Receiver I/Q Imbalance: Tone Test, Sensitivity Analysis, and the Universal Software Radio Peripheral

• Published in: IEEE transactions on instrumentation and measurement Vol. 59; no. 3; pp. 704 - 714
• Main Authors: Handel, P., Zetterberg, P.
• Format: Journal Article
• Language: English
• Published: New York IEEE 01-03-2010; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: algorithm; Baseband; Broad topic : 5. Instrumentation for and Measurement in Communications; compensation; Computer simulation; Cramer-Rao bound (CRB); Cramér-Rao bound (CRB); Estimates; frequency; gain imbalance; ieee-standard-1057; imbalance; in-phase/quadrature (I/Q); in-phase/quadrature (I/Q) imbalance; Leakage; Least squares approximation; local oscillator (LO) leakage; Local oscillators; Monte Carlo methods; Noise levels; Oscillators; parameter-estimation; Particle measurements; phase; quadrature skew; Radio frequency; Receivers; Sensitivity analysis; signals; Software radio; Software testing; Studies; systems; Time measurement; tone tests; universal software radio peripheral (USRP)
• ISSN: 0018-9456; 1557-9662; 1557-9662
• DOI: 10.1109/TIM.2009.2025989

The problem of determining the gain imbalance, quadrature skew, and local oscillator leakage of contemporary radio frequency receivers by tone test is considered. A least squares approach for indirect estimation of the sought-for parameters is proposed, which is linear in six out of its seven parameters. The performance of the method, particularly its accuracy as a function of measurement time, imbalance parameters, and signal-to-noise ratios, is investigated. The theoretical predictions of the performance are illustrated by Monte Carlo simulations and experimental data that are obtained from testing several universal software radio peripheral (USRP) receivers. This paper shows that, for the studied off-the-shelf receivers, gain imbalance and quadrature skew may accurately be predicted (i.e., < 0.1 dB and <, respectively) by employing baseband data covering only a handful of full periods of the excitation stimuli. Using short records of data, this paper also shows that local oscillator leakage adding a bias term in the receiver baseband data may suffer from a systematic bias error on the order of 15 dB in the estimate. To obtain local oscillator leakage estimates with an uncertainty on the order of 1 dB, the measurement time has to be increased by two orders of magnitude.