Measurements & Instruments

Measurement map

1. Vector network analyzer

A VNA measures complex traveling-wave ratios. For a two-port device,

Practical workflow

1. Choose frequency span, IF bandwidth, power, averaging and number of points.
2. Warm up cables and fixtures if precision matters.
3. Calibrate at the desired reference plane: SOLT, TRL, ECal or another suitable method.
4. Verify calibration with a known standard or thru.
5. Measure the DUT without changing cable geometry unnecessarily.
6. Save complex data, not just screenshots.

2. Signal generator

A modern vector signal generator is a calibrated RF transmitter used to create CW, modulated, pulsed, multicarrier and arbitrary waveforms. Its internal chain combines precision references, PLL/DDS synthesis, digital baseband, interpolation, DACs, vector or direct-RF conversion, filtering, ALC and switched attenuation.

See Modern Signal Generator — RF & Vector Signal Generation for the detailed architecture.

3. Spectrum analyzer

For a detailed explanation of modern FFT/real-time spectrum analyzers—including the Signal Hound BB60D superheterodyne/preselector architecture, SP145 low-IF architecture, and the much wider-band R&S FSW platform—see Modern Spectrum Analyzer Signal Processing.

Use a spectrum analyzer for carrier power, harmonics, spurious emissions, occupied bandwidth, phase-noise-related skirts and interference searches.

Key settings:

• RBW: frequency selectivity and noise bandwidth;
• VBW: post-detection smoothing;
• reference level / attenuation: overload protection versus sensitivity;
• detector: peak, RMS, sample, quasi-peak depending on task;
• sweep time: tied to span and filters.

A narrower RBW reduces displayed noise power approximately with $10\log_{10}B$, but also slows measurement and may hide fast events.

4. Vector signal analysis and EVM

A VSA demodulates a known modulation format. Error vector magnitude is conceptually

EVM is a compact metric that can include amplifier nonlinearity, IQ imbalance, phase noise, frequency error, noise, compression and channel effects. Interpretation depends strongly on the equalization and reference definitions used.

5. Oscilloscope

An oscilloscope connects electromagnetic systems to time-domain behavior: pulses, envelopes, beat notes, transients, switching, detector signals and TIA outputs.

Useful rules:

• analog bandwidth should exceed the highest spectral content you want to preserve;
• sample rate should comfortably exceed Nyquist and support desired timing resolution;
• averaging improves uncorrelated noise but can erase nonstationary behavior;
• a 10× probe or active probe changes circuit loading differently than a 50 Ω input.

For a nominally Gaussian system, a rough rise-time relation is

6. RF power measurement

Power meters are often more accurate than spectrum analyzers for absolute average power. Directional couplers allow forward/reflected measurements in high-power paths.

Mismatch uncertainty matters because source and sensor reflection coefficients interact. For pulsed/high-PAPR signals, ensure the sensor supports the waveform’s crest factor and bandwidth.

7. Noise figure

For the Y-factor method,

combined with the calibrated excess-noise ratio of the source. Noise-figure work is sensitive to source ENR calibration, connector loss, gain variation, image responses and analyzer noise floor.

8. Antenna patterns, gain and efficiency

Far-field range

A common criterion is

Measure amplitude and phase versus angle with a calibrated source/receive geometry.

Near-field scanning

Planar, cylindrical or spherical near-field measurements can be transformed computationally to the far field. This is invaluable when the required far-field distance is impractically large.

OTA systems

Modern wireless devices are often evaluated over the air because the antenna, enclosure, transceiver and beamforming algorithms behave as one coupled system.

9. EMC and compliance measurement

EMC measurements usually separate into:

• conducted emissions;
• radiated emissions;
• conducted immunity;
• radiated immunity;
• ESD / EFT / surge depending on product class;
• coexistence / desense / self-interference investigations.

Instrumentation can include LISNs, current probes, CDN/BCI fixtures, antennas, preamplifiers, spectrum/EMI receivers, RF power amplifiers, field probes and chambers.

A compliance result is not just a spectrum: detector type, bandwidth, distance, antenna factor, cable loss, preamplifier gain and chamber/site validation all matter.

10. Electric-field measurement

Possible methods include:

• calibrated dipole/monopole probes;
• isotropic E-field probes;
• electro-optic sensors;
• antenna-factor conversion;
• Rydberg-atom electrometry.

For an antenna-factor method,

under the calibration convention used. Probe perturbation, polarization, near-field structure and spatial averaging can dominate uncertainty.

11. Magnetic-field measurement

• Hall sensor: DC to moderate-frequency local field, compact but offset/temperature sensitive.
• Fluxgate: precise low-frequency vector field.
• Search coil: AC magnetic field; induced voltage scales with $d\Phi/dt$.
• Optically pumped magnetometer: atomic Larmor/Zeeman response, potentially extremely sensitive.
• NMR probe: accurate field magnitude in suitable ranges/materials.

12. Optical readout for atomic systems

A common chain is

Measure and record:

• detector responsivity and transimpedance gain;
• 3-dB bandwidth;
• optical power and saturation margin;
• laser linewidth and frequency reference;
• beam waist and overlap;
• scan linearity;
• RF source calibration and cable loss;
• lock-in/reference phase if demodulating.

13. Calibration hierarchy

A strong experiment separates instrument calibration from physical-model calibration.

Examples:

• VNA calibration moves the electrical reference plane.
• Power-sensor calibration establishes absolute RF power.
• Antenna calibration establishes gain/antenna factor.
• Atomic AT splitting can provide field calibration through a known transition dipole.
• A photodetector transfer function converts current to measured voltage.

14. Measurement uncertainty

For independent input quantities $x_i$, first-order uncertainty propagation is

In practice, distinguish:

• random repeatability;
• calibration uncertainty;
• systematic bias;
• drift;
• model uncertainty;
• spatial variation;
• bandwidth / filtering uncertainty.

15. A reusable measurement checklist

Inertial-navigation measurements

INS characterization connects sensor physics directly to precision measurement. Important observables include gyro/accelerometer bias, scale factor, axis misalignment, vibration sensitivity, turn-on repeatability, temperature coefficients, Allan deviation, timing latency, and error growth during GNSS outages.

See Modern Gyroscopes — How They Work for gyro-specific rate-table, bias, vibration, scale-factor, and Allan-deviation measurements, and Modern Inertial Navigation Systems for the complete INS context.

Arduino and laboratory control

Arduino-class microcontrollers are useful for low-level experimental control: reading sensors, generating triggers, switching relays, programming SPI/I²C devices, controlling motors, and collecting slow telemetry. They should be treated as real embedded hardware with voltage/current/timing constraints rather than as generic USB peripherals.

See Arduino Boards — What You Need to Know.

ADC and DAC characterization

Data converters require both static and spectral measurements. ADC testing commonly uses code-density/ramp methods for INL/DNL and low-distortion sine/FFT tests for SNR, SINAD, SFDR and ENOB. DAC testing includes code sweeps, settling/glitch measurements, and spectral analysis of dynamic outputs.

See ADC & DAC — Data Converters.

3GPP RF and conformance measurements

For cellular RF work, measurement conditions and limits should be traced to the applicable 3GPP requirement and conformance specification rather than treated as generic instrument settings. The key UE chain is typically 38.101-x → 38.521-x → 38.508-1, while gNB RF uses 38.104 → 38.141-1/-2.

See 3GPP Standards — What an RF & Wireless Engineer Should Know.

Hardware-interface bring-up

Board-level interface debug connects protocol behavior to physical measurements. PCIe, USB, Thunderbolt, HDMI, I²C, SPI and UART bring-up may require a combination of oscilloscope, logic analyzer, TDR/TDT, VNA, BERT and protocol-analyzer measurements.

See Hardware Interfaces & PCB/Flex Design for interface-specific signal, layout and debug considerations.

Radio-astronomy receiver measurements

Radio-astronomy hardware characterization includes receiver noise temperature/Y-factor, gain and bandpass stability, cryogenic LNA/mixer performance, LO phase noise, polarization isolation, ADC headroom and RFI linearity. See Radio Astronomy — RF Circuits & Receiver Systems.

Related pages

EMI/EMC · Signal Integrity & Power Integrity · High-Speed SerDes & IO · Analog Mixed-Signal IC Design · RF & Microwave · Antennas · Ground-State Magnetometry · Rydberg Semiclassical Optics · Theory ↔ Experiment

Advanced topic — uncertainty covariance and calibration reference planes

For measured parameters (\mathbf x) with covariance (\mathbf U_x), the first-order uncertainty of derived results (\mathbf y=f(\mathbf x)) is

For complex VNA data, treat real and imaginary parts consistently. Errors due to standards, cable flexure and instrument drift are often correlated across frequency. Consequently, independent error bars can misrepresent the uncertainty of a fitted resonance or group delay.

Calibration choice matters: SOLT, TRL and multiline TRL impose different assumptions about known standards and transmission lines. De-embedding is only as valid as the fixture model and reference-plane definition. Validate using a device not employed in the calibration.

For modulated signals, EVM is not unambiguous without specifying the reference waveform and permitted gain, phase, timing and equalization corrections.

Advanced exercise: fit a resonator to complex (S_{21}) data with an explicitly correlated measurement covariance; compare extracted (Q) uncertainty to an unweighted magnitude-only fit.

Further reading: NIST covariance-based VNA uncertainty; NIST mm-wave VSA calibration.

Advanced measurement pathways

For uncertainty-aware end-to-end instruments, continue with Advanced Antenna Metrology, Radiometry, Magnetic Cleanliness, Payload Readout Electronics and Timing & Synchronization.