Switching digital servo with mosfet / dead servo

I'm using some Hitec 7950's in a robotics application with arduinos. The digital servos in general will not release their last position when the signal stops - they will either hold it or return to a "fail-safe" position. This is not necessarily desirable, and so I've added a logic level mosfet driven by one of the Arduino pins to switch off/on power to the servos completely .

This works but one of the servos died completely after about 5 minutes of testing. No warmth or anything like that, just went dead. I somewhat imagine that spikes from the switching may have killed something in the electronics of the servo. The mosfet circuit is still fine and will run other servos & motors.

I already had a kick-back diode across the servo power lines, so now I'm thinking a capacitor might be useful. But I'm not sure of the best place for it. This morning I added a 47uF cap to the Arduino pin to soften the rise to +5 on the mosfet. The alternative would be I guess to put the cap across the servo power leads, parallel to the diode. I'll find out when I get home tonight if this thing survived, it will be running on its own all day to test:)

Just wondered if there are any thoughts about this, whether the cap is a good idea, where a cap is best placed, what size makes sense, other ideas of how the 7950 servo might have died... Thanks!
-Mark M.
http://www.markmalmberg.com

I'd want to look carefully at the spec sheet for those servos to find whether there are any gotchas. For example, it's conceivable that they don't like having a signal line driven higher than the power line, and what you're doing could easily cause that. In other words, to me this feels as if you're using the servo incorrectly and not just a noise issue.