conversation.
| Surface | Machine class | Typical coverage | Main risk |
|---|---|---|---|
| Glass roof, 0-15 deg | Tracked suction robot | 60-100 m2/h | Roof-edge falls |
| Atrium skylight | Compact suction robot | 20-40 m2/h | Tight access, public below |
| Faceted atrium glass | Compact + manual edges | 20-40 m2/h | Seams and gaps |
| Curtain wall, vertical | Winch or rail system | 80-150 m2/h | Wind on the face |
| Interior mall ceiling | Light compact unit | 15-35 m2/h | Public safety, drips |
Why does one robot rarely fit all three?
- Roof machines optimise for area per hour, which makes them heavy and wide. That weight is a liability on fragile atrium glazing.
- Atrium machines optimise for size and manoeuvrability, which caps tank capacity and coverage.
- Curtain wall machines optimise for vertical hold and wind tolerance, using hardware that has no place on a shallow roof.
- Site access, not glass type, often decides which of the three you can even deploy.
Who needs which, and who is overbuying?
A single-site facility with one large flat skylight needs one roof-capable machine and nothing else. Bigger portfolios with mixed assets should not buy a machine for every surface. They should map each glass type, count cleaning cycles, and put the robot where the cycle count and area justify it. The rest stays manual or on a contract.
The common overbuy is a company that sees a vertical curtain wall and buys a roof robot hoping it will cope. It will not. Talk through your actual surfaces via the contact page, and read this comparison of curtain wall and skylight machines before choosing.
How tall can each robot work, and does height change the choice?
Height changes the access plan more than the machine. A glass roof at 12 metres is reached from a plant room or a maintenance deck. An atrium skylight at 30 metres above a shopping floor needs a crane or a permanent lift point, and that single factor can add tens of thousands to the setup before a robot is even chosen. Ask your structural team where a lift point can legally go, because retrofitting one later is far more expensive.
Wind is the other height variable. Curtain wall robots face gusts on an open facade, which is why their winch systems carry wind-speed cutouts. Roof robots on shallow glass are more sheltered, but a machine caught by a gust at 20 metres with a light chassis can slide. Check the rated wind limit and stop the job above it.
What about interior mall ceilings?
Interior ceilings over public walkways are their own niche. The glass is often high, access is difficult, and the risk is less about falling machines than about dripping water and startled shoppers. Compact units with tight spray control fit here, and cleaning is usually scheduled out of trading hours. Coverage is modest, often 15 to 35 m2/h, because the machine spends as much time repositioning as cleaning.
Key Takeaways
- Roof, atrium and curtain wall jobs call for three different robot classes.
- Roof machines favour throughput and tank size; atrium units favour small size.
- Vertical curtain walls need winch or rail systems, not suction alone.
- Fragile atrium glazing is a point-load risk, so keep the machine light.
- Site access often decides the machine before the glass does.