Noniterative WLS design of digital FIR filters for aliasing errors cancellation in bandwidth interleaving digital-to-analog converter

Abstract

This paper presented a noniterative weighted least squares (WLS) design of digital finite impulse response (FIR) filters for canceling the aliasing errors caused by the non-ideality of analog mixers and filters in bandwidth interleaving digital-to-analog converter (BI-DAC). Taking a BI-DAC with M sub-DACs as an example, we firstly deduced the approximation error function. Then, based on this approximation error function, the optimum coefficients of all digital FIR filters were obtained by using the noniterative WLS method. Further, we deduced the computational complexity of our presented noniterative WLS design. Several design examples were given to verify the performance of our presented noniterative WLS design containing their effectiveness and computational. The simulation results showed that, by using our presented noniterative WLS design, the minimum distortion error was 0.021dB, and the maximum aliasing error was -67dB which satisfied the desired spurious free dynamic range (SFDR) in a 11-bit BI-DAC system. Finally, we compared the computational complexity among our presented noniterative WLS design and another three optimal designs, and satisfactory simulation results were obtained.