Since there's been no response to Lawrence's post regarding process, material, and thickness, I'll respond by saying that in my mind the weld crater is a defect. As we all know, D1.1 allows unfilled craters with the exception of those outside their effective length in intermittent welds. I guess in this instance the crater is not considered a defect by AWS. Even though this is allowed by D1.1, I just have our shop guys fill them all in as they go. This eliminates any questions about whether the welders can leave craters on their welds or not. It also eliminates me having to worry about crater cracks, which can form in the weld pool as it solidifies and shrinks, causing the crater to tear due to shrinkage stresses. I don't know if there are any "textbook" methods for filling craters or not, but they can easily be eliminated by holding the arc at the end of the weld for a short time instead of quickly removing it, which is what our guys do in the shop. By holding the arc at the end of the weld, it allows the crater to fill up and helps reduce shrinkage forces. If there is a situation where craters are left and need to be filled after the fact, and if no cracks exist, the crater can be filled by welding from slightly beyond the crater back into the crater and towards the weld. I've seen single plate shear connections (AISC shear tabs) welded to supporting beams by other fabricators with unfilled craters at the tops of the welds. Very dangerous.