Thanks guys for the kind words but I wouldn't describe it as "all the info".
Here is what I know... JB weld is a GREAT repair to get back out of the woods or to finish a race. You have to clean the cases very well but it can take to the aluminum and hold for a short period. It will eventually separate from the aluminum and begin to leak again. JB weld (or any of the other epoxy based "repairs") is only temporary.
Tig welding is an absolute repair method by melting the base metal and adding an alloy rod to build up over the crack and permanently repair it. Unfortuntely there are a few issues with this method... First, Tig welding requires argon sheath flow to keep the base aluminum from oxidizing as the plasma melts it into a weld pool. Aluminum is slightly porous and absorbs petro chemicals (particularly heavy oils such as 2 stroke oil and transmission fluid) and that oil must be baked out of the aluminum before a repair can be made. Second, the aluminum that yamaha used to cast the cases out of is particularly poor. It's an alloy blend which does not readily take the filler rod alloys without some coaxing. It can be made to take the tig weelding rod but only after some work. Third, the resulting weld is considerably harder than the base material. If the welder slips and catches the mating surfaces and hardens that area, lapping the cases may be inneffective in smoothing the case back for a good seal.
There is a third repair method which is equally as permanent as tig welding without some of the drawbacks of Tig welding (but with a few drawbacks of it's own of course). Aluminum "brazing" using an alloy rod and a moderate heat source (Mapp gas or oxy-fuel) to repair the base metal. The material is marketed under the names "alumiweld" and "duraweld" and many other marketing names. The material isn't aluminum but a proprietary alloy which will take to pretty much any non-ferrous metal.
It is extremely sensitive to heat and not cheap to use. It needs between 800°F and 1,000°F to bond properly with the aluminum. Anything under that temperature and it simply won't melt, anything above and the alloy oxidizes and balls up. The trick to the material is heat control. The final product is slightly harder than the case aluminum (marginally so) and slightly more brittle (again only slightly so) but it matches the color and strength well. Machining it is easy and it sands and burrs well once it's cooled down. It also responds well to sand paper.
The material is available at harbor freight under the name "alumiweld". The package contains 8 rods and costs $15. A similar material is also available at this website:
Aluminum Welding - Aluminum Repair - Aluminum Brazing - AluminumRepair.com
Of course they claim their product is better than the others... I will be able to quote about that this week (I ordered a pound of it for testing purposes) but I believe it probably works about the same as the other products...the only hiccup is that you have to order 1# (34 rods)at a time for $65 plus $15 shipping and handling. The alumiweld can be had for $15 (granted the price per rod is nearly the same but it doesn't require you purchase a huge pile you may never use)
Another trick with both materials is the "tinning" you have to do with it. There is no flux, you actually use the rod itself as "flux" by "rubbing" the rod with a wire brush to force it into the pores of the aluminum. Once some of the alloy has taken to the aluminum, the repair can be made by pooling some of the alloy up carefully heating the nearby metal so you don't burn the alloy.
I've had satisfactory results patching all but the largest of holes (king size sharpie marker sized hole between the crankcase and the balancer shaft area.