[AdamHeard]

I view weight as a limited resource on the robot, you can only have so much of it.

If you could buy drive gearboxes that worked great, were reliable, and fit your application perfectly for $100, why pay $200 for something that does the same thing? I feel the same way about weight, if you can cut weight WITHOUT killing reliability (which should be obvious, you should be making parts strong enough regardless), I see no reason not to.

Being lightweight in component design allows you many benefits; you can fit more on a robot as it all weighs less, you can keep the robot weight light for better driving & decreased battery use, and you can make up that weight with ballast real, real low on your robot to put the CG exactly where you want it.

I've seen a recent trend on CD that many people feel that removing weight through machining operations is unnecessary, and even a foolish waste of resources. It is a worthwhile operation, and doesn't even *need* to be a CNC'd accurate operation, much of the cosmetic pocketing you see on gearboxes with fancy curves and such can be approximate with a manual mill (if you know integrals, imagine you're approximating the pocket with a bunch of rectangles X units wide. X being an endmill small enough for a decent amount of material to be removed, but large enough as to not take forever). You can even draw the pattern you'd like, and drill or dremel material out; just don't take more material out than you designed in the drawing.

Before trying to machine weight out, assuming your fabrication resources are limited, you should consider thinner and lighter materials or just try to make parts smaller overall.