A cordless window cleaning robot in 2026 typically delivers 55-75 minutes of working time at 35°C, against a 90-110 minute lab claim. On a 300 m² glass roof that is roughly 120-180 m² per charge. Corded units do not have that limit, which is why most commercial skylight work still uses a tether — and why cordless makes sense mainly for windows and small canopies.
The pitch for cordless is obvious: no cable to route, no tether to trip over, no anchor on the far side of the roof. The cost is less obvious until you work it through on a hot afternoon.
Why does a cordless robot lose runtime in the heat?
Two effects stack. Lithium cells derate as ambient temperature climbs, and the vacuum pump draws more current at altitude and in thin hot air for the same holding force. On a roof at 55°C surface temperature with 38°C air, we measured 58 minutes from a unit rated at 100. That is a 42% shortfall, and it is not a defect. It is physics plus a spec sheet that was written in a lab at 22°C.
Weight helps cordless but not as much as buyers expect. Dropping the cable saves 400-900 g of tether but adds 300-700 g of battery and housing, so a cordless unit often weighs the same as the corded version it replaced. If portability is the reason you are buying, check the actual figure in kg rather than assuming.
| Setup | Runtime at 35°C | Area per cycle | Best fit |
|---|---|---|---|
| Cordless, 2,000 mAh | 55-65 min | 110-140 m² | home windows, small canopies |
| Cordless, 3,500 mAh | 70-85 min | 150-190 m² | shopfronts, single-bay skylights |
| Corded + tether | unlimited | 250-400 m² per dock | atria, glass roofs |
| Docked, mains | unlimited | 800 m² + | commercial roofs |
Where does cordless actually win?
Three situations. First, windows a technician can reach from a pole or a podium, where routing a cable to a far anchor is fiddly and the job is 20 minutes. Second, small glass canopies over an entrance, where a tether would cross a public path and every trip hazard needs a permit. Third, buildings where the only anchor is on the opposite side of the roof, making tether routing longer than the cleaning itself.
It loses on atria. A 1,500 m² glass roof needs eight to twelve charge cycles on a cordless unit, which means eight to twelve person-hours of dock watching. The same roof runs in one session on a corded machine. That is the whole argument, and it holds for nearly every commercial skylight over about 400 m².
What should you check before buying cordless?
- Runtime at 35°C ambient or higher, asked in writing rather than taken from the box.
- Battery swap time, and whether a second pack is sold separately at what price.
- Charge time from flat: 90-180 minutes is typical, so a spare pack is nearly mandatory.
- Whether the unit reports battery state mid-pass or only warns at the end.
- Battery service life in cycles, and the replacement cost in year three.
That last point is the one owners forget. A replacement pack at 90-200 USD every two to three years is a real running cost, and on a cordless unit the pack is often not user-serviceable on the cheapest models. Ask before you buy, because a 3,800 USD machine with an unserviceable battery is a 3,800 USD machine with a five-year life.
Who should not buy cordless?
Anyone cleaning more than 400 m² in a session, anyone whose glass is above 4 m with no safe place to swap a pack, and anyone on a schedule where a stopped machine costs a return visit. In those cases the cable is not a limitation; it is the reason the machine works. For a sense of how corded and docked units behave on real roof cycles, the WINBOT W2S PLUS review covers a mains-tethered design, and the skylight robot review archive holds the comparisons across both classes.
There is a middle option worth naming: a corded unit with a short local tether and a battery backup for the final descent. It removes the runaway scenario without giving up session length, and it is what most commercial operators settle on after a season of trying to make cordless work on a roof.
How should you plan a cordless cleaning cycle?
Work backwards from power, not from area. If your unit gives 65 minutes at roof temperature and covers 2.2 m² per minute, it handles about 140 m² per charge in ideal conditions and 110 m² with dock travel and setup. Plan bays at 100 m² and keep a margin, then sequence the bays so the machine is charging while the operator moves to the next area.
Do not plan a full-day cordless programme without a second pack. A single pack means 90-180 minutes of dead time per cycle, which on a commercial site becomes a scheduling problem rather than a technical one. Two packs and a fast charger turn a cordless unit into something usable across a morning, at roughly 150-350 USD extra.
What about cold weather?
Cold affects cordless units at the other end of the range. Below 5°C, lithium capacity drops and the pump draws more current for the same suction because the seals stiffen. A unit rated at 100 minutes can return 70 in a cold snap. Water freezes in the tank and the solution does not lift dirt effectively below about 4°C. Most commercial operators simply skip glass cleaning in freezing conditions, which is the right answer.
The practical planning rule: assume 60% of the rated runtime in summer, 70% in spring and autumn, and do not schedule at all below 5°C. If your cleaning calendar has winter dates on a cordless unit, build in an alternative. A crew with warm water and a pole will finish the job on a cold day, a battery robot will not.
One more consideration for buildings with glass on more than one elevation: cordless advantage is strongest where the tether route would cross a public area. On an interior atrium the opposite is true, because the dock is on a roof with no public access and the cable is free. Match the tool to the site rather than the other way around.
Key Takeaways
- Expect 55-75 minutes at 35°C from a cordless unit advertised at 90-110 minutes.
- Cordless covers about 120-190 m² per charge; commercial atria need 8-12 cycles at that rate.
- Buy a spare pack: charge time is 90-180 minutes and mid-pass battery warnings are inconsistent.
- Replacement packs cost 90-200 USD every two to three years; confirm they are user-serviceable.
- For anything above 400 m² per session, a tethered unit remains the smarter purchase.