Diagnosis
The Roborock S7 relies on a combination of LiDAR (the spinning turret) and infrared cliff sensors to navigate. When the robot avoids large, open sections of a room despite no physical obstacles, it is almost always experiencing "phantom cliff detection." This occurs when the infrared sensors on the bottom of the chassis receive a signal that mimics a drop-off, triggering a safety override that prevents the robot from entering that specific zone to avoid falling down stairs. This phenomenon is usually caused by one of two things: physical debris buildup on the sensor windows or "light absorption" from the flooring. Dark-colored rugs, black tiles, or deep navy carpets absorb the infrared light emitted by the sensors rather than reflecting it back. Because the sensor receives no return signal, the onboard logic assumes there is a void (a cliff) and marks that entire section of the map as a "no-go" zone in real-time, causing the robot to pivot and avoid the area entirely. You can confirm this by placing a white piece of cardboard over the dark area; if the robot suddenly cleans over it, the issue is the floor color.
How to Fix It
To resolve sensor interference and restore full room coverage, follow these precise steps:
- Power down the Roborock S7 completely using the physical power button on the side of the chassis to prevent accidental activation during cleaning.
- Flip the robot upside down on a soft surface (like a towel) to avoid scratching the top LiDAR turret.
- Locate the six infrared cliff sensors. These are the small, transparent plastic windows arranged around the perimeter of the bottom plate.
- Use a dry, lint-free microfiber cloth to wipe each sensor window. If there is stubborn grime or a "film" of dust, dampen a Q-tip with 70% isopropyl alcohol and gently scrub the interior of each sensor window. Ensure no streaks are left behind, as streaks can refract the IR beam.
- Inspect the side-brush and main brush rollers for any tangled hair or threads that might be hanging down and obstructing the line of sight of the sensors.
- Flip the robot back over and perform a "Map Reset." Open the Roborock app, go to Settings > Map Management, and delete the current map. This forces the S7 to re-scan the room without relying on previous "phantom" obstacle data.
If That Didn't Work
- Dark Floor Workaround: If the robot is avoiding a black rug or dark hardwood, the sensors are physically unable to "see" the floor. The only permanent fix is to create a "No-Mop Zone" or "No-Go Zone" in the app for that specific area and manually clean it, or use a lighter-colored rug runner to bridge the gap.
- LiDAR Obstruction: Check the LiDAR turret on top. If there is a piece of hair or a small insect trapped inside the rotating laser housing, the robot may perceive a "ghost" wall. Use a can of compressed air to blow out the turret from the top and sides to ensure the laser is spinning freely without interference.
- Firmware Conflict: Ensure your S7 is updated to the latest firmware version. Roborock frequently releases "navigation optimization" patches that adjust the sensitivity of the cliff sensors to reduce false positives on dark surfaces.
- Hardware Failure: If the robot consistently avoids the same spot regardless of floor color or sensor cleanliness, a specific cliff sensor may have failed. Check the app for a "Cliff Sensor Error" notification. If a sensor is dead, the entire bottom sensor array module typically needs replacement.
Heads Up
- Avoid using glass cleaners or ammonia-based products on the sensor windows, as these can leave a chemical residue that interferes with infrared light transmission.
- If you have a multi-story home with high-contrast flooring (e.g., white tile transitioning to black slate), be prepared for the robot to occasionally struggle at the transition line.
- Clean the sensors every 2-4 weeks as part of your regular maintenance routine to prevent dust buildup from triggering false cliff alerts.
- Never paint or put stickers over the cliff sensors to "trick" the robot into cleaning dark rugs; doing so removes the safety mechanism and will cause the robot to fall down actual stairs.