Question: CurveFitting of Data

Given Datas:

xd := Vector([0, 25, 50, 75, 100, 125, 150, 175, 200, 225, 250, 500, 750, 1000, 1250, 1500, 1750, 2000, 2500, 3600])

yd1 := Vector([0, 6.24, 11.26, 14.9, 17.58, 20.01, 22.0, 23.56, 24.85, 26.03, 27.36, 35.07, 39.4, 42.61, 44.36, 45.79, 46.62, 46.9, 47.79, 48.85])

Given function: (looks a little different in Maple, but copy and paste did not work)

<maple>F2 := proc (x) options operator, arrow; K2*(1-(T12-Tv2)*exp(-x/T12)/(T12-T22)+(T22-Tv2)*exp(-x/T22)/(T12-T22)) end proc</maple>

If I call

N2 := NonlinearFit(F2(x), xd, yd, x, initialvalues = [K2 = 50, T12 = 600, T22 = 70, Tv2 = 260], output = parametervalues, parameternames = [K2, T12, T22, Tv2])

I don't get the optimal fitting, although I'm rather close to the solution. Has anyone an idea for this?