During the QPD connection check, I found that the segment #4 of the MC2_TRANS showed a big offset of approximately 4300 counts. According to the 10 min trend, this offset has been there from the beginning (Jan. 16th [1]). I disconnected the signal chain at various points to find out which electronics is causing the offset. It seems that the AA board [1][2] is the one. I disconnected the SCSI (which connects the AA board and ADC) to see if it removes the offset and indeed the offset went to zero with the SCSI off. I will have a close look at the AA board tomorrow and hence may need to pull the chassis out of the rack if necessary. Note that because of this investigation the IMC was intentionaly unlocked from approximately 8PM to 9PM.
[1] LHO alog 5127
[2] D0902783
[3] S1102790