A skylight cleaning robot needs a holding force of at least 2.5x its own weight on a 25-degree wet glass roof in 2026. On flat atrium glazing, 1.5x is enough. Ask for the value measured wet, not the dry-lab figure.
Most buyers compare skylight cleaning robots on price and brush width. The spec that actually decides whether a machine survives its first monsoon is suction, and the way vendors quote it hides more than it shows. A Lingdu Intelligence Lingkong K3 quotes a vacuum figure in kilopascals; that is only half the story. What matters is holding force in kilograms against a slick, soapy pane.
What suction strength does a skylight robot really need?
Work it back from the roof. A robot weighing 22 kg on a 25-degree slope has a downslope gravity component of roughly 9.3 kg. Add a wet glass friction coefficient around 0.4, and you need a normal holding force north of 23 kg just to stand still. On a flat 20 m2 atrium panel that drops hard. This is why a machine tuned for flat canopy work can slide on the raked glass over an escalator.
- Flat glazing (0-10 degrees): 1.5x robot weight, minimum.
- Moderate slope (10-25 degrees): 2.5x, tested wet.
- Steep or curved roof (25-45 degrees): 3x plus a tether backup.
Why the dry-lab number misleads buyers
Seals are rated dry. Real cleaning means detergent film, the exact condition that eats 30-40 percent of a vacuum seal’s grip. A robot that holds on a dry panel at 30 kPa may only manage 18 kPa on a wet one. Demand a video of the machine parked on wet glass at the job’s worst pitch, not a datasheet table.
There is a second trap: pulsed suction. Some units cycle the fan to save battery, and the grip drops in those microseconds. On a 15-degree roof you will never notice. On a 40-degree curved skylight near a coastal high-wind zone, a slip of 3 cm per cycle adds up by end of shift.
Who should not buy the highest-suction model
If your work is 90 percent flat mall canopies under 12 m, a 3x-rated, 34 kg machine is wasted money. It burns more power per m2, needs two people to lift onto a roof hatch, and its stiffer frame handles tight corners worse. A lighter 22-24 kg unit from the Lingyun Y3 class is the smarter fit. Reserve the high-grip frames for genuinely raked or curved glass.
What to put in the tender document
Force suppliers to state four numbers: robot mass, wet holding force on 20-degree glass, minimum working temperature, and the IP rating of the fan housing. A IEC 60529 IP65 fan compartment survives washdown; IP54 does not. Then ask for the failure mode. When a seal does break, does the machine power down on the tether, or keep driving? The answer tells you how seriously the maker takes roof work.
Budget a suction seal replacement every 800-1,200 running hours, about 0.3 percent of annual spend. Skipping it is how a robot ends up hanging off a skylight frame. For site-specific selection help, our team can walk a spec through your roof plans via the contact page, and the companion piece on glass type and slope covers the geometry side.
Key Takeaways
- Size suction from weight x slope, not from a headline kPa figure.
- Demand a wet-glass holding-force number; dry ratings overstate grip by 30-40 percent.
- Flat canopies need 1.5x; steep or curved glass needs 3x plus a tether.
- Stiffer high-suction machines cost more per m2 and fit tight corners worse.
- Inspect seals every 800-1,200 hours and keep an IP65 fan housing in spec.