An Improved Frequency-Interleaved DAC Model Based on the Frequency Matching and Modeling Optimization Algorithm for AWG Instrumentation

The frequency-interleaved digital-to-analog converter (FI-DAC) could enhance the bandwidth of the arbitrary waveform generator (AWG). At present, the formulation of the FI-DAC model mainly relies on the empirical design, lacking of sufficient theoretical basis. To address this problem, a frequency matching and modeling optimization (FMMO) algorithm is proposed. First, the alternating direction method of multipliers (ADMMs) and sequential quadratic programming (SQP) algorithm are utilized to optimize the model parameters of the FI-DAC system’s digital- and analog-domain filter groups. In this way, the accuracy of the model is enhanced. Meanwhile, it calibrates aliasing errors during the model design. Second, it addresses the channel energy mismatch and impedance mismatch issues in the hardware design by optimizing the model architecture. Finally, a high-bandwidth and high-signal-quality FI-DAC model is formulated. By conducting simulation, discrete module verification, and hardware evaluation, results show that the model can enhance the maximum effective output bandwidth of the 5-GSa/s DAC to 3 GHz. In addition, it achieves full-band flatness superior to 0.28 dB and effectively transmits 16 quadrature amplitude modulation (QAM), 64 QAM, 256 QAM, and 512 QAM communication signals at center frequencies of 1 and 2.5 GHz, respectively. With a center frequency of 1 GHz and 16 QAM modulation, the error vector magnitude (EVM) is reduced to 1.3%, and the SNR is increased to 35.3 dB.