Before buying a skylight cleaning robot in 2026, ask five questions: what suction data exists for your slope, how long spare cups and filters take to ship, who trains the operator, can you trial on your worst bay, and can they produce CE and IP paperwork. Weak answers on any one are a warning.
Most buyers spend their time comparing price and throughput. The deals that go wrong go wrong on the boring questions – spares, training, and paperwork – asked too late.
1. What suction data do you have for my slope?
Not a datasheet number. Ask for the test result on a slope and glass type close to yours, including the glass surface finish. Suction performance changes between coated, tinted and plain glass, and between dry and wet surfaces.
If the supplier only has a lab number measured on dry float glass at 5 degrees, and your roof is 45 degrees with a low-E coating, you are buying an untested claim. Ask for the actual test conditions.
2. How fast can you ship spare cups and filters?
A robot grounded for two weeks because a cup set is stuck in customs costs more than the cup. Ask for warehouse location, typical lead time, and whether they hold stock locally. A supplier with no regional spares holds your uptime hostage.
| Question | Good answer | Red flag |
|---|---|---|
| Suction data | Slope and glass specific | Only a lab number |
| Spares lead time | 3-10 days regional stock | 6-8 weeks from factory |
| Training | Onsite, 1-2 days, included | PDF only |
| Trial | On your site, your bay | Showroom only |
| Certification | EC DoC + IP report | CE logo, no document |
3. Who trains the operator, and what does it cover?
Training should be onsite, hands-on, and cover the failures, not just the happy path. A good session includes what to do when a cup loses seal mid-cycle, how to recover a snagged tether, and how to inspect cups before each shift.
If the training is a PDF and a video link, expect calls in month three. Budget two days for the first crew and refresher time when staff turn over.
4. Can you trial on my worst bay?
Your worst bay is the deepest, steepest, dirtiest part of the roof. A supplier confident in the machine will bring it to your building. One who only offers a showroom demo is showing you a controlled environment, not your reality.
Watch the trial for the moments that matter: launch, edge return, and slope transition. Those are where problems surface. A smooth run across flat glass tells you very little.
What if the supplier answers well but the price is high?
Price a machine against the cost of failure, not against a cheaper competitor. A cheap robot with no regional spares and no training can cost more in the first year than a pricier, properly supported machine costs over three. Buy the support, not just the hardware.
What is the one question buyers forget?
How do we get it onto the roof in the first place? A machine that will not fit through the roof access hatch, or cannot be carried up the stairs, is useless on a building with no goods lift. Check the door widths, the hatch size, and the lift capacity before you fall in love with a model.
It sounds trivial until a 60 kg machine is stranded on the ground floor with no way up. Measure first, buy second.
How do you run the comparison without wasting weeks?
Give each supplier the same roof drawing and the same five questions, and ask for written answers. The supplier who replies with specifics and admits the limits of the machine is usually the one to trust. The one who answers everything with a brochure quote is not.
A short list of two or three machines is enough. More options past three just delay the decision, and the differences between the third and fourth best machines are rarely material to the outcome.
5. Can you produce the certification file?
EC Declaration of Conformity, IP test report, risk assessment, and a maintenance schedule with wear limits. Ask for all four before signing. In 2026, insurers and clients increasingly want this on file, and a missing document can stall a deployment.
A supplier who dodges this question is telling you something. Legal marking is cheap to do properly and expensive to fake when an incident happens.
Ask Skylight Cleaning Robot these five questions on the Lingkong K3, Lingyun Y3 or Lingfeng S1 and compare the answers honestly.
Key Takeaways
- Demand suction data for your actual slope and glass finish.
- Confirm regional spare-parts stock and real lead times.
- Training should be onsite and cover failure recovery, not just the basics.
- Insist on a trial on your worst bay, not a showroom panel.
- Get the full certification file before you sign anything.
How should you compare two suppliers fairly?
Score the five answers on a simple sheet and weight them by what matters to you. If uptime is critical, weight spares lead time highest. If your crew is thin, weight training hardest. The price difference between two machines is rarely larger than the cost of a bad spares pipeline.
Ask each supplier for a reference site with a similar roof, and call it. Ten minutes with a real owner tells you more than an hour of sales slides. Ask what has broken and how fast it got fixed.