The jump you see in 2014 is due to the CPS 2014 ASEC experimental design. When analyzing the 2014 data, you will need to select either the 3/8 file or the 5/8 file using HFLAG. Otherwise, if both parts of the file are analyzed together and weighted using ASECWT, the resulting estimates will be twice the US population, hence the large increase in your sum of weights.
As for 1970, the difference is likely caused by some of the comparability issues prior to 1980. If you haven’t done so already, I would recommend comparing your computed aggregate household income to HHINCOME to see how it compares.