Hi Torch
Attached is a picture of the key width which I measure as 3.20 - likely the 3 mm you refer to. I also tried to measure how proud the key was from the shaft at 2.03 mm. That's pretty suspect if the shaft is 10 and the aggregate is 11.5 as shown earlier something does not add up but sounds like you have that all figured out.
I'm not surprised. Digital calipers tend to lull one into a false sense of accuracy. The display shows a precision of .01mm, but the reality is the best accuracy one can hope for under ideal conditions is +/- .03mm -- and that's between the jaws, the one aspect that can be zeroed out first, when measuring clean flat things.
The accuracy of the internal jaws is worse (but improves somewhat as the hole diameter increases) and the accuracy of the depth gauge is really more of a suggestion than a real measurement when measuring out-of-position against a surface that's curving away from you. It is
extremely difficult to hold the calipers perfectly perpendicular in two axis simultaneously. So let us assume they are using the standard listed in Machinery's Handbook for a 10mm shaft -- 3mm.
I should note the 17 x 19 mm shaft seems to ride in a shallow recess on the actual transmission which you can see when you look at the shaft measurement pic earlier.
Yes, I noted that earlier. I allowed 9mm (IIRC) in the above sketch. You should measure that depth to be sure -- the caliper depth gauge accuracy will be sufficient if we add a mm or two for safety.
I am assuming the fender washer lugbolt was talking about was riding on the top of the pulley under the nut.
Yes. His concern is the length of the threads may end up a bit shy of the pulley hub. The washer ensures there's still a thread or two left when the nut is tightened.
So the only other measurement required is the thickness of the actual weld-on pulley where it meets the hub. At a guess, I'd say 3/16" to 1/4" from the photos on PA's website but I've been wrong about such things before -- twice already in just this thread alone!