Dreame L10s Ultra: test of a latest generation vacuum cleaner that does (almost) everything

 The latest generation of robot vacuum cleaners have started to arrive en masse in 2022, like the Roborock S7 MaxV Ultra and Ecovacs Deebot X1 Omni . Devices equipped with stations capable of doing what the user still had to do, that is to say empty the dust from the small internal bin or fill the bin with water and clean the mop after the robot mop.

It is this type of model that Dreame offered us to test, a brand that follows trends and which already has numerous models to its credit such as the D10 Plus or the Bot W10 . This time the robot is called Dreame or DreameBot L10s Ultra and it also promises to limit the tasks that owners must carry out between each cleaning or vacuuming session.

Like its counterparts, the size of the station is quite imposing, but Dreame has chosen to place the elements rather high up and to keep a more reasonable size in width. The robot is of classic design and several accessories come with it such as a cleaning brush for the bins and a detergent cartridge.

The tanks are placed at the very top and are accessible by simply lifting the hood. One is for clean water, the other for dirty water from cleaning the brushes. We also see three buttons which are used to trigger the cleaning or vacuuming session, the return to the station or the drying of the brushes.

It is necessary to remove the trays to access the location of the cleaner cartridge which will be diluted with water. As this model is brand new, we have not yet been able to find the price of the recharge on the web. However, this remains optional.

Below is the location of the dust collection bag, one of which is already in place in addition to the one delivered.

At the very bottom, where the robot takes place, the system used for cleaning the brushes can be removed for easy washing. This task is the responsibility of the user, but is not systematic at each pass.

The robot also has 3 buttons to be able to act without the application. Its sensors for LIDAR navigation are aided by a camera and artificial intelligence.

The rear cover lifts to provide access to the internal dust tank. Although it will be emptied at each session by the station, it will probably be necessary to clean the filter occasionally. QR Codes redirect to the DreameHome application.

Before installing it, it is necessary to put in place the side brush which will remove the dust and the two main brushes for washing. These are attached magnetically and are very easy to position or remove.

Once placed on its charging station, the robot wakes up and automatically exits to reposition itself. Installation at the application level is child's play, with the robot only needing access to the home WiFi network. Advice on preparing the house and on the functionality of the robot are also given.

In the application we discover the settings accessible from the main page which you must browse to adjust to your convenience. The most important is the cleaning method: vacuuming, cleaning or both. We can also influence the power and humidification of the brushes.

In the settings many entries are accessible with even more settings. For example, you can choose what behavior the robot will have when it recognizes a textile floor, rug or carpet.

The Dreame L10s Ultra has a quick mapping system to obtain a first map without having to start a first cleaning, and the function is rather practical for setting limits or personalizing the map first. It only took him 7 minutes to cover 45m2 of the rooms in which he was tested. On the next pass he refines his cartography and tries to divide it into different pieces as best as possible.

Of course the result does not correspond exactly to the real parts. It's up to the user to correct by merging or splitting pieces. It is advisable to attend the first pass to identify places where it can get stuck or where it takes too long to go through to optimize washing. We then place virtual barriers and areas to avoid in the application. It is also possible to designate places where he should only vacuum and not wash.

This is also one of the good surprises of the first passage: low-pile carpets such as doormats or carpets are not dampened when you pass. It detects the texture and stops the mop while stopping applying pressure. A good point, but which will have its limits on long-haired carpets which will have to be avoided by placing an exclusion zone via the app.

We also find a fault common to all robot washers: the edges are not washed as well as the rest. Even if it can get through places as tight as it does, the brushes can't reach the ends.

The default suction, the quietest, is pleasant to the ear, but a little weak and we selected the intermediate power. On a floor that was intentionally very dirty and filled with leaves and animal hair, it would even have been necessary to move up a gear. However, it manages to suck up the vast majority.

But what little remains then gets stuck in the washing brushes. For a better result, in the case of the presence of a lot of material like here, it will be necessary to make two passes or first carry out a vacuuming session without the brushes.

In terms of navigation, we don't have much to complain about. Only the most vicious trap in the house required our intervention: the edge stuck to the bottom of the stairs. But just about every model got caught in it.

The top of a staircase posed no problem for him. He stopped in time and, after some hesitation, managed to get out of the trap of the small ledge to overcome.

Same scenario regarding the imposing open bay window with its rail which could trap the mechanism. Head on or across, it never got stuck and continued on its way.

The forest of chair legs under the kitchen table didn't stop him either, although the legs were too thin for his sensors to detect. He then proceeded methodically using his physical bumpers to make his way there and clear everything that was reachable.

We also appreciated the recognition of objects, in particular that of the cables which he carefully avoided. This is the objective aided by artificial intelligence which allows it to identify possible obstacles.

If all the cables have been recognized, as well as the lighting bases, he still sometimes makes a mistake by confusing, for example, the bay window rail with a bathroom scale. That didn't stop him from crossing it.