Help in writing code for ordinal model with latent classes using gllamm

Hi,

I'm new to STATA and have been attempting to write out a latent class model that has 1) W1 and W2 as covariates in the class membership model 2) X1 X2 X3 and X4 as covariates in the ordinal model conditional on class membership.

I have written the following commands and the program has not converged in more than 500 iterations:

eq m1: W1 W2
gllamm Y X2 X3 X4 , i(id) l(oprobit) f(binom) peqs(m1) ip(f) nip(2)

Just to be clear, I do not have any random coefficients in the model - I'm just interested in specifying different coefficients across the two latent classes. Could anybody help me in getting the syntax right for this? I have pored over the manuals for over a day and have not made much headway. I'm not sure why any quadrature methods would be applied in this estimation but removin the ip(f) option resulted in the estimation of just one class.

Thanks,
Padma