Plaster dust is sifting onto my shoulder by the second mounting bracket, and I still haven't found a stud that isn't buried under a century of horsehair lath. This is the DIY install I put off far longer than I should have — swapping every standalone smoke unit in our craftsman bungalow for an interconnected system of X-Sense smart detectors, room by room, before old-home maintenance turned into old-home negligence. Smart home safety is an easy thing to shrug off until the one alarm you own has gone a jaundiced yellow and you can't remember the last time you tested it.
The push to actually finish the project traces back to a morning a few weeks earlier, watching a test robot vacuum stall out in the kitchen. The dock light was blinking red, the error beep cycling every so often through the house, and Beans — our senior beagle — never lifted his head off the dog bed three feet away. If a beeping robot that close couldn't get a rise out of him, a single smoke alarm two rooms over during an actual fire wasn't going to do much better. The short version of what I landed on: in a balloon-framed two-story like ours, you place detectors for the path smoke and carbon monoxide actually travel — basement to attic, hallway junctions, high-load rooms — not just one per bedroom door the way the box tells you to.
Old-Home Wiring Makes This Bungalow a Chimney for Smoke
Houses built in the 1920s like ours weren't framed the way houses are framed now. Builders often skipped fire blocking between the studs, so if something starts burning in the basement — the same basement where my old Roomba i3 died on the steps — smoke can travel straight up through the wall cavities to the attic before a single ceiling-mounted alarm ever catches a whiff of it. That's what got me thinking about detectors as a network instead of gadgets scattered around the house. Our lath and plaster walls create a mild Faraday cage effect on the 2.4 GHz signal the base station uses, so a detector that can't reliably reach the hub isn't really part of the smart system — it's just a loud box on the wall.
Why the NFPA 12-Inch Rule Gets Messy on Horsehair Plaster
NFPA 72 sets the placement standard for residential smoke alarms, and one line matters most here: wall-mounted units need to sit within 12 inches of the ceiling. That's a five-minute job on drywall. On century-old horsehair plaster it's a different project — every hole releases a fine grey dust that's heavier and grittier than anything a robot vacuum has ever picked up off our hardwood, and it keeps turning up in the floorboard seams for weeks afterward.
Finding the sweet spot meant balancing that 12-inch rule against the oak trim Sam would never forgive me for drilling into. Go too high and I'm putting a hole through solid molding; too low and the detector misses smoke pooling at the ceiling. Most units ended up ceiling-mounted, roughly two feet clear of any HVAC vent so airflow wouldn't push smoke past the sensor before it registered. It's the same logic as where to put an air purifier for the best indoor air quality — you're chasing the dead zones where air doesn't actually move, not just picking a wall and hoping.
The CO side of the system got its own placement logic entirely. Our furnace and water heater live in the basement, so a CO detector went in near the top of the stairs — the same open stairwell off the kitchen that's never had so much as a gate on it, robot vacuum or otherwise. Smoke rises and pools near the ceiling; carbon monoxide mixes into the air more evenly, which is the real reason one detector per floor was never going to cover a stacked, boxy layout like this one. You need coverage at the source and at the places people actually sleep, not just a single unit splitting the difference.
Bedrooms Aren't the Priority I Expected
Standard advice says prioritize every bedroom first, and I get why — that's where people are asleep and slowest to notice smoke. But in a house this old, the bigger risk sits in the rooms with the worst wiring, not the rooms with the most pillows. Our dining room sideboard and the living room outlet strip — the one running robot vacuum docks, an air purifier, and Sam's gaming rig off wiring meant for a single floor lamp — worried me more than the guest room we open twice a year.
Outside the kitchen went a detector too, far enough back that a seared dinner wouldn't set it screaming, plus one over the living room where all that outdated wiring converges. Between the two, that's where I'd bet on trouble starting if it ever does — not the quiet bedroom at the end of the hall.
Getting the X-Sense Base Station to Hear the Back Bedroom
The hallway junction mattered more than any single room. Instead of mounting a unit generically "in the hallway," I put it right where the bedrooms meet the open living space, just past the built-in linen closet — close enough to the base station's line of sight that the signal didn't have to fight through the kitchen's tile backsplash to register. That hallway runs hardwood the whole way, with one long runner rug muffling the floor near the bedroom doors, and it's the corridor every signal in this house eventually has to cross.
For the living room I skipped the wall entirely and went with a central ceiling mount, clear of the drafts off our original double-hung windows and with a straight sightline over the sectional — the low-slung one with maybe three inches of clearance underneath, just enough for a robot vacuum to squeeze under and not much else.
What Changed by Week Six
The X-Sense app itself was a relief — clean, none of the multi-screen maze I've come to expect setting up an ECOVACS unit. The real test was whether the units actually talked to each other. Each alarm puts out somewhere around 85 dB on its own, loud in a small room but easy for Beans to sleep through if it's three rooms away and he's deep in one of his afternoon naps. Interconnection means every unit fires at once, and six weeks into running the system, the base station hadn't dropped the back bedroom detector once — no missed pings, no manual re-sync, just a green all-clear sitting on my phone every time I checked it from the coffee shop down the street.
A friend of mine who breeds Bernese mountain dogs swears her whole litter sleeps through alarms that would send most people bolting for the door — big dogs, deep sleepers, apparently it's a breed thing. Beans isn't a hundred-pound dog, he's just old, but the simultaneous-alarm design solves the same problem either way: it's less about raw volume and more about making sure the sound actually reaches the dog instead of just the room he happens to be in.
DIY Placement Lessons for Smart Home Safety, Room by Room
If you're mapping this out in your own old house, the biggest lesson is that placement is a system problem, not a room-by-room checklist. The basement entry mattered most to me personally — the top of the stairs, right by that gate-less stairwell off the kitchen, is where the furnace and water heater live, and it's also close to the spot my old Roomba used to lose its way trying to find the dock. Now that corner is doing something that actually matters if the furnace ever has a bad day. The hallway junction, the living room ceiling, and the spot outside the kitchen do the rest of the work, but none of it counts for much if the base station can't hear the unit at the far end of the house.
Most weeks my attention is on dustbin weights and how fast a new bot maps a floor plan, not on wall-mounted battery units that just sit there until they're needed. I'll argue suction against battery life until Sam changes the subject, and I've got opinions about whether a mop mode is worth the extra noise floor it usually brings — pet waste avoidance logic and HEPA filtration grades feel more urgent to me most days than a detector I forget exists until it beeps low-battery at 2 a.m. None of that tracking prepared me to take a smoke detector as seriously as I now do, and running sticky lint rollers over the couch cushions between every vacuum cycle taught me the same lesson from the other direction — some fixes only look like they're solving the problem.
Yusuf, my neighbor across the alley, still texts me reliability updates on the Dreame L20 Ultra he bought after watching me argue through the Roborock S8 Pro Ultra vs Dreame L20 Ultra for homes with pets comparison in his driveway for twenty minutes. He hasn't asked me a single question about the detectors, which might mean I finally got something right on the first try.
Sam stopped rolling his eyes at my "tech projects" the day the app pinged during a test and he said, "Okay, that one's actually useful." He's right that this isn't the part of the house that impresses anyone — it doesn't vacuum a floor, it doesn't map a room, it just sits there. But if there's one thing carrying over from six weeks of watching this system work, it's that detector placement is a network problem before it's a shopping decision: every wall, every joist bay, every closed door between the base station and the farthest bedroom is signal you have to design around ahead of time, not something you get to fix after the fact.