This morning during maintenance time we have complitely unplugged from mains supply all the devices located into the half-size rack (the one at the entrance of CEB, see left photo); they are now powered via a dedicate mains AC power generator, set at 230Vac 53Hz (see right photo).
The plan is to leave it until next maintenance, if no problems arise.
The goal is to better understand the origin and coupling of the 150Hz and 350Hz (harmonics of 50Hz) seen in many signals.
Preliminary (and promising?) results.
we focused looking at LSC_DARM and SSFS_B4_Err-pre-50kHz for two ITF locks:
we zoomed around the fundamental and up to the 7th harmonic (350Hz --> 371Hz)
There are interesting reduction effect (mainly in SSFS but also in DARM) even if it is unclear whether a reduction in DARM is due to the new mains frequency or to other reasons (different lock conditions?).
Well, in all cases the 50Hz (and harmonics) line is REDUCED in DARM, not increased.
Worth noting that there are notches in some signals (e.g. SSFS), so the height of the 50Hz and harmonics lines could change if the frequency shift slightly, as happens during the day.
There are two way in which the mains can enter some signals:
For the first case it would not be difficult to move almost all power supplies away from the rack (the magnetic field falls with a square law, at least); if this displacement does not have evident results, it could be a ground coupling: but this is more difficult to manage.
Attached: FFTs around some interesting frequencies
Having some locking period after the AC Mains moved from 50 to 53Hz, we now have more statistics.
Unfortunately it seems that the height of the fundamental (50Hz) and its harmonics change from one lock to another: what seemed to be a reduction at the beginning, now it is no longer true.
Attached a plot of FFT around 350Hz (seventh harmonic, just an example) with the RED curve of May the13th (before the 50 --> 53Hz change) and all the remaining colors AFTER the change.
Sometimes the 350Hz line is lower (for example, the BLACK curve), but sometimes it has the same amplitude.
More locking period, more statistics ...
50Hz line changes a factor 25 during some locks of May the 18th
(see attached plot)
Today at 09:42:00 UTC we have moved the generator frequency from 53.00Hz to 53.14Hz
This morning we have roll back to the original Mains UPS distribution, switching OFF and unplugging the power Mains generator that was set at 53.14Hz
Please note: while powering again devices, one AC-to-DC power supply emitted a bad sizzling noise, we left it unplugged. The unpowered devices are "New Focus" picomotor drivers, normally not used and moreover polluting DARM with a line at 107Hz
