Machine balancing is the process of detecting and correcting uneven mass distribution in rotating parts so they spin without excessive vibration. Every rotating component, from a small electric motor to a massive industrial turbine, generates centrifugal force that pushes each particle of mass outward from the axis of rotation. When that mass is evenly distributed, the forces cancel out and the part spins smoothly. When it isn’t, the heavy side pulls the entire rotor off-center with every revolution, producing vibration that accelerates wear on bearings, seals, and surrounding equipment.
Why Rotating Parts Become Unbalanced
A perfectly balanced rotor is more of an ideal than a reality. Even brand-new parts carry some degree of imbalance due to the stack-up of manufacturing tolerances. Small imperfections in casting, such as internal voids, porosity, or air pockets trapped in the metal, shift the center of mass away from the geometric center. Machining errors, where a bore isn’t perfectly centered or a surface is cut slightly off-axis, add to the problem. Poor assembly, where components like fan blades or impellers are mounted with slight eccentricities, contributes as well.
Imbalance also develops over time during operation. Temperature changes cause thermal distortion, which shifts mass distribution, and this effect can compound itself: the resulting vibration generates more heat, which causes more distortion. Rotational stresses gradually deform components. Corrosion, erosion, and material buildup (like dust accumulating unevenly on fan blades) all shift the balance point. This is why balancing isn’t a one-time event for most industrial machinery.
Static vs. Dynamic Balancing
The two fundamental approaches to balancing differ in how many correction planes they address, and the choice between them depends largely on the shape of the rotor.
Static balancing (also called single-plane balancing) resolves all the unbalance forces into one plane and corrects them by adding or removing mass in that single plane. It’s called “static” because the imbalance is detectable even when the part isn’t spinning: a statically unbalanced disc placed on frictionless bearings would roll until the heavy spot settled at the bottom. This approach works well for disc-shaped parts where the diameter is large relative to the width. The common rule of thumb is that if a rotor’s diameter-to-length ratio is greater than about 2:1, single-plane balancing is sufficient.
Dynamic balancing addresses both the primary forces from unbalanced mass and the secondary force couples that arise when imbalance exists at different points along the rotor’s length. These couples only reveal themselves when the part is spinning, which is why the process is called “dynamic.” Corrections are made in two or more planes, typically near each support bearing. After dynamic balancing, the rotor is balanced in both static and dynamic conditions. Any part with appreciable width, like a multi-stage pump rotor or a long drive shaft, requires this two-plane (or multi-plane) approach because imbalance at opposite ends can create a wobbling motion that single-plane correction can’t fix. A rotor can be perfectly balanced statically yet still vibrate dynamically if equal heavy spots exist on opposite sides at each end, creating a rocking couple.
How Imbalance Is Measured
Unbalance is expressed as a weight at a given distance from the center of rotation. If you have 5 grams of excess mass sitting 100 millimeters from the shaft center, the unbalance is 500 gram-millimeters (g-mm). The most common unit combinations are gram-millimeters, gram-centimeters, ounce-inches (oz-in), and kilogram-meters. For couple unbalance, where opposing forces create a twisting effect along the rotor’s length, the units add another distance dimension: g-mm² or oz-in².
The sensors used to detect this imbalance are typically piezoelectric accelerometers, small devices that convert mechanical vibration into electrical signals. They’re mounted on or near the bearing housings where vibration is most easily measured. For situations where direct contact with the vibrating surface isn’t practical, non-contact eddy-current sensors (proximity probes) can measure shaft displacement without touching the rotor. Both types feed vibration amplitude and phase data to the balancing instrument, which calculates how much weight to add or remove and where to place it.
Balancing Machine Types
Dedicated balancing machines fall into two categories based on their bearing design, and each measures unbalance differently.
Hard-bearing machines have rigid supports that barely move during operation. They measure the centrifugal force generated by unbalance directly, using vibration sensors mounted on the machine’s structure to detect and plot force magnitude and direction while the rotor spins. Because they measure force rather than displacement, hard-bearing machines don’t need a calibration run for each new rotor type. The rotor sits on rollers directly on its bearing mounting surfaces.
Soft-bearing machines use flexible supports (ladder bearings) that allow each end of the rotor to oscillate freely. They measure the amplitude and direction of that oscillation, typically using vibration pickups combined with a stroboscope to identify the angular position of the heavy spot. The rotor is placed on stands with its own anti-friction bearings installed. Older soft-bearing machines measure dynamic runout rather than residual unbalance, which is a less direct measurement. Some rotors are designed in a way that prevents them from being supported on rollers, making soft-bearing machines the only option.
Field Balancing in Place
Not every rotor can be pulled from its housing and placed on a balancing machine. Field balancing, also called in-situ balancing, corrects imbalance while the machine remains installed in its operating position. This is common for large equipment like industrial fans, generators, and turbines where disassembly would mean days of downtime.
The most basic field method is single-plane balancing using a graphical approach, and it follows a logical sequence. First, the machine runs at its normal operating speed while a sensor records the vibration amplitude and phase angle. This initial reading, the “original” vector, gets plotted on polar graph paper. Next, a trial weight of known mass is attached to the rotor at a known position, and the machine runs again at the same speed. The new vibration reading, called the “original plus trial” vector, is plotted on the same graph.
Drawing a line between the tips of these two vectors reveals the “effect” vector, which shows how much and in what direction the trial weight changed the vibration. From this, the technician can calculate two things: the angle the trial weight needs to be rotated so its effect directly opposes the original imbalance (180 degrees from the original vector), and the final correction weight, scaled up or down by the ratio of the original vibration to the effect vector’s length. If the original vibration is twice as long as the effect vector, the correction weight needs to be twice the trial weight.
After applying the calculated correction, the machine runs again to verify the result against acceptable vibration limits. If vibration is still too high, the process repeats using the first correction as a new reference point. Two-plane and multi-plane field balancing follow the same logic but require measurements and corrections at multiple locations along the rotor, sometimes at multiple operating speeds for rotors that pass through critical speeds during startup.
Why Balancing Matters for Equipment Life
Imbalance is one of four common mechanical faults in rotating machinery, alongside looseness, misalignment, and bearing wear. What makes it particularly destructive is that centrifugal force increases with the square of rotational speed. A small imbalance that causes minor vibration at low speed can produce punishing forces at full operating speed. The vibration transfers through bearings into housings, foundations, and connected equipment, shortening bearing life, loosening fasteners, fatiguing structural components, and generating noise.
Precision balancing reduces these forces to within acceptable tolerances, which vary by application. A grinding spindle running at tens of thousands of RPM demands far tighter balance than a slow-turning concrete mixer. Industry standards define permissible residual unbalance based on rotor mass and operating speed, giving maintenance teams a clear target rather than chasing perfection. The goal isn’t zero imbalance, which is physically impossible, but imbalance low enough that its effects fall below the threshold where they cause damage or interfere with the machine’s function.