What is the difference between insertion loss and return loss and how are they measured?
IL and RL Measurement Techniques
Insertion loss and return loss are the most frequently measured parameters in RF engineering, used to evaluate cables, connectors, filters, amplifiers, antennas, and every other RF component.
Measurement Setup
Insertion loss measurement: (1) Calibrate the VNA with a THRU connection (normalizes out cable and connector losses). (2) Insert the DUT between the calibrated ports. (3) Read |S21| on the VNA display. The insertion loss is the negative of |S21| in dB. (4) Frequency sweep: measure IL across the entire operating bandwidth. For a bandpass filter: IL varies from the passband value (minimum loss, e.g., 1 dB) to the stopband rejection (maximum loss, e.g., 60 dB). Return loss measurement: (1) Calibrate Port 1 with SHORT/OPEN/LOAD (one-port calibration). (2) Connect the DUT to Port 1. (3) Read |S11| on the VNA display. The return loss is the negative of |S11| in dB. (4) For a two-port device: measure S11 with Port 2 terminated in 50 ohms (or the system impedance). Then measure S22 with Port 1 terminated. Full two-port calibration provides S11 and S22 simultaneously.
Sources of Measurement Error
Insertion loss errors: (1) Connector repeatability: each mating of a connector pair introduces ±0.02-0.05 dB variation (precision connectors) or ±0.1-0.3 dB (general-purpose connectors). For measuring a 0.5 dB attenuator: connector repeatability of ±0.1 dB represents a 20% measurement uncertainty. (2) Cable loss drift: cables change loss with temperature and flexing. Phase-stable cables: ±0.02 dB/°C. Standard cables: ±0.1 dB/°C. (3) Impedance mismatch: if the DUT has poor return loss, the insertion loss depends on the source and load impedances (which are 50 ohms only if the VNA port match is perfect). Mismatch uncertainty: delta_IL = ±20×log10(1 ± |S11_source| × |S11_DUT|). For source match = 20 dB and DUT return loss = 10 dB: delta_IL = ±20×log10(1 ± 0.1 × 0.316) = ±0.27 dB. Return loss errors: (1) VNA directivity: the corrected directivity (after calibration) determines the minimum measurable return loss. If corrected directivity = 40 dB: return loss measurements > 40 dB are unreliable. Typical corrected directivity: 40-50 dB for coaxial calibration, 30-40 dB for on-wafer or waveguide calibration. (2) Connector quality: a worn or damaged connector can have 20-30 dB return loss, which is the measured value regardless of the DUT.
Converting Between IL, RL, and Other Parameters
Key relationships: RL to VSWR: VSWR = (1 + 10^(-RL/20)) / (1 - 10^(-RL/20)). RL to reflection coefficient magnitude: |Gamma| = 10^(-RL/20). RL to mismatch loss: ML = -10×log10(1 - |Gamma|^2). For RL = 10 dB: |Gamma| = 0.316, ML = 0.46 dB. For RL = 15 dB: |Gamma| = 0.178, ML = 0.14 dB. For RL = 20 dB: |Gamma| = 0.1, ML = 0.04 dB. Total through loss = IL + ML (insertion loss includes the mismatch loss component). For low-loss components: the mismatch loss can be a significant fraction of the total insertion loss. Example: a cable with 0.2 dB conductor/dielectric loss and 15 dB return loss: total IL = 0.2 + 0.14 = 0.34 dB.
RL = -20log₁₀|S11| dB
VSWR = (1+|Γ|)/(1-|Γ|)
Mismatch Loss = -10log₁₀(1-|Γ|²)
|Γ| = 10^(-RL/20)
Frequently Asked Questions
What return loss is considered acceptable?
Depends on the application: for antennas: RL > 10 dB (VSWR < 2:1) is the common specification. RL > 14 dB (VSWR < 1.5:1) is considered good. RL > 20 dB (VSWR < 1.2:1) is excellent. For connectors and adapters: RL > 20 dB is typical for SMA at < 18 GHz. RL > 26 dB is expected for precision connectors. For amplifier ports: input RL > 10-15 dB is typical (some low-noise amplifiers sacrifice input match for noise performance). For filters: passband RL > 15-20 dB is typical (determines passband ripple: RL = 20 dB corresponds to ±0.04 dB ripple). For cable assemblies: RL > 20 dB at each connector and > 26 dB along the cable itself.
Can insertion loss be negative?
Yes. Negative insertion loss means the device has gain (output power > input power). An amplifier with 20 dB gain has IL = -20 dB. The VNA displays S21 = +20 dB. A negative insertion loss in a passive device indicates a measurement error (the calibration reference is wrong, the cables have changed, or the DUT is oscillating). Some passive devices appear to have slightly negative IL at certain frequencies due to mismatch effects: if the DUT provides a better impedance match than the calibration THRU, the mismatch loss decreases, making the DUT appear to have slight gain. This is a measurement artifact, not real gain.
How do I measure insertion loss of a very low-loss device?
Measuring IL < 0.5 dB requires care: (1) Use the best available calibration (SOLT with precision standards or TRL calibration for the lowest uncertainty). (2) Minimize connector effects: use the same connector type on the DUT as the calibration reference (avoid adapters that add their own IL). (3) Repeat the measurement multiple times and average: the statistical uncertainty decreases with the number of measurements. (4) Control temperature: a 0.1 dB drift in cable loss equals the entire DUT loss for a 0.1 dB device. Use phase-stable cables and allow thermal equilibrium. (5) Use high IF bandwidth averaging on the VNA (reduce IF BW to 100 Hz or lower) to improve trace noise. (6) Consider TRL calibration (which uses airlines instead of lumped loads) for the highest accuracy at the reference plane.