Cheryl, Jenne, Rahul
On Jan 8, 2020 we had a lockloss due to rung up violin modes. When the operators attempted to re-lock the interferometer, we suffered another lock-loss (at Guardian state 430, Engage_ASC_FOR_FULL_IFO). Hence, we analyzed the typical noise floor during lock acquisition and then tried to define a threshold before attempting full IFO lock. Figure attached below shows the DARM spectrum for 08 Jan 2020, focussed on the 1st and 2nd harmonics for the violin modes. In the plot (Darm.jpg, during the 2nd attempt for re-locking the IFO), the noise floor at Guardian state 430 is shown in red line (UTC 00:02:32). As one can see, several of the modes are rung up, ranging from 10^-15 m/sqrtHz to 10^-14m/sqrtHz (while the typical rms of the floor was around 10^-17m/sqrtHz). The IFO remained at this state for 4 hours, during which the rung up violin modes were damped (1st harmonic for this case). Later, several of the peaks subsided, except for a couple of them, following which a full IFO lock was achieved as shown in the ndscope.
Based on this experience we can set some guidance for the violin mode noise floor during locking attempts. Once Guardian state reaches PREP_ASC_FOR_FULL_IFO (429), at this point if no more than 1-2 modes are above 10^-15m/sqrtHz, while the others are below 10^-16m/sqrtHz then the operators can proceed to the next Guardian state. If at all the rung up modes are required to be damped then during this time the Guardian can remain at the same state.
Rahul-
We should discuss this more in person before issuing guidance to operators, to avoid causing or spreading confusion. In this case there is already guidance available, which I believe should be more reliable than what you outline above. Perhaps we just need to spread the word better.
Operators-
There is a dtt template that Jeff Kissel made several years ago which is linked on the violin medm screen. If operators run it they can check if any modes are rung up enough to saturate the AS_C QPD, and as the text in the template says, if the rms on that QPD is low enough that we aren't saturating it, we know that it is safe to engage the ASC loops which use that QPD.
It seems as though part of the dtt template that Jeff made for checking if the AS_C QPD is saturating has been deleted at some point in the editing that has happened. Jeff K notes in his alog that he put them in the svn so we should be able to recover: 37921
There's a need for two types of templates for the violins, one to use when they are very rung up to assess the success of moving past Engage_ASC_FOR_FULL_IFO, and one to use in low noise to evaluate individual peaks, or groups of peaks, that are ringing up during the lock, or modes that we are damping for the first time.
This alog has one example of the first type of template, and as an example of the second type, I've updated the 2nd harmonic template with the current H1 spectra, I've also been keeping it up to date as we damp more modes, and I've modified it to look all violin frequencies.