I thought some of this may be of interest to someone.
I suspect if you made several measurements at each station where you measured diamter and then used the mean measurement, the "bumps" on your graphs would smooth out a bit.
I think the bumbs are real. I think the way lines are made it is possible that with sub par qa/qc it could happen.
The DT10, I only measured one end and made a mirror for the other end, someday I will measure the other end, but you have to take the whole thing off of the reel>