Glossary term

IQ Imbalance

Engineering definition of IQ imbalance covering quadrature gain and phase mismatch, image rejection, EVM impact and receiver validation evidence.

Branch: Telecommunications Engineering
Glossary type: phenomenon
Content: Glossary term
Updated: Jun 26, 2026
Revision: v1.0.0 · reviewed

Definition

phenomenon

IQ imbalance is mismatch between the in-phase and quadrature signal paths of a complex transmitter or receiver, usually seen as gain error, phase error, DC offset or poor image rejection.

IQ imbalance affects direct-conversion and low-IF receivers, quadrature modulators, digital upconverters, software-defined radios and vector signal measurements. It distorts the complex plane by stretching, skewing or offsetting the constellation. It can increase EVM, degrade image rejection, leak mirror frequencies, bias channel estimates and create misleading diagnostics if it is mistaken for phase noise, frequency offset or thermal noise.

IQ imbalance is mismatch between the in-phase (I) and quadrature (Q) paths of a complex signal chain. Ideal I and Q paths have equal gain, 90 degree phase separation, matched bandwidth, low offset and consistent timing. Real transmitters and receivers can have gain mismatch, quadrature phase error, DC offset, filter mismatch, converter mismatch or calibration drift.

The result is not just a cosmetic constellation distortion. IQ imbalance can raise EVM, reduce image rejection, leak mirror-frequency content, bias channel estimates and make a receiver look interference-limited even when the root cause is inside the RF or mixed-signal chain.

Complex Signal View

An ideal complex baseband signal is:

x(t)=I(t)+jQ(t)

With IQ imbalance, the measured signal can be approximated as:

y(t)=\alpha x(t)+\beta x^*(t)

where x^*(t) is the complex conjugate. The conjugate term represents image leakage. In an ideal quadrature system:

\beta=0

Real systems have nonzero beta, so mirror-frequency content appears.

Gain And Phase Error

For small gain mismatch Delta g as a fractional amplitude error and quadrature phase error theta in radians, a rough EVM contribution screen is:

\displaystyle EVM_{IQ}\approx\sqrt{\left(\frac{\Delta g}{2}\right)^2+\left(\frac{\theta}{2}\right)^2}

This is a screening relation, not a substitute for vector signal measurement. It is useful because it shows why small gain and phase errors can be important for high-order modulation.

Image Rejection

Image rejection ratio can be stated as:

IRR_{dB}=P_{desired}-P_{image}

where both powers are measured at comparable reference planes and bandwidths. A higher value means better rejection of the mirror component.

When image rejection is poor, an interferer or noise from the image side can fold into the wanted complex baseband and degrade demodulation.

Worked Example

A quadrature receiver has measured I/Q gain mismatch:

\Delta G=0.8\ \text{dB}

Convert the amplitude ratio:

r=10^{0.8/20}=1.096

Use fractional mismatch:

\Delta g=r-1=0.096

Measured quadrature phase error is:

\theta=3.0^\circ=0.0524\ \text{rad}

The screening EVM contribution is:

EVM_{IQ}\approx\sqrt{(0.096/2)^2+(0.0524/2)^2}

EVM_{IQ}=0.0547=5.5\%

If the modulation release budget allows only 3.5% EVM contribution from IQ mismatch, the impairment fails by about 2.0 percentage points.

Image Rejection Example

In a tone test, the desired component is:

P_{desired}=-20\ \text{dBm}

and the image component is:

P_{image}=-48\ \text{dBm}

Image rejection is:

IRR=-20-(-48)=28\ \text{dB}

If the requirement is 35 dB, margin is:

M_{IRR}=28-35=-7\ \text{dB}

After calibration, the image component falls to:

P_{image}'=-60\ \text{dBm}

so:

IRR'=-20-(-60)=40\ \text{dB}

The corrected margin is:

M_{IRR}'=40-35=5\ \text{dB}

Difference From Phase Noise And CFO

Phase noise tends to create rotational spreading or random phase motion. Carrier frequency offset creates systematic constellation rotation over time. IQ imbalance skews, stretches or mirrors the constellation because the I and Q axes are not matched. These impairments can all increase EVM, but their corrective actions differ.

Constellation inspection is useful. Elliptical or slanted point clouds suggest IQ gain or phase mismatch. Rotational smear suggests phase noise or carrier recovery. A steady rotating constellation suggests residual frequency offset.

Engineering Use

IQ imbalance matters in quadrature modulators, direct-conversion receivers, low-IF receivers, SDRs, vector signal generators, vector signal analyzers, radar receivers and digital predistortion loops. Calibration may be analog, digital or mixed. Some systems estimate correction coefficients at startup; others track temperature and frequency dependence continuously.

Mitigation can include gain calibration, quadrature phase correction, DC offset removal, image-rejection calibration, matched filters, improved layout symmetry, better LO distribution, temperature compensation or digital compensation.

Validation Evidence

A defensible IQ imbalance review includes gain mismatch, phase mismatch, image rejection, DC offset, test frequency, bandwidth, power level, temperature, receiver gain state, calibration state, modulation mode, EVM before and after correction, image-tone test, spectrum capture and uncertainty.

The evidence should be taken at relevant frequencies and gain states. A one-frequency factory calibration may not hold across a wideband receiver or a hot field installation.

Common Mistakes

Common mistakes include treating EVM as a root cause, confusing IQ imbalance with phase noise, checking only one frequency, ignoring DC offset, applying calibration at one gain state and using it at another, hiding image leakage under averaging, and forgetting that transmitter and receiver IQ imbalance can both contribute to the same measured constellation error.

The practical rule is to separate gain mismatch, phase mismatch and image leakage, then verify that the correction improves EVM or image rejection in the real operating mode.

REF

Disciplines