Розмір шрифту: 

There was proposed a new method of estimation of multiprobe microwave multimeter algorithms precision by mean of accumulation of partial error. The least square solution is used for variance and covariance matrix definition. The weighed coefficient is obtained from algorithms derivatives with respect to intermediate variable. Substituting the expression for standard deviation (variance) end weighed coefficient in formula for partial error accumulation we can compare different algorithms, study frequency properties of algorithms and make conclusion about its applicability in multiprobe microwave multimeter for passing power and reflection coefficient definition. It was proved that solutions (i.e. algorithms) are the same for analytic and numerical (least square) method. So it is possible apply least square solution for three equation with three unknown variable.

multiprobe microwave multimeter; least squares solution; weighted coefficient; accumulation of partial error