In mechanical strength test stands (resonant fatigue, modal analysis, hydraulic-shaker excitation), a spectrum analyzer measures the frequency-domain content of force, strain, and acceleration transducers, while a signal calibrator injects a known reference (often 1 kHz, 1 g, or a defined voltage) to verify the entire signal chain end-to-end.
The two instruments are not substitutes: per the Agilent amplitude-accuracy application note [S2], a spectrum analyzer's internal mixer compresses by only 0.01-0.06 dB at -10 dBm drive, whereas a signal calibrator's purpose is to deliver a precise, traceable stimulus that the analyzer or DAQ then quantifies. Agilent's primer [S1] makes the receiver side explicit, stating the instrument "can measure either the absolute value of a signal or the relative amplitude difference between any two signals" with scale calibration annotated on the graticule.
Defining the two instruments in a test-rig context
A spectrum analyzer is a swept or FFT-based receiver that displays amplitude versus frequency; in a mechanical strength lab, that means resolving the harmonic structure of a shaker table, the sidebands of a cracked shaft, or the resonance peaks of a coupon under cyclic load. Agilent's note [S2] sets the design target by noting that "a spectrum analyzer's RF and IF circuits will tend to exhibit some small (about 0.01 to 0.06 dB) amounts of compression at this level, so it is desirable to maintain the level to the mixers well below -10 dBm."
A signal calibrator (precision signal generator or transducer calibrator) is a stimulus source. Anritsu's guide [S4] classifies it separately as "an essential item of equipment for any communications test laboratory or workshop" and notes that "the cost of a signal generator largely depends on the additional" modulation and level-control capability. For mechanical work, the calibrator typically drives a shaker, an ICP-style accelerometer, or a load-cell conditioner with a known sine, random, or shock profile traceable to a national metrology institute.
Selection criteria for mechanical strength applications
Four parameters govern spectrum-analyzer setup, per Rohde & Schwarz's operations guide [S1]: center frequency, span, resolution bandwidth (RBW), and amplitude/reference level. For fatigue-test monitoring of a steel coupon at 1-2000 Hz, an RBW of 1-10 Hz is typical, while a shaker-control calibrator must hold total harmonic distortion below -80 dBc to avoid contaminating the device under test.
Material and environmental selection matters as much as electrical spec. LISUN's construction note [S6] states that "spectrum analyzer construction requires selection of materials that ensure RF performance stability, mechanical durability, and environmental protection across the intended operating conditions," and that "low-loss dielectric materials are essential for minimizing signal attenuation." Benchtop units used near a 50 kN hydraulic press need a chassis rated for continuous mechanical vibration, which handheld analyzers typically satisfy through integrated elastomeric mounts but some low-cost USB dongles do not.
Who should buy which (and who should buy both)
Standalone modal-analysis shops that only need to identify natural frequencies and damping ratios can justify a single mid-range spectrum analyzer (10 Hz to 26.5 GHz coverage) and rely on the shaker's own internal calibration. A calibration house that issues ISO/IEC 17025 accredited certificates, by contrast, needs a signal calibrator traceable to a national metrology institute, because the spectrum analyzer alone cannot prove the integrity of an unknown transducer [S5].
A full-scale structural-test laboratory running fatigue, proof-load, and acoustic-emission tests almost always needs both: the analyzer to characterize the response, the calibrator to verify the accelerometer, load cell, and pressure sensor chain before each shift. Skipping the calibrator step is one of the most common audit findings in ISO/IEC 17025 surveillance, per the calibration-services notes from Mchale [S5] and Applied Technical Services [S5]. Techmaster [S3] lists aerospace and defense, electronics manufacturing, medical devices, and academic R&D as the four core user communities, all of which demand routine, documented calibration intervals.
Criteria-based comparison of the two instrument classes
On four decision criteria the two diverge sharply. (1) Frequency range: a general-purpose spectrum analyzer typically covers 9 kHz to 26.5 GHz or higher, while a mechanical-strength signal calibrator rarely needs more than 0.1 Hz to 50 kHz. (2) Amplitude accuracy: Agilent [S2] targets the analyzer's mixer at well below -10 dBm to keep compression under 0.06 dB and sets the input attenuator reference at 10 dB; a calibrator, conversely, is specified at its output connector (often ±0.05% of setting) rather than at an internal node. (3) Calibration interval: Applied Technical Services [S5] argues that a 12-month cycle is the maximum for ISO/IEC 17025 work and that high-utilization labs should drop to 6 months. (4) Mechanical ruggedness: per LISUN [S6], benchtop analyzers typically meet commercial shock and vibration profiles, while handheld units often add IP54 sealing and 1 m drop tests for field modal surveys on bridges and towers.
Test-rig integration also differs. A modern servo-hydraulic frame is built around a servo motor and proportional valve; the spectrum analyzer is a passive observer, while the signal calibrator is part of the closed loop that the PLC cross-checks against the reference load cell. Buying one without understanding its place in that loop leads to wasted capital and rejected test reports.
Real use cases, failure modes, and limitations
Case 1 - wind-turbine blade modal survey: handheld spectrum analyzer plus impulse hammer with built-in load-cell calibrator; failure mode is aliasing of the high-frequency hammer pulse if RBW is set too wide [S1]. Case 2 - aircraft component fatigue test: benchtop spectrum analyzer with tracking generator feeding a shaker driven by a calibrator, monitored by pressure sensor arrays in the hydraulic supply; failure mode is analyzer overload (compression above 0.1 dB) when the shaker radiates switching harmonics into the measurement antenna [S2]. Case 3 - production-line proof load: a load-cell calibrator applies a known mV/V stimulus, no spectrum analyzer required; failure mode is relying on the indicator's internal reference instead of an external calibrator traceable to ISO/IEC 17025 [S5].
Both instrument classes share a documented limitation, captured verbatim by Tektronix [S1]: "in real life, however, the nature of signals is not known in advance, so having both instruments allows proper characterization of the signal." That statement is doubly true on a modern structural test stand, where variable-frequency drives emit broadband EMI that a strain gauge will pick up and a spectrum analyzer is the only way to separate from the mechanical response.
Standards, calibration, and sourcing
Calibration discipline follows ISO/IEC 17025 in nearly every commercial lab [S5]; aerospace and defense programs additionally impose their own test-equipment acceptance plans. The 0.01-0.06 dB compression figure quoted by Agilent [S2] is the de-facto acceptance threshold for amplitude accuracy in modern swept analyzers, while signal calibrators are typically specified against the same ISO/IEC 17025 framework at their output connector rather than at an internal mixer. The Mchale calibration guide [S5] flags three reasons to stay on schedule: measurement accuracy, compliance with standards (ISO/IEC 17025), and device longevity.
For procurement, the practical signal to track in 2026 is whether your calibration provider's scope of accreditation already lists your specific analyzer and calibrator models under ISO/IEC 17025 - if not, expect a 6-12 month onboarding delay before certificates become defensible in a customer audit. Second signal: as servo-hydraulic rigs add Ethernet-APL and IO-Link sensors, the spectrum analyzer's role is shifting from a periodic checker to a continuous on-line monitor, which will reshape calibration interval assumptions documented in current ISO/IEC 17025 scopes.