Time, Frequency, Light, Power, and Energy

Time and Frequency Metrology

Time interval and frequency represent the most precisely realized physical parameters in metrology. The SI base unit of time—the second ($s$)—is defined by fixing the unperturbed ground-state hyperfine transition frequency of the cesium-133 atom ($\Delta \nu_{\text{Cs}}$) at exactly $9,192,631,770\text{ Hz}$.

Hierarchy of Frequency Standards

The SI second refers to the unperturbed cesium-133 transition. Commercial atomic references are physical devices with frequency shifts and performance limits; they are not automatically evaluated primary standards. A GNSS-disciplined reference requires a documented time-system link, receiver and cable delays, frequency averaging, and holdover assessment. GPS system time is not simply UTC(NIST).

Frequency Counters and Timing Measurements

• Direct Counting vs. Reciprocal Counting:
  • Direct Counters: Count input signal cycles during a fixed timebase gate (e.g., $1.0\text{ s}$). Resolution is $\pm 1\text{ count}$. Measuring a $60\text{ Hz}$ line yields $60 \pm 1\text{ Hz}$, an unacceptable $1.67\%$ error.
  • Reciprocal Counters: Measure one or more input periods against a timebase. They can give relatively uniform fractional resolution over a broad input-frequency range, but actual performance depends on gate duration, interpolation, trigger noise, timebase uncertainty, and the counter’s design.
• Trigger Noise and Timing Jitter: In time interval and period measurements, electrical noise on the input signal causes random premature or delayed threshold crossings, generating trigger noise error:

$$\sigma_t = \frac{e_n}{dV/dt}$$

Where $e_n$ is RMS noise voltage and $dV/dt$ is signal slew rate across the trigger threshold. Technicians maximize slew rate by using fast rise-time edges and correct input coupling (DC coupling to prevent baseline wander).

---

RF and Microwave Power Calibration

High-frequency power metrology operates from kilohertz to millimeter waves. In high-frequency transmission lines, voltage and current vary with spatial position due to standing waves, making power the fundamental parameter.

RF Power Sensors

• Thermistor Mounts (DC Substitution Primary Standard): A matched pair of negative temperature coefficient thermistors is placed in a self-balancing DC Wheatstone bridge. Prior to applying RF, precision DC bias power ($P_{\text{DC1}}$) heats the thermistors to maintain bridge balance (e.g., $100\text{ or }200\ \Omega$). When RF power is introduced, it adds heat; the feedback circuit decreases DC bias power to $P_{\text{DC2}}$ to maintain identical operating resistance. The RF power is calculated directly:

$$P_{\text{RF}} = \frac{P_{\text{DC1}} - P_{\text{DC2}}}{\text{Effective Efficiency}}$$

• Diode Sensors: Their low-level response can be approximately square-law. Transition level, linearity correction, pulse response, and range depend on the sensor design and calibration; −20 dBm is not a universal boundary.

Impedance Mismatch and Reflection Coefficients

When connecting an RF generator (source reflection coefficient $\Gamma_g$) to a power sensor (load reflection coefficient $\Gamma_l$), impedance mismatch creates standing waves:

$$\Gamma = \frac{Z - Z_0}{Z + Z_0} \quad \text{and} \quad \text{VSWR} = \frac{1 + |\Gamma|}{1 - |\Gamma|}$$

For the simple two-port model, source and load reflection coefficients can bound a first-order mismatch interaction. The approximation below neglects higher-order terms; use the model required by the actual power measurement:

$$M = 1 \pm 2 |\Gamma_g| |\Gamma_l|$$

Mismatch can be a significant RF power uncertainty contributor. Its magnitude depends on reflection coefficients, setup, frequency, and other contributions; it is not invariably the largest. Evaluate the adopted mismatch model and source stability.

---

Optical and Luminosity Measurements

Optical metrology is divided into radiometry (pure physical radiation power across all wavelengths) and photometry (optical radiation weighted by the human eye's spectral sensitivity).

Photometers and Lux Meters

Photometers employ silicon photodiodes equipped with custom multi-layer optical glass absorption filters designed to match the CIE Standard Photopic Observer curve $V(\lambda)$, which peaks at $555\text{ nm}$ (green light).

• Spectral Mismatch: A photometer’s spectral response differs from the ideal visual weighting. Evaluate the source spectrum and detector correction for the task; one f₁′ limit is not a universal requirement for all laboratory photometers.
• Cosine response: A photometer’s angular response should approximate the required cosine weighting. Diffusers and optical design are possible means; opaline or PTFE is not a uniquely mandatory material. Characterize angular and spectral-response deviations for the intended illumination geometry.

Spectrophotometers

Spectrophotometers measure spectral transmittance ($T$) or reflectance ($R$) as a function of wavelength. A monochromator utilizes an adjustable entrance slit, collimating optics, a reflective diffraction grating, and an exit slit to isolate narrow wavelength bands (spectral bandwidth $\Delta \lambda$).

• Wavelength Calibration: Verified using emission lines from low-pressure discharge arc lamps (mercury, neon) or certified rare-earth liquid/glass absorption filters (Holmium Oxide and Didymium).
• Stray Light Error: Spurious radiation outside the nominal bandpass hitting the detector. Stray light compresses measured absorbance ($A = -\log_{10} T$), causing serious non-linear under-reporting at high absorbance values ($A > 2.0$).

Laser Power Meters

• Thermopile Sensors: Laser energy strikes a broad-spectrum ceramic/metallic absorber disk, generating an axial or radial heat flow measured by an array of thermocouples. Handles milliwatts to tens of kilowatts, continuous wave (CW) or pulsed, but exhibits slow thermal response times (1 to 5 seconds).
• Semiconductor Photodiodes (Si, InGaAs): Fast response (nanoseconds), high sensitivity (picowatts to milliwatts). Photodiode responsivity $R(\lambda)$ (amperes per Watt) is strongly wavelength-dependent, requiring selection of the laser operating wavelength in firmware. High power densities saturate the junction, distorting readings.

Integrating Spheres

An integrating sphere is a hollow spherical enclosure coated internally with a highly reflective, nearly perfect Lambertian diffuse coating (barium sulfate $\text{BaSO}_4$ or sintered PTFE / Spectralon, reflectance $\rho \approx 98\%\text{ to }99\%$).

Light entering or generated inside the sphere undergoes multiple diffuse internal reflections, distributing radiant flux uniformly across the interior surface. A shielded photodetector measures irradiance proportional to total luminous flux (lumens) or radiant power (Watts).

• Baffles: Internal shields must be positioned to prevent direct unreflected line-of-sight light from the source reaching the detector.
• Self-Absorption Correction (Auxiliary Lamp Method): When a physical lamp fixture or test device is inserted into the sphere, its body absorbs a portion of the circulating light. An auxiliary stable lamp mounted on the sphere wall is illuminated with and without the DUT present to calculate a mathematical absorption correction factor.

Power and energy measurement

For DC, $P=VI$. For sinusoidal AC, real power is $P=V_{RMS}I_{RMS}\cos\phi$; apparent power is $S=V_{RMS}I_{RMS}$. For a distorted waveform, real power is the average of $v(t)i(t)$; a simple phase-angle model may not be adequate. A true RMS voltage measurement alone does not establish real power into an arbitrary load.

For an ideal zero-mean sine, $V_{RMS}=V_p/\sqrt{2}$. Crest factor is peak magnitude divided by RMS. The general RMS operation is the square root of the time-average of squared values. Check bandwidth, coupling, crest factor, loading, frequency, and instrument rating.

Energy integrates power over time. One watt-hour is 3600 joules. In a hypothetical constant-load meter test, real power is 1.20 kW over 15.0 minutes, so reference energy is $1.20\times15/60=0.300\text{ kWh}$. A meter indicating 0.303 kWh has error $0.003/0.300\times100=+1.00\%$. If its pulse constant is 1000 impulses/kWh, the expected count is 300 impulses; confirm whether the constant is impulses/kWh or its reciprocal before converting. Include power-reference, timing, pulse resolution, stability, and connection contributions in the measurement uncertainty. A display with more decimals does not replace those controls.