Thanks for the input.
I think where we are stuck is the fact that the problem is very intermittent. We run a series of parts with absolutely no problems and then run another set where these defects show up. The machining, prepping, cleaning, etc - all those steps are followed to the t. In regards to the resulting weld bead, even on the ones that show defects, the bead appearance actually looks good. The resulting bead isn't black, gray, or anything that would indicate contaminated shielding gas.
To back up a little bit, when we first saw this problem, we were using a WT-27 torch with a conversion kit that allowed us to use the smaller front gas lens parts. We had been using this setup for a while with no problems. Once we started seeing these defects, we switched the front end parts back to the large diameter gas lens - the one that is about 90 bucks from weldcraft. (part 49V24). This didn't seem to improve things much. However, I just looked at the mesh screens on the gas lens and they do look a little worn. I'm not really sure if this is causing the problem or not, but I'm going to buy a replacement screen kit from arc-zone.com and put some new ones in. At least that's cheaper than spending 90 bucks for a new lens. I'd be surprised if this made a big difference because i'd assume that if the screens were worn, you'd see problems with the previous weld passes. However, I guess I can understand from the previous comments that for the initial passes, the groove holds the gas coverage pretty well, but once you start to get closer to the cover passes, if there is any restriction in the gas flow from the gas lens, it might create a problem for those passes since there's no joint to hold the gas. Is that basically the idea?
In regards to cup sizes, stick out, etc this is what we are running..
7/16" min cup size
30CFH minimum
7/16" max stickout, but use 3/8" when groove dimensions allow