Daniel sent us this one, and it's a deeply practical question about a tension a lot of renters are living with right now. He's built a home security system using Home Assistant and Alarmo — door sensors, PIRs, Zigbee wall switches — and it works. But almost everything is running on coin-cell batteries because, as a renter, he can't just cut into walls and run mains power. The question is: what's the actual risk margin of relying on batteries for security, and if you wanted to wire these things to mains anyway, what exactly do you say to an electrician who's never heard of a Zigbee sensor? What are the voltage, current, and wiring parameters that turn this from a vague idea into a conversation a tradesperson can actually work with?
This is one of those questions where the gap between what the hardware can do and what the rental market lets you do is the entire problem. You can build a surprisingly capable security system for under two hundred dollars with off-the-shelf Zigbee sensors and a Raspberry Pi running Home Assistant. But the moment you depend on it for actual security, the battery question stops being an inconvenience and starts being a liability. A CR2032 that dies at three in the morning during a break-in isn't a maintenance issue — it's a failure of the entire system.
Daniel's instinct is right — you can configure battery alerts in Home Assistant, you can set automations that ping your phone when a sensor drops below twenty percent, but that's a mitigation, not a solution. The failure mode on these coin cells isn't always a nice gradual decline.
No, and this is the first thing people get wrong about battery-powered security sensors. Let's get into the actual numbers, because they're illuminating. A typical Zigbee door and window sensor — something like the Aqara model, which is probably the most common one in Home Assistant setups — draws roughly zero point zero five milliamps in standby. That's fifty microamps. When it actually transmits a state change, it bursts up to about ten milliamps for maybe fifty milliseconds. That's it. A CR2032 coin cell has a nominal capacity of about two hundred twenty-five milliamp-hours. So on paper, you're looking at well over two years of battery life.
On paper being the operative phrase.
Real-world factors eat into that aggressively. Temperature swings — if that sensor is on a doorframe that gets direct sun or is near a draft, the battery chemistry behaves differently. Reporting frequency matters — if your Zigbee mesh is having a bad day and the sensor retries transmission three or four times per event, each retry is another ten-milliamp burst. And then there's the self-discharge rate of the battery itself, which accelerates as the voltage starts to sag. In practice, most people see six to twelve months on a door sensor, not two years. PIR sensors like the Aqara RTCGQ11LM are similar — about zero point zero three milliamps standby, eight milliamps on trigger — but they trigger more often, so they tend to drain faster.
The battery life is shorter than the spec sheet promises, fine. But the real security problem isn't that you have to replace batteries more often — it's that the death isn't predictable.
And this is the misconception I want to nail down. Home Assistant can read the battery level from most Zigbee sensors — it shows up as a percentage, and you can set an automation to alert you when it drops below, say, twenty percent. The problem is that what the sensor reports as battery level is derived from voltage, and the discharge curve of a CR2032 is not linear. It'll sit at around three volts for most of its life, then fall off a cliff. You can go from twenty percent reported to dead in a matter of hours, sometimes less. And if there's any corrosion on the contacts — which happens in humid environments — the voltage reading becomes unreliable. You could have a sensor reporting eighty percent battery while it's already intermittently failing to transmit.
The alert system gives you a false sense of security. The sensor says it's fine, then it's not, and the gap between those two states might be overnight.
That's the scenario that should keep you up at night if you're depending on this for security. A door sensor that silently dies at two in the morning is a door that's no longer monitored. Alarmo won't know the difference between "door closed" and "sensor offline" unless you've specifically configured it to treat offline sensors as a fault state — and even then, you're getting an alert after the fact, not preventing the gap.
That's the case for wiring these things. But "wiring to mains" for a device that runs on three volts DC from a coin cell — what does that actually mean? You're not running a hundred-twenty-volt line directly into a door sensor.
No, and this is where most people's mental model breaks. What you're actually doing is installing a low-voltage DC power supply — essentially a transformer that converts a hundred-twenty or two-hundred-forty volt AC wall power down to five volts or three point three volts DC — and then running thin, low-voltage wire from that supply to each sensor location. The sensor itself either needs to have screw terminals for external power or a micro-USB port. Most battery-only Zigbee sensors don't have either. So step one is choosing sensors that actually accept external power.
Which is a smaller pool than people expect. The Sonoff SNZB-01P can be powered via USB, but the standard Aqara door sensor? You're either soldering wires to the battery contacts — which I don't recommend unless you really know what you're doing — or you're buying different hardware from the start.
That hardware exists, it's just less visible in the consumer market because the entire Zigbee sensor ecosystem is built around the convenience of stick-on battery devices. But if you're going wired, you're looking at industrial or semi-industrial sensors — things like the Xiaomi Aqara wired door sensor, or certain models from Centralite and Visonic that have terminal blocks for external power. They're not harder to set up, they're just not what Amazon's algorithm shows you first.
Let's say Daniel finds sensors with power inputs. He's now got to talk to an electrician. What does he actually say? What are the numbers?
Here's the spec sheet you hand to the electrician. Voltage: five volts DC is the sweet spot — it's the USB standard, it's what most of these sensors accept, and it's low enough to be completely safe. Current: budget about a hundred milliamps per sensor, even though most draw far less. That gives you headroom. So if you're wiring eight sensors, you need a power supply that can deliver at least eight hundred milliamps — call it one amp to be safe. A five-volt, one-amp wall-wart power supply costs about ten dollars and can handle ten sensors without breaking a sweat.
The wire itself?
Twenty-two AWG solid-core wire is the standard for this kind of low-voltage run. At twenty-two gauge, the resistance is about sixteen ohms per thousand feet. Over a fifty-foot run — which is longer than most apartments — at a hundred milliamps, the voltage drop is about zero point zero eight volts. You don't need anything thicker. For longer runs, say over a hundred feet, you'd step up to eighteen AWG, but in a typical two-bedroom apartment, you're never going to hit that.
We're talking about wire that's basically doorbell wire. Thin, cheap, easy to work with.
And this is where the rental situation gets interesting, because Daniel's constraint isn't technical — it's the lease. You can't cut into drywall and fish wire through studs without landlord permission, and most landlords will say no or charge you for restoration. But here's the thing: low-voltage wiring doesn't have to go in the wall.
Surface-mount raceway.
Wiremold, or any equivalent brand. It's a plastic channel with an adhesive backing that sticks to the baseboard or runs along the ceiling line. You snap the wires inside, snap the cover on, and it looks neat enough that most people won't notice it. When you move out, you pull it off, spackle a few tiny nail holes if you used the screw-mounted version, and you're done. It's completely renter-friendly, and it's explicitly allowed under most lease agreements because it doesn't modify the structure.
Here's a subtle point most electricians won't volunteer: under the National Electrical Code, specifically NEC article seven twenty-five, low-voltage wiring — defined as under sixty volts — doesn't require a licensed electrician in most jurisdictions. You can do the low-voltage side yourself. The only part that needs a licensed electrician is installing the receptacle that the power supply plugs into, if you need a new one.
Which most people don't — you just plug the five-volt supply into an existing outlet. So the electrician conversation Daniel needs to have is actually much narrower than he thinks. It's not "wire my entire smart home." It's "install a Class Two power supply near my Home Assistant server, and I'll handle the rest." A Class Two supply, per UL thirteen ten, is specifically designed for low-voltage applications and has built-in current limiting for safety. Any electrician will know exactly what that means.
Let's walk through the actual conversation. Daniel calls an electrician. What does he say?
"I need a Class Two power supply installed, five volts DC output, rated for at least one amp. I'll be running twenty-two AWG solid-core wire from the supply to each sensor location, daisy-chained. Each sensor draws fifty to a hundred milliamps max. The sensors themselves are Zigbee door and window contacts with screw terminals for external power. I'm handling the low-voltage wiring and sensor installation myself under NEC seven twenty-five. I just need you to install the receptacle and the transformer if it's not a plug-in type.
That's a script that makes you sound like you know what you're talking about. The electrician hears "Class Two," "five volts DC," "NEC seven twenty-five," and they immediately understand the scope. They're not being asked to learn about Zigbee or Home Assistant. They're being asked to provide clean five-volt power — which is about as routine as it gets.
Here's the thing: if the electrician pushes back and says they need to do the whole job, you can politely point to NEC seven twenty-five and say you're comfortable with the low-voltage portion. Most will be fine with it. Some might even be curious about what you're building.
We've got the technical parameters and the conversation script. But there's a cost question hanging over all of this. Daniel's current setup — battery sensors taped to the wall — is cheap and fast. What's the premium for going wired?
Let's run the numbers for an eight-sensor apartment over five years. Battery sensors: you're looking at about thirty dollars each for something like an Aqara door sensor — that's two hundred forty dollars upfront. Batteries: eight sensors, each needing a CR2032 roughly twice a year, at about a dollar per battery — that's sixteen dollars a year, eighty dollars over five years. Total cost of ownership: about three hundred twenty dollars.
The wired approach?
Wired sensors with external power inputs tend to run a bit more — call it forty dollars each for something comparable, so three hundred twenty dollars for eight. A good five-volt, two-amp power supply: fifteen dollars. Twenty-two AWG wire: maybe ten dollars for a hundred-foot spool. Surface-mount raceway: about twenty dollars for enough to do a whole apartment. And the electrician visit to install the receptacle if you need one: a hundred to a hundred fifty dollars, though you may not need it at all. Total: roughly four hundred seventy-five dollars on the high end.
The premium is about a hundred fifty to two hundred dollars for zero battery anxiety and a system that's genuinely reliable as security infrastructure.
That premium shrinks if you take a hybrid approach, which is what I'd actually recommend for most renters. Wire the perimeter — the doors and windows that are actual entry points. Those are your security-critical sensors. Leave interior PIRs and those four-gang wall switches on battery. The PIRs are for convenience automation, not security — if one dies, your lights don't turn off automatically, which is annoying but not a safety issue. And the wall switches, those Zigbee four-gang remotes taped to the wall with VHB tape? They're not security devices at all. They're convenience. Let them run on CR2032s.
That cuts the wiring job in half. Four or five perimeter sensors instead of eight or ten. The cost premium drops to maybe seventy-five dollars, and you're covering the highest-risk points.
There's a second-order benefit here that's worth mentioning. When you wire the perimeter sensors, you eliminate not just the dead-battery failure pattern but also the radio-retry problem. A battery sensor that's struggling to reach the Zigbee coordinator will retry multiple times, burning through battery faster. A wired sensor doesn't care — it's got unlimited power, so it can retry as many times as it needs to. Your Zigbee mesh becomes more reliable at the points that matter most.
The sensor isn't just more reliable because the battery won't die — it's more reliable in operation because it's not conserving power.
Now, there's one more failure pattern that wiring introduces, and we should be honest about it. When you move from distributed battery power to a centralized power supply, you create a single point of failure. If that five-volt supply dies, every wired sensor goes dark simultaneously.
Which is arguably worse than one battery dying at a time.
The mitigation is straightforward: put the power supply on a small UPS, or use a power supply with a battery backup module built in. Something like a five-volt UPS hat for a Raspberry Pi can also power your sensor network. If you're already running Home Assistant on a UPS — which you should be for security — you just tap the sensor supply into the same backup. Problem solved for about thirty dollars.
We've covered the technical specs, the electrician conversation, the cost comparison, and the failure pattern. But there's an elephant in the room that Daniel's question hints at: why is this still so hard? Why isn't there a standard for power-over-data in the smart home sensor world?
There sort of is, and it's called Thread. Thread is the networking protocol that Matter runs on, and it's designed from the ground up to support low-power devices — including ones that can run on batteries for years. But Thread also supports what are called "router-capable" devices that are mains-powered and can forward packets for battery-powered "end devices." The vision is that your light bulbs and smart plugs — things that are already mains-powered — become the backbone of the mesh, and your battery sensors just talk to them. In that world, the sensor battery lasts much longer because the radio doesn't have to transmit as far.
We're not there yet.
We're not there yet. The Thread and Matter ecosystem is still maturing. Device support is spotty. And critically for Daniel's use case, Alarmo and Home Assistant's security capabilities are still primarily built around Zigbee and Z-Wave. Thread support in Home Assistant is improving, but it's not the thing you bet your home security on today.
For the next two to three years at least, the choice is what we've been describing: battery with alerts, or wired with a power supply.
I think we'll see a wave of products aimed directly at this gap. Renter-friendly wired sensor kits — sensors with micro-USB ports, bundled with adhesive raceway and a plug-in five-volt supply, marketed as "no-electrician-required." The pieces all exist, they're just not packaged together yet. Someone's going to put that kit on Amazon and clean up.
What does Daniel actually do on Monday morning?
First, before touching any wiring, make sure the Alarmo hub and any sirens are on a UPS. Those are the highest-draw devices and the most critical for security. If the power goes out, you want Alarmo still running and the siren still capable of triggering. That's step one, and it's independent of the sensor question.
Set up the battery alert automation properly. Don't just rely on the default battery percentage entity in Home Assistant. Create an automation that triggers when any security sensor's battery drops below twenty percent, and have it send a critical notification to your phone — the kind that bypasses Do Not Disturb. Make sure it works. That's your safety net while you figure out the wiring plan.
Step three is the wiring plan itself.
Identify which sensors are security-critical — doors and windows that are actual entry points. That's your wiring scope. Source sensors with external power inputs — look for models with screw terminals or micro-USB. Buy a five-volt, two-amp Class Two power supply, a spool of twenty-two AWG solid-core wire, and enough surface-mount raceway to reach each sensor location. If you need a new receptacle near your Home Assistant server, call an electrician and use the script we laid out. If you don't, you can do the whole thing yourself under NEC seven twenty-five.
The hybrid approach means you're not rewiring the entire apartment. Four or five sensors, an afternoon of running raceway, and you've eliminated the battery-anxiety problem at the points where it actually matters for security.
The total cost for that hybrid approach — wiring just the perimeter — is probably under two hundred dollars including the sensors. And you get to keep the convenience of battery-powered PIRs and wall switches everywhere else. It's the pragmatic middle ground.
One thing we haven't touched on: those four-gang Zigbee switches Daniel mentioned, taped to the wall with VHB. Is there even a wired alternative for those in a rental?
Honestly, not a good one. Those battery-powered scene switches are designed to be stuck anywhere — they're not replacements for hardwired light switches, they're supplementary controllers. If you wanted a hardwired version, you'd need to replace the actual in-wall switch, which a renter absolutely should not do without the landlord's involvement. And even then, you're talking about a Zigbee or Z-Wave in-wall relay, which is a different product category entirely. For those, I'd say embrace the battery. They're not security devices. If the battery dies, you lose the convenience of a scene controller, not the ability to detect a break-in.
The distinction we keep coming back to is: security sensors versus convenience sensors. Wire the security ones, manage the batteries on the convenience ones, and don't conflate the two.
That's the principle. And I think that distinction is what most of the online discussion misses. People treat all their smart home sensors as one category, and then they either wire everything — which is overkill and expensive — or they leave everything on battery and worry. Separating security from convenience is the mental model that makes the whole problem tractable.
On the battery side, there's an interesting development worth watching. Lithium thionyl chloride batteries — the kind used in industrial sensors — have a ten-year shelf life and a much flatter discharge curve than CR2032s. They're not drop-in replacements for coin cells, but some newer Zigbee sensors are starting to use them. If that trend continues, the battery-versus-wired debate might look very different in a few years.
That's the open question, isn't it? Will battery chemistry improve faster than the Thread and Matter ecosystem matures? If we get sensors with reliable ten-year battery life, the whole wiring conversation becomes moot for anyone except the most paranoid. But we're not there yet, and for someone depending on this system for security today, wired is the only way to eliminate that failure pattern entirely.
The rental market specifically is underserved here. The assumption baked into most smart home products is that you own the walls. The reality for millions of renters is that you don't, and the industry hasn't caught up. The kit we're describing — sensors with power inputs, adhesive raceway, plug-in supply — it doesn't exist as a single SKU. Someone's going to build it, and they're going to sell a lot of them.
Until then, you piece it together yourself. The numbers are on your side — the power requirements are so low that the wiring is trivial, the cost premium is modest, and the reliability gain is immediate. The hardest part is just knowing what to ask for.
Now: Hilbert's daily fun fact.
Hilbert: In the nineteen thirties, a marine biologist claimed to have discovered a species of giant tube worm in the Aleutian Trench that generated its own light through bioluminescent bacteria. The finding was celebrated for decades until a reexamination of the specimens in the nineteen eighties revealed they were actually fragments of a common North Pacific sponge that had been colonized by phosphorescent microorganisms. The "giant tube worm" never existed.
A sponge with a glow-in-the-dark infection passed for a new species for fifty years.
Marine biology in the thirties was apparently just pointing at things and guessing.
So the future of renter-friendly smart home security is either going to be solved by better batteries, better power-over-data standards, or someone finally packaging the pieces we described into a single product. My money's on the product — the demand is too obvious to ignore. If you found this useful, leave a review on your podcast app. It helps other renters who are staring at a pile of Zigbee sensors and wondering if they're actually secure.
This has been My Weird Prompts. Thanks to our producer Hilbert Flumingtop. I'm Herman Poppleberry.
I'm Corn. We'll catch you next time.
