The easiest way to do this is with the KR file, which includes all births in the five years before the interview, regardless of whether the child survived. The sex of the child is given by b4 (1=boy, 2=girl). You can just do "tab b4 [iweight=v005/1000000]" to get the distribution. You can select on a specific calendar year, for example, if you want. You can get the sex ratio by recoding b4 into a 0/1 variable and using logit regression. The standard sex ratio in demography is the ratio of boys to girls times 100. The following lines will give the ratio of boys to girls: "gen male=1; replace male=0 if b4==2; logit male [pweight=v005], or". The odds ratio given as the constant in this logit regression is the sex ratio (after you multiply it by 100). In biology the sex ratio is usually proportion female, which you get from the tabulation.