A hard-bearing balancing machine measures centrifugal force, not motion, by way of force transducers mounted under rigid pedestals that lock the rotor journals in place [S1][S2]. Because the supports are stiff, the rotor is fully constrained, and the unbalance reaction that the shaft delivers into each pedestal is captured as a force vector with magnitude and angular position, then divided by the calibration constants to produce the residual unbalance [S2][S4].
That single-run behaviour contrasts with the soft-bearing approach, in which the rotor is allowed to swing on a flexible suspension and the resulting displacement is what the instrument reads, typically requiring a trial weight and several runs per rotor [S1][S2]. For shop-floor throughput on repetitive rotor families, hard-bearing stands became the default by the 1990s once permanently calibrated electronics replaced the per-rotor trial-weight workflow [S3].
How the Force Signal Is Produced at the Pedestal
The journal sits in a fixed bearing block coupled to a force sensor, so any unbalance reaction from the spinning rotor is transmitted straight into the transducer instead of displacing the support [S2][S5]. Force is recovered as F = m·r·ω², where m is the unbalance mass, r is its radius, and ω is rotational speed in rad/s; this is the same relationship used to convert the measured reaction back into a residual unbalance in g·mm [S3].
Because the signal scales with ω², doubling the test speed increases the measurable reaction force by a factor of 4, which is why hard-bearing accuracy improves as run speed rises and degrades on very slow rotors [S1]. A 30 RPM steam turbine example cited for soft-bearing stands on rail tracks would yield roughly 0.04 percent of the unbalance force seen at 1500 RPM on a hard-bearing stand, which is one practical reason slow, heavy rotors often stay on soft-bearing machines [S1].
Permanent Calibration Through the "abc" Geometry
Hard-bearing machines are permanently calibrated using a test rotor and a known trial weight in three runs: a reference run with no test mass, a run with the test weight on the left correction plane, and a run with the test weight on the right correction plane [S3]. The instrument then stores influence coefficients that map the measured support forces to unbalance at each correction plane, indexed by the rotor's "abc" dimensions (the axial distance from each support to the two correction planes and between supports) [S3].
Setup for a production rotor only requires entering those geometric distances and the rotor weight, with no trial weight attached to the part itself [S3]. This geometry-driven approach is what makes a single production run sufficient in most cases, although overhung rotors, correction planes too close together, or unusually large diameters can still demand an extra check run [S3].
Unbalance Reduction Ratio and Why It Matters
A 95% unbalance reduction ratio (URR) means the first run drops the initial unbalance to 5% of its starting value, so any rotor starting at 20x its tolerance is theoretically acceptable in one run, while an 80x initial unbalance would need a second pass [S3]. A machine rated at only 90% URR reduces unbalance to one tenth per run, so the same 20x starting rotor would still need a follow-up run to reach tolerance [S3].
This is why procurement specs that quote only "95% URR" can mislead buyers on hard, narrow-tolerance rotors, and why experienced operators consolidate multiple correction weights into a single angular sector rather than scattering small corrections around the rim, a workmanship issue still common with novice operators [S3]. Comparison passages across bearing platforms, including the role of ball bearing rollers in the work-support stands themselves, make it easier to see why a 95% URR figure is the right baseline to demand on a modern shop-floor hard-bearing machine.
Where the Direct-Force Method Beats the Soft-Bearing Method
Hard-bearing machines are centrifugal-force measuring systems that can balance rotors of virtually any type without per-setup calibration, provided the geometry is entered correctly [S3]. They also handle correction planes close to the support bearings more cleanly than the displacement method, because the force vectors are read at the supports themselves and resolved mathematically to the correction radius [S3].
The soft-bearing approach still wins on three practical fronts: rotors with very low achievable test speeds (such as large steam turbines balanced around 30 RPM), on-site or portable work where the foundation is not a controlled mass, and low-volume, high-mix maintenance shops that cannot justify a permanently calibrated stand [S1][S2]. A force gauge reading the same kind of direct-load signal on a shop bench is the handheld analog of what a hard-bearing pedestal does at production scale, which is why both work on the same F = m·r·ω² principle once rotor mass and speed are known.
Sensing Limits, Suspension Stiffness, and Common Failure Modes
Because the supports are rigid, the smallest measurable unbalance is set by the transducer noise floor and the residual compliance of the pedestal, not by the resonance of a soft suspension [S2][S5]. On the flip side, any drift in pedestal stiffness, mounting bolt preload, or sensor zero will show up directly as a force-reading error, which is why hard-bearing machines are much less tolerant of being moved than soft-bearing stands and often need re-verification after a relocation [S2].
Roller-bearing work supports introduce their own asynchronous motion, and that unwanted vibration can mask or distort the force signal if the rotor is run too slowly or if the support rollers are worn; air-bearing work supports are commonly specified when this floor becomes the limiting factor [S5]. Process engineers comparing machine classes should also note that the roller bearing stack itself contributes measurable error to the balance result, an effect that grows as the required balance grade tightens.
Single-Run Procedure and Operator Pitfalls
With geometry and rotor mass loaded into the controller, a single run delivers unbalance magnitude and angle for both correction planes, after which the operator adds or removes weight in the indicated sectors [S1][S3]. The classic novice error is to apply many small scattered correction weights, which improves the displayed number but produces a rotor that runs roughly in service because the residual is fragmented; consolidating the same total mass into a single angular sector gives better real-world behaviour at the same displayed URR [S3].
For a deeper look at how the heavy-spot angle is actually computed from these force signals, the related walkthrough on how a dynamic balancing machine finds the heavy spot angle is a useful complement to this force-measurement view. Track next: ISO 1940-1 balance grade updates for fan and pump rotors in the 2026 revision cycle, and any movement by the major balancing-machine OEMs toward in-line correction-plane verification on hard-bearing stands.