How to calculate dynamic forecast standard errors

[petrenko]

Could anybody give me an example how to calculate dynamic forecast standard errors manually?
I have a simple AR(1) regression. I can calculate forecast errors of static forecast at particular value of X0 in a sample of regressor Xt in Excel using a simple formula like:
SE(X0) = SE of regression * square root of ( 1 + 1/number of observations + (X0 - average X) squared / sum (Xt - average X) squared )
The results of these manual calculations perfectly match E-Views calculations. But I can not reproduce the results of E-Views dynamic forecast. Help...