Dave,
This may be a classic case heat-to-heat penetration variation due to Marangoni fluid flow. This is a very common and widespread problem, mostly associated with 300 series stainless steels, that is the subject of hundreds of research articles over the past 20+ years. It is well known in many orbital and automated applications in aerospace, semiconductor and food process tubing installation, tube mills, etc, but I don't believe it is widely known in general industry. Here is a link to one good paper published by AWS:
http://files.aws.org/wj/supplement/Pierce/ARTICLE3.pdfSo what is it? In a nutshell:
1. If the sulfur content of your stainless steel is less than about 30 ppm (0.003 wt%), the surface tension driven fluid flow of the weld will be outward from the center of the weld, very efficiently transferring the bulk of the heat outward, producing a wide and shallow weld. Increasing current or decreasing travel speed is largely ineffective at increasing penetration because the strength of the outward flow is simply increased.
2. If the sulfur content of your stainless steel is above 100 ppm (0.010 wt%), the surface tension fluid flow will be inward converging and downward, diving the heat and penetration deeper, producing a narrower and deeper weld for the same given weld settings at the low sulfur weld.
3. Sulfur is the major "surface active" element affecting Marangoni flow, but other elements like O2, Se, Ca, Mn, and so on, may influence the surface tension gradient, either because they are also surface active or they form compounds with the surface active elements and thus negate the effect. So sulfur content is not a hard fast rule, but at extremes like 10 ppm versus 200 ppm, the effect is usually clear.
4. EWI produces and sells an oxide mixture that is applied to the weld joint as a light paste, and this ensures inward Marangoni flow and deep penetration. See this link:
http://www.ewi.org/products/ewi-deeptig.asp5. High oxygen content does the same as sulfur, and it has been shown that controlled aggressive wire brushing in air can produce enough surface oxide on stainless steel to improve penetration, however there may be other drawbacks from too much surface oxide in the weld.
6. In wire feed welds, a filler wire with high sulfur content can improve penetration in low sulfur base metal.
7. Sorting your materials by sulfur content, then characterizing the penetration and further sorting, may allow you to develop suitable welding parameters based on the given heat you are welding that day.
8. Again, you can try to improve penetration by holding the shortest possible arc length, or using 95/5 Ar/H torch gas, and these may help a bit, but the major tendency of tha low sulfur steel is go wider and wider as arc power or arc concentration is increased. See the conduction EB welds in the first reference.
This is not an April Fools joke, google it, it's very real.
Let us know what you find out with your problem.
Stan