There are plenty of robo-vacuum & robo-mop review on the web, but none of them tests their WiFi implementation and especially, how much RF they emit, i-e how much they unnecessarily irradiate you while they do their job, and worse: while they are sleeping / charging in the dock.

This is an UNSPONSORED test, so I only bought two models. They are from the two biggest contenders in the high-end vacuum+mop robot segment. The two tested models are:

Dreame L10s (generation 1)
RoboRock Qrevo Master

Most IoT companies will use the same WiFi hardware and software for their whole line of products. Therefore it may very well be the case that the following applies to all devices of Dreame and RoboRock. That is, until, as we’d hope they would do, they make some firmware updates to lower their unnecessary emissions.

Why do these devices emit WiFi?

The robots can be configured in great detail through an app on a smartphone. It is arguable necessary to have an advanced UI in order to edit/configure the map of your home and the desired cleaning settings for each room.

Because the robot is moving through the house, it is understandable that it uses a wireless network connection and not a wired one ;)

In an ideal world, we could dream of a docking station connected through ethernet that would upload whatever programming we want into to the robot (though physical contacts). The robot would then not be connected while it wanders through the house.

This would only fall short in a few areas:

when connected, robots can update the map of what is already cleaned or not in real time
when connected, robots can alert you if they get stuck somewhere and can’t finish cleaning nor return to the docking station
when connected, robots can be used a realtime mobile surveillance cameras

Moreover, the robots are actually built to work without WiFi! Because indeed, they foresee that when wandering around the home, they may end up in an area where there is no WiFi reception. They will carry on with the cleaning task and simply reconnect to WiFi when they find the network again. So the only place where they actually need a WiFi signal is when they are at the docking station and awaiting a command to start a new cleaning task.

The way it is implemented: all the computing logic is in the robot and it’s probably even the robots that commands functions of the docking station (empty bin, rinse mops, etc.). Therefore all WiFi is received and emitted by the robot and nothing by the docking station.

The problem is, like many other IoT devices, the robots emit way too much and way too strong a WiFi signal than necessary

Needless to say: they only work on 802.11b / 2.4 GHz.

Test 0: RF baseline levels in my basement

RF levels with WiFi OFF and Roborock OFF

As you can see, levels are < 1 µW/m², which is extremely low (I can only get this little RF noise in the basement).

Note: all following measurements are made without moving the meter (it’s on a tripod).

Test 1: Turn on the robot, WiFi still off

This simulates the case where you turn off your WiFi at night because you think it’s good but you actually leave your devices on, and they will be desperately searching all night for a WiFi network to connect to…

Roborock searching for a WiFi network...

The signal is not constant. It’s more like bursts of RF every couple seconds. You need to look at the “MAX” readings on the screens. After a few minutes I get:

The RoboRock emits bursts up to 39 200 µW/m²
The Dreame emits bursts up to 6 100 µW/m²
That is 6.4 X more RF power emitted by RoboRock.
Reminder: Lower is better!

Test 2: Turn the WiFi on

This simulates normal operation of the robot, when connected to WiFi and being able to communicate with its cloud based “home” server (the server from the manufacturer).

Note: in reality we’d need to follow the robot around while cleaning, but it’s very difficult to do that and guarantee that measurements are made at the same distance for both robots. It would also matter to make measurements on the exact same cleaning path (because of signal reflections & absorptions). I deemed this too complicated for now (but let me know if you would like to see such a test.)

After a few minutes, I observe this:

The RoboRock emits bursts up to 50 000 µW/m²
The Dreame emits bursts up to 5 000 µW/m²
That is 10 X more RF power emitted by RoboRock.
Reminder: Lower is better!

Note: it is a pretty rare coincidence that the numbers are so perfectly round.

Test 3: Turn robots OFF, keep WiFi on

This is to verify that we are measuring RF from the robot and not other WiFi activity.

I forgot to change the Post-It note for that photo ;)

A few things that guarantee there is no interference from my WiFi:

The base station (the “WiFi router”) is behind 2 walls
The base station power is at 6% of a standard base station (FritzBox setting)
I have no other device turned on that can communicate with that WiFi network

As you can see, even when my 6% WiFi is on, in this room, the RF levels are roughly equivalent to when my WiFi is off.

Test 4: WiFi ON bur robots “sleeping”

After a few minutes of doing nothing (and not being on the charging station), the robots go to “sleep” or “standby mode”, as shown in their app:

Note: I use the app over Ethernet, not WiFi, of course…

RoboRock still talking really loud while it's supposed to sleep

After a few minutes:

The RoboRock emits bursts up to 43 800 µW/m² while “sleeping”
The Dreame emits bursts up to 5 990 µW/m² while on “standby”
That is 7.3 X more RF power emitted by RoboRock.
Reminder: Lower is better!

IMPORTANT Notes:

The results of this test are globally consistent with the previous one (Test 2), however:
Dreame emits more in standby than in normale mode. This may be real but more likely, I should have waiting a bit longer in Test 2.
I am not 100% sure the robots don’t have a “deeper sleep” mode after a longer wait time, but for now, I think this is sufficient to form an opinion on the difference between both robots

Conclusion

While the Roborock has some advantages on the Dreame (especially a better phone app UI), when it comes to the WiFi implementation and RF pollution, the conclusion is very clear:

The Roborock vacuum emits 6 to 10 times more WiFi RF power than the Dreame
Both operate at the same range from the base station so there is absolutely no need for the Roborock to emit this “loud”. Up to 90% of the emission power of the Roborock is unnecessary waste and RF pollution.
Dreame is the clear winner of the #LowFi #LowRF award!

Bonus

The Dreame docking station has a physical power switch on the back of the station. When you turn on the docking station and the robot is docked, it turns on the robot automatically.

However, to turn off the robot, you need to long press the button on the robot. The robot can be turned off when the station is off (but needs to be removed from the station if the station is on).

The Roborock station has no switch. If you want to turn it off, you need to unplug it. If you want to turn off the Roborock robot, you need to remove it from the station or unplug the station.

Note: none of the docking stations emit RF on their own, only the robots do. I think they do emit (harmless) infrared signals for helping the robot align when trying to dock.

Here again, the Dreame dock is more convenient (even if I’d prefer the power button to be more accessible).

Why do these devices emit WiFi?#

Test 0: RF baseline levels in my basement#

Test 1: Turn on the robot, WiFi still off#

Test 2: Turn the WiFi on#

Test 3: Turn robots OFF, keep WiFi on#

Test 4: WiFi ON bur robots “sleeping”#

Conclusion#

Bonus#