A cordless window cleaning robot typically runs 60 to 90 minutes per charge, enough for 25 to 40 square metres of glass. That covers a residential skylight or a small atrium. It does not cover a 200 square metre commercial glass roof without a spare battery and a plan.
The cordless pitch is simple and it is genuinely good: no cord dragging across a wet roof, no tripping hazard on a ladder, no extension reel to haul up six flights. The catch is that the same battery powers the vacuum pump, which is the thirstiest part of the machine. Suction is not a nice-to-have. Lose it and the robot leaves the glass.
How long does a cordless window cleaning robot really run?
Manufacturer claims cluster around 90 minutes on eco mode. Independent tests land closer to 55 to 70 minutes on a dusty exterior panel, because the robot uses higher suction and more scrubbing passes. Cold weather knocks another 15 percent off lithium cells below 5 degrees Celsius.
If your skylight is 6 square metres and the robot covers roughly 0.4 square metres per minute including turns, you have about 95 percent of a cycle inside one charge. Comfortable. If you manage a 60 square metre atrium, you need two batteries minimum and a charging break in the middle of the job.
Is a tether still better than a battery?
For anything above four storeys or larger than 50 square metres, often yes. A tether gives constant power and a physical retention line in one. The trade is setup time and a wire on the glass. Cordless wins on speed-to-start and anywhere you cannot safely route a cable.
My honest position: cordless for roof hatches and residential skylights, tether for curtain wall and atrium. The hybrid approach, where a cordless unit accepts a tether as backup, is the design that makes sense commercially and it is rarer than it should be.
If you want to see how retention is handled on a tether-first machine, our Lingkong K3 review walks through the anchor and cable management in detail.
What battery technology do these robots use?
Almost all current units use lithium-ion packs between 5,000 and 6,500 mAh at 14.4 to 24 volts. A few use lithium-polymer for a flatter shape in thin bodies. Expect 300 to 500 full cycles before capacity drops to 80 percent, which at twice-weekly use means two to three years.
Care matters more than chemistry. Heat is the killer. A robot that sits in a rooftop equipment room at 45 degrees in summer will lose capacity in a year. Store it indoors, charge at room temperature, and do not leave it on the charger for weeks at 100 percent.
How do you plan a job around battery limits?
Split by zone, not by time. Clean one pane group, return to the charger, swap. What you must not do is stop mid-pane, because the robot’s path memory on most units resets and you get an overlap band that collects residue.
Build in a swap buffer. If a full charge gives 65 real minutes and a zone takes 50, you have 15 minutes of slack. That is thin in wind. On a skylight with a 20 degree pitch and gusts of 8 metres per second, the robot works harder and drains faster, so treat anything above 5 m/s as a stop-work condition regardless of battery.
Who should not buy cordless?
Facility managers with glass roofs over public circulation areas. The failure mode is not the battery, it is the downtime between swaps while the unit sits on a roof edge charging. Corded or tether machines keep running. Also skip cordless if your only access is a roof hatch with no nearby power, because charging becomes a logistics problem you did not budget for.
Homeowners with one skylight over a stairwell? Cordless is excellent. No cable over the stairs, no reel, and 60 minutes is plenty for one panel every couple of weeks.
What you should check before paying for cordless
How do you test run time before buying?
Run it yourself if you can. Ask a dealer for a demo on a dirty panel, not a clean one, and time it. Failing that, find a reviewer who measured a stated panel size and worked out square metres per minute from that. Two numbers, panel size and minutes, are enough to check any claim. If a review gives neither, it has not been tested.
What should you not assume?
Who should not trust the run time number
- Real run time on high suction, not eco mode, from a review that measured it.
- Battery swap time under 20 seconds, ideally hot-swappable without powering down the robot.
- Whether the unit accepts a tether as fallback for larger jobs.
- IP rating on the battery compartment. IPX4 minimum for roof work, IPX5 better.
- Spare battery price. Some are 40 percent of the robot cost.
Read the specification sheet, then read a real review. The gap between claimed and measured run time on these units is one of the widest in small appliances. Our ECOVACS WINBOT W3 review is a useful comparison point if you are benchmarking against a mainstream name, and the full review index covers the rest.