To add to the flexible radio tray/servo failure discussion.....
Please note that
IF flex was a problem you
should some of the following:
elongated servo mounting holes on the servo mounting ears AND/OR the radio tray,
the servo mounting screws
should be loosening seemingly "on there own"
you
should see shiny, polished spots on the radio tray under the servo mounting ears.
I think it's rather unrealistic to think that the servo case (small, effective load bearing box section after all) is absorbing the flexural load without the case to tray interface showing evidence of movement.
Regardless, bottoming a chassis out (pancaking a chassis, etc) would not be "flexing" the chassis as is theorized here as the failure mechanism of the servo. Bottoming out of the chassis would be the most efficient method of transmitting shock to ALL shock critical devices (servo’s, receivers, glow plugs, etc.). Now, end-over-end cartwheeling would much more likely throw a chassis in an out of maximum flex.
Personally, I'm leaning toward the incorrect linkage spring/ball cup-coil bind issue as the cause of the failures.