A Method for ADC Dynamic Parameter Testing Under Non-coherent Sampling

The accuracy of dynamic parameter testing for high-precision analog-to-digital converters (ADC) is often compromised by spectral leakage arising from non-coherent sampling. This paper proposes a correction method that eliminates leakage through signal reconstruction. The framework first employs Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) and the Hilbert Transform (HT) to obtain a high-accuracy initial frequency estimate. This estimate then seeds a cascaded threeand four-parameter sine fitting algorithm to precisely determine the signal’s parameters, which are used to reconstruct an ideal, leakage-free coherent signal for analysis. Simulations confirm that the proposed method accurately corrects even severe spectral leakage, with SNDR and SNR errors remaining within $\pm 0.2 \mathrm{~dB}$, and THD and SFDR errors constrained to $\pm 1.5 \mathrm{~dB}$ compared to ideal coherent benchmarks across a wide range of non-coherent factors. By fundamentally avoiding the main-lobe broadening and amplitude loss associated with traditional windowing techniques, this approach relaxes stringent hardware requirements, enabling high-precision ADC testing with lower-cost equipment.