It would be better if you post motion related questions under motion.
I don't know who wrote the statement you point to on the web, and while it is sort of true, it is not directly available.
The gear mate defines the gear ratio coupling between to rotating components. If you apply a torque on one of the shafts, it will scale the torque correctly to give you the reactionary torque on the other shaft. Based on your gear pitch circle, you can then work out what the gear tooth force is (on average)
In reality for most gears, the tooth force varies with the rolling contact point (radial contact position changes) so you get cyclic variation in the force for each tooth. The only way to accurately get that function is using precise 3D contact since it correctly accounts for the geometry.
Even in ADAMS where there is a mathematical gear joint, you define the pitch circles and all it does is give you the resulting tooth force at that location (not the detailed force variation with contact point).
Also remember that mathematical joints ignore gear backlash so if you're concerned with that, 3D contact is the way to go.
Thanks Ian for the help,I will try this.