I am certain it has to be mechanical. So put in Hiconn BOB with closed loop to motion controller system. This a mach 4 system. problem is that I don't get same steps per inch anywhere in any of the axis' . Originally thought maybe ballscrews just sitting there for long time and no lube maybe so did this .
https://www.youtube.com/watch?v=9gpITCJxcKo
https://www.youtube.com/watch?v=97ytxrFVkQs
problem was not cosine error with DTI. I have a problem where steps per can vary from 100,070 to 100,050 steps per inch on X axis for example.
Here is torn X axis belt from last video. Torn about .045-.05 from edge of belt
Attachment 471092 Here is mismatch from pulleys.
Attachment 471094 ballscrew pulley .562" inside of the edge of that mounting box. Servo pulley almost right on the edge of plate that mounts on that face in previous picture.
Attachment 471096. What really stinks is that I have no way of checking stretch of that belt when it is all put together. I am suspecting that I need to make 3/16" thick steel servo adapter plate instead of the 1/2" thick aluminum ones already on there. That might give me enough play so I can align my collet mounted servo axis pulley. Same situation on my Y axis, but at least there is a way to be able to measure belt stretch when it is all assembled. Now mind you the Y axis does have more stretch than it had before these servo adapter plates. Maybe that is where I am losing some accuracy. Here are pics of Y axis
Attachment 471100 Attachment 471102 Attachment 471104
Maybe timing pulley mismatch doing this problem or is it the stretch? Maybe something else?
Ballscrew pics
Attachment 471106 Attachment 471108