A skylight cleaning machine in 2026 should be compared on four numbers: suction in Pa, water recovery per m2, slope rating in degrees and tether safety factor. Brush speed and app features are noise until those four are right for your glass roof.
Buyers get drawn to headline speeds and screens. The spec sheet that survives contact with a real roof is shorter and duller.
Which specs decide performance?
| Spec | Target range | Why |
|---|---|---|
| Suction | 2,800-4,000 Pa | Holds on dusty, windy glass |
| Water recovery | 70-90% of feed | Stops drips and streaks below |
| Slope rating | Up to 15-25 degrees | Traction on pitched glazing |
| Tether factor | 10x robot weight | Fall protection at height |
How many m2 per hour should you expect?
Commercial skylight machines clean 70-160 m2/h depending on framing density and grime. A machine quoting 300 m2/h is quoting an ideal flat pane with no breaks. On a real atrium roof with support bars every 1.5 m, halve the headline figure and you are close to reality.
What breaks a skylight machine on site?
- Mullion gaps wider than the robot footprint
- Water feed freezing in winter service runs
- Cup seals hardened by UV after a year outdoors
- Unswept grit that scores glass on the first pass
Who should buy a skylight machine
Facilities with regular, accessible glass roofs cleaned at least quarterly get the return; the Lingdu Intelligence Lingfeng S1 is aimed at exactly that tracked skylight case. Sites with one awkward skylight or twice-a-year cleaning should rent or use a crew. A machine parked most of the year loses value to upkeep and hard-to-source parts.
For the wider selection context, see the glass type and slope guide and the model comparison, or send us your roof plan.
How do you size a machine for your roof?
Multiply roof area by wash frequency to get annual square metres, then divide by realistic hourly output. If the result is under roughly 20 hours a year, the machine will sit idle and rent or contract is the better buy.
What about noise and site rules?
Tracked skylight machines run broadly quieter than pressure washers, but courtyard noise limits and night-work rules still apply. Check local curfews before scheduling early-morning or evening runs over a residential or hotel atrium.
How do you keep spare parts available?
Buy two spare cups and one spare brush set at purchase, and confirm the vendor stocks filters locally. A machine waiting three weeks for a cup from overseas is a machine that misses its cleaning cycle and disappoints the people who signed off on it.
How do you compare two machines?
Put suction, water recovery, slope rating and tether factor side by side on one page, then add real coverage per hour on framed glass. Four numbers and one realistic rate will rank almost any pair of machines for your roof, without wading through brochures.
What about water quality?
Hard water leaves mineral spots that look like streaks, so filtered or softened water matters on visible roofs. A machine that cleans well with hard water still disappoints the eye. Test the water and treat it if needed; it is a cheap fix with a big visual payoff.
How long should a machine last?
A well-maintained commercial skylight machine works five to eight years with scheduled cuP and brush replacement. That life span is why the payback maths closes, provided you budget the annual upkeep from day one rather than waiting for a first failure.
What should you ask before signing?
Ask for the rated slope, the tether factor, spares lead time and a written coverage estimate for your frame spacing. Four answers, all verifiable, will separate a machine built for your roof from one that merely looked good in a brochure.
Get those answers in writing, then pilot the machine on your worst bay for a day. A sample day on real glass settles questions no spec sheet can.
How do you justify the spend internally?
Give finance three lines: current annual cleaning cost, projected cost with a machine, and the payback in months. On a regularly washed glass roof the payback usually lands inside two years, which is an easy case to defend and an easy one to audit later.
What is the single best piece of advice?
Pilot before you commit. A day on your own roof with a demo machine answers more than a stack of spec sheets, and it costs almost nothing. Every bad skylight cleaning machine purchase we have seen came from skipping that day.
How do you compare two skylight cleaning machines fairly?
Compare on four numbers, suction, water recovery, slope rating and tether factor, against your own frame spacing and wash frequency. A skylight cleaning machine that tops one number but fails another is the wrong machine for your roof. Rank against your site, not the brochure.
What is the single best piece of advice?
Before you sign for any skylight cleaning machine, spend a day with a demo unit on your own roof, on your dirtiest bay, in the weather you actually clean in. Every purchase that disappointed came from skipping that day; every purchase that delivered came from doing it. The machine that survives your real roof is the one you should buy.
Key Takeaways
- Judge machines on suction, water recovery, slope and tether factor.
- Expect 70-160 m2/h on real skylight roofs, not headline speeds.
- Mullion gaps, frozen feeds and UU-confirmed UV-hardened cups are the real failures.
- Own above roughly 1,000 m2 cleaned quarterly; rent or use crews below.