Welding Tips and Tricks

I have a project of welding some 17NiCrMo6-4 (AISI4317) steel.
The idea is to weld a motorcycle engine sprocket to a machined bar of annealed AISI4317 bar.
Problem is that it's difficult to manufacture the internal spline on the bar itself, so we get the ready made spline from the sprocket and weld it on the annealed bar.
The sprocket is case hardened. I will machine the teeth of the sprocket on the lathe and also i will make a chamfer for the weld. Look at the attached CAD image to see what the sprocket will be like after machining.



Yellow areas are untouched/hardened areas and red are the machined areas.

Next we will get the bar and machine it in the shape you see in the attached CAD image (adaptor-silver part).



The two will fit together and will be welded all the way round, like this:



Of course i will preheat the assembly @300degC prior starting welding it.

1) Do you think i can weld the 2 parts without annealing the sprocket?
I assume that the machined areas of the sprocket (red areas) will unveil the "soft" core of the sprocket so i will be able to weld the two with success.

2) Do you think that i can make a decent stress relief treatment by putting the welded assembly in a simple oven for 6 hours at 160 deg.C ? That was the reccomendation of a local heat treatment shop.
But i read that AISI4317 need to be stress relieved at 630-650 degC.

3) Finally, what type of filler rods do you recommend me to use? ER80S-D2?

from what I remember from machine shop way back in high school. to case harden, low carbon steel is packed and heated in a high carbon material where the carbon migrates into a thin top layer of the low carbon steel. This makes a thin layer of high carbon steel on the outside that can be hardened. The depth of the layer depends on how long it was kept at heat. Check on the web but I would think only .020" would be hard and the rest parent metal.

I would expect that the tooth area was where they wanted the hardness most so machining off the teeth should be difficult. Try grinding .005" to.020" off the surface if you can't break the surface on the lathe.

Mark