“…Looking at these papers it seems that except of some simpler cases there is no general result concerning the convergence of the constructed Tau approximants. starting from the operational approach given in [1], we show most interesting cases, using Chebyshev polynomials the Tau method is in fact a method Petrov type, thus convergence results follow theorems given in [2]. To determine y R we take from G only the first n+m+1 rows and columns, i.e., we solve the linear system < 6 > ân G n = y n j^v is a s-called Tau approximant, because The solution it generates the polynomial defect (7) (Dy n )(x) = a n ny = (a^Jv,, (a n n n+n+h ) v n+m+h , where h=max(a^ii) is the so-called height of the differential operator D. He notice that in view of (4) and (6) the components of the defect vanish with respect to the first n+1 elements of the base v.…”