Curtain wall robots climb vertical glass using cables and winches; skylight robots work horizontal and pitched roofs on suction pads. One machine rarely does both well because the safety cases are completely different.
Buyers keep asking for a single unit that handles the facade in spring and the atrium roof in summer. It sounds efficient. In practice you end up with a machine that is mediocre at both and a risk assessment nobody can sign off.
What is the physical difference between the two jobs?
A facade robot fights gravity every second. Everything about it is designed around not falling: redundant cables, a top anchor, a winch with a brake, and a tether that would survive a full power loss. Its suction is a secondary hold, and it moves in long vertical passes.
A skylight robot has gravity on its side until the pitch exceeds roughly 30 degrees. It can use lighter pads because the glass is under it, not beside it. Its risk is different: slipping downslope on wet glass, dropping through an open vent, or overloading a single laminated pane.
| Factor | Curtain wall robot | Skylight robot |
|---|---|---|
| Primary hold | Cable + winch anchor | Suction pads on glass |
| Typical pitch | 70-90° | 0-25° |
| Fall consequence | Catastrophic, needs redundancy | Panel damage, lower fall risk |
| Cleaning pattern | Long vertical passes | Raster over panel field |
Can one robot really do both?
Some vendors claim it. Ask them one question: if the suction fails at 60 metres on a vertical wall, what holds the machine? If the answer is a cable, then the machine is a facade robot with a gimmick roof mode. If the answer is nothing, walk away.
The honest split is: buy a skylight machine for glass roofs, atria and pitched industrial glazing, and hire or buy a separate facade system for vertical curtain walls. Lingdu Intelligence’s Lingkong K3 and Lingyun Y3 are skylight machines first; that focus is why their load and slope numbers are credible.
What about curtain wall robots that work from inside?
These exist and they are genuinely useful for double-skin facades and high atria with internal access. They still need a defined track or rail, and they still cannot reach external mullion faces. For a building with a sloped atrium roof above a public concourse, an internal access route plus a skylight robot is often the cheapest safe combination.
Who should buy which
If 70 percent or more of your glazing is roof glazing — factory sawtooth, mall atria, station canopies — buy a skylight robot. If your building is a tower with a uniform vertical facade, buy or contract a facade system. If you have both in equal measure, run two contracts rather than forcing one machine to cover two safety cases.
The one thing not to do: put a skylight robot on a vertical wall because it has suction pads. The pads are sized for a glass surface beneath the machine, and downslope hold on a vertical pane is a different calculation entirely.
Selection questions
- What percentage of your glazing is pitched versus vertical?
- Is there a fixed anchor rated for a facade winch?
- Can water and power reach the roof edge, or only ground level?
- Who signs the fall-arrest risk assessment for each surface?
- How many separate access windows do you get per year?
Send your glazing breakdown through the contact page and we will tell you honestly which half of the problem our machines solve. The surface-by-surface cleaning guide walks through each glass type in more detail.
What does a mixed contract look like?
Many buildings split the work by surface and by season: facade in spring when pollen lands, roof glazing after rainy months when the film builds. That split lets each machine do the job it was designed for and keeps the risk assessments separate and simple. It also makes each contractor accountable for a defined surface rather than a general cleaning promise.
Does one machine mean one operator?
Mostly. A skylight machine runs with one trained operator plus a spotter if there is a drop below. A facade system usually needs a ground crew at the winch anchor as well. That difference in headcount is a quiet part of the cost case, and it is one of the reasons roof robots pay back faster than facade robots on the same area.
How do you compare two quotes fairly?
Strip both quotes to the same scope: machine, spare pads and brushes, filtration, tether and anchor, training, warranty and spare-part lead times. Vendors quote wildly different packages, so a lower price often just means fewer spares and shorter support. Ask each to state coverage per shift against a roof like yours, in writing. The vendor who hesitates on that number is the one to leave out of the shortlist.
What does a typical cleaning cycle look like?
A full cycle is more than the time the machine spends cleaning. It is the walk to the roof, the water fill, the boundary setup, the run itself, and the pack-down. On a well-mapped roof the setup is 20 minutes and the run dominates. On a difficult roof with restricted access it can be 45 minutes before the machine lays down its first pass. Budget the whole cycle, not the pass, when you plan a shift.
How does glass type change the brush and pad choice?
Coated glass, laminated units and textured or fritted panels all behave differently under a rotating brush. Fritted glass wears pads faster and needs lower brush pressure. Coated low-E glass scratches easily, so a clean pad and a fine brush are not optional. Textured or prismatic glazing traps dirt in the pattern and often needs more passes, which quietly doubles your clean time on those sections.
Key Takeaways
- Facade robots are held by cables; skylight robots by pads against gravity.
- Skylight machines work 0-25 degrees; facade systems cover 70-90 degrees.
- If suction fails on a vertical wall, something else must hold the machine.
- Interior-facing robots suit double-skin facades and high atria.
- Two contracts beat one compromised machine when the split is close to even.