A roof cleaning robot for skylights works on framed glass up to about 35° of pitch when a rated tether is used. It fails on frameless structural glazing, on roofs with deep penetrations, and on glass older than the seal life of the machine.
Skylight cleaning is the awkward job in every building. Too high for a pole, too small to justify a rig, and directly above whatever is underneath. Robots fit that gap well, which is why the category grew. They also fail in predictable ways that a spec sheet will never mention.
What roof cleaning robot fits a skylight in 2026?
Match three numbers: pitch, pane size and fall height. Pitch above 35° is where most portable units start slipping, because the suction chamber loses contact pressure at the uphill edge. Pane width under 400 mm leaves the body overhanging the frame, which breaks the seal. Fall height over 6 m means you need a rated anchor point, not the small cord that ships in the box.
Skylight type matters as much. Unit skylights with aluminium frames are the friendly case: a clean edge gives the machine a stop. Frameless structural glass roofs are the hostile case, because there is no physical edge and some models will drive straight off.
Why does a skylight cleaning solution matter more than the robot?
Because residue is what you see, not dust. Skylight glass is usually low-e or solar-control coated, and those coatings are soft. A dry pass with a stiff pad on a dusty low-e pane leaves a fine scratch grid that shows up in raking light within a month. Cleaners that leave mineral film do the other half of the damage.
The practical answer for 2026: demineralised or RO water with a neutral pH, and a pad that is spun or squeezed before the dry pass. Never run the first pass dry. Our guide to water, soap and pH goes deeper, but the short version is that most streaking complaints are water hardness, not machine performance.
What are the real failure modes on roof glass?
A short honest list, drawn from field reports:
- Seal lift on sun-warmed glass: a black rubber seal at 60°C softens and loses shape.
- Sliding at the downhill edge when the tether has slack.
- Cable snag on roof penetrations, vents and cable trays, which stalls the unit mid-pass.
- Water pooling at the frame and dripping into the building, not on the glass.
- Battery cut-off mid-pane on larger skylights, leaving a visible stop line.
None of these are exotic. All of them are avoidable with site planning, which is why a two-hour survey before purchase saves a lot of money.
Is a robot or a pole kit better for roof glazing?
Pole kits reach about 8 m and cost a fraction of a robot, but they need a stable standing position below the glass and they do not remove dirty water, only spread it. On a walkable flat roof with low skylights, a pole is fine and we have said so before.
Above 8 m, or where the only access is from inside the atrium, a robot is the practical option. If you are weighing the two directly, our robot versus pole comparison sets out the numbers. The Lingfeng S1 review covers the obstacle-crossing class that suits framed roofs with step changes.
Who should not put a robot on their roof?
Buildings where the skylight glass is laminated and damaged, or where the frame has visible deflection. Suction force on a compromised pane is a bad idea. Also anyone who cannot define an anchor point without drilling into waterproofing they do not want to touch. And heritage buildings with leaded or single-glazed lights, where the glass itself is the asset.
Roof layouts usually fail the plan in the same place: the transitions. A skylight run that changes pitch between two sections, or steps down 150 mm at a wall junction, forces a stop and restart, and every forced stop is a chance to lose the seal. Sketch a plan of your roof and mark where a machine would have to stop. If the answer is more than two per run, a robot is competing with a pole kit on labour, and probably losing.
Timing matters as much as equipment. Glazed roofs are hottest between noon and four in summer, and surface temperatures on dark glass roofs in the Gulf regularly exceed 70°C. Rubber seals take a permanent set at those temperatures. Operators we work with clean between 06:00 and 09:00 and accept the earlier start as part of the job.
One structural caution: never assume the frame can carry a tether anchor just because it carries the glass. On older skylights, the fixings holding the glazing bars may be far weaker than the glazing itself. A structural engineer’s two-line note on anchor capacity is cheap; discovering the limit during a cleaning job is not.
The maintenance rhythm that keeps skylights clean
Most skylight dirt is not uniform. Rain leaves a mineral ring at the frame, birds leave deposits on the high side, and pollen collects in the lower corner of each bay. A robot running a fixed pattern handles all three about equally well, which is the wrong distribution. Two passes a year on the whole roof, plus a targeted manual wipe of four corners per bay every autumn, gives better results than four robot passes and no corners.
Keep a small kit at roof level: a water-fed pole for the corners, two microfibre cloths, a litre of demineralised water, and a spare seal. That kit costs less than one service visit and covers the jobs a robot cannot reach, which are the ones clients notice when they look up from inside.
Key Takeaways
- Portable skylight robots handle framed glass to roughly 35° pitch with a rated tether.
- Frameless structural glazing is the failure case; there is no physical edge to stop the unit.
- Residue and coating scratches come from water chemistry and dry passes, not motor weakness.
- Sun-warmed seals soften; run early morning on exposed roofs where possible.
- Pole kits win under 8 m on walkable roofs; robots win above that or from interior access.