It seems I may have figured out one (possibly both) of the problems in my last update. Using about 1/3 to 1/2 full speed on the weapon motor, I was able to throw my wood dummy a respectable distance. The key here is collision speed; in yesterday's tests I "crept up" on the dummy so all that happened was the weapon would scrape at the sides. Today I pushed it into the dummy at a speed comparable to combat movement which apparently is enough for the blade to "catch" instead of "scrape."
As I said, this was with the weapon running much less than full speed, which cut down on vibration a good bit. I'm headed to McMaster-Carr in a few to pick up some bearings, which should help, too.
Since I don't need full speed for effectiveness, I need to investigate programming a max RPM into the ESC. Failing that, I'll put some sort of mechanical block on the transmitter's weapon throttle so I don't redline it.