I use the SPICE model downloaded from the ADI website of AD8065 and configured as a G=+1 buffer and simulate for the ouput nosie density. It is supposed to be the same with the equivalent input noise density. But the part of 1/f noise is missing and the white noise is much lower than that in datasheet, the noise above 10MHz is too large compared with the value in datasheet. Is the SPICE model of AD8065 so inaccurate on noise performance? The model of AD8033 also has the same kind of problem.