cbarnett - you can remove the parts in bold and my statement remains valid. You have ignored the whole point of my post, but I will frame it in the context of yours in the interests of civil discussion.
You are correct that accelerometers are essentially cantilevers or springs (either macroscopic of MEMS, that part is irrelevant to everything other than their accuracy and size). Are you trying to say that in vacuum free fall an accelerometer can measure gravity? This is completely untrue - as I said if this were possible then an accelerometer would be all you would need to determine an absolute Universal rest frame, which is forbidden by relativity. Even if you allow such a frame to exist, an accelerometer doesn't have some magic mass bound to that frame to measure against.
An accelerometer
can not measure gravity in free fall. It
can measure the acceleration caused by drag, but in vacuum free fall it will show
exactly zero acceleration, even if you are in orbit.
If you don't believe me, then wiki also supports this line of reasoning:
Conversely, the device's output will be zero during free fall, where the acceleration exactly follows gravity. This includes use in an earth orbiting spaceship, but not a (non-free) fall with air resistance, where drag forces reduce the acceleration until terminal velocity is reached, at which point the device would once again indicate the 1 g vertical offset.