Again, my theory is that the windows simply get jammed somehow when in full up or full down position (or by sticky weatherstripping). For example when they reach the full up position, what stops them is nothing more than the top of the window frame -- by the same action as when someone's arm is caught in the window, thus preventing the window from lopping off the arm. If you continue holding the switch in the up position, current is automatically cut to the motor after a few seconds, and becomes inoperable for a few more seconds, by a heat-sensitive contact strip inside the motor. For some reason, something occasionally jams in the window drive mechanism (i.e., the "regulator" -- and NOT anything inside the motor), in effect duplicating the window's full up (or down) position, and until the regulator mechanism is somehow relieved -- by either partial disassembly, or vibration, or simply waiting -- the motor senses that continued non-overcomable resistance, and cuts the current after a few seconds of switch application. Try my method of always backing off full-up and full-down positions by briefly tapping the reverse button to relieve that stress on the mechanism. If that doesn't work, then I think it's likely caused, as Canyonero suggests, by weatherstrip binding -- which would be essentially the same problem (and if his isn't the permanent fix, then I don't know what is, because you'll likely never get everyone to properly perform my "reverse-switch-tapping" routine). And again, I suspect the landfill is full of expensive, fully operable window motors that mechanics were perfectly happy to replace, and often the problem cleared up due to nothing more than regulator relubrication, or "freeing up" the system by simply working on it..., or maybe the whole problem recurred after the car left the service bay with new window motors.