You have somewhat less reason to be concerned about P2O5 in materials added post-kiln. The key principle is to target the % of readily available soluble sulfate as a set retarder. As a 1st step, just work on the % sulphate that you are adding. Next,consider that quite small amounts of soluble phosphate can retard setting compared to the times observed with natural gypsum (even though remaining within the range permitted by national standards), though the degree varies with the particular phosphate compound concerned. For prudence, to avoid customers experiencing a sudden step change, various reports recommend starting off by using a 50:50 mix of phosphogypsum with your traditional supply and then adjusting this by monitoring the effects experienced in practice by customers - not just those recorded for standard laboratory tests.