Still another possibility. Put the splined gear (and a precise ruler or, better, square) on your scanner and scan it in at high resolution. Enlarge it several times its real size (you will have the scales to make sure everything scales correctly in both directions and there is no directional distortion). You can then analyze all the parameters you want. Might want to input the scan into Inkscape (free drawing program available for both Linux and Windows) to do some of the measuring/drawing, etc. You have the number of splines, of course. What the OD is, ind the inner OD is, tooth width, tooth angle, curves/chamfers etc. could all be determined. At least pretty close.
You could include one spline and one arm (so you have both a positive image and a negative image) to compare. I think that is what I would do. Easier to get the Max OD from the spline, the min OD from the arm.
I have sent out a query to my guy, and will let you know if I hear anything back from him.
Buy the way,
http://www.emachineshop.com/
has a gear design utility (their software is free) that might make your visualization and generation of the specifications a tad easier.
Of course, if you were to use the Servo City hubs or shafts, or, for that matter, the molding technique, you would not be stuck with Futaba or futaba clones as your only source of servos.
http://www.strikemodels.com/ for instance, carries some very interesting servos, including a standard size servo that is a sail-winch servo, getting 6 turns, as I recall. About twice the number of rotations as the similarly priced, but almost twice as large, Hitec. I am not sure what spline they are using.
Again, I will let you know if I find out anything.
Cheers,
Wreno