On 25 Jul 2006 12:49:36 GMT, Bob Quintal Gave us:
They are all constantly moving. Even once "set" the wiper location has to "settle". It can be tracked too. It exhibits more noise until such settling has taken place. Different resistance mediums have different settling times. Wiper material cause variances as well.
Wirewound precision potentiometers do not exhibit this "problem". Only if they get set between two turns of the resistor. The wiper design typically won't allow this though.
Note that the pot that uses a metal wiper contacting a metal resistance medium (wire wound) is far less noisy than the metal wiper contacting a resistance film of "metal" or carbon slurry, etc.
So all of your parallels of thousands of locations don't make as intimate of contact as does metal on metal. That alone should tell you that ten thousand tiny pathways are going to be noisier than a few hundred strong, hard conducting, metal on metal pathways. That is where the noise comes from. Such tiny pathways are constrictive on an individual basis, even though there are enough to make "conduction" "easy", events occur in those pathways that can generate other than the original signal. We call it noise.
In the metal on metal, the path is not constricted in any way, and there are less collisions taking place to make noise.