I can never remember which one to use but I see you have the result, left-justified. Try right-justifying it instead and see if that helps (ADCON1.7=1).

The resistance of your pot is 5 times the recommeded maximum impedance but I don't think it should make the readings unstable. Try increasing the sample time by changing the DEFINE ADC_SAMPLEUS 50 to something higher and see what happends.

By the way, what do you mean by jumping from value to another? Is that a count or two or how much is it "jumping"? A little filtering might be all that is needed.

And finally, since you have the ADC configured for external VRef, make sure you have proper decoupling on your VRef input.

/Henrik.