Another thing that happened during the maintenance is that the north end building has been cleaned: https://logbook.virgo-gw.eu/virgo/?r=64413 . Could have something been damaged, or a cable disconnected during the cleaning that causes the 48Hz bump to be high?

Unfortunately, a dedicated BruCo run with all the channels highlighted no significant coherence for Hrec at the frequency of the line and bump: link to results. Also for the cross-bar resonance mode itself, we usually don't see any coherence during usual operation conditions.

About the red calibration flags on WI, since yesterday slightly after 16h UTC: they indicate that there is 50 µs munsynchronisation of the WI DSP (see the jump in the bottom left plots of figures 1 and 2, and figure 4). This is the case since when Valerio synchronized the WI DSP to fix the red timing flag (see figure 3).

--> for calibration point of view, the WI DSP must be resynchronized before the next Calibration data taking (tonight).

--> WI  is also actuated for the 50 Hz damping loop, but it is not clear if this issue of synchronisation  could generate such a bump around ~48 Hz.

The excited peak pair, 48.08 and 70.02 Hz, best match the crossbar modes of the WE, as measured with magnetic injections.

Here the table from VIR-0300A-24:

I remind the direct measurement done by Paolo in O3. They roughly match the above, except that NE (47.75 Hz here, around 48.4 Hz in O3).

We found the source of the 48Hz bump.
It comes from a WE motor crate, the one that drives the MH.MA.1 and MH.MA.2 motors. We don't know why this new coupling mechanism has occured since Tuesday, but removing the mains cable from an unpowered crate was sufficient to eliminate the noise.

We leave the power cord reconnected but the front panel switch in the OFF position (it is equivalent).

I'm just adding a few pointers to help understand why this issue occurred. In fact, the same exact problem already happened on October 11 2023. These are the relevant logbook entries: #62060, #61792, #61959. 

During yesterday's investigation, we injected a far-field magnetic sweep with the aim of cross-checking and identifying each resonance frequency of the crossbars.

Here is what came out:

It is interesting to see that after the “disconnection” of the WEB motor crate (and disappearance of the noise) there remains something in Hrec that seems to be associated with CEB (47.52 Hz and 47.74 Hz) and WEB (48.09 Hz). This is an indication that we might gain something more in terms of Mpc if we could find and reduce some other sources.