A failing rolling-element bearing does not announce itself in the ordinary velocity spectrum until it is already well along. Long before then, each pass over a tiny defect produces a faint, high-frequency ring, too weak and too high in frequency to show up next to the machine's dominant low-frequency vibration. Acceleration envelope detection is the signal-processing technique that pulls those weak, repetitive impacts out of the high-frequency noise and makes them measurable. It is why enveloping typically gives the earliest warning of a bearing or gear fault in predictive maintenance.
Envelope detection in one line: Acceleration envelope detection, also called enveloping or demodulation, is a technique that band-pass filters high-frequency acceleration around the resonant ringing caused by bearing or gear defects, then extracts the amplitude envelope of that signal. The result reveals the low-frequency repetition rate of the impacts, exposing early-stage bearing and gear faults that are invisible in the standard velocity spectrum.
When a rolling element passes over a small defect, a pit or spall on a raceway, it delivers a sharp mechanical impact. That impact excites high-frequency resonances in the bearing and surrounding structure, which ring briefly and die away, only to be struck again on the next pass. The energy in these impacts is small and sits at high frequencies, far above the machine's running-speed vibration, so in the ordinary velocity spectrum it is buried under the much larger low-frequency peaks. The information is there, but it is drowned out.
Envelope detection recovers it in two steps. First, a band-pass filter isolates a high-frequency window around the resonance the impacts excite, deliberately discarding the loud low-frequency vibration and keeping only the region where the faint ringing lives. Second, amplitude demodulation extracts the envelope, the outline that traces how the amplitude of that filtered signal rises and falls. Because the ringing is triggered once per impact, the envelope rises and falls at the impact repetition rate. Taking a spectrum of the envelope then reveals a clear peak at that repetition frequency, where before there was only broadband noise.
The elegance is that the technique converts a high-frequency, low-energy phenomenon into a low-frequency, clearly measurable one. The actual ringing might be at several kilohertz, but its repetition rate, the thing that identifies the fault, is a low frequency tied to how often the defect is struck. Enveloping strips away the carrier and leaves that repetition rate exposed. Variants of the method exist under names such as demodulation, envelope analysis, and PeakVue, and while the implementations differ, they share the same core idea of isolating and demodulating high-frequency impact energy.
The reason enveloping leads the velocity spectrum comes down to when each becomes sensitive. In the earliest stage of a bearing failure, the defect is a microscopic pit that produces sharp, high-frequency impacts but adds almost nothing to the overall vibration level. The velocity spectrum, weighted toward the mid-frequency range and dominated by running-speed vibration, simply does not register that tiny, high-frequency contribution. The envelope method, tuned precisely to the high-frequency band where the impacts ring, sees them clearly while they are still small.
This gives envelope detection a long lead time. It can flag a developing bearing or gear defect stages before the fault grows enough to raise the overall velocity or appear as a defect-frequency peak in the standard spectrum. In a typical failure progression, enveloping detects the impacting first, the velocity spectrum picks up defect frequencies as the damage spreads, and only in the late stages does the overall vibration climb enough to trip broadband alarms. By then the bearing may be close to failure; the value of enveloping is buying the weeks or months of warning that let maintenance be planned rather than reactive.
The trade-off is that enveloping is a targeted tool, not a general severity measure. Because it deliberately discards the low-frequency vibration and focuses on a specific high-frequency band, its amplitude does not map onto standards like the overall-velocity zones, and it does not judge imbalance or misalignment. It is best understood as an early-warning specialist for the impacting faults, bearings and gears, that ring at high frequency, used alongside the velocity spectrum rather than in place of it. The two together cover both the early impacting stage and the later, broadband growth of a fault.
Because the envelope produces a clean, low-frequency repetition rate, it lends itself to automatic alarming just as the velocity spectrum does. A monitoring system can track the envelope amplitude, or the amplitude at the specific bearing defect frequencies within the envelope spectrum, and trend it over time. A rising envelope level is one of the earliest quantitative indicators that a bearing has begun to spall, and setting an alarm on that trend gives predictive maintenance its longest possible lead time before a machine reaches a damaging condition.
Bringing envelope measurements into a SCADA platform lets that early warning reach the people who plan the work, wherever they are. Merobix can trend the envelope value from each monitored bearing alongside the overall vibration and the process data, so an operator sees an envelope alarm rising well before the broadband vibration moves, and can schedule a bearing change on the next planned outage rather than after a failure. For fleets of pumps, fans, and gearboxes across remote and unmanned sites, this remote early warning is exactly what makes condition-based maintenance workable at scale.
The historized envelope trend is what separates a genuine developing fault from a transient. A one-off high reading can come from a passing event; a steadily climbing envelope level over successive measurements is the fingerprint of a defect growing. By keeping that history and establishing each bearing's normal envelope baseline, the platform can distinguish the two and raise a confident alarm on the true upward trend. Correlated with load and speed, the envelope history also helps confirm the fault is in the bearing and tied to operation, sharpening the decision of when to intervene.
A defect on a bearing raceway produces sharp high-frequency impacts each time a rolling element passes over it, and these impacts ring at frequencies far above the machine's normal vibration. Envelope detection band-pass filters that high-frequency ringing and demodulates its amplitude, exposing the low-frequency repetition rate of the impacts. Because it targets exactly the high-frequency band where early defects live, it detects them while they are still too small to affect the overall vibration.
In the earliest stage of failure a bearing defect is tiny and produces high-frequency impacts that add almost nothing to overall vibration, so the mid-frequency-weighted velocity spectrum does not register them. Enveloping is tuned to that specific high-frequency band, so it sees the impacts clearly while they are still small. This lets it flag a developing fault stages before defect frequencies appear in the velocity spectrum or the broadband vibration climbs.
No. Enveloping deliberately discards the low-frequency vibration to focus on high-frequency impact energy, so its amplitude does not map onto overall-severity standards like the ISO velocity zones, and it does not assess imbalance or misalignment. It is an early-warning tool specialized for the impacting faults of bearings and gears, used alongside the velocity spectrum rather than as a replacement for a general severity measure.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.