Pitch diameter plus voodoo.voodoo, n., A method of adjusting the actual centre distance from the theoretical amount of (D[sub]1[/sub] + D[sub]2[/sub]) ÷ 2 to something close, but not exactly the same.
So how’s that done? It depends on the gears themselves. For FIRST gearboxes, depending on the tolerances that I’m sure that I can maintain, I tend to say that I’d hope that the worst-case scenario would run freely without binding (or missing entirely!). That means, if you’re drilling gearbox plates on a drill press, without the benefit of a good solid vise, you may be looking at the possibility of missing by 0.020" or so, even if you centre-drill (how accurate is that, if it’s on the same drill press?), and exercise great caution. With the probable exception of the 32 pitch and 0.7 module gears, most everything *should *work under those conditions, but not necessarily well. If working with a good mill, perhaps 0.005" is more reasonable (actually, more accuracy is possible, but I wouldn’t count on it, without some decent expertise involved). Also, sometimes the gears have unusual shapes: motor pinions have undercut teeth in many cases, for instance, and some have extended-length teeth.
In any case, I tend to add an extra 0.005" to the centre distance, as a rule; but for the Bosch drills, I think that I added 0.007"; others (Joe Johnson, if I remember correctly, likes 0.003". He’s got better equipment than I did, I’m sure.) I’m not at all the most experienced one here, so you may want to consult a practicing engineer who may deal with this on a more regular basis. He’ll have his own voodoo reasons for the way he does it.
Now, if you’re just sketching, forget this. But if you’re going to build it, you might as well consider it now, rather than try to insert it in later. But: it may mean that the tangency constraints don’t quite work correctly in Inventor anymore (since the theoretical pitch circle circumferences aren’t necessarily tangent anymore). I like to set up an array of points, and dimension the separation between these, and then reference my gears based on these locations–I can then manipulate the points, rather than fighting with the tangency constraint (which is tough enough in Pro/E, and awful in Inventor). One last thing. If you fiddle with centre distances, make sure to check that the separation between parts is what you expect. A missed or incorrect constraint can cause things to miss or overlap unexpectedly, since they aren’t being constrained in tangency.