The Machine

(Note: I started writing this about a year and ago and am only now getting around to publishing, so it starts as exploration then turns more into a review…)

Now that we’ve got a large tiled area to keep clean, and a Labrador to work against our efforts to do so, it seemed the right time to enlist the services of a robotic minion to help shore up our side of the endeavour.

Since the introduction of the Roomba in 2002 things have moved along apace and now the market’s awash with models at every price point and from a variety of manufacturers with different degrees of reputation. The shopping & selection process reminded me a lot of the cyberdeck selection process in William Gibson’s “Neuromancer”, although one hopes with less dramatic outcome. And the decision to make for your first purchase of a robot vacuum is “Do I go in and spend a big wedge on something with all the features, despite having no proof it’s going to improve my life – or do I buy a cheapy to test the process and then upgrade when we know we like it?”.

Of course, in evaluating the high end the low end suddenly becomes much more conspicuously shit. Do we spank £800 on a Dyson Heurist 360? Or £170 on a Robit V7S Pro? What about the Deenkee, or the Coredy? Or the Bagotte? Or the Kyvol, or Eufy? They all look the same. Time to dig into the reviews…

Can you imagine having this thing autonomously trawling round your house?

Incidentally, if there’s one thing I’ve learned about comparing reviews, it’s if you don’t have a strong field of options to eliminate from then it’s a GREAT way of making your shortlist longer rather than being in any way useful. We had “mopping” on the wishlist too, which doesn’t make the selection process any easier – I liked the look of the Bissell Spinwave, with its dual rotating mopping pads instead of the much more passive “dragging a damp cloth around in its wake” that lots of other models seemed to favour… however as it’s not available in the UK and the company’s UK website doesn’t even acknowledge its existence, that one seemed a non-starter.

Most of the “Best Robot Vacuums of 2021” type listicles seemed to favour the Roomba s9+ – and its self-emptying bin sounded like the business. However the pricetag comes along with the featureset, so that’s borderline useless.

Very annoyingly – according to reviews, unless you’re prepared to splash a decent wedge on one of these things there’s very little “range consistency” between different models in a line and the reviews. A “next iteration” model from a manufacturer with the same featureset and pricetag can go from a “Must Buy” to an “Absolute Dud” without warning. And do you think it’s easy to track down the specific models mentioned in reviews?! AAAARGH!

The only hard bit of data I really had to go on was “not to get the same model as I’ve got”, contributed by one Billy Abbott – who said that his robot slave Derek seemed to come out periodically and negotiate the room, but contribute very little to any sort of cleaning endeavour. Not that I guess it’d be that easy to tell. I know that Liz’s requirements are a little more demanding than those placed on Derek.

We settled on the Ecovacs Deebot Ozmo U2pro. For some reason. It purported to have good reviews, an acceptable cloth-dragging mop function, a generous dust chamber and pet-specialised brush, and was in the sort of pricerange where it made you wince a bit but was acceptable if it turned out to be utter shit.

Initial forays were less than stellar, but that was more a combination of optimistic expectations and our lack of awareness on how to use the thing than anything else. We’re now several months in to our custodianship and have learned a lot more about what to expect from one of these gadgets. Which is another way of saying “made peace with how shit they are”, although that’s not *completely* fair.

So, the FIRST thing is – these contraptions can only ever be good for “maintenance” cleaning. They’ll never replace a human cleaning properly with intent. Mopping adequacy aside, they can’t see where the dirt/dust/hair is, so they rely on a comprehensive floor coverage pattern to cover all the bases of being likely to catch all the bits. On THAT basis I think the mopping mode does a reasonable job… it’s more of a damp-cloth dusting than a mop, but it does shift a certain amount of easily-persuaded dirt off the floor.

Fundamentally, we only use it to do the tiled dining room/kitchen area – usually once or twice a week in between appearances of our magnificent and wonderful cleaners: maintainers of our domestic happiness.

The biggest hassle is preparing the room for a Robot Lap. This means moving chairs & stools out of harm’s way, as chair legs present a difficulty level between annoyance, and insurmountable for the poor thing. Ensuring thin wires are tucked away is important too, lest they get wrapped around a wheel. Closing critical doors/blocking doorways to keep it off the carpet completes the exercise.

The Deebot seems to work on a pattern basis – it’s very clear that it has no concept of building a picture of the room that it’s in, what it’s already covered, and where its base station is. This is probably why it’s in the pricerange it’s in… however should we upgrade I think that feature will be critical. Models with mapping facility can then be excluded from certain areas, which reduces the preparation hassle. I have enjoyed watching it try to negotiate its way out of tricky spots or around corners – however I can tell that it’s a fairly simple edge-detection at work, and for an irregular shaped room with a kitchen island I just end up feeling sorry for the damn thing: clearly having no idea where it is at any given moment. And then if you decide you’re tired of it bimbling back & forth, there’s the small matter of waiting for it to find its way home.

As the mopping pattern above demonstrates, it covers *most* of the ground. Small brushes probe further into the corners to brush dust in however it definitely misses stuff. I’m always amused that its front-facing sensor can seem to detect the white cabinetry kickplate and stop just in front of it to turn, but then on the more shady black island cabinetry it doesn’t see the edge and just bumps into it. So from the sofa you can hear “Whzzzzzzz… turn…turn…. whzzzzzzzzzzzzzz[bump]turn… turn…whzzzzzzz… turn…turn…. whzzzzzzzzzzzzzz[bump]”, etc. all the way along the island.

The bigger issue is how it responds to labrador hair. Of which we have PLENTY. Everywhere.

I’d say about 70% of the time the hair’s picked up sensibly and without issue. The rest of the time, however, there can be “issues”. Sometimes the dust chamber just fills up. Other times a couple of longer human hairs might catch across a gap providing enough of a barrier that labrador hair starts piling up rather than going through to the dust chamber. In the case above, some of the hair caught on the plastic flap on the dust chamber, forming a barrier to anything going in – and backed up the vacuum channel through to the brush.

The main downside to this is having to pay attention to the robot and to intervene when necessary. This sounds like a question of convenience, but primarily if I wanted to be involved I wouldn’t have bought a frigging robot to do the job for me, would I?

I’m interested that LarryBStanding doesn’t pay it any attention at all – he’ll generally stay out of its way and go lay on his bed, and it’ll approach the rug the bed is on, encounter the edge and then turn around and he’ll stay laying there and not feel threatened by it. Sometimes he’ll stand in a spot and it’ll bump into his leg or something, where it’ll politely back up turn around, and head off somewhere else – and he barely pays it any mind.

Overall I think the concept has proven itself in value, and demonstrated to me that if we were prepared to spend a bit more we ought to be able to find a model that addresses most of the shortfallings of the current setup. If that included better negotiation of chair legs then I think the benefit would outweigh the effort one had to spend preparing the room. But unless things have changed a lot in the last 12 months I doubt that there’s been an improvement in the availability of a more “active” mopping facility (better than dragging a soggy cloth around), so that may remain untenable. Similarly I’d want to understand how a “self-emptying” model would fare where pet hair was concerned.

Oh, and the white one – when viewed from a certain angle – looks like it could pass for a member of Daft Punk. When it doesn’t have a big labrador-merkin trailing around after it.