An ECOVACS window cleaning robot works on skylights only after you check three things: pitch under about 40 degrees, a frame lip thick enough to stop the machine at the bottom edge, and laminated glass rather than frameless sealed units. The hardware is capable; the roof geometry is what decides whether it stays useful.
What does the ECOVACS range actually cover?
The company sells three window machines with different footprints. The WINBOT W2S Plus is the everyday square-window unit. The WINBOT W3 is the newer body aimed at the same job. The WINBOT mini 2 shrinks the footprint for smaller panes and tighter frames. None of them was designed around skylight work; skylight work is something owners try after buying one for windows.
That is not a criticism. It simply means the fit question is yours to answer before you mount the machine on a sloped roof.
At what pitch does a vertical-window robot stop making sense?
Most vertical units assume gravity pulls the machine down the pane, so the safety design is a bottom stop and a tether. Rotate the same machine onto a 35 degree skylight and gravity now pulls it across the glass as well as down. The cup array still holds, but the shear load on the rims rises, and the machine drifts on the slope between cleaning passes.
A practical ceiling is around 40 degrees for a vertical-window machine with a dry, well-sealed surface. Beyond that, the gripping envelope of a machine built for facade work is the better answer.
Why does frame width decide the job?
Two reasons. First, edge detection: most machines stop when a sensor no longer sees glass. A narrow mullion can be invisible to that sensor. Second, the bottom stop. On a vertical window the machine rests against the sill. On a skylight it rests against the bottom frame bar, and that bar is often only 20-30 mm deep.
Which skylights are a hard no?
Frameless laminated glass with no perimeter stop is the obvious one. Old single-glazed panes with visible seal failure are the second, because the machine’s spray pushes water where it should not go. Sealed units with a 40 mm frame are workable but tight. Our ECOVACS WINBOT W3 review covers how the newer body handles those tight margins.
What should a buyer check on site?
Three measurements and one photo. Pitch in degrees, frame lip width at the bottom, glass thickness. Then a photo of the walkway directly under the skylight. The photo answers the fall-risk question that specs never do.
How does the machine behave on wet, dirty skylight glass?
Worse than on a clean vertical window, for two reasons. First, a wet film lowers friction, so the cup rims have less mechanical resistance to sliding. The vacuum is doing all the work rather than sharing the load with friction. Second, skylights collect a mix of dust, pollen and bird residue that turns into a slippery paste once sprayed.
Run a dry pass first on heavily soiled glass, then a wet pass. Two passes cost more time but prevent the drift that shows up as a slow slide toward the frame during the wet cycle. If the machine has a water volume setting, start at the lower end on skylights and increase only after watching one full pass.
The practical consequence: a machine bought for vertical windows may work fine on a ground-floor skylight and frustrate on a rooftop one. Same machine, different surface condition.
What does maintenance look like after a roof season?
Cup rims first, then sensors, then the tank. Silicone cups pick up grit and harden with UV exposure, and a roof season puts a machine in stronger sun than a window job ever does. Inspect the rims for a flattened profile or a hard, shiny surface; either means the seal is no longer conforming to glass texture.
Edge sensors are the second wear point, because roof glass carries more dust and a dusty sensor misreads a mullion as glass. Wipe sensors before each day of use on a roof. This single habit prevents most of the pauses that owners describe as random.
The water path is the third. Filters clog with roof dust faster than with window grime, and a clogged filter reduces spray volume, which owners interpret as weak cleaning rather than a maintenance issue.
Is there a case for mixing brands on one building?
Only if the roofs are genuinely different. A tower with both vertical curtain wall and a glass atrium roof is a reasonable case for one facade machine and one roof machine, because the failure modes differ so much. Two identical machines on two differently shaped roofs usually means one is fighting its geometry every visit.
Mixing brands on the same roof is a training problem. Operators develop habits tied to one body shape, one remote layout, one alarm vocabulary. Swapping between units mid-season produces handling errors that look like hardware faults.
The safer pattern is one model per surface type, with spares held for each. Spare cups are cheap; a machine sitting idle waiting for a cup from overseas is not.
Key Takeaways
- 3 models is the number that decides most jobs, ahead of any feature on the box.
- Plan around 15 degrees as a practical limit; beyond it the machine stops being the right tool.
- Maintenance routines beat spec sheet comparisons over a two-year horizon.
- Test on one pane with the safety device fitted before committing a full roof.
- Compare against a rope-access quote before buying, not after.