LH popped off before I even stated to suspect anything... But looking my datalogs, knowing now what to look for, from autumn '14 vs spring '15 I might know the day when the soon-to-pop started leaking. Now plastic tanks are off but I will test ICs after aluminium tanks are on.

 

My issue with the 997.2s was not the end tanks. After 4 hard years at the track they were tight as a drum. I actually like the composite end tanks as they have much better insulating properties compared to aluminum. I think the problem with the end tanks is when the tabs are not fully engaged, many of which are not right out of the box. I crimped each tab individually when they were new. It's the cores that were leaking in my case, both in the same spot as you can see in the link below. Not enough to show up on data logs but they leaked. My friend's 997.2s did the same after about 10 track days with VTGs, again leaking in the same areas. Remove your ICs and pressurize them to 25psi. They should hold without any leak down for several minutes. The leak in the core will be on the intake side where the heat gets concentrated. I would make sure your cores are good before spending time and effort welding on aluminum end tanks onto leaky cores. Good luck..

https://www.youtube.com/watch?v=QECR...ature=youtu.be

 

Thanks for that link, very educational. Yes, you're correct about those loose tabs. If tanks have room to move, they'll wiggle themselves away with boost surges. Glue acts like additional safety wire on top of them.