Transient Carrier Feedthrough

The primary mechanism is a momentary DC bias voltage appearing at the baseband input of the IQ modulator during the power-up or power-down sequence. This offset is caused by differential amplifier settling and bias network charging in the digital-to-analog converter (DAC) and reconstruction filter stages. Unlike steady-state DC offset, this transient component has a distinct time-varying profile that reflects the specific time constants of the transmitter's baseband circuitry.

In the frequency domain, transient carrier feedthrough appears as a narrowband spectral line precisely at the local oscillator (LO) frequency. This occurs because the DC offset does not modulate the carrier but simply passes it through to the output. The amplitude envelope of this spectral line during the transient period is a direct trace of the DC offset's temporal evolution, creating a unique 'power-on signature' visible on a spectrum analyzer in zero-span mode.

While all direct-conversion transmitters exhibit some degree of steady-state LO leakage due to permanent component mismatches, transient carrier feedthrough is a dynamic phenomenon. Key differences include:

• Duration: Exists only for microseconds during burst onset/offset.
• Amplitude Profile: Has a distinct attack, decay, or ringing shape rather than a constant level.
• Origin: Driven by time-varying circuit settling, not static manufacturing tolerances. This dynamic profile provides a richer feature set for fingerprinting than a static leakage measurement.

The specific shape of the transient carrier feedthrough is heavily influenced by:

• DAC Architecture: Current-steering vs. voltage-output DACs exhibit different glitch impulse responses during code transitions from zero to the bias point.
• Modulator Topology: Gilbert cell mixers and passive ring modulators have distinct LO-to-RF isolation recovery profiles.
• Baseband Filtering: The group delay and settling time of the reconstruction low-pass filter directly shape the DC offset transient as it reaches the modulator input.

To isolate this fingerprint, a receiver can perform tuned envelope analysis at the known carrier frequency. The process involves:

• Digital Down-Conversion: Tuning precisely to the LO frequency.
• Narrowband Filtering: Isolating the carrier leakage from the modulated signal sidebands.
• Hilbert Transform: Extracting the instantaneous amplitude envelope of the filtered leakage. The resulting amplitude vs. time profile—including its rise time, overshoot, and settling behavior—serves as a robust feature vector for emitter identification.

Transient carrier feedthrough is inherently robust against multipath fading and Doppler shift. Because the feature is a temporal amplitude profile at a single frequency, frequency-selective fading does not distort its shape—it only scales the overall amplitude. This makes it a highly reliable physical-layer identifier even in challenging non-line-of-sight (NLOS) environments, provided the receiver has sufficient dynamic range to capture the low-level leakage signal above the noise floor.

The root cause of transient carrier feedthrough. A momentary direct current bias appears at the modulator output during the turn-on transient due to baseband amplifier settling and local oscillator leakage. This non-zero DC voltage shifts the carrier suppression balance, causing the unmodulated carrier to leak through to the output. The magnitude and duration of this offset are unique to each device's amplifier input offset voltage and bias network time constants.

The physical mechanism enabling feedthrough. An ideal IQ modulator perfectly suppresses the carrier when baseband inputs are zero, but transient gain and phase mismatches between the I and Q paths break this suppression. During the rapid power-up sequence, the differential pairs in the mixer have not yet reached thermal equilibrium, creating a temporary carrier suppression degradation that manifests as a spectral line at the carrier frequency.

The direct coupling of the local oscillator signal into the RF output path, bypassing the modulator. During the transient, this leakage is often elevated due to VCO pulling and unstable bias conditions. The leaked carrier appears as a distinct spike in the spectrum, and its amplitude envelope during the transient reveals the settling behavior of the oscillator's buffer amplifier and isolation characteristics of the mixer.

A key feature for quantifying feedthrough. The center of mass of the transient's short-time Fourier transform spectrum shifts dramatically when carrier feedthrough is present. A strong unmodulated carrier component pulls the centroid toward the carrier frequency, creating a measurable metric. The trajectory of this centroid over the transient duration reveals the time constant of the DC offset decay as the modulator stabilizes.

A time-domain technique for detecting carrier feedthrough. By measuring the precise intervals between consecutive zero-voltage crossing points of the captured waveform, the instantaneous frequency is extracted. During carrier feedthrough, the zero-crossing pattern becomes highly regular at the carrier frequency, in contrast to the irregular pattern of a modulated or noise-dominated transient. This regularity serves as a robust detection feature.

A higher-order spectral analysis tool that reveals quadratic phase coupling within the transient. Carrier feedthrough creates a deterministic phase relationship between the leaked carrier and any intermodulation products generated by the non-linear turn-on. The bispectrum isolates these non-Gaussian interactions, effectively suppressing Gaussian noise and highlighting the non-linear hardware signature of the modulator's carrier suppression failure.