I had a servo (a Futaba S148) that used to "twitch" like this alot. It was a little disconcerting..... I found out that if I was using a battery that was "close" to the 4.6V mark it would flip out OR if it was mounted NEAR the servo. I also found that the wiring in the unit was pretty twisted up (the "electronic spaghetti" look...). I took my time and a little scotch tape + zip-ties and made sure to put a good bit of distance between all the wires and tried to avoid any "snaggles" with some small diameter conduit.

It solved my problem with it. But then again, sometimes servos just twitch like a meth junky straight out of the box.... that's when you send it back for a new servo....... in my experience quality control at the factory can only go SO far (i'm an industrial mfg engineer and I can tell you that near the end of an 8hr shift, our quality issues CONSISTENTLY spiked due to worker fatigue) but MOSTLY b/c they just got SICK OF IT, started "clock watching", and really quit giving a hoot if it was correctly assembled or not! I imagine servo building is alot of AOS chinese mfg (AOS=*$$-on-stool mfg!)

Let me give you an example (and, by the way, this is TRUE!): You just went to a MAJOR dept store and spent $450 on a hot-snot brand new, shiny red lawnmower. You took it home, unpacked it, cranked it up, made a pass on your front lawn, and THEN noticed that something was "not quite right"..... yep, we ACTUALLY sent out NOT ONE, BUT MANY very well known and respected brand name mowers WITHOUT BLADES!!!!!!!!!!!!!!! I would have given $100 JUST to see the look on that guy's face when he realized that his BRAND NEW mower HAD NO BLADE!!!!!!!!!! [X(] I would have pee'd myself laughing! I almost cracked up when the plant manager went into an apoplectic rage during a meeting when HE found out about it! It was HYSTERICAL !!!!!!!!