A skylight cleaning robot covers 300-600 m2/h on open flat glass in 2026. On framed roofs expect 200-350 m2/h because of mullion crossings, and on a vertical curtain wall 150-250 m2/h. Battery life of 3-4 hours and an 8-12 L tank are the practical limits on a single shift.
Vendors quote the top number. Planners should use the middle one, because the factors that cut throughput are not exotic.
What limits a shift, battery or water?
Usually water. A machine with an 8 L tank using 0.6 L per 100 m2 runs dry after roughly 1,300 m2, which at 400 m2/h is a little over three hours. A lithium pack giving 3.5 hours lines up almost exactly, so on wide open glass the two constraints hit at the same time.
| Surface | Coverage | Limiting factor |
|---|---|---|
| Flat atrium glass | 400-600 m2/h | Water tank |
| Pitched glass roof | 300-450 m2/h | Reservoir + access moves |
| Framed industrial roof | 200-350 m2/h | Mullion crossings |
| Vertical curtain wall | 150-250 m2/h | Tether management |
The curtain wall number surprises people. The robot itself is not slower vertically. The crew is, because every 10-15 metres of descent needs tether attention and the launch point must be repositioned.
What slows a robot down in the real world?
- Framed glass: each mullion crossing costs a second or two of hesitation, and on a grid every metre it adds up fast.
- Hard water spots: a second pass to remove dried minerals can double the surface time.
- Strong sun: glass heats up, water evaporates before the brush passes back.
- Rooftop clutter: vents, cable trays and antennas force repeated manual repositioning.
- Loose dust: on very dusty glass, the brush skates and needs a pre-wet pass.
Dust is the underrated one. On a roof near a construction site or a cement plant, the first pass over dry glass can spread dust into streaks. A quick pre-wet with the onboard spray, then a normal pass, is faster than cleaning the streaks out afterwards.
How do you plan a realistic daily output?
Take the surface coverage rate, then apply a 65% efficiency factor for setup, moves, refills and battery swaps. A crew planning 400 m2/h should schedule about 260 m2/h of actual delivery over a working day. On a 3,000 m2 glass roof that is a little over 11 hours of machine time, or two days with one robot.
Two robots do not simply halve the time either. On a single roof they get in each other way at the launch point, and the second operator still needs the same safety setup. Plan for around a 40% gain, not 100%.
Which specification should you optimise?
If your roof is flat and open, buy the largest tank and the longest battery life you can afford. If your roof is framed, prioritise roller clearance over raw coverage, because a robot that stalls on every mullion will never hit its rated speed. Our guide on how to choose a skylight cleaning robot walks through that trade-off, and skylight cleaning robot maintenance covers the consumables that quietly erode throughput.
Ask for a time-and-motion study on your own surface type before you commit to a productivity figure. contact the lingdu intelligence team to request one.
What does a shift plan actually look like?
For a 2,500 m2 pitched glass roof at 350 m2/h delivered after the efficiency factor, the machine needs a little over seven hours of surface time. Split across a morning and an afternoon block to avoid midday sun, with a 20-minute setup each block and one battery swap, and the job fits inside a single working day for one operator.
That is the number to put in a tender. A figure of 600 m2/h from a brochure translates to less than half that once setup, repositioning and refills are counted, and a contractor who bids on the brochure number will lose money on the job.
Frequently asked question: does cleaning frequency change throughput?
Yes, and in a way that is easy to miss. A surface cleaned monthly has light soil, so the robot runs one pass at full rated speed. A surface cleaned twice a year carries baked-on film that needs a slow pass and often a second one. The same machine can deliver 400 m2/h on the first building and 220 m2/h on the second.
This is why frequent cleaning is not just about appearance. It is cheaper per square metre and gentler on glass, because heavy film requires more brush pressure and more water. Schedule the interval from the soiling rate, not from a budget cycle.
What maintenance keeps the throughput honest?
A glazed, clogged brush reduces contact and slows effective cleaning even though the motor still runs at speed. Rinse brushes after every shift and replace them at 300-500 hours. A worn suction skirt also costs speed, because the machine pauses to re-grip more often on sloped glass.
Track actual delivered square metres per operating hour in a simple log. When the number drops by more than about 15% against the baseline, inspect the brush, the seal and the filter before assuming the machine has lost power. Our maintenance notes cover the intervals in detail.
Key Takeaways
- Plan on 300-600 m2/h on flat glass, 200-350 m2/h on framed roofs.
- Water capacity, not battery, is usually the binding constraint.
- Apply a 65% efficiency factor for setup, refills and repositioning.
- Two robots on one roof deliver about a 40% gain, not double output.