Does anyone have measured movement tolerances of V-slot when using metal wheels? I don't have accurate enough measurement equipment to find this on my own.
It's effectively zero because it's a symmetrical set of 2 planes sitting flat on top of one another and you pre-load the wheels with the eccentric spacer. The main play/slop will come from the bearings and that would be well documented by the manufacturer and you can get more precise bearings if you need less slop. The actual variance in the thickness of the extruded aluminum is the one thing you can't do much about and it's the same as the 80/20 T-Slot aluminum specs so you can find them there. I can almost certainly guarantee you the variance is on par with extremely expensive steel rods and linear bearings. Other tolerances will be determined by the "run-out" of your router(different for every router), the stepper motor degrees per step resolution(different for every stepper and is 1.8 deg/step on many motors but you can use micro-stepping to get finer resolution if your board supports it), and then there is the overall design rigidity and material properties that can change based on temperature(see coefficient of thermal expansion), force applied(see flexing), and acceleration(see backlash). So for example, you might have a router that has a loose bit-holder that contributes 0.1mm of error into everything you cut... and then you have loose bearings that add another 0.1mm, and finally your aluminium rail is "bumpy" which can make things off by 0.01mm. So theoretically the most accurate you can be is to 0.21mm. The rest of the things I mention could be theoretically compensated for in software or by the operator.
