estimating an ARMA model including a dummy

[yussel]

Dear all,
Could anyone give some suggestions how to solve the following problem I got in Eviews (version 3.1 at home, 5 at the university)

I need to estimate an ARMA-model for volatility. As there are some periods with huge (abnormal) volatility I created a dummy which should now be incorporated in the ARMA model. In theory I'd like to estimate the follwing equation: vol c ar(1) ar(1)*dummy ar(2) ar(2)*dummy ma(1) ma(1)*dummy ma(2) ma(2)*dummy. As Eviews can not estimate this combination of ARMA and dummy I tried to model it this way: vol=C(1)+C(2)*vol(-1)+C(3)*vol(-1)*dummy+C(4)*vol(-2)+C(5)*vol(-2)*dummy+C(6)*resid+C(7)*resid*dummy+C(8)*resid+C(9)*resid*dummy. But writing down resid in the equation for the MA terms also does not work.

Does anyone have a suggestion how to estimate this the easiest way?

Thanks in advance!