According to Fig1 in the original post, DC values of IM horizontal signals were jumped significantly while the other signals were not.
So, I checked the signals of each OSEM and found that only H3 OSEM signals were jumped (fig1).
Since it is unlikely due to PR3 real motions because it is difficult to move only one OSEM signal, there may be some issues with the H3 OSEM.
This DC value jump was recovered about 8.5 minutes later after the jump (fig2).

We continued to investigate this phenomena is hardware trouble or not.