![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
Confused, distressed Roomba?
Jun. 30th, 2009 10:40 pmOur Roomba has been developing signs of what appears to be madcow: wandering vaguely, getting stuck, hooting mournfully; the final symptom being an inability to dock. The poor thing would approach in a decisive fashion and then, just as it hit the ramp, get confused and back off again, whirl a bit, and trundle off in the wrong direction.
Sunday, I stripped a beast long overdue for a full decoke, and not before time. It's been [mumble] since I cleaned out its geartrain and it was quite clear from the teeth that Insufficient Flossing has been taking place. Yuck. Much gunk elsewhere under the covers, of course, and I stripped the buffer assembly for the first time ever on the educated guess that close-range navigational difficulties might be down to sensor issues. What I found under there certainly seemed to support my hypothesis.
I couldn't help feeling that there were some sort of dodgy cross-species relationship issues with using the little hand-held vacuum to go suckety at the internals of the Roomba.
Reassembled, I pressed the button. Sadly, if anything, the thing was even more confused: it actually worked now, after a fashion, for some time but then got into a snit and decided it had an invisible (invisible to us anyhow: I expect the Tribe could see it) barrier in front of it, and reversed first one way, turned, reversed another way, turned again, and finally mounted the armchair base and stopped.
I finally got round to further investigation tonight. It was clear enough that the buffer sensors were the prime suspects; one of the microswitches sticking down intermittently would cause exactly the navigational symptoms that were now distilled into unambiguity in the newly cleansed machine. Sure enough, closer inspection showed that one end of the spring metal strip that the buffer acts against had come adrift from its little plastic entrapment: it's not obvious, as the strip was still in the right place and sproinged pretty much as it should to the touch, but the lack of proper location was enough to börk the fine balance of reality avoidance.
So, with the buffer assembly reassembled in a reverse of disassembly sort of way, a final trial. And with a metaphorical w00t! at least, the Roomba headed across the floor into Ron, then trundled under the armchair and with an air of insouciance over the base where it would normally get stuck; not just that, but lifted into the hallway and dock button pressed, the machine headed straight toward its power base, gave a little wiggle, and settled with a satisfied and melodious little toot.
Success! But a final note: if you're tempted into more and deeper surgery of your Roomba by this, please remember that the Roomba is mostly held together by screws into plastic. If you have any lack of mechanical empathy, or have a tendency to give screws just a little bit more torque to make sure they're properly tight, I respectfully suggest you leave well alone and find someone who's at least read Zen and the Art of Motorcycle Maintenance to do the job :)
Sunday, I stripped a beast long overdue for a full decoke, and not before time. It's been [mumble] since I cleaned out its geartrain and it was quite clear from the teeth that Insufficient Flossing has been taking place. Yuck. Much gunk elsewhere under the covers, of course, and I stripped the buffer assembly for the first time ever on the educated guess that close-range navigational difficulties might be down to sensor issues. What I found under there certainly seemed to support my hypothesis.
I couldn't help feeling that there were some sort of dodgy cross-species relationship issues with using the little hand-held vacuum to go suckety at the internals of the Roomba.
Reassembled, I pressed the button. Sadly, if anything, the thing was even more confused: it actually worked now, after a fashion, for some time but then got into a snit and decided it had an invisible (invisible to us anyhow: I expect the Tribe could see it) barrier in front of it, and reversed first one way, turned, reversed another way, turned again, and finally mounted the armchair base and stopped.
I finally got round to further investigation tonight. It was clear enough that the buffer sensors were the prime suspects; one of the microswitches sticking down intermittently would cause exactly the navigational symptoms that were now distilled into unambiguity in the newly cleansed machine. Sure enough, closer inspection showed that one end of the spring metal strip that the buffer acts against had come adrift from its little plastic entrapment: it's not obvious, as the strip was still in the right place and sproinged pretty much as it should to the touch, but the lack of proper location was enough to börk the fine balance of reality avoidance.
So, with the buffer assembly reassembled in a reverse of disassembly sort of way, a final trial. And with a metaphorical w00t! at least, the Roomba headed across the floor into Ron, then trundled under the armchair and with an air of insouciance over the base where it would normally get stuck; not just that, but lifted into the hallway and dock button pressed, the machine headed straight toward its power base, gave a little wiggle, and settled with a satisfied and melodious little toot.
Success! But a final note: if you're tempted into more and deeper surgery of your Roomba by this, please remember that the Roomba is mostly held together by screws into plastic. If you have any lack of mechanical empathy, or have a tendency to give screws just a little bit more torque to make sure they're properly tight, I respectfully suggest you leave well alone and find someone who's at least read Zen and the Art of Motorcycle Maintenance to do the job :)