Skylight cleaning machines fall into three tiers by suction class and water recovery: light consumer units for single panes, mid-tier roof-capable machines for atria, and tethered commercial systems for glass above occupied space. A five-minute test on your own roof tells you which tier you need before any sales call.
Most buyers start at the wrong end. They look at machines, then try to fit the building to the spec sheet. Reverse it. The building decides the tier, and the tier narrows the shortlist to about four models.
What is the five-minute roof test?
- Measure three panes, not the whole roof. Note length, width and pitch with a phone level.
- Count obstacles in one square metre: mullion ridges, gutters, sealant beads, vents.
- Photograph the perimeter margin between glazing and parapet, with a tape measure in frame.
- Note what is below the glass. A car park is different from a food court.
- Check where a tether could anchor, and whether anyone will approve drilling into it.
That is it. Those five answers map directly onto tier, and they cost nothing. Every buyer who skipped this step and bought on brand recognition either overpaid or underbought, and underbuying is the more expensive mistake because it usually gets discovered mid-job.
How do the three tiers actually differ?
| Tier | Suction class | Recovery | Tether | Typical price (USD) |
|---|---|---|---|---|
| Light | Single pump, 240-280 mm seal | Wiper and absorbent strip | Optional, short | 400-900 |
| Mid | Dual pump, 300-340 mm seal | Vacuum lip recovery | Rated, 15-30 m | 900-2500 |
| Commercial | Dual pump plus back-up, 340 mm+ | Vacuum recovery and filtered tank | Rated, 30 m+, documented | 2500-7000 |
The jump from light to mid is mostly about recovery and obstacle handling. The jump from mid to commercial is about documentation: load cases, spare parts commitments, service network, and a tether that an engineer will sign off on. You are paying for liability transfer as much as hardware.
Does a bigger pad always mean faster cleaning?
No. On a grid glazing roof, the pad is wider than the pane often enough that the machine spends its time repositioning rather than cleaning. A 320 mm pad on 900 mm wide panes makes sense. The same pad on 600 mm panes means constant edge work.
Productive cleaning rate is area divided by cycle time, including turns, repositioning and refills. In one field log we saw a wide-pad machine return 82 m2/h on open panelling and 41 m2/h on a bayed atrium roof with the same unit and operator. Same machine, half the throughput. Spec sheets quote the first number.
What about the glass you can see through?
Visibility is the tier selection criterion people forget. On an opaque roof, residue is invisible. On an atrium skylight, it is the whole point of the job. If occupants stand under the glass, tier 1 machines will disappoint within one season regardless of how carefully they are run.
The Lingyun Y3 sits in the mid tier and was designed with that constraint. Our Lingyun Y3 review covers the visible-glass case and the obstacle handling, which matters more on an atrium than raw coverage rate.
What does a first-year budget look like per tier?
Light: 600 USD machine, 80-150 USD consumables, zero support. Practical lifespan under heavy use, 18-24 months. Mid: 1,800 USD machine, 200-350 USD consumables, one spare battery. Lifespan 3-4 years. Commercial: 4,500 USD machine, 500-900 USD consumables, service contract at 8-15 % annually. Lifespan 5-7 years with servicing.
Divide total first-year cost by cleaned square metres per year to get your real rate. A light machine on 400 m2 is competitive. A light machine on 4,000 m2 is a false economy, because it will be replaced twice before a commercial unit wears out once.
Who should buy which tier, plainly
Single residential skylight under 6 m2, accessible with a pole: light tier, no debate. Atrium or glass roof of 50-300 m2 with gantry access: mid tier, plus accessories for margins. Glass above occupied space, above 15 m, or any building with an insurance-driven maintenance programme: commercial tier, and ask for the load case document before signing.
Anyone who wants to understand which specific machines land in which tier, including on pitched and stepped glazing, should start with the consolidated skylight robot reviews rather than individual product pages.
Key Takeaways
- The building picks the tier. Measure three panes, obstacles and tether anchor points first
- Light tier suits single panes; recovery and obstacle handling define the mid tier.
- Commercial tier is largely about documentation and liability transfer, not raw hardware.
- Wide pads can halve throughput on bayed roofs. Cycle time, not pad width, decides rate.
- Divide first-year cost by annual cleaned area to compare tiers honestly.
Does the tier depend on the building or the occupant?
Both. A glass roof over a warehouse and a glass roof over a hotel lobby may have identical geometry and completely different cleanliness standards. The tier test is about acceptable results, not just area. Write the standard down before the vendor visits.
Can you upgrade a tier later?
Not really. Suction class, recovery system and tether rating are structural. Buying one tier lower than you need and planning to upgrade is the most expensive route in the market. Get the tier right once.
How long does the roof test take once you are on site?
Twenty minutes with a phone, a tape and a spirit level. Pitch on three panes, obstacle count in one square metre, perimeter margin measurement and two photographs of the area below the glass. That is enough to eliminate half the shortlist before any sales visit.
If the roof is not accessible safely, that fact alone changes the recommendation. No machine solves an access problem, and a job that needs a rope access crew to reach the roof also needs a crew to run the robot.