Tom,
Each time you weld another layer the sulpher in the previous layer can evolve up through the layers increasing in concentration to the point that it becomes visibly manifest. And it will be worse with higher currents, greater depth of fusion, and greater dilution. So, its not a question of interpass, its a quesiton of the beginning content of S (you're gonna have high dilutionin your first layer), how much of each layer is liquified, and becomes a part of the subsequent layer (in other words dilution), causing the volume percent to increase with each pass. The thing is, if you're welding Ni on carbon steel you will always have NiS (or maybe NiS2-or Ni2S-or some variant thereof-I don't remember).
The question is, how much and can you see it. If this is the problem (and keep inmind this is only brainstorming) then by going to a cleaner material like 516 the problem would likely disappear.