Federico's reply to the old entry suggests that it might be due to a switching power-supply of crate 39 that operates around 37kHz, which somehow aliases down to 3 kHz. Looking at some random channels, I found the line in at least one channel of every ADC that is housed in that crate (Bs_MoniCB01, MoniCB01, Bs_MoniCB02, MoniCB02, MoniCB03 and Det_Env01), see fig 2. The line is seen best in Ca_PR_RM_CoilD, but since Gc_PR_z does not show the bump, it is probably picked up somewhere in the cable or in the ADC. How the noise couples to the dark-fringe is not clear. We could use one of the portable magnetometers to obtain more evidence against the suspect.
