Ask five questions before buying a skylight cleaning robot in 2026: measured m2/h on dirty glass, suction holding margin over the rated slope, water recovery method, spare-part lead time, and how many operators get trained. Vague answers on any of the five are a warning.
A spec sheet answers none of these properly. The point of the questions is not to hear a number, but to see whether the supplier has a real number at all.
Question 1: What is the measured m2/h on dirty glass?
Not peak, not clean-pane. Ask for a figure on a first pass over dirty glass, with the pitch stated. A confident supplier names a range like 100-150 m2/h and explains what changes it. A weak one repeats the spec-sheet peak of 350-400 m2/h and cannot say what happens on a slope.
Question 2: What is the suction margin over the rated slope?
Suction is a safety function. Ask for the holding force and how it compares on the steepest glass the robot is rated for. A margin below 1.3x on the rated slope is thin. Watch the answer for honesty about wet or dusty glass, where grip drops. Suppliers who claim no degradation anywhere are telling you they have never tested it.
Question 3: How is water recovered, and where does it go?
Ask how the machine removes dirty water. A recovery squeegee and tank is the answer for atria and any glass above people. A spray-and-drag design leaves streaks and drips. Get the tank capacity in litres and the m2 per fill. If the supplier talks only about cleaning quality and never about the water going somewhere, that is the answer.
Question 4: What is the spare-part lead time?
Suction cups, brushes and filters wear out. Ask the lead time for a cup set and a brush, in writing. A 2-4 week lead time means you need a spare set on site from day one. Ask, too, whether parts are the manufacturer’s own or a rebadged third-party set, since that decides grip and safety.
Question 5: How many operators are trained, and does it include refresher training?
Two operators minimum, and a refresher when the robot’s firmware or brush setup changes. A supplier who trains one person and leaves hands you a single point of failure. Ask what happens when that person leaves.
| Question | Good answer | Warning sign |
|---|---|---|
| Measured m2/h | range on dirty glass, pitch stated | only spec peak |
| Suction margin | 1.3x+ with wet/dust noted | claims no degradation |
| Water recovery | squeegee + tank, litres given | no mention of water |
| Spare lead time | 2-4 weeks, parts stated | vague or silent |
| Training | 2+ operators, refreshers | one trainee, no follow-up |
Who can skip these questions?
Nobody buying a robot for glass above people. Even a small Lingfeng S1 on one skylight deserves questions 2 and 3, because suction and drips are the two failure modes that cause real damage. Price-led buyers skip them and discover the cost in year two.
One trap: a supplier who answers only by email and never offers a site visit. Suction and tether routing are site-specific. Push for a visit before you buy. The team here arranges them through contact.
How do you test a supplier instead of trusting the answers?
Ask for one reference from a site like yours, then call it. Ask that reference what went wrong in year one, not whether they like the robot. Every deployment hits something: a tether snag, a stubborn jet, a part delay. A reference who reports the problem and the fix is more useful than one who reports only praise, because it tells you how the supplier behaves when things break.
The second test is a site visit. A supplier who walks your roof, checks anchor options and warns you about the awkward glass is worth more than a discount. One who quotes from a photo and never visits is pricing a machine, not your problem.
What should be agreed in writing before purchase?
Five items: the measured performance basis, the suction margin on your steepest glass, the water recovery method and tank capacity, the spare-part lead time and first stock, and the training scope. Put them in the contract, not an email. When a dispute appears in month eight, the written basis is the only thing that settles it.
| Item | Write it as |
|---|---|
| Performance | measured m2/h range, pitch stated |
| Suction | margin over rated slope, wet/dust noted |
| Water | recovery method, litres, m2 per fill |
| Spares | lead time and opening stock |
| Training | operators count and refresher terms |
When is the cheap offer actually fine?
For a small indoor skylight with easy access and a buyer who will run the robot themselves, a lower-cost unit can work, provided suction and water recovery are sound. The expensive parts of this market are weather exposure, slope and scale. If you have none of those, skip the premium and put the savings into spare parts, which you will still need.
What answer should make you walk away?
Any variation of we will figure it out after you buy. Suction margin, water recovery and spare-part supply are not details to settle later; they are the reasons a robot program succeeds or fails. A supplier who deflects them has either not tested the machine or does not want the comparison. Either way, the risk lands on your roof.
The subtler warning is a supplier who answers everything perfectly and never asks about your roof. A good supplier wants to know your glass type, pitch, access and water point, because those change the recommendation. A pitch that ignores your site is a sign the answer is a template.
What does a strong supplier relationship look like after the sale?
Two signals. First, a first-year check-in to review the maintenance log and cup condition, not just a warranty reminder. Second, a named contact for parts, so an order does not disappear into a general inbox. Neither costs the supplier much, and both predict whether year three goes smoothly. For buyers, ask directly: who do I call for a cup set, and how fast does it arrive? If the answer is vague before the sale, it will be vaguer after.
The best protection is a short written performance agreement tied to the five questions above. It turns a sales conversation into a specification, and a specification is what survives a change of staff on either side.
Key Takeaways
- Demand a measured m2/h figure for dirty glass, with pitch stated.
- Insist on a suction margin of 1.3x or more on the rated slope.
- Confirm water recovery method and tank litres per fill, especially above people.
- Get spare-part lead time in writing and keep a spare set on site.
- Two trained operators minimum, with refresher training after setup changes.