Pick a skylight cleaning robot by matching glass type, roof slope and reach first. Flat laminated skylights suit a lightweight tracked robot; steep or curved glass roofs need a vacuum-adhesion unit with at least 25 kg of holding force and an IP65 body rated for wet work.
Most buyers start with price. That is backwards. A robot that slips on a 25-degree glass roof is useless at any price, and one that weighs too much for a 40-year-old roof structure creates a liability you will carry for years. The Lingdu Intelligence range divides neatly by these constraints: Lingkong K3 for flat and low-slope skylights, Lingyun Y3 for larger flat glass roofs and atriums, Lingfeng S1 for tighter or sloped panels.
What glass types will the robot actually touch?
Glass type decides brush pressure and cleaning method more than any other factor. Coated and tinted glass scratches when you push a stiff brush with grit under it. Laminated glass can carry a heavier robot. Single-glazed panels, common on roofs built before 1995, flex under load and usually rule out anything above 8-10 kg.
- Laminated or toughened glass, 8 mm and above: most robots fit, 12-18 kg is fine.
- Tinted or low-E coated glass: soft brushes only, avoid abrasive pads, keep water pH between 6 and 8.
- Single glazing under 6 mm: lightweight units only, check deflection before you buy.
- Polycarbonate or acrylic roofs: not suitable for suction robots at all, the surface scratches and the seal fails.
How steep a skylight roof can a cleaning robot handle?
Slope is where cheap robots fail. A vacuum-adhesion robot holds by generating negative pressure against the glass, and the holding force has to beat gravity plus the shear force pulling it downhill. On a flat roof, the cup only fights the robot’s own weight. On a 30-degree slope, the downhill pull roughly doubles.
Practical limits we see in the field: tracked robots with slip control manage 0-15 degrees safely. Vacuum-adhesion units like the Lingyun Y3 hold 30 degrees without drama and can be pushed to 40 degrees on dry laminated glass. Beyond 45 degrees, nothing on the market is worth the risk, and you should look at rope access with a fixed anchor instead. Rain changes everything. Wet coated glass can drop adhesion by a third, so always derate the slope limit when the forecast is unsettled.
Why does robot weight matter for older glass roofs?
Point load is the number nobody asks about and later regrets. A 15 kg robot sitting on four small contact patches concentrates force. If your roof was designed to a 0.75 kN/m2 live load and you cannot confirm the purlin spacing, a heavy robot parked in one spot is a real problem. Ask a structural engineer before you sign, especially on retrofit projects. Wet cleaning adds water weight, and a 20 litre tank plus the robot can push a single panel close to 40 kg.
Can a skylight cleaning robot reach the whole roof?
Reach is the last filter. Robots clean well but not far. Most units work a 1.2 to 1.8 m wide path per pass, so a 30 m long skylight run needs either a guide rail or repeated repositioning by an operator on a walkway. On atrium glass roofs with a ridge at 12 m, you need a tether or a remote line rated for that drop. If the roof has no walkway, no parapet and no hatch access, budget for access equipment on top of the robot price.
Who should not buy here: single-person property managers cleaning one small domestic skylight twice a year. The setup time kills the payback. Who should: facilities teams running 800 m2 or more of glass roof across several buildings, where one operator with a robot replaces a two-person rope crew.
Do you need a rail, and when does it pay off?
On a skylight run longer than about 15 m with a consistent width, a fixed rail changes the job. The robot tracks the rail, so the operator no longer steers it back after every pass, and output climbs because ramp turns disappear. Rails cost money and need fixings designed for the roof, but on a building you will clean four times a year for a decade, they usually pay back inside three years. On an irregular glass roof with stepped panels, skip the rail and accept the slower hand-guided pass.
One detail people miss: the rail has to survive the roof movement. Thermal expansion on a 20 m aluminium rail across a hot roof can reach several millimetres, and a rigid fixing without a sliding joint will work loose within a season. Ask the installer how the rail handles expansion before you sign.
How much does the wrong choice cost?
More than the machine. If you buy a unit rated for 15 degrees and your roof runs at 25, you discover the problem the first wet morning, not in the showroom. The supplier may take the machine back, but you have lost the cleaning season and paid a crew anyway. If you buy too heavy for a fragile roof, the cost is structural and possibly uninsurable.
The cheapest mistake to avoid is the impulse buy at a trade fair, where the demo floor is flat, dry and spotless. Insist on a trial on your own roof, on your own glass, at your own slope. A supplier confident in the machine will agree. One who will not is telling you something.
Key Takeaways
- Match glass first: laminated and toughened glass take a heavier robot, single glazing does not.
- Slope derates holding force; derate again for wet glass, roughly a third.
- Check point load on retrofit roofs before buying anything above 12 kg.
- Reach, not price, decides whether one operator can cover the whole roof.
- Small single-skylight jobs rarely pay back; multi-building portfolios usually do.
Need a spec check against your actual roof drawings? Send the panel layout and slope to our team via the contact page and we will map it to K3, Y3 or S1.