A vibration signal from a running machine is a rich, messy waveform full of frequencies, but you cannot alarm on a whole waveform. The overall vibration level distills that entire signal into one number: how much the machine is vibrating overall, across a defined frequency band. It is the value a monitoring system trends day to day and alarms on when it climbs, the first thing a reliability program looks at, and the number that stands behind published severity charts. This guide explains what the overall level is, why it is the workhorse of machine-health screening, how it differs from spectral detail, and how to set it up as a simple scalar tag in a cloud monitoring platform.
Overall vibration level in one line: An overall vibration level is a single broadband number that represents the total amount a machine is vibrating over a defined frequency band, most often expressed as RMS velocity in millimeters per second across roughly ten hertz to one kilohertz. It is calculated by combining the energy of all the frequencies in that band into one scalar value that a monitoring system can trend and alarm on. Because it collapses the whole signal into one figure, it is ideal for screening machine health and setting alarm thresholds, but it cannot by itself tell you which fault is present, which is what spectral analysis provides.
The overall vibration level is a summary statistic of the vibration signal. The instrument samples the vibration over a short window, and rather than reporting the full spectrum, it computes a single value that reflects the combined magnitude of everything happening in the chosen band. The usual choice is the root-mean-square value, or RMS, because RMS reflects the energy content of the vibration and correlates well with how damaging the vibration is, which is why severity standards are written in RMS terms. The result is one number that rises when the machine as a whole vibrates harder and falls when it settles down.
The band over which the overall is taken matters, and the common band for general machine health is roughly ten hertz to one kilohertz. This band captures the frequencies where the dominant machine faults live, such as unbalance, misalignment, looseness, and the running-speed family, while excluding both very low frequencies that are hard to measure cleanly and very high frequencies that belong to specialized bearing detection. Measuring the overall in velocity units across this band is the convention that the widely used severity guidelines are built on, so an overall taken this way can be compared directly against those charts.
Because it is a single scalar, the overall level is easy to store, trend, and act on. Every measurement produces one comparable number, so a plot of overall level over weeks or months tells at a glance whether a machine is stable, slowly deteriorating, or suddenly worse. That simplicity is the whole appeal. It reduces a complex, continuously changing waveform to a value that a technician, an alarm system, and a manager can all understand without any spectral training.
The overall level is the workhorse because it answers the first and most important question cheaply: is this machine getting worse. Most reliability programs cannot afford to analyze a full spectrum from every machine every day, but they can afford to trend one number, and that one number reliably flags a machine that needs a closer look. A stable overall means the machine is behaving; a rising overall means something is developing. This screening role, catching the machines that deserve attention and leaving the healthy ones alone, is where the overall earns its keep.
It is also the natural quantity to alarm on. Severity guidelines assign ranges of overall velocity to broad condition zones, from good to acceptable to unsatisfactory to unacceptable, so an operator can set warning and danger thresholds directly against a machine's overall level without needing to interpret frequencies. When the overall crosses the warning threshold, the machine is flagged for investigation; when it crosses the danger threshold, it may warrant shutting the machine down. That direct mapping from a single number to an action is why the overall drives the alarm layer of nearly every vibration program.
The trade-off is that the overall tells you that something is wrong but not what. Because it sums all frequencies together, a rise could come from unbalance, misalignment, a bearing, looseness, or resonance, and the overall alone cannot separate them. That is by design: the overall is a screening and alarming tool, and once it flags a machine, spectral analysis takes over to identify the specific fault. The two work as a pair, the overall for continuous watching and the spectrum for diagnosis when the overall says to look closer.
In a cloud monitoring platform such as Merobix, the overall vibration level is naturally represented as a scalar tag, a single value that updates on a schedule just like a pressure or temperature reading. The sensor or vibration transmitter computes the overall on the machine and reports one number, which the platform stores as a time series. This makes vibration monitoring look and behave like the rest of the instrumentation on the platform, so an operator does not need special vibration tooling to keep an eye on machine health alongside everything else at the site.
Setting it up is a matter of defining the tag, its units, and its alarm thresholds. The tag is configured in velocity units to match the severity guidelines, its band is set to the standard health-screening range, and warning and danger setpoints are entered based on the machine class and the applicable severity zones. Once those thresholds are in place, the platform trends the value continuously and raises an alarm when it crosses a setpoint, exactly as it would for any other process measurement, so a developing machine problem produces a notification without anyone watching the trend by hand.
The strength of handling the overall as an ordinary scalar tag is that it brings machine condition into the same remote picture as the process. An operator responsible for unmanned sites can see the pump's overall vibration next to its flow, pressure, and motor current, trend them together, and correlate a rise in vibration with a change in operating conditions. When the overall alarms, the site can be prioritized for a route measurement or a spectral pull to diagnose the fault, so the cheap continuous number on the platform decides where the expensive detailed analysis is spent. That is the overall doing its screening job at fleet scale.
For general machine health it is most often measured in RMS velocity, in millimeters per second, taken over a band of roughly ten hertz to one kilohertz. Velocity is used because the widely accepted severity guidelines are written in velocity RMS, so an overall measured this way can be compared directly against those charts. Vibration can also be expressed in displacement or acceleration for particular purposes, but velocity is the standard for the broadband health-screening overall.
The overall level is one number that sums the energy of all frequencies in a band, so it tells you how much the machine is vibrating overall but not which frequencies are responsible. A spectrum breaks the same signal into its individual frequencies, showing where the vibration is coming from and therefore which fault is present. The overall is used for cheap continuous screening and alarming, and the spectrum is used for diagnosis once the overall flags a machine. They work together rather than replacing each other.
You set warning and danger thresholds based on the machine's class and the applicable severity guidelines, which map ranges of overall velocity to condition zones from good through unacceptable. The warning threshold flags the machine for investigation, and the danger threshold indicates the vibration is high enough to consider shutting the machine down. In a monitoring platform the overall is configured as a scalar tag in velocity units with those setpoints entered, and the platform alarms automatically when the value crosses them.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.