论文部分内容阅读
当取样时间τ和取样周期T_2可以比较时,取样电路传递函数不再是一个常数,产生频率非线性失真.在带有CPU的仪器中,可利用数字信号处理方法,对这种失真进行补偿.当τ/T_2=0.5时,未补偿时,传递函数从1降到0.6366。采用频率取样法设计的FIR数字滤波器时域滤波补偿后,传递函数最小为0.997614。最大为1.00230.当τ/T_2=0.8时对幅度为100左右模拟信号取样时。其误差的方差为1.7915,信噪比只有35.15db,利用DFT计算方法在频域滤波补偿后。使噪声方差从1.7915降到1.175×10~(-4)。信噪比从35.15db提高到118.81db。取样精度提高约15000倍.
When the sampling time τ and the sampling period T_2 can be compared, the transfer function of the sampling circuit is no longer a constant, resulting in non-linear frequency distortion. In CPU-equipped instruments, digital signal processing can be used to compensate for this distortion. When τ / T_2 = 0.5, the transfer function decreases from 1 to 0.6366 when not compensated. FIR filter using frequency sampling method of time domain filter compensation, the minimum transfer function of 0.997614. The maximum is 1.00230. When τ / T_2 = 0.8, the amplitude is about 100 when the analog signal is sampled. The variance of the error is 1.7915, SNR is only 35.15db, the use of DFT calculation method in the frequency domain filter compensation. The noise variance is reduced from 1.7915 to 1.175 × 10 ~ (-4). Signal to noise ratio increased from 35.15db to 118.81db. Sampling accuracy increased by about 15000 times.