I had a chance to get into the frozen tundra of the great white north, I mean the shed that I pretend is my shop. It was the sketchiest clamping I'd ever done. I had a 5" angle plate, the base of the indexing head is 6", so I picked up some 1-2-3 blocks. I did a rear clamp, and an upper clamp (the Harbor Freight mini mill forced me to install it at an angle on the table, which isn't a big deal since I'd be milling on a circle.
I calculated the outside to the inside edges for the bolt circle, found the edge, and the end mill for the pocket loosened the backplate and spun it off, so I had to reset again. The next attempt, I clamped to the backplate, as well. I only had 6 clamps on the mini mill. It was clamp the back plate, mill the pocket, unclamp the backplate, and rotate 120 degrees (twelve hole index ring meant move four holes in between each pocket). On the Harbor Freight mini mill, I didn't have enough space to use the drill chuck, so I could only install a center drill in a collet and get the holes started. Those holes were finished on the drill press.
The chuck register surface was not quite small enough (I'd rather go large than small, actually). I threw the back plate into the freezer for an hour and a half. I put the chuck into the oven after that for 15 minutes or so. Those temperature differences should give enough deviation in expansion/contraction based on the heat differential. I grabbed the parts out of the oven and freezer, and bolted them together.
Next step to make this perfect is to chuck up a rod in the lathe, then bear down on that with this chuck so I can take some material off the back and not have so much overhang. It will also give me a chance to do some thread clearance so it will fit perfectly on the dividing head.