Expect 90-160 m2/h of actual cleaned glass from a skylight cleaning robot in 2026, not the 200 m2/h on the brochure. The gap is setup, repositioning, refills and rewash. Measure a full cycle, not a straight run.
Cleaning speed is the easiest spec to inflate and the hardest to defend. A robot gliding in a straight line on a clean pane can hit 200 m2/h. Put it on a real roof with a hatch, a corner, a refill and two rewash passes, and the true number falls by a third or more.
What counts in a real skylight robot cleaning cycle?
Four blocks. Travel and setup from the hatch to the first panel. The cleaning passes themselves. Repositioning between separated glass sections. And refill or battery swaps. On a 400 m2 mall canopy, setup and repositioning can eat 35-45 minutes before the robot has cleaned a useful metre. A Lingyun Y3 with a 3.5-hour battery and 12-litre tank covers most of a 500 m2 flat roof in one shift if the layout is open.
| Phase | Share of shift |
|---|---|
| Setup and travel | 10-20 percent |
| Active cleaning | 55-65 percent |
| Repositioning | 10-15 percent |
| Refill / rewash / pack | 10-15 percent |
How do you benchmark a robot fairly?
Time one full cycle on your worst panel, not your best. Record three numbers: total cleaned area, total shift time, and rewash area. Effective rate is cleaned area divided by total time. If the supplier demos only on flat interior glass, ask them to replicate your corner and your hatch. A machine can look twice as fast on a demo roof and land identical on yours.
Where does speed actually come from?
Three levers. Brush width: a 1.2 m swathe beats a 0.8 m one by roughly 50 percent per pass. Path overlap: dropping overlap from 20 to 10 percent saves a fifth of the time with no visible quality loss on flat glass. And refill design: a robot fed by a lightweight hose avoids tank stops entirely, which on a large roof can add 15 percent to a shift.
Who chases speed and loses?
Sites that push overlap down too far on visible atrium glass. The streaks show under downlights and you rewash anyway, so the time saving evaporates. Similarly, chasing m2/h on a hard-water site leads to skipping the rinse step, and sun-dried spots cost a full re-clean. Speed matters on large, plain, routine roofs; on premium glass, quality per pass wins.
The risk of over-optimised numbers
If your contract is priced on a brochure m2/h, you carry the shortfall risk. Write service agreements against measured effective rate after a two-week pilot, with a defined rewash threshold. Then budget consumables against hours, not area. Our efficiency, battery and tank guide carries the runtime numbers, and the contact page can help scope a pilot.
Key Takeaways
- Plan on 90-160 m2/h effective, not the 200 m2/h headline.
- Setup and repositioning eat 35-45 minutes on a real 400 m2 roof.
- Benchmark your worst panel, and count rewash area.
- Wider brush and tighter overlap are the real speed levers.
- Price service contracts on measured effective rate, not brochure figures.