Things Worth Wanting / iRobot / Roomba Max 875 Combo
Roomba Max 875 Combo: A House Is a Moving Target
A chair moves. A crumb falls. A rug interrupts the tile. For a household robot, an ordinary afternoon is a demanding engineering assignment.
A very small piece of evidence
The cereal had something to teach the engineers.
One of my favorite details in the Roomba story involves a glass tabletop and crushed cereal.
In his firsthand account of developing Roomba, inventor Joe Jones describes watching debris from beneath a transparent surface. The brush was throwing particles out of the cleaning path. Seeing the motion helped the team change the surrounding shroud so escaping material could be recaptured.
That is a wonderful engineering scene. The problem is small enough to fit beneath a kitchen chair, but understanding it requires a different point of view. Get underneath it. Watch what actually happens. Let the cereal explain the flaw in the theory.
I can appreciate a sophisticated algorithm. I can also appreciate somebody building a way to look at the underside of a mess.
Both belong in the history of useful technology. A household robot has to turn a plan into contact with real material. Dust does not move because the product presentation was persuasive.
The Roomba Max 875 Combo with AutoWash Dock is the current subject of this entry. The original assignment was the Combo j9+; I have followed the same brand forward to the flagship iRobot introduced on September 2, 2026.
I have not owned or personally tested it. My interest comes from its documented engineering and the unusually rich story behind the task it performs.
A cleaning robot brings together perception, movement, airflow, water and the demands of a place where people are busy living. That makes it a compelling addition to Things Worth Wanting. The story begins with a chore and expands into a question that has occupied roboticists for decades: how do you make a machine useful in an environment you do not completely control?
The room keeps changing
A floor plan leaves out most of the interesting parts.
The walls may stay put. Daily life rarely does.
Imagine giving a robot an architect’s drawing of a dining room. It would show the walls and doorways beautifully. It would say much less about the chair somebody pulled out after breakfast.
It would also omit the bag beside the table, the slippers near the sofa and the little collection of crumbs that appeared while no one was officially making a mess.
Those details make a house an interesting place for robotics. The environment has enough regularity to learn, along with enough variation to require fresh information.
A person entering the room brings years of experience to the scene. We recognize objects, infer what can move and step around things almost without discussing the process with ourselves. We can also pick up the chair.
A floor-cleaning robot has a narrower set of available actions. It must sense, choose a route and move its body through the space while keeping its cleaning equipment usefully engaged.
That is why the familiar phrase “robot vacuum” understates the assignment. Vacuuming is one operation inside a larger sequence. Reaching the debris and returning home afterward are also part of the job.
This is the side of advanced robotics that I find especially interesting: translating a human request into a chain of physical actions. “Clean the room” sounds simple because we already know so much about rooms.
Start specifying every necessary decision and the request becomes a remarkably substantial piece of work.
From a proposal to the living room
Before Roomba, there was DustPuppy.
Jones’s account reaches back to MIT’s Artificial Intelligence Laboratory and his 1989 RugWarrior project. In 1999, he and Paul Sandin proposed a cleaning robot at iRobot under the name DustPuppy. Management gave them $10,000 and two weeks to produce a prototype. Roomba ultimately launched in September 2002.
The early design used a round body with wheels positioned so it could turn in place. Its behavior-based control combined simpler responses to circumstances rather than requiring an elaborate understanding of everything in the room.
Those historical choices are worth remembering without assigning the original control system to today’s model. The current robot belongs to a later stage of the story.
What stays interesting is the discipline behind the early project. A prototype has to make an idea concrete enough to evaluate. A commercial product has to repeat its useful behavior in somebody else’s house.
Between those two stages live the details that rarely become household vocabulary: the clearance beneath furniture, the way debris enters a brush enclosure, the space available for a battery, the ability to recover from an awkward position.
Each decision touches the others. Give one component more room and another component has less. Increase the demands on the battery and the rest of the design has to account for that energy. Change the body’s geometry and you change where the machine can go.
I like this history because it gives the company a practical identity. The Roomba story involves learning how to package robotic behavior into an object that can perform an ordinary domestic task.
There is something unusually ambitious about choosing ordinary life as your operating environment. People will not arrange every chair to flatter the machine. Breakfast will proceed on its own schedule.
The useful robot has to meet that world at floor level.
Light, distance and a decision
Measuring a room is the beginning of navigating it.
The Max 875 combines ClearView Pro Dual-Line LiDAR mapping with PrecisionVision AI obstacle recognition. The product names describe related capabilities, but measuring geometry and recognizing an object answer different questions.
LiDAR uses light to obtain distance information. NOAA’s explanation of the technology describes how light-based measurements can support detailed representations of physical surroundings. The same broad idea of measuring with light has applications at very different scales.
For a household robot, geometry helps establish where surfaces and boundaries are. Recognition adds another kind of interpretation: what might this thing in the path be?
Then comes action. A measurement does not turn a wheel by itself. The system has to use its information to select movement.
Three questions inside one trip across the room
Where is the available space? What is occupying part of it? What movement should happen next?
Those questions connect sensing, interpretation and control. The robot’s visible motion is the result we get to watch.
A useful way to picture the distinction is to imagine a chair moving between cleaning sessions. The wall behind it has not changed. The route through the room has.
A map provides context, while fresh observations help the machine work with the current scene. That is a general robotics principle, not a claim that every object will always be recognized correctly.
The appeal for me is the cooperation among disciplines. Optics produces information. Software organizes and interprets it. Motors carry out a chosen motion. The cleaning assembly needs that motion to put it in the right place.
Watch the machine from across the room and you see a compact object making a turn. Inside that apparently simple event is a chain of decisions about the physical world.
It gives an everyday floor a surprising amount of intellectual depth.
A mechanical idea with consequences
Give the air a useful path.
One of the Max 875’s most interesting features is SealForce. iRobot describes a skirt that lowers from the chassis against carpet fibers to concentrate the vacuum’s airflow.
I like this feature because it makes the relationship between machine and surface part of the mechanism. The carpet is where the cleaning work has to happen. The way the vacuum meets it matters.
Think about air moving through a passage. Pressure differences drive the movement, while the shape and resistance of the available path affect how the air flows. Openings around a cleaning head become part of that system.
Pressure and airflow are related, but they are different quantities. Pressure is force per unit area. Volume flow rate describes how much volume passes a location per unit time.
OpenStax’s treatment of flow rate makes another useful distinction: flow rate and flow speed are not interchangeable either. The area of the passage is part of the relationship.
That gives us a better way to appreciate the design than treating one large specification as the whole story. A cleaning system has to move debris, and the route available to the moving air is one of its important physical conditions.
SealForce is an attempt to shape that condition at the carpet itself. Its appeal is the directness of the idea: change the interface where the work is being done.
That returns me to the cereal experiment. The generations differ, but the engineering question remains wonderfully concrete. What is the material actually doing when it reaches the machine?
A brush, a shroud and a seal are physical answers. I find that every bit as interesting as the software that brought the robot to the spot.
A different kind of mess
A dry crumb and a dried spill ask different questions.
A loose particle can be swept or carried away. A dried residue is attached to the surface. The second situation asks the cleaning system to do more than collect something already free to move.
The Max 875’s ThermaMist applies heated ultrasonic mist ahead of the PowerSpin Max roller’s scrubbing action. iRobot specifies 140°F water and a roller speed of 300 revolutions per minute.
There are several separate ideas inside that short description: wetting the material, bringing thermal energy to the task and moving a cleaning surface against the floor.
The mist is heated liquid, not a claim of steam cleaning. At ordinary atmospheric pressure, 140°F is well below water’s boiling point. Fine droplets and water vapor are different physical things, even when everyday language makes a visible cloud sound like steam.
That distinction helps keep the engineering story accurate. The feature is interesting on its own terms.
Consider the familiar act of letting water sit on a dried food spill before wiping it. You are giving the liquid time to interact with the residue. Scrubbing then adds mechanical action. Different materials respond differently, which is why the process and the surface both matter.
The robot turns a version of that sequence into coordinated hardware. It has to deliver liquid, bring the roller into contact and move through the area while managing the material it removes.
A roller also has a geometry worth considering. As it rotates, different parts of its circumference reach the contact region. Rotation and forward travel together determine how the cleaning surface passes over the floor.
Three hundred revolutions per minute is five turns per second. That arithmetic does not establish a cleaning score; it simply makes the motion easier to imagine.
There is considerable activity underneath a machine that, from a standing adult’s perspective, appears to be making a calm trip across the kitchen.
At the boundary
The rug changes the assignment.
A useful combination tool needs a way to change how it meets the world.
A mixed-floor room introduces a transition. A surface suitable for the current mopping action can end at the edge of a rug.
DriLift responds to detected carpet by raising the mop and lowering a protective roller shield. PerfectEdge addresses another boundary by extending the roller toward walls and baseboards.
These are two different examples of the robot changing its physical relationship with the surroundings. At one boundary, the cleaning element needs separation. At another, it needs additional reach.
I find the distinction satisfying. A room has open areas, edges and transitions. Each presents a different geometric problem.
Take a round body beside a straight wall. The center of the machine cannot occupy the wall, and its main cleaning path has a defined width. Reaching toward the boundary lets an attachment work beyond the route of the body itself.
The same general design idea appears in many useful tools: keep the main structure where it can travel, and give the working part a controlled way to reach its task.
Over carpet, the question reverses. The wet component needs to be managed as the robot moves into a different material environment.
These details make a combination vacuum and mop interesting as a coordinated machine. The value comes from arranging distinct operations around the surface underneath it.
The robot is doing more than carrying two tools. It is being asked to make their use fit the room.
The job continues at home base
The dock is part of the robot’s working life.
The AutoWash Dock empties debris and dirty mop water, washes the roller at 176°F, dries it with 131°F air and mixes cleaning concentrate with water for refilling.
Those are distinct stages. Washing addresses material on the roller. Drying addresses the water left afterward. Emptying and refilling prepare the system’s storage for further work.
I find the dock compelling because it expands the design question beyond the moment the machine is moving across the floor. A reusable cleaning tool has a before and an after.
- Prepare
- A cleaning system needs usable supplies, available capacity and a working surface ready for the task.
- Clean
- Movement and contact bring the tool to the floor while collected material goes somewhere inside the system.
- Reset
- Debris and used water leave the robot, and the cleaning components are prepared for another session.
That reset stage matters to the larger idea of automation. If a machine performs the central action but leaves every supporting action to a person, the person remains closely involved in each cycle.
Adding support functions changes the rhythm of that involvement. It moves more of the repeatable sequence into the product’s design.
Ownership still includes care for tanks, brushes, filters and contacts according to the owner’s instructions. Floors and cleaning solutions also need to be compatible with the intended use. An automated routine remains a physical system with supplies and components.
That does not diminish the interesting part. It makes the assignment visible. Useful automation often consists of identifying the repeated steps around a task and deciding which of them a machine can handle.
The dock gives the robot a place in that cycle. Returning home becomes another working stage, rather than simply the end of a route.
Why this belongs here
The future has crumbs in it.
There is a version of technological progress that exists in spotless renderings. Nobody has dropped anything. The chairs are aligned. Every surface seems to have been selected by the same architect.
I am more interested in the version that can participate in an actual home.
That is why Roomba belongs in my digital time capsule. It makes the ambition of robotics legible through a familiar task. You do not need to understand a control system to understand why somebody would want help cleaning a floor.
Spend a little more time with the object and the deeper story emerges. A scientist’s prototype becomes a household category. A problem with escaping cereal becomes a lesson in observation. Light measures a room. A moving skirt changes the path of air. A roller and its dock divide the work of cleaning and preparation.
What appeals to me is the density of thought inside the ordinary object. There are decisions about matter, motion and human routine packed into something intended to pass beneath furniture.
I would want a future reader to understand that practical ambition. We were trying to give machines useful jobs in the places where life actually happened.
There is also an appealing humility to the task. However sophisticated the technology becomes, it must eventually deal with what landed on the floor.
Somewhere in that requirement is the enduring lesson of the glass tabletop. Look closely. Follow the crumb. Improve the part that touches the problem.
Useful details
Questions about Roomba Max 875 Combo.
Why does this entry cover Max 875 instead of Combo j9+?
The collection follows newer same-brand models. This entry follows the original j9+ assignment forward to Max 875 Combo, introduced by iRobot on September 2, 2026.
Is ThermaMist steam cleaning?
No. The specified 140°F heated mist is below water’s boiling point at ordinary atmospheric pressure. This feature does not establish a steam-cleaning or disinfection claim.
What is the purpose of SealForce?
It lowers a skirt against carpet fibers to concentrate airflow at the cleaning surface.
How does it handle carpet while mopping?
DriLift raises the mop and lowers a protective shield when carpet is detected.
What is the difference between mapping and obstacle recognition?
Mapping establishes spatial information about the surroundings. Object recognition interprets things in that space. Navigation uses available information to choose movement.
Does the dock eliminate all owner maintenance?
No. Automated washing, drying and emptying support the cleaning cycle. Owners still follow the care instructions for tanks, brushes, filters and contacts.
Has Rob personally tested this Roomba?
No. This is a researched feature about the documented design, its history and the underlying science, without claims of ownership or measured cleaning results.