Power decides more than people expect
When people plan a connected device, they usually start with what it should sense or control. Power tends to come second, and then it quietly dictates a lot of the design: how often the device can report, which radio it can use, how big it is, where it can be mounted, and how much maintenance it needs once it is installed.
There are four broad ways to power a device: mains electricity, batteries, power delivered over a data cable, and energy harvested from the surroundings. Many real products combine two of them, such as mains power with a small backup cell.
Mains power
A device plugged into a wall outlet, or wired into a building’s supply through a small adapter, has the simplest life. Energy is effectively unlimited for the purposes of a small electronic gadget, so the designer can use a fast processor, a bright display, a camera or a radio that stays on all the time.
The trade-offs are practical rather than technical:
- The device must be near an outlet, which limits where it can go.
- Installation may need an electrician if it is wired in rather than plugged in.
- A power cut turns the device off unless it has a backup.
- The adapter or internal supply is itself a part that can fail or waste a little energy as heat.
Mains suits things that already sit near power: speakers, hubs, plugs, thermostats, lighting and most appliances. It also suits anything that must be always listening, such as a video doorbell or a camera that records continuously.
Batteries
Batteries free a device from the wall, which is why they dominate sensors in awkward places: a door contact, a leak sensor under a sink, a soil probe at the edge of a field. The cost is that every design choice now spends a limited budget.
There are two main families. Primary (non-rechargeable) cells are installed and replaced when empty; they are common in low-power sensors because they have a long shelf life and need no charging. Rechargeable cells suit devices that are used heavily or handled daily, such as wearables, but they need a charging circuit and usually have a limited number of charge cycles.
Battery-powered design pushes you toward a specific style:
- The processor is a small, frugal one, typically one of the microcontrollers designed for low power, rather than a full computer.
- The device sleeps most of the time and wakes briefly to measure and report.
- The radio is chosen for low energy use per message, which is why low-power networks and Bluetooth Low Energy are popular for battery sensors, while always-on Wi-Fi connectivity is often a poor fit.
- Someone must replace or recharge the cells, which becomes a real cost when there are hundreds of devices.
How long a battery lasts depends on the design, the environment and the reporting pattern, covered in more detail in separate articles on what drains battery life, on sleep behavior, and on chemistry.
Power over a data cable
Some wired devices receive power and data through the same network cable. This is common for cameras, wireless access points and some desk phones. A single cable run replaces both a data connection and a separate power outlet, which is helpful on ceilings and outdoor poles where no outlet exists.
The limits are cable length (typically around 100 meters for standard network cable) and the amount of power available, which is modest compared with a wall outlet but enough for many devices. The switch supplying the power becomes a central point: if it fails or loses power, everything it feeds goes down together. A dedicated article covers this in more detail.
Energy harvesting
Harvesting collects small amounts of energy from light, heat differences, vibration or radio waves and stores it in a tiny battery or capacitor. The attraction is obvious: no battery changes, and a device that can in principle run for years.
The reality is that the energy collected is small and irregular. Harvesting works only for devices that do very little, report occasionally and tolerate gaps. It is usually paired with a small storage element, and the design must assume that on some days there is almost no energy at all. Solar is by far the most common source outdoors and has its own sizing problems, discussed in a separate article.
Comparing the options
| Power source | Typical strength | Typical weakness |
|---|---|---|
| Mains | Plenty of energy, always on | Needs an outlet; fails in a power cut |
| Primary battery | Easy to place, simple | Finite life; replacement visits |
| Rechargeable battery | Reusable, good for frequent use | Charging needed; aging over time |
| Data cable power | One cable, central control | Cable length and power limits |
| Harvesting | Little or no maintenance | Small, variable energy |
Questions to ask before choosing
- How much energy does the device really need? Estimate an average, not a peak. A short radio burst can use far more than the rest of the day combined.
- Where will it live? Outdoors, in the cold, in a ceiling void or in a sealed enclosure all change what is practical.
- Who will maintain it, and how often? A battery change that is easy on one device is expensive across a large fleet; this is a core part of the total cost of a deployment and of how well it can scale from ten devices to thousands.
- What happens when power fails? Safety sensors and anything tied to locks or alarms may need a backup source.
- What is the environmental cost? Frequent cell replacement creates waste, and disposal deserves planning from the start.
A sensible default
For most homes, mains power for anything near an outlet and long-life batteries for small sensors is a workable pattern. Outdoor and remote sites tend to lean on batteries, solar with storage, or cable-delivered power where a network cable already runs. Whichever you choose, treat power as a design input from day one rather than something to bolt on at the end, because it is far cheaper to adjust a plan than to revisit a deployed device.