I must say that I also find it difficult to give a proper interpretation of your results, but here are a number of things that could help you understand the situation:
1)When welding the "button", you will have very close to an ideal weld metal reading because with 5-6 layers, you will have negligible dilution from the base metal. In addition, you probably have at least a thickness of 10 - 15mm of weld metal. This means that your ferritscope measurement will not be significantly influenced by the base metal.
2)When performing the "actual joint", you will no doubt have significant dilution from the base metal. In addition, the readings from the ferritscope will be influenced by the base metal.
3)I am not exactly sure what your material specification is, but looks like a casting spec. If this is the case, then there can be significant variations in the composition of this base material if it was not properly heat treated. This can significantly influence the weld composition and ferritscope readings.
4)An interesting piece of info. that I found from Avesta literature: Their 317L TIG/GMAW wire has a typical FN of 5, while their 317L SMAW electrode has a typical FN of 10 even though the nominal chemical composition is the same. (If someone can explain this, then please tell me!)
Hope that there is a usefull idea that can help you in the above.
Regards
Niekie Jooste