linda sim - static sphere - updates H line optical depth VERY slowly

[gary]

the linda.in sim reported by linda strubbe is very slowly converging the H line optical depths. it is an optically thin static sphere which has not been well tested - all we test are either optically thick or not static/sphere.

initial H optical depth guess is large, but first iteration is thin. thereafter code uses the "far side" optical depth from previous iteration. it should double the face to face optical depth instead.

any optically thin static sphere will show this problem - look at old, new H line optical depths in attached sim which runs very quickly