I have been studying my 640-3 pictures and I think I've spotted the mechanism for the failure when it occurs.
The lug on the flag that I ground off lays in a track in the hammer. That groove has a detent in it. When the locking cam pushes the flag to the locked position, the lug engages in that detent.
When you rotate the cam to unlocked a small spring pulls the flag back down and returns the lug to the track. This small spring is the only thing keeping the flag in the unlocked position.
I think what may be the mode of failure is that teeny little spring coming unseated and letting the lug pop up where it's not intended to go. The angles are such that the flag has to be moving around a bit every shot and heavy hitting rounds will make it move more. Because its intended to have enough mechanical advantage to keep the hammer from moving while locked, if it gets to where its not supposed to be, it will lock it up.
I would be curious to see how many revisions of this feature there are and the drawing histories at S&W to see if they've been quietly addressing it or ignoring it.
Viewing or use of this blog or any contents or links contained herein by any person or entity within the confines of the states of Arizona and/or Tennessee is prohibited .
Violators of this policy agree to hold the owner of this blog, its contents and all links contained herein, harmless to any harm or offense taken or perceived.
Permission to use any content on this site is explicitly denied to Robert Farago, his family, his friends, his associates, his pets, and his employees and/or employer and/or their employees either in part or whole.
This blog is not Strunk and White compliant; read with all due caution, diligence and pity.
Due to ITAR regulations, Godless Foreigners™ are required to avert their eyes.