A good analysis was to see at, what particularly moment, did the cards "explode". Maybe the problem was the airflow. Again, the problem is not the airflow itself the the amount of airflow needed to cool down the VRM's that are already at it's limit of temps/efficiency.
Sometimes... 2ÂºC cooler or warmer make the damn think blow off or just keep "screaming with pain"
I know that 2 of them were at post screens ... so fan rpm down to minimum ... if VRM's temps were alarming ... and the fan rpm going to min ... that could cause the temps to ramp up since at post, theres a small increasing of power draw and that could be the "switch" to make it short-circuited ...
Again, this is just me thinking out loud
eheheh ... i'm really intrigued on how this will be solved Edited by Pedros - 3/30/11 at 9:46am