A wireless vibration sensor is a small, battery-powered device with a MEMS accelerometer that bolts or glues onto a machine, measures vibration on a schedule, and transmits the readings over a radio link to a gateway, no signal cable and no technician walking a route to collect the data. It turns periodic condition checks into a continuous, always-on trend.
These sensors sit between two older approaches: route-based data collection, where a technician carries a portable analyzer machine to machine once a month, and permanently wired online systems that guard the most critical assets. This guide explains how wireless sensors work, how they compare to route-based collection, what you give up by going wireless, and where they belong in a condition monitoring program.
What is a wireless vibration sensor and how does it work?
It is a self-contained node that senses, processes, and transmits vibration without wires. Most share the same signal path:
- A MEMS accelerometer often triaxial, reading vibration in three directions at once, measures acceleration at the mounting point. MEMS sensors are cheap, rugged, and low-power, which is what makes battery operation practical.
- An onboard processor computes summary features on the device: overall RMS velocity, peak acceleration, temperature, and often an FFT spectrum, so the sensor sends compact results instead of raw waveform.
- A radio commonly Bluetooth Low Energy for short hops or a sub-gigahertz protocol like LoRaWAN for long range and low power, sends the data to a gateway, which forwards it to software.
- A battery powers all of it, typically for three to seven years depending on how often the sensor wakes up, samples, and transmits.
The sensor spends almost all its life asleep. It wakes on a schedule, every few hours is common, takes a measurement, sends it, and sleeps again. That duty cycle is the whole battery-life game, and it is the root of every tradeoff below. For how the sensor is fixed to the machine, which matters more than most buyers expect, see accelerometer mounting methods.
How do wireless sensors compare to route-based data collection?
Wireless sensors trade signal depth for coverage and frequency; route-based collection does the reverse. Neither is strictly better, they answer different questions.
| Wireless sensors (continuous) | Route-based (walk-around) | |
|---|---|---|
| Measurement frequency | Every few hours, automatically | Typically monthly, when the route runs |
| Coverage | Every instrumented point, always on | Only what the route reaches, when it runs |
| Signal depth | Summary features, often FFT; battery-limited resolution | Full high-resolution waveform and spectrum |
| Labor | Install once; no recurring collection | Ongoing technician hours every cycle |
| Cost per point | Low hardware, low labor | No hardware, high recurring labor |
| Best at catching | Faults that develop between routes | Detailed diagnosis of a known problem |
The gap that continuous monitoring closes is timing. A bearing fault can go from first detectable to functional failure in less than a month, inside the window between two route visits. On the P-F curve route-based collection can walk right past the point where the problem first became visible and only catch it on the next pass, sometimes too late. Continuous sensing shortens that blind spot to hours, which is the whole reason machine monitoring is moving from monthly snapshots to always-on streams. That is also why wireless data feeds naturally into predictive maintenance: a trend sampled every few hours is enough to model where a fault is heading, where a monthly dot is not.
There is a labor story underneath the timing one. A walk-around route is recurring cost forever: someone carries the analyzer, collects the points, and uploads the data every cycle, whether or not anything is wrong. Wireless sensors move that cost to a one-time install. On a plant with dozens of monitored points, the hours a technician used to spend collecting readings become hours spent on the repairs the readings call for, which is the more valuable use of a scarce, hard-to-hire skill set. The sensor hardware is often the smaller line item; the labor it frees is the real return.
What do you give up by going wireless?
Three things, all traceable to that battery. Know them before you buy, because a sensor sold on convenience can quietly under-deliver on diagnosis.
- Resolution and bandwidth. Capturing a long, high-sample-rate waveform costs power and radio time. Many battery sensors send summary features and a limited-resolution spectrum, which is plenty for trending but can miss the fine detail a specialist uses to distinguish, say, an outer-race defect from a cage fault. Read the spectral range and lines-of-resolution spec, not just the marketing.
- Sampling frequency. More frequent measurements drain the battery faster. A sensor set to wake every four hours will not catch a fault that appears and fails within one hour, rare, but real on fast-degrading assets. Truly critical machines still warrant continuous wired monitoring.
- The human read. A wireless system flags a threshold breach; it does not always tell you what the fault is. You still need the diagnostic knowledge, or software with it built in, to turn an alarm into a work order. Matching alarms to ISO 20816 vibration zones gives you defensible thresholds instead of guesses, and bearing defect frequencies turn a spectrum into a named fault.
Where do wireless vibration sensors make sense?
On the wide middle of your equipment, assets too important to run to failure, but not important enough to justify a permanent wired system. Rank your machines by criticality and the answer usually sorts itself into three tiers.
How do you deploy wireless sensors well?
A pile of sensors is not a program. The plants that get value follow a sequence.
- Rank assets by criticality first. Use an equipment criticality analysis to decide which machines get wireless sensors, which get wired monitoring, and which stay on routes. Instrumenting everything wastes money on machines you would happily run to failure.
- Choose the sensor to the fault you fear. Match spectral range and resolution to the failure modes on those assets, low-speed bearings and gearboxes need different specs than a standard motor. Buy for the diagnosis, not the datasheet headline.
- Mount consistently and repeatably. Bad mounting corrupts the signal above a few kilohertz. Standardize location and mounting method per machine so readings are comparable over time.
- Set thresholds from a standard, not a guess. Start alarm limits from ISO 20816 zone boundaries for the machine class, then tune to each asset's baseline once you have a few weeks of data.
- Wire the alarm to an action. An alert that lands in an inbox nobody owns is worthless. Route every threshold breach into your CMMS as a work request so it becomes a planned job, and feed confirmed findings back to sharpen the thresholds. This closed loop is where the reliability gains in equipment reliability actually come from.
What the numbers say
- Continuous condition data is what moves failures from unplanned to planned. The U.S. Department of Energy's FEMP O&M guidance, maintained by PNNL, reports condition-based programs save 8–12% over preventive-only maintenance, with the opportunity versus reactive operation reaching 30–40% (PNNL, O&M Best Practices: Maintenance Approaches). Cheaper sensing widens the set of assets where that math works.
- Vibration alarm limits should come from a standard. ISO 20816-1 sets the general framework, evaluation zones A through D by vibration magnitude and change, that defensible wireless thresholds are built on (ISO 20816-1:2016, Mechanical vibration, Measurement and evaluation of machine vibration).
- Sensors free up scarce people. The U.S. Bureau of Labor Statistics projects 13% growth (2024–2034) for industrial machinery mechanics and millwrights, much faster than average, with about 54,200 openings a year (BLS Occupational Outlook Handbook). Eliminating walk-around routes puts those hours on repairs instead of readings.
Wireless vibration sensors are not a replacement for expertise or for wired monitoring on your worst-case assets, they are the cost-effective way to put eyes on the many machines that used to get checked once a month or not at all. The value comes when their data lands in the same place as your work orders and machine signals, not in a separate dashboard nobody opens. That single-layer view is what Harmony builds; see how it looks on a real floor in the CLS case study or on the features overview.