Vibration Meter

What is a Vibration Meter

A vibration meter is a measuring instrument that converts the mechanical oscillation of a machine surface into an electrical signal, processes that signal, and displays a single overall amplitude value. The sensor, almost always an accelerometer, is coupled to a non-rotating part such as a bearing housing. The meter then integrates and filters the raw acceleration into the chosen output quantity and reports its magnitude as RMS (root mean square), true peak, or peak-to-peak over a defined frequency band.

The defining word is "overall." Unlike a vibration analyzer, a meter does not resolve the signal into its frequency components. It answers "how much" but not "why." For a healthy or sick verdict against an ISO severity zone, that is exactly enough, which is why the overall meter is the workhorse of operator-driven and route-based condition monitoring programs. The cost of one overall meter is a small fraction of a full analyzer, so a plant can equip many operators rather than one specialist.

Functionally a vibration meter has four stages: (1) the transducer, typically an IEPE accelerometer that outputs a low-impedance voltage proportional to acceleration; (2) signal conditioning, which supplies the constant-current bias to the sensor and applies anti-alias and band-pass filtering; (3) the processing path, which integrates acceleration to velocity, and velocity to displacement, and computes RMS or peak detector values; and (4) the display, alarm, and data-storage interface. Many modern units add wireless upload to a maintenance app so that readings flow into a trend database.

Vibration measurement as an engineering discipline matured through the twentieth century. The piezoelectric accelerometer, commercialized in the mid-1940s, made compact, rugged vibration sensing practical. The first severity guideline that engineers still recognize, VDI 2056, arrived in 1964 and was later folded into the ISO 2372 and ISO 10816 series, which standardized the broadband velocity criterion in millimeters per second. In the 2010s the ISO 20816 series began to consolidate machine-vibration measurement under one umbrella, and the IEPE constant-current sensor interface became the near-universal physical layer.

The scale of the measured quantity is wide. On a slow-turning kiln support roll the meaningful signal may be a few micrometers of displacement at well under 10 Hz, while a high-speed compressor gear mesh produces meaningful acceleration energy past 10 kHz. No single setting covers that span, which is why a competent meter offers selectable detectors, filters, and units, and why understanding the measurement quantity (Chapter 4) is the heart of correct use.

A vibration meter sits inside the larger discipline of condition-based maintenance, where vibration is the most information-rich of the common condition indicators alongside temperature, oil analysis, and motor-current signature. The overall reading from a meter is the trigger that decides whether a machine needs the deeper attention of a vibration analyzer. In practice a plant builds a route, a fixed list of measurement points and machines, and an operator walks it on a schedule, logging one value per point. The trend of those values over weeks and months, not any single reading, is what reveals a developing fault, so consistency of sensor, mounting location, and machine operating condition between visits matters as much as the absolute number.

Types of Vibration Meters

Vibration meters fall into distinct classes by purpose and capability. Choosing the wrong class is the most expensive selection mistake: an overall screening meter cannot diagnose a bearing fault, and a machine-health meter cannot certify a worker's daily exposure. The table below separates the main classes by the measurement they perform and the question they answer.

Overall severity meters are the simplest and most common handheld instruments. They report a single broadband number, usually velocity RMS in the 10 to 1,000 Hz band that ISO 10816 specifies, and compare it against an alarm. An operator can survey dozens of pumps and fans per shift. These instruments trend a machine over time but cannot tell imbalance from misalignment.

Bearing-screening meters add a high-frequency channel above the severity band to catch the early, low-energy impacts of rolling-element bearing and gear defects. The Fluke 805, for example, measures overall vibration from 10 to 1,000 Hz and adds a Crest Factor Plus bearing assessment using the sensor tip from 4,000 to 20,000 Hz, with an acceleration range of 0 to 50 g. This class trends an enveloped or crest-based index that rises long before the overall velocity does.

Vibration analyzers and data collectors are a separate tier, technically meters with a spectrum engine. Route-based collectors such as the SKF Microlog and the Emerson CSI 2140 store multiple measurement points, capture FFT spectra and time waveforms, and run single-plane or two-plane balancing. They diagnose the cause that an overall meter only flags, but they cost far more and demand a trained analyst.

Human-vibration meters are a regulatory instrument class governed by ISO 8041-1. They apply frequency weightings to triaxial acceleration and report frequency-weighted RMS, the daily exposure A(8), and the vibration dose value, for hand-arm and whole-body exposure assessment. They share an accelerometer with machine meters but are not interchangeable with them. Online fixed monitors are the permanently installed cousin, continuously watching overall vibration on critical machines and tripping protection when a level is exceeded.

Sensor Technologies

A vibration meter is only as good as its transducer. Three sensor families dominate machine measurement, each with an optimal frequency band and amplitude window. The table compares the key engineering parameters, after which each technology is described in detail.

IEPE accelerometers (Integrated Electronics Piezo-Electric, also sold as ICP by PCB Piezotronics and Isotron by Endevco) are the default sensor for almost every modern vibration meter. A piezoelectric crystal generates charge under acceleration, and a built-in amplifier converts that high-impedance charge into a robust low-impedance voltage at the sensor. The electronics draw a constant current of 2 to 20 mA over the same two-wire coaxial cable that returns the signal, sitting on a DC bias of roughly 7 to 14 V. This lets ordinary coaxial cable run long distances without a separate charge amplifier. The industry standard sensitivity is 100 mV/g, with a usable amplitude window of about plus or minus 50 g.

Charge-mode accelerometers output the raw piezoelectric charge directly, with no built-in electronics, and require an external charge amplifier. Their advantage is temperature: without internal semiconductors they tolerate hot environments well beyond the roughly 120 C ceiling of IEPE units, reaching into gas-turbine and exhaust applications. The penalty is cost, the need for low-noise cable, and sensitivity to cable handling and contamination, so charge mode is reserved for high-temperature and specialty service.

Velocity transducers of the classic moving-coil type contain a coil moving relative to a permanent magnet inside the case. They are self-generating and output velocity directly without integration, which suits simple overall-velocity meters and legacy turbine monitoring. Their usable band is narrow, roughly 10 to 1,500 Hz, bounded by the suspension resonance at the low end, and the moving parts wear over time, so piezoelectric accelerometers with electronic integration have largely displaced them in new designs.

Eddy-current proximity probes are non-contact sensors that measure the relative displacement between a shaft and its bearing housing. A driver oscillator energizes a coil at the probe tip and senses the gap to the conductive shaft, giving a direct micrometer reading typically around 8 mV per micrometer over an 80 mil range. They are the correct sensor for fluid-film (sleeve) bearings, where the shaft floats on an oil film and casing-mounted accelerometers cannot see the shaft motion. They are central to turbomachinery protection systems and standards such as API 670, and they read displacement directly from DC, unlike casing accelerometers.

Standards, Units, and Severity Zones

The single most important concept in vibration measurement is choosing the right measurement quantity for the frequency content of the fault. Acceleration, velocity, and displacement are mathematically linked by integration, but each emphasizes a different part of the spectrum, so the same machine can look healthy in one unit and alarming in another. The table maps quantity to its best use.

Velocity in mm/s RMS is the universal severity unit for general rotating machinery and the basis of the ISO machine-vibration standards. It is roughly flat in the 10 to 1,000 Hz band where imbalance, misalignment, and mechanical looseness live, which makes it a fair single number for comparing different machines. When someone says a pump is at "4.5 mm/s," they almost always mean broadband velocity RMS measured per the ISO method.

ISO 10816 and its successor ISO 20816 classify that broadband velocity into four evaluation zones, measured on the bearing housing in the 10 to 1,000 Hz band. ISO 20816-1 (2016) replaced ISO 10816-1 and is gradually consolidating the older ISO 10816 series for non-rotating parts together with the ISO 7919 series for rotating shafts. The four-zone scheme is consistent across both. The boundaries below are the widely used ISO 10816-3 values for medium machines; the exact figure depends on machine group (power) and on rigid versus flexible foundation.

Other standards govern adjacent uses. ISO 2954 set the requirements for broadband vibration-severity measuring instruments, which is the historical basis for the meters used in ISO 10816 surveys. For protection of large turbomachinery, shaft displacement is monitored per the ISO 20816 (formerly ISO 7919) shaft-vibration approach and the API 670 machinery protection standard.

A wholly separate family addresses people, not machines. ISO 8041-1 defines the performance of human-vibration meters, which apply frequency weightings (Wh for hand-arm, Wd, Wk, and Wj for whole-body) and report exposure metrics such as A(8) and the vibration dose value. The underlying assessment methods come from ISO 2631 for whole-body and the hand-arm series, and feed regional limits such as EU Directive 2002/44/EC. Confusing a machine-severity meter with a human-vibration meter is a common and consequential error.

Finally, instrument and sensor accuracy is anchored by calibration. ISO 16063 defines the methods for calibrating vibration and shock transducers, producing the sensitivity-versus-frequency curve that every credible accelerometer datasheet should reference. In the field, a portable shaker calibrator, commonly running 159.2 Hz at 1 g RMS, verifies the mounted sensor and meter before a survey. The 159.2 Hz figure is not arbitrary: it corresponds to an angular frequency of 1,000 radians per second, which makes the integration constants between acceleration, velocity, and displacement convenient round numbers, so 1 g (9.81 m/s squared) at that frequency yields a clean velocity and displacement to check all three quantities at once.

A practical caution on units closes this chapter. Velocity may be quoted as RMS or as peak, and the two differ by a factor of the square root of two (about 1.41) for a pure sine; mixing them silently doubles or halves an apparent severity. Acceleration may be quoted in g or in m/s squared, which differ by 9.81. Displacement may be peak-to-peak or zero-to-peak. Before comparing any two readings, or a reading against a standard limit, confirm the detector and the unit, because the ISO severity zones are defined specifically in millimeters per second RMS over the 10 to 1,000 Hz band, and a value in any other convention must be converted before the comparison is valid.

Key Specification Parameters

A vibration-meter datasheet can list dozens of lines, but only a handful drive selection. The decisive parameters are measurement quantities supported, frequency range, amplitude range and resolution, detector type, accelerometer sensitivity, mounting interface, accuracy, and environmental rating. Each is decoded below.

Measurement quantities and detectors. Confirm the meter outputs the quantities you need, acceleration, velocity, and displacement, and the detectors you need, RMS, true peak, and peak-to-peak. RMS is the energy-based value the ISO standards use; true peak captures impulsive bearing and gear events; peak-to-peak is the convention for shaft displacement. A meter that reports only velocity RMS cannot screen bearings; one that reports only acceleration cannot be compared against ISO velocity zones without integration.

Frequency range. The severity band is 10 to 1,000 Hz, but bearing and gear faults need a high-frequency channel reaching 5 to 20 kHz. The Fluke 805, for example, covers 10 to 1,000 Hz for overall measurement and 4,000 to 20,000 Hz for its bearing index. Match the high end to the highest gear-mesh or bearing-defect frequency you expect, and remember that the sensor and its mounting, not the electronics, usually set the true upper limit.

Amplitude range and resolution. Typical handheld ranges run to about 50 g acceleration, several hundred mm/s velocity, and a few millimeters of displacement, with resolution such as 0.01 g. The operating point should sit comfortably inside the range, not at its extreme, and the resolution must be fine enough to trend small changes on a quiet machine.

Accelerometer sensitivity and mounting. The bundled or specified accelerometer sensitivity, usually 100 mV/g, sets the amplitude window (about plus or minus 50 g at 100 mV/g). Higher sensitivity (500 mV/g) lowers the noise floor for slow machines; lower sensitivity (10 mV/g) extends range for shock. Mounting sets the usable upper frequency: stud mounting reaches 10 kHz or more, adhesive close behind, a flat magnet roughly 2 kHz, a two-pole magnet near 1 kHz, and a probe tip below 1 kHz with poor repeatability.

Accuracy and calibration traceability. Look for a stated amplitude accuracy (often a few percent of reading) and a frequency-response tolerance, with calibration traceable to ISO 16063. For ISO 10816 survey work the broadband instrument should meet the requirements historically defined by ISO 2954. For human-vibration work the meter must hold a valid ISO 8041-1 pattern evaluation. A meter without traceable calibration cannot defend its readings in an audit.

Environmental and interface parameters. Ingress protection (IP54 to IP67), operating temperature, drop rating, battery life, and data interface (on-board memory, USB, Bluetooth to a maintenance app) determine field usability. The table below summarizes the spec lines that actually drive a decision.

Selection Decision Factors

To turn the preceding chapters into a specific instrument, follow the decision sequence below. Most selection errors come not from a single wrong line item but from skipping the first question, what job the meter must do, and jumping straight to brand and price. These seven steps work as an RFQ template.

1. Define the job: screening, diagnosis, or compliance. Screening needs an overall or bearing-screening meter; diagnosis needs an FFT analyzer or data collector; occupational compliance needs an ISO 8041 human-vibration meter. These are different instrument classes, not different price tiers of one product.
2. Pick the measurement quantity and band: velocity RMS at 10 to 1,000 Hz for general severity against ISO 10816 and ISO 20816, displacement for slow shafts and fluid-film bearings, acceleration to 5 to 20 kHz for rolling-bearing and gear-mesh faults. Confirm the meter supports the detector (RMS, peak, peak-to-peak) you will alarm on.
3. Select the sensor and sensitivity: 100 mV/g IEPE for general machinery, 500 mV/g for low-speed and low-amplitude machines, 10 mV/g for shock, charge-mode for high temperature, eddy-current proximity probes for sleeve-bearing shaft motion. Verify the mounted resonance sits well above the highest frequency of interest.
4. Choose the mounting method: stud or adhesive for bearing and gear diagnostics that need a wide band, flat magnet for repeatable route trending, probe tip only for quick overall-velocity checks. Standardize mounting locations so trends stay comparable over years.
5. Set severity criteria: identify machine group and foundation type, then apply the correct ISO 10816 or ISO 20816 zone limits, or derive baseline-plus-percentage alarms when the standard does not cover the machine. Document the zone boundaries used so different technicians read the same number the same way.
6. Verify standards and calibration: ISO 2954 conformance for broadband severity instruments, ISO 16063 traceable sensor calibration, ISO 8041-1 pattern evaluation for human-vibration meters. Plan a 12 to 24 month recalibration interval and a portable shaker calibrator (159.2 Hz, 1 g RMS) for in-situ checks before each survey.
7. Total cost of ownership: instrument price plus sensors and cables, software and database subscription, training, calibration, and battery and spare-part availability. A cheap meter that no operator trusts, or one whose data never reaches a trend database, costs more in missed failures than the price difference to a credible instrument.

One dimension is easy to overlook at purchase but decisive after a few years: serviceability and ecosystem. Local calibration service, spare accelerometers and cables, firmware and software updates, and integration with the plant's CMMS or condition-monitoring platform determine whether the meter is still useful in year five. PCB Piezotronics, Wilcoxon, and Hansford supply the IEPE accelerometers most meters expect; Fluke and SKF cover handheld screening; SKF Microlog and Emerson CSI 2140 cover route-based diagnosis; Svantek and Bruel and Kjaer cover human-vibration compliance. Choosing within an established ecosystem keeps the instrument calibrated, supported, and connected to the rest of the maintenance program.