I've come to the realization that this topic is spread out over 40 pages now and probably too much material to digest at one time. There is also some incorrect information that makes it confusing. Furthermore, this has been a slow learning process as all information and data were not readily available in the beginning. Based on input from you all, pictures and research, I have come to an understanding of the problem and wanted to share my conclusions with everybody. I may be wrong on a few points, but overall I think I'm pretty close. At any rate, this should help keep the discussion going in the right direction. Here's my summary:
The problem is metal fatigue caused by too much stress, heat, and movement in a small area during expansion & contraction cycles (this occurs as the metal heats and cools). To simplify things, it's similar to the kind of fatigue you see by bending a paperclip back and forth several times in the same spot. Eventually the metal "work hardens", fatigues, and breaks. As far as the refer goes, it's a little more complicated because there are several factors at work here. The metal of the element pocket together with it's weld are much thicker than the boiler tube metal. The result is the boiler tube expands and contracts at a different rate than the adjacent metal (the weld and element pocket combination). Over time, the metal right next to the weld gets an extreme work-out (just like bending the paperclip back and forth), and will eventually work harden, fatigue and crack. The lower wattage elements do not stress the metal beyond it's capabilities, like the higher wattage ones do, but unfortunately, changing the element to the lower wattage one after the pattern has been set probably won't stop the process of failure.
Another issue is the way the element heats. From testing I have concluded the element has "hot spots". The most extreme one being near the area in question. The reason gas operation puts less stress on things is because the flu tube has a more uniform temperature throughout the heat transfer area (the 5 1/2" weld between the flu and the boiler). This puts the stress over a larger area instead of focusing extreme heat in one small area. I may be wrong on a few points, but this is my understanding of the problem. Hopefully this will help explain things better. Tim
