So how can this happen? The primer plate must be misaligned when the primer drops into place, causing the primer to rotate. More investigation revealed that the primer feed is not indexing correctly all the time. The black marks show that the primer wheel is not always stopping in the correct place, see offset circles.
There is no user adjustment for this condition that I know of.
In any case I talked to Dillon and they sent me a new primer system which I will be using for this evaluation. In addition to documenting all the adjustments I will try to put a video camera on the primer indexing bar to see if there is a detectable problem there. There should be, but it might be so small that it isn't easily observable. The primer disk only has to be off a few thousands for the primer to get tipped. The next question is how does the primer stay in that position with the shell plate and other things that it has to rotate between? Must be enough clearance to permit it to happen. Even if the shell plate were grossly under tightened (or in this case not even present) the primer feed should have taken care of it, but it didn't. Could it be something pushing down or up on the primer disk, along with a bit of powder that sets up enough clearance?
This is actually encouraging in that identifying the actual problem is the first step in fixing it. And this is NOT an operator problem if it is what I have seen so far.
This is still just speculation at this point. Much more testing will reveal whatever is going on with more certainty.
Also I don't see this primer problem reported for the 550 or 1050. They have differen't primer feeds than the 650, so whatever there is that is differen't must be responsible.