Glass roofs, atrium skylights and curtain walls need different cleaning robots in 2026. A skylight runs shallow (5-20 degrees) on lower suction, an atrium glass roof mixes slopes and needs tether anchoring, and a vertical curtain wall needs the highest holding force plus a fall-arrest line.
Buyers often lump these three into one category called “glass cleaning robot” and then wonder why a machine that works on a mall skylight fails on a tower facade. The geometry is different, and geometry decides the machine.
How does slope change the robot spec?
A skylight sits between 5 and 20 degrees on most buildings. The robot feels mild gravity pull and can run a lighter chassis with onboard water. An atrium glass roof can tilt from 10 to 60 degrees across a single span, which forces a heavier anchor and tether. A curtain wall is near-vertical, so holding force is everything.
| Surface | Slope | Min suction | Power | Anchor |
|---|---|---|---|---|
| Skylight | 5-20 deg | 2200 Pa | Battery OK | Edge stop |
| Atrium glass roof | 10-60 deg | 2800 Pa | Hybrid | 2 edge anchors |
| Curtain wall | 75-90 deg | 3000 Pa+ | Tether | Fall-arrest line |
Why do atrium roofs need more anchoring than plain skylights?
Because the slope varies. A robot crossing from a 12 degree panel to a 55 degree panel experiences a sudden load change. Without tether tension the machine can slide on the steeper side, especially when the glass is wet during cleaning.
The Lingyun Y3 is built around this mixed-slope atrium case. The Lingfeng S1 leans toward flatter, tight-bay skylights where agility beats raw holding power. The Lingkong K3 handles taller, more exposed roofs where reach and tether safety dominate.
What breaks on curtain wall that does not break on skylights?
Wind. A vertical facade at 40 m catches gusts that a horizontal skylight never sees. The 2026 rule of thumb: stop operations above about 12 m/s sustained wind on facade work, and log the gust reading before each cycle. Edge cups lose seal first when wind gets under the chassis.
Water drainage also changes. On a skylight, runoff flows to the gutter. On a curtain wall, cleaning water runs down the glass face and can streak lower panels if you use too much detergent. Spot-free rinsing matters more on facades than on roofs.
What about glass with coatings or films?
Low-E coatings and applied films are softer than plain float glass and scratch under grit. Any robot running on coated glass should use a soft brush and filtered water, and the squeegee blades need checking for nicks before each cycle. A single embedded grain can leave a permanent line across a whole panel.
Tinted and reflective glass hides dust less well than clear glass, which is why buildings with mirrored facades tend to clean more often. The visual threshold for “dirty” is lower when the glass mirrors the sky.
How do you plan the first clean on a new building?
Post-construction glass carries cement dust, sealant residue and finger marks that a normal cycle will not shift. Run a dry pass first to lift bulk grit, then a wet pass with a mild detergent, then a rinse. Skipping the dry pass drags abrasive dust across the surface and dulls both the glass and the brush.
How do you map the roof before choosing?
Sketch it bay by bay. Mark slope changes, mullion spacing, and every obstacle. That single drawing tells a supplier more than a dozen photos, because it captures the numbers that decide the machine: worst slope, longest run, tightest bay, and highest point.
Once you have the drawing, the choice between a light skylight machine and a tethered facade unit becomes obvious. Buyers who skip this step end up with a machine that is wrong for the roof and right for nothing.
How often should each surface be cleaned in 2026?
A working rule for temperate cities: curtain walls every 6-10 weeks, atrium glass roofs every 8-12 weeks, and interior skylights twice a year. Rain resets the clock a little on exterior glass, but pollen and city dust bring the film back faster than most owners expect.
Coastal sites should shorten every interval by roughly a third. Salt film is invisible for a while, then it dulls reflections suddenly, and by then you are cleaning more often anyway. Plan for the extra cycles rather than reacting to them.
- Curtain wall: every 6-10 weeks.
- Atrium glass roof: every 8-12 weeks.
- Interior skylight: twice a year.
- Coastal sites: cut intervals by about a third.
Which surface gives the best payback?
Curtain walls, on most 2026 projects, because they are the largest glass area and the hardest for humans to reach. Atrium glass roofs come second. Small decorative skylights rarely justify the hardware on their own.
The exception is a building with all three. Running one robot family across skylight, atrium and facade cycles spreads the training and spares cost and lifts the return on the whole set.
Send Skylight Cleaning Robot your elevation drawings – the team will tell you which surface to automate first.
Key Takeaways
- Slope, not glass size, is what separates skylight, atrium and facade robots.
- Steeper mixed-slope atrium roofs need tether anchoring, not just suction.
- Facade work stops around 12 m/s sustained wind in 2026 practice.
- Runoff and detergent control matter more on vertical glass than on roofs.
- Curtain walls usually pay back fastest because they are the largest glass area.
Should one robot cover all three surfaces?
Rarely with a single unit, unless the roofs are modest. A machine tuned for a vertical facade carries extra holding force and weight that make it clumsy on a shallow skylight. A light skylight machine would not survive the wind load on a 40 m facade.
The practical approach is one family, two machines – a lighter unit for skylight and shallow atrium glass, and a heavier tethered unit for the facade. Shared consumables and shared training keep the overhead down even with two bodies.