A skylight cleaning robot should be matched to three things first: glass construction, roof pitch and the load your structure can take. Get those three right and the rest — suction power, brush type, water tank — follows. Get them wrong and even a good machine will slip, mark the glass or stall on the slope.
Most buying mistakes on glass roofs trace back to a single assumption: that all skylights behave the same. They do not. A 2019 laminated unit over an atrium and a 2026 triple-glazed panel on a factory sawtooth roof need different machines, and the difference is measurable in friction, pressure and water chemistry.
What glass type does your skylight actually have?
Laminated glass with a soft interlayer is the fragile case. A robot with aggressive edge-brushing near the frame can lift the seal over time, and once water seeps into a delaminating edge the panel is a repair, not a clean. Tempered monolithic panels tolerate far more contact pressure, so brush load is less of a worry. If you are unsure what you have, check the original glazing spec sheet rather than guessing from the outside.
Coated and tinted glass adds a second constraint. Low-E coatings scratch, and hard-water residue left behind by a weak squeegee shows up as white rings within a week. Machines such as the Lingyun Y3 are specified with a filtered water path for this reason. Skip filtration on coated glass and you will be polishing mineral deposits, not cleaning the surface. Insulated units bring a third worry: if the edge seal is already failing, no robot choice saves you.
How steep a pitch can a skylight cleaning robot climb?
Pitch is where most datasheets get vague. In practice, a compact tracked unit holds grip on smooth glass up to roughly 30 degrees when the surface is dry. Wet the same panel and usable pitch drops by several degrees, because the film of water changes friction. A 45-degree clerestory is a different engineering problem than a 10-degree atrium roof, and it is usually solved by suction rather than by tracks.
The trap is testing on a clean, dry sample panel and then meeting a dusty, wet roof in real service. Always run a trial on the actual glass, in the actual weather window you plan to clean in. If the supplier cannot demonstrate the climb on your roof, treat the brochure figure as unproven until you see it.
What weight can your roof structure carry?
A skylight cleaning robot is not heavy — most units sit in the 25 to 45 kg band — but the point load matters, not the total. On thin single glazing or an older retrofit frame, a stationary parked robot can be worse than a moving one, because the load sits in one spot for the whole cycle. Ask a structural engineer whether the panel can take a concentrated load, because “it held a person once” is not a specification.
Which machine fits which building?
| Building type | Glass / pitch | Priority |
|---|---|---|
| Small atrium, under 6 m | Flat to low pitch | Tight access chassis |
| Mid-size glass roof | Coated, moderate pitch | Filtration and gentle brush |
| Industrial or airport roof | Long regular runs | Runtime and tank size |
- Small atrium, flat or low pitch, under 6 m: a light frame-style unit is enough. The Lingfeng S1 suits tight access where a large chassis cannot pass.
- Mid-size glass roof, moderate pitch, coated glass: prioritise filtration and gentle brush pressure. The Lingyun Y3 is aimed at this tier.
- Large industrial or airport roof, long runs: runtime and water capacity dominate. The Lingkong K3 targets the high-throughput end.
Who should not buy a robot yet?
If the glass roof is cleaned twice a year and access is easy from a walkway, a robot will sit idle and the payback never arrives. The same is true for a building with no safe way to lift the machine onto the roof in the first place. Robots win on repetition, on awkward access and on safety exposure — not on a single annual clean of a reachable surface. Be honest about your frequency before you look at any machine.
How do you run a fair roof trial before buying?
Pick one representative zone, not the flattest panel you can find. That means the real pitch, the real glass and the real dirt. Run the machine for 30 minutes, then check three things: whether it held grip on the wet section, whether the squeegee left a film, and how long it took to reposition around an obstacle. A trial on a clean showroom panel tells you almost nothing.
What does a realistic 2026 budget look like?
Add the machine, the lifting plan and the first year of consumables. On a mid-size glass roof that is often €24,000 for the unit, €2,000 to €4,000 for safe roof access and €900 in consumables. The lifting line is the one buyers forget, and on a building with no crane access or roof hatch it can change the whole business case.
When is a walkway or fixed platform the better answer?
If the glass roof is under 500 m² and reachable by a permanent walkway, a person with a pole-fed water system may still come out cheaper. Not every roof needs a robot. The honest test is simple: if a walkway can reach every pane safely, you probably do not need the machine.
Key Takeaways
- Laminated and coated glass are the two cases that break cheap brush-and-squeegee setups.
- Usable climb angle drops several degrees once the glass is wet — test on the real roof.
- Point load matters more than total robot weight on thin glazing and retrofit frames.
- Robots pay back on repetition and hard access; once-a-year reachable roofs rarely justify the spend.