Transfer function determination from impulse response via walsh functions

Abstract

LETTERS ro THE EDITOR 85For carrying out the experiment above, the following components have been used: *15 V op-amp type ML 741 having a measured unity gain bandwidth of f t = 700 kHz, Rf = 56 kR,The compensating components used are r = 68 s1 and C = 3.3 nF.Five per cent capacitors and precise resistors have been used. Thus the errors due to passive component tolerances are quite small.The experiment has been repeated with the filter designed at a higher value of natural frequency f o = 18.279 kHz by using a value of C1 = C2 = 68 nF. Figure 4 shows plots for the compensated and the theoretical responses. In this case it has been observed experimentally that the filter with the ordinary uncompensated VCVS introduces oscillations with large amplitudes at a frequency near f o .It is worth noting that the above design is directed towards the compensation of the phase error (which is of a first-order magnitude equal to --WT for the uncompensated VCVS). The compensated structure proposed, however, is very general and may realize different responses by proper choice of the time constant Cr. For example a maximally flat magnitude response may be obtained by choosing R = R 1 = 1 kR, C1= C 2 = 178 nF.