Jeremy, just so you understand, sparking is too hot. Welding is just under that. That said, the atmosphere of the forge will have to be hot enough to spark, but you will have to gauge the heat in the particular work piece with the amount of time it's in there. A thicker piece will take longer to heat up as would a billet. Testing the heat with a small demension scrap might be misleading since it would spark much sooner than your intended billet sized piece. Time is the element that you control along with the intensity of the forge.
Karl says it good when he says something to the effect of "heat X Time". But, if the heat is not intense enough, it wont reach welding heat no matter how long it's in there. After the atmosphere gets hot enough, then all that's left is time.