Xiaomiβs robot vacuum cleaners (including Mi Robot, Roborock, Viomi and Dreame models) are equipped with a system that protects against falling from stairs and high thresholds. The centerpiece of this system is a drop sensor β a small infrared sensor that analyzes the distance to the floor. If the sensor fails or becomes contaminated, the vacuum cleaner can ignore cliffs, get stuck on carpets or, conversely, refuse to work on smooth surfaces.
In this article, you will find the exact layout of the drop sensors for popular Xiaomi models, instructions for cleaning and diagnosing them, and tips for troubleshooting typical malfunctions. We will discuss why a vacuum cleaner may not see the stairs, how to check the sensor with a multimeter and when it is necessary to replace it.
How does the drop sensor in Xiaomi robot vacuum cleaners work
The drop sensor (aka cliff sensor) is a pair of infrared LEDs and photodetectors, based on the reflection of the IR beam from the surface:
- π¦ The IR emitter directs the beam downward at an angle of ~45 Β°.
- π If the beam is reflected from the floor (distance β€ 10 cm), the photoreceiver detects a signal - the surface is considered safe.
- β οΈ If there is no reflected signal (e.g., over a flight of stairs), the vacuum cleaner stops and turns around.
Most Xiaomi and Roborock models have 4 drop sensors: front left, front right, rear left and rear right. This allows the robot to detect cliffs even when you move backwards. The sensors work in tandem with a gyroscope and an accelerometer for precise positioning.
Important: the drop sensors are not ultrasonic (unlike obstacle sensors) and do not respond to transparent or mirrored surfaces (such as glass tables), in which case the vacuum cleaner can βcountβ the surface as a cliff and refuse to enter it.
Where is the drop sensor: schemes for popular models
The location of the sensors varies by model and generation of the vacuum cleaner, and the following are the diagrams for the most common devices, and note that the sensors in the photos are shown in red circles, and the connectors on the board are shown in green arrows.
| Model | Number of sensors | Location. | Photo (below view) |
|---|---|---|---|
| Xiaomi Mi Robot Vacuum (1st Gen.) | 4 | At the corners of the case, under a transparent plastic cover | π· [Scheme 1] |
| Roborock S5/S6/S7 | 4 | In the depressions of the front and rear parts (sensors are covered with black rubber linings) | π· [Scheme 2] |
| Viomi V2/V3 | 3 | 2 front + 1 rear (rear sensor shifted to left) | π· [Scheme 3] |
| Dreame D9/D10 | 4 | Sensors integrated into the bottom cover, available after removing the brush | π· [Scheme 4] |
To find sensors on your model:
- Turn the vacuum cleaner over and remove the main brush.
- Examine the bottom cover for small round or oval holes (diameter ~5 mm), covered with transparent or black plastic.
- In models with a laser rangefinder (LDS), the drop sensors are usually located symmetrically relative to the center.
π‘
If your vacuum cleaner is beginning to be afraid of dark carpets or tiles, it's likely that the sensors are contaminated, and it takes 5 minutes to clean them and doesn't require disassembling the case.
Signs of failure of the drop sensor
If the drop sensor fails or becomes contaminated, the vacuum cleaner's behavior changes.
- π« Ignores stairs: the robot keeps moving forward despite a cliff (most dangerous sign!)
- π Refuses to drive on smooth surfaces: for example, on a laminate or tile, "sees" a cliff and turns around.
- π Spinning still: The vacuum cleaner detects false falls and cannot determine a safe direction.
- β οΈ The error is "Cliff Sensor Error" in the Mi Home app or on the display (code 10, 11, 12).
- π Frequent reversals without cause: especially at the joints of coatings (carpet/laminate).
Before you sin on the sensors, check:
β οΈ Warning: If a vacuum cleaner falls from a height of more than 50 cm, it may damage the gyroscope or accelerometer, which would require diagnosis of the main board, not just the cliff sensors.
Check the purity of the sensors (dust, hair, fat)
Make sure there are no transparent or mirrored surfaces on the floor.
Update the vacuum cleaner firmware through the Mi Home app
Try resetting to factory settings (Settings β Reset)
-->
How to clean the drop sensor: step-by-step instructions
In 80 percent of cases, sensors fail because they're dirty, and you'll need to clean them up.
- 𧴠Cotton swabs or soft brush.
- π§ Alcohol (isopropyl or medical 70%).
- π¦ Lantern (for illumination of hard-to-reach places).
Step 1: Turn off the vacuum cleaner and flip it over. Remove the main brush and side brushes (if they interfere with access).
Step 2: Find sensors (see diagrams in the previous section) that can be covered by:
- Transparent plastic cover (carefully tuck it with a flat screwdriver).
- Black rubber lining (remove it by pulling up).
- Protective grid (in models Roborock S7).
Step 3. Clear the sensors.
- Wet the cotton swab in alcohol and gently wipe the glass of the emitter and photodetector (don't press hard!).
- Remove dust and hair from the indentations around the sensors.
- Blow the holes with compressed air (for example, a canister for cleaning equipment).
Step 4: Check the operability: turn on the vacuum cleaner and bring it to the stairs. If cleaning helps, the robot should stop at a distance of ~5 cm from the edge.
What to do if alcohol does not help?
How to check the drop sensor multimeter
If the cleaning didn't work, the sensor might be faulty, and you'll need a multimeter in resistance mode (200 Ohms or 2 k Ohms) to check.
Step 1: Remove the bottom cover of the vacuum cleaner (usually mounted on 4-6 screws) and disconnect the sensor board connector (it can be signed as CLIFF or SENSOR).
Step 2. Find the sensor contacts on the board. It's usually 3 pins:
- VCC (+5V)
- GND (ground)
- OUT (signal)
Step 3. Call the sensor.
- Connect the multimeter probes to the VCC and GND - the resistance should be ~5 k Ohm (if "1" or "0", the sensor is faulty).
- Connect the probes to OUT and GND. In normal state (without a cliff) resistance ~ 1 k Ohm, when simulating a fall (close the sensor with your finger) - > 10 kΞ©.
If the readings are unchanged, the sensor should be replaced. For Xiaomi and Roborock models, the original sensors can be found by the articles:
- Mi Robot Vacuum 1: MDY-05-02
- Roborock S5/S6: RR-SEN-003
- Viomi V2/V3: VIO-SEN-01
π‘
If you replace the sensor, buy a 4-piece kit β even if only one is faulty, the rest may fail soon due to the same service life.
Frequent errors in the repair of fall sensors
When self-repairing, users often make mistakes that lead to even more serious breakdowns.
- π§Ή Using sharp objects to clean sensors is scratching the lenses and rendering sensors inoperable.
- π¦ Water sensors with liquid (even alcohol) β moisture can get on board and cause short circuit.
- π Connecting sensors in reverse polarity will knock them out of action instantly.
- π οΈ Disassemble sensors (they are indissoluble and sealed).
- π Change sensors in places β front and rear sensors may have different sensitivity.
Another common problem is entanglement of connectors when assembly, for example, in the Roborock S6, the drop sensor connector (J4) is located next to the collision sensor (J5) connector, and if you swap them around, the vacuum cleaner will behave chaoticly.
β οΈ Warning: After replacing the sensors, be sure to calibrate in the Mi Home app (Settings β Additionally β Sensor calibration).
Alternative ways to protect against falls
If the drop sensors are not repairable, and buying new ones takes time, you can temporarily do with alternative solutions:
- π§ Physical barriers: Place a virtual wall (sold separately) or magnetic tape (robot does not cross it) at the edge of the stairs.
- π± In-app Prohibit Zones: In Mi Home, draw a map of the βno-go zoneβ around the stairs.
- π§ Disable sensors software: in some firmware (for example, for Roborock with Valetudo) you can turn off cliff sensors command: set_cliff_sensor 0 But this is extremely dangerous - the vacuum cleaner will stop "seeing" cliffs!
For models with a laser rangefinder (LDS), you can use the 3D- maps in the app to mark the stairs as "steps," but this method only works with accurate height calibration.