Vibration analysis has a reputation problem.
Not because it doesn't work. It does. When applied correctly, on the right equipment, with the right context behind it, vibration analysis is one of the most powerful tools in a predictive maintenance program.
The problem is that it gets applied everywhere, on everything, by people who were handed a data collector and told to "do vibration" without being told what vibration can and can't see.
The result is a program that generates a lot of numbers, produces a lot of confidence, and occasionally misses the failure it was supposed to catch — which is a uniquely expensive kind of wrong.
If your motor PM program treats vibration as the answer to every question, this post is about the questions it can't answer. If you want the broader picture of where vibration fits inside a complete motor PM strategy, start here.
What Vibration Analysis Actually Does
Before covering where it fails, it's worth being clear about where it works.
Vibration analysis measures the mechanical motion of a machine — the amplitude, frequency, and phase of oscillations at specific points on the equipment. When a bearing begins to wear, when a rotor becomes unbalanced, when misalignment develops between a motor and its driven load, the vibration signature changes in characteristic ways that an experienced analyst can identify and trend over time.
Done well, route-based vibration collection on a monthly or quarterly schedule can detect developing bearing faults weeks or months before failure. It can catch imbalance conditions before they accelerate bearing wear. It can identify misalignment before it works its way through the coupling and into the driven equipment. It can find looseness conditions that generate impact events with every revolution.
These are real, valuable findings. Vibration earns its place in a serious PM program for exactly these reasons.
The mistake is assuming it earns a place everywhere, for everything, unconditionally.
When Vibration Works Well
Vibration analysis performs best under a specific set of conditions that don't always exist in real plants.
Rotating equipment with rolling element bearings is where vibration delivers most reliably. AC motors, pumps, fans, gearboxes, and compressors with consistent operating speeds and loads give vibration analysis the stable baseline it needs to detect change meaningfully. The bearing failure modes that vibration catches earliest are exactly the ones that produce characteristic frequency signatures as rolling elements pass over developing defects on raceways and races.
Consistent operating conditions matter enormously. A motor running at the same speed, same load, and same temperature every time data is collected gives you a clean baseline to trend against. Deviations from that baseline mean something. The analysis has context.
High criticality equipment justifies the investment in route-based collection because the cost of the program is small relative to the cost of an unplanned failure. A 200 horsepower motor driving a critical production process is a reasonable vibration candidate. A quarter horsepower conveyor motor driving a non-critical transfer conveyor probably isn't.
When Vibration Misses the Failure Entirely
This is the part vibration programs don't advertise.
Electrical failures are largely invisible to vibration analysis until they're advanced enough to produce mechanical symptoms. Insulation breakdown, phase imbalance, and winding faults develop without producing meaningful changes in vibration signature. These failure modes require entirely different tools to detect. A motor with progressive insulation degradation will produce clean vibration data right up until the winding fails and the rotor locks up — at which point vibration analysis has nothing useful to add to the conversation.
Bearing electrical erosion from VFD-driven motors is another blind spot. EDM damage accumulates on bearing raceways and produces a frosted, pitted surface that eventually causes bearing failure. In the early and middle stages of EDM damage, the vibration signature may show nothing unusual. The bearing is degrading. The plots look fine. The program generates confidence that isn't earned.
Slow speed equipment is notoriously difficult to assess with standard vibration analysis. Motors and driven equipment operating below approximately 120 RPM produce vibration amplitudes so low that standard data collectors struggle to distinguish developing faults from background noise. Specialized low-frequency analysis techniques exist but aren't part of most route-based programs.
Variable speed equipment presents a different challenge. A motor running at different speeds each time data is collected makes trending nearly impossible without speed normalization. Raw vibration data collected at inconsistent speeds tells you very little about whether the machine is getting better or worse.
Common Mistakes in Route-Based Vibration Programs
Most vibration programs that fail don't fail because vibration analysis doesn't work.
They fail because the program was built and operated in ways that guarantee it won't.
Collecting data without baselines is the most common and most damaging mistake. A single vibration reading has almost no diagnostic value without something to compare it to. The first round of data collection on any machine establishes the baseline — everything after that is about detecting change from that baseline. Programs that start collecting data without establishing clean baselines under known good conditions are generating numbers without context. Numbers without context are not analysis. They're just data storage.
Infrequent collection intervals defeat the purpose of trending. Route-based programs that collect data quarterly on fast-moving, high-load equipment may miss the entire development window of a bearing fault. Bearing failures on heavily loaded motors can progress from detectable to catastrophic in weeks. A quarterly route that happens to fall after the detectable window and before the catastrophic failure contributes nothing.
Inconsistent measurement locations corrupt data over time. Vibration readings are sensitive to exactly where on the machine the sensor is placed and how it's attached. Programs that rely on technicians to find approximately the right spot each time introduce measurement variability that looks like machine variability. Marked measurement points and consistent attachment methods — magnetic mounts at minimum, stud-mounted accelerometers for critical machines — are not optional details.
Treating alarm thresholds as pass/fail without trending is how programs generate false confidence. A machine running at 0.3 inches per second velocity when the alarm is set at 0.5 is not necessarily a healthy machine. It may be a machine that was at 0.1 six months ago and is trending upward. The trend is the finding. The threshold is just a backstop.
Vibration Is a Tool, Not a Program
The most expensive mistake a maintenance program makes with vibration analysis is treating it as a complete predictive maintenance strategy rather than one tool within one.
Route-based vibration collection on appropriate equipment, with clean baselines, consistent measurement locations, adequate collection frequency, and an analyst who understands what the data can and can't see — that's a valuable program component.
Route-based vibration collection deployed everywhere, on everything, without baselines, by technicians who were handed a data collector and told to collect — that's a liability dressed up as a program.
The PM program underneath the vibration data still matters. Vibration tells you something is changing. A structured PM program tells you what to do about it before the bearing makes the decision for you.
Used together, they're significantly more capable than either one alone.
Where Vibration Fits in Your Motor PM Program
Route-based vibration collection belongs in your motor PM program under specific conditions — critical equipment, rotating machinery with rolling element bearings, consistent operating speeds, and collection intervals matched to the failure development window of the equipment.
It does not belong everywhere by default, and it does not replace the electrical condition monitoring that most motor PM programs are already missing.
These task lists are built around the full range of motor failure modes — mechanical and electrical — so your program isn't blind to the failures vibration can't see:
- AC Motor preventive maintenance tasks
- DC Motor preventive maintenance tasks
- VFD preventive maintenance tasks
Vibration analysis is honest about what it finds.
The mistake is assuming it's looking at everything.