I've tried connecting up the center tap of the servo pot to an analog input of the arduino and measured it with the oscope of MRL.  fun...

here's some images of the with and without the ground line connected.. i'll leave it up to you to see which is which.  (I was manually turning the pot myself.. watching what happened on the analog input pin 0 of the arduino that i connected up...)

looks like the PWM comes through pretty clearly as noise if you don't connect the ground..