Direct, Indirect, and Ratio Methods

In calibration metrology, the method of measurement dictates how an unknown quantity is compared to a traceable standard, directly governing the structure of the measurement uncertainty budget. Technicians must select between direct, indirect, ratio, transfer, differential, and substitution methods based on required Test Uncertainty Ratios (TUR), instrument geometry, physical domain boundaries, and systematic error cancellation requirements. Mastering these six classical techniques allows metrologists to isolate the true measurand from environmental and instrumental artifacts.

---

Classification and Taxonomical Framework

According to the International Vocabulary of Metrology (VIM, JCGM 200:2012), a measurement method is a generic description of a logical organization of operations used in a measurement. In the 2024 ASQ Certified Calibration Technician (CCT) Body of Knowledge, measurement methods are categorized into six foundational techniques:

1. Direct Measurement (including absolute and fundamental methods)
2. Indirect Measurement
3. Ratio Measurement
4. Transfer Measurement
5. Differential Measurement
6. Substitution by Unit Under Test (UUT)

Each method addresses specific physical limitations of measuring equipment, reference standards, and human observation. Selecting the appropriate method minimizes the number of uncorrected systematic biases and prevents the unnecessary compounding of measurement uncertainties.

---

Direct Measurement (Absolute and Fundamental Techniques)

Definition and Physics

A direct measurement provides a measurement result obtained directly from the measuring instrument without supplementary calculations involving functional relationships with other physical quantities. The measurand is compared directly with an internal or external graduated scale or physical standard of the same physical dimension.

• Fundamental / Absolute Measurement: A subset of direct measurement where the value is determined directly from base physical definitions (e.g., length, mass, time) without reference to secondary working standards of the same physical quantity. Classical examples include measuring time intervals via atomic hyperfine transitions ($^{133}\text{Cs}$) or determining length using vacuum laser interferometry referenced to the speed of light ($c = 299,792,458\text{ m/s}$).

Metrological Implementations

• Dimensional Metrology: Measuring the length of a machined pin using an outside micrometer, a vernier height gage, or an optical glass scale. The displacement of the measuring spindle maps directly to graduations representing the SI base unit (meter).
• Mass Metrology: Determining mass on a direct-reading electronic analytical balance where electromagnetic force compensation balances the gravitational pull on the pan directly against a factory-calibrated internal reference.
• Electrical Metrology: Measuring DC voltage by connecting a digital multimeter (DMM) across a circuit, where the internal analog-to-digital converter (ADC) compares the unknown input directly against an internal bandgap or zener voltage reference.

Inherent Limitations and Error Mechanisms

While direct measurement is procedurally straightforward, it provides no self-cancelling error mechanisms:

1. Instrument Gain and Linearity Errors: Any systematic offset, scale stretching, or non-linearity across the entire measured span translates directly into measurement error.
2. Contact and Fixture Deformation: Direct mechanical contact induces elastic Hertzian contact stress deformation ($w \propto (F^2 / R)^{1/3}$), requiring force control via ratchet thimbles or regulated probe droop.
3. Thermal Expansion: Both the measuring tool and the workpiece expand with temperature, requiring manual or mathematical corrections to the standard reference temperature of $20^\circ\text{C}$ ($68^\circ\text{F}$).

---

Indirect Measurement

The elementary power equations below apply to DC resistive operation or the equivalent unity-power-factor case. General AC real power is the time-average of $v(t)i(t)$; for sinusoidal voltage and current it is $V_{\mathrm{rms}}I_{\mathrm{rms}}\cos\phi$. RMS voltage times RMS current alone is apparent power when power factor differs from one.

Mathematical Modeling ($Y = f(X_1, X_2, \dots, X_n)$)

An indirect measurement determines the value of a measurand by measuring one or more other quantities that are functionally related to the measurand through an established physical law. The measurand ($Y$) is not directly accessible or is impractical to measure directly, requiring mathematical transformation from input variables ($X_i$):

$$Y = f(X_1, X_2, \dots, X_n)$$

Common Metrological Applications

• Density Determination: Calculating mass density ($\rho$) from gravimetric mass ($m$) and geometric or pycnometric volume ($V$):

$$\rho = \frac{m}{V}$$

• Electrical Power Determination: In DC and low-frequency AC systems, electrical power ($P$) is rarely measured by a single primary sensor. Instead, it is computed indirectly from independent measurements of potential difference ($V$) and current ($I$), or resistance ($R$):

$$P_{\text{DC}} = V \cdot I = \frac{V^2}{R} = I^2 \cdot R$$

• Flow Rate via Differential Pressure: Determining volumetric flow ($Q$) through an orifice plate by measuring the differential pressure drop ($\Delta P$), fluid density ($\rho$), and upstream pipe geometry ($D$):

$$Q = C_d \frac{A_2}{\sqrt{1 - \beta^4}} \sqrt{\frac{2 \Delta P}{\rho}}$$

• Modulus of Elasticity ($E$): Determined indirectly in materials calibration by measuring applied tensile load ($F$), instantaneous elongation ($\Delta L$), initial cross-sectional area ($A_0$), and initial gage length ($L_0$):

$$E = \frac{\sigma}{\varepsilon} = \frac{F / A_0}{\Delta L / L_0} = \frac{F \cdot L_0}{A_0 \cdot \Delta L}$$

Uncertainty Propagation (GUM Framework)

Per the Guide to the Expression of Uncertainty in Measurement (GUM, JCGM 100:2008), the combined standard uncertainty $u_c(y)$ of an indirect measurement is calculated using the law of propagation of uncertainty:

$$u_c^2(y) = \sum_{i=1}^n \left( \frac{\partial f}{\partial x_i} \right)^2 u^2(x_i) + 2 \sum_{i=1}^{n-1} \sum_{j=i+1}^n \frac{\partial f}{\partial x_i} \frac{\partial f}{\partial x_j} u(x_i, x_j)$$

Where $c_i = \frac{\partial f}{\partial x_i}$ is the partial derivative (sensitivity coefficient) of input quantity $X_i$, and $u(x_i, x_j)$ represents the covariance between correlated input quantities.

Calibration Trap: Correlated Variables in Indirect Metrology

Shared equipment or environmental influences can correlate input estimates, but do not guarantee nonzero correlation. Identify common influences and evaluate covariance where significant. In a DC resistive model $P=V^2/R$, ignoring an actual covariance can increase or decrease the calculated uncertainty depending on signs and sensitivities.

---

Ratio Measurement

Theoretical Principle

A ratio measurement determines the magnitude of an unknown quantity by establishing its numerical ratio to a known reference standard of the same physical quantity, or by determining a dimensionless proportionality constant. The fundamental relationship is:

$$X_{\text{uut}} = K_{\text{ratio}} \cdot X_{\text{standard}}$$

Where $K_{\text{ratio}}$ is a calibrated, highly linear division ratio. The dominant advantage of ratio measurement is common-mode cancellation: any amplitude fluctuations, low-frequency drift, or noise in the excitation power supply affect both the unknown and the reference equally, cancelling out of the ratio calculation:

$$\frac{V_{\text{uut}}}{V_{\text{std}}} = \frac{I_{\text{ex}} \cdot R_{\text{uut}}}{I_{\text{ex}} \cdot R_{\text{std}}} = \frac{R_{\text{uut}}}{R_{\text{std}}}$$

Core Ratio Instruments and Circuits

1. Kelvin-Varley Voltage Dividers (KVD):
   • Consist of cascaded decades of matched precision resistors with a traversing double-pole switch that shunts two adjacent resistors in each decade.
   • Delivers linear ratio division from $0.0000000$ to $1.0000000$ with linearity uncertainties reaching $< 0.1\text{ ppm}$ ($0.00001\%$).
   • Used to calibrate high-accuracy calibrators, digital multimeters, and potentiometers.
2. Inductive Voltage Dividers (IVDs) / Ratio Transformers:
   • Employ multi-winding precision autotransformers wound on high-permeability toroidal tape-wound cores (Supermalloy or nanocrystalline).
   • The ideal voltage ratio follows the turns ratio. Real inductive dividers also exhibit winding, leakage, burden, frequency, and phase effects; evaluate their calibration and operating conditions.
   • Input impedances exceed $100\text{ k}\Omega$ while output impedances remain $< 1\ \Omega$, achieving ratio accuracies better than $0.01\text{ ppm}$ ($10^{-8}$) at audio frequencies ($400\text{ Hz to }1\text{ kHz}$).
3. Wheatstone and Kelvin Bridge Networks:
   • Four-arm resistive bridges where an unknown resistor $R_x$ is balanced against a certified standard resistor $R_s$ using ratio arms $R_1$ and $R_2$. At bridge null ($V_{\text{null}} = 0$):

$$R_x = R_s \left( \frac{R_2}{R_1} \right)$$

• Kelvin double bridges introduce auxiliary ratio arms that prevent lead and contact resistances from affecting low-resistance measurements ($10^{-6}\ \Omega\text{ to }1\ \Omega$).