RF and Microwave Calibration Mistakes and Myths That Put Compliance at Risk

The most common RF and microwave calibration mistakes are trusting a plausible-looking display as proof of accuracy, neglecting connector and cable condition, using the wrong calibration standards for the connector interface, treating "calibrated" and "accredited" as interchangeable, and skipping calibration on equipment that "seems fine." Each is a genuine and recurring gap in RF measurement programmes, and each carries real compliance risk, because RF error is rarely visible until something downstream fails.

Myth 1: If the reading looks normal, the instrument is accurate

This is the single most consequential myth in RF calibration. A spectrum analyser with an amplitude offset still shows a clean, readable trace. A signal generator with a frequency offset still outputs a stable-looking signal. Neither fault announces itself visually. Only comparison against a traceable reference standard reveals it. Relying on "it looks fine" instead of a calibration due date is how out-of-tolerance RF equipment stays in service for months.

Myth 4: "Calibrated" automatically means "accredited"

As with every other measurement discipline, a certificate that says "calibrated" is not automatically an accredited ISO/IEC 17025 certificate. For RF/microwave equipment used in compliance testing (EMC/EMI, type approval, defence or aerospace acceptance), an auditor or regulator will look for accredited (SAC-SINGLAS in Singapore) calibration with stated measurement uncertainty at the relevant frequencies, not a generic in-house or manufacturer check.

15 further mistakes worth checking your programme against

• Neglecting connector and cable condition. Worn, dirty or damaged RF connectors introduce error that no instrument calibration can compensate for, since it's downstream of the instrument's internal reference. Precision connectors have a finite mechanical mating cycle life; used far beyond it, they're a realistic source of variation a technician might otherwise blame on the instrument.
• Using generic calibration standards for the wrong connector type. Calibration kits are connector-specific (N-type, SMA, 3.5 mm each have distinct characteristics). A mismatched or worn kit introduces systematic error, particularly for VNA calibrations where the standards define the measurement reference plane.
• Ignoring frequency-band-specific behaviour. An instrument can be within tolerance at one part of its range and drift at another. Interpreting a certificate as if performance were uniform across the whole range can miss a real problem in the band you actually use.
• Letting compliance-critical equipment run past its due date. Because RF equipment often underpins regulatory or contractual compliance work, a lapsed calibration means every result since the due date becomes questionable — yet because the instrument "still works," it's exactly the equipment most likely to be overlooked without a rigorously enforced recall system.
• Overlooking calibration-kit and cable calibration status. Calibration standards degrade with use like any other precision component, and an aging kit silently introduces error into every measurement, even on instruments that are themselves perfectly calibrated. Cable insertion loss and phase characteristics also change with flexing and age.
• Assuming a repair restores calibration. A repair touching any part of the signal path (a replaced attenuator, a swapped front-end module) can change performance characteristics. Recalibrate after any repair rather than trusting "it powers on and looks normal."
• Treating a self-calibration routine as equivalent to external calibration. Built-in self-alignment routines correct certain internal drift sources but cannot detect errors in the instrument's own fundamental reference — a useful complement to, not a replacement for, periodic external calibration.
• Ignoring cable and adapter loss in the measurement budget. Cables and adapters between the instrument and device under test introduce their own insertion loss and mismatch, uncorrected unless the calibration reference plane was established at the actual connection point. This error grows with cable length and frequency.
• Not verifying accessories and antennas used with EMC/field equipment. An antenna or preamplifier has its own calibration factor that must be applied for a correct reading — a common gap when equipment is shared across teams and less visible than the receiver it's connected to.
• Trusting a single reference measurement instead of a full multi-point calibration. RF performance is often not uniform across amplitude range either; an instrument accurate at mid-range power can behave differently near the top or bottom of its range. A single-point check is a useful interim tool, not a substitute for full calibration.
• Not accounting for temperature effects. Oscillator stability, amplifier gain and cable characteristics shift measurably with temperature, particularly outdoors. Equipment calibrated in a controlled lab but deployed in Singapore's hot, humid outdoor environment carries a real, if smaller, error source.
• Underestimating cable flexing as an error source. A cable bent or moved between calibration and measurement can change its electrical length and loss slightly. Keep cable routing as consistent as possible between the two, particularly with VNAs at higher frequencies.
• Assuming a higher-specification instrument needs a less rigorous programme. All RF/microwave instruments drift with time, temperature and use regardless of price point; a high-end analyser left uncalibrated for years is not more trustworthy than a budget one, it simply started from a better baseline.
• Treating calibration as a one-off compliance task rather than an ongoing programme. Reactive calibration (scrambling ahead of an audit) produces rushed turnaround at premium cost, equipment discovered significantly out of tolerance, and gaps in as-found data. A steady, proactively scheduled programme consistently outperforms an intensive but reactive one.
• Assuming institutional knowledge survives staff turnover. A programme depending on one person's memory of which instruments need extra care is fragile. Documented procedures, interval rationales and equipment-specific notes protect continuity that memory does not.

The highest-leverage fixes remain the fundamentals: trust the calibration due date over a plausible-looking display, maintain connectors and cables as part of the measurement chain, match calibration standards to the actual connector interface, confirm accredited scope for compliance-critical work, and recalibrate after any repair rather than assuming it restored a known-good state.