Abstract:  Frequency-domain adaptive algorithm has the benefits of fast convergence speed and low computational burden, which makes it a good choice in the implementation of active noise control systems. However the commonly used bin-normalized frequency domain algorithm suffers from deterioration of least mean error in non-causal conditions. Active noise control system has its inherent electric delay, in the application of active noise control in a short duct, non-causal conditions often occur. It is a meaningful work to research or improve the performance of frequency domain adaptive algorithms in non-causal conditions. This paper analyzes the steady state solution of bin-normalized frequency domain algorithm in non-causal conditions. It is proved that the steady state solution of the bin-normalized frequency domain algorithm is related with the flatness of reference signal power spectrum. A new frequency domain filter coefficients updating method is also proposed in this paper, and theoretically the new algorithm can converge to the optimal solution in non-causal conditions. The performance of the bin-normalized frequency domain algorithm is illustrated by simulations. Furthermore, the performances of related algorithms are also compared in a real time active noise control DSP system. Simulations and experiments verify the theoretical analysis.