I am estimating an Interval Regression model. I wrote my likelihood function (see intreg, STATA) as stated bellow, and compared my results with STATA. I am finding almost exactly (within 0.001) same parameters as STATA but the BFGS algorithm seems still not to converge.