\chapter{Reducing differential motion of aLIGO seismic platforms}
\chapter{Control of seismic platforms motion and LSC offloading}
\label{CPSdiff}
During 2019, I spent some months working at the LIGO Hanford site (Washington, USA). This experience allowed me to be critically involved in the complicated life of a gravitational-wave interferometer. In particular, I was given the opportunity to study how to improve LIGO performances at low-frequency, focussing on the reduction of seismic motion of the platforms where the optics are located.\\
In this chapter I will demonstrate how we can modify seismic control configuration of LIGO: in particular, this study should help reducing the differential motion between the chambers, making them move in sync, and help reducing and stabilizing the rms motion of the auxiliary sensors, through an LSC offload. The final goal is to obtain different and possibly better performance for seismic motion stabilization, faster and longer locking mode and, ultimately, more gravitational waves detections. The detailed computations included in this chapter are original and partially presented to the LIGO community and stored in LIGO DCC \cite{proposal}\cite{technote1} .\\
...
...
@@ -325,7 +325,7 @@ In order to compute the platform motion for the single chambers in isolation and
\centering
\includegraphics[scale=0.45]{images/bscblend.png}
\includegraphics[scale=0.3]{images/hamblend.png}
\caption[Blending filters]{Plots of all possible costs built with different combinations of blending filters. The orders of magnitude are indicated by the low and high pass indices l and h of the binomial filter in \ref{binomial} and the are going between l =[1,4] and h = [1,4]. The plateau on both plots is given by the fact that the SC filter is dominating over those frequencies, which induces issues in the choice if the blending frequency. This makes a new way to evaluate blending filter in presence of the SC filter necessary.}
\caption[Blending filters]{Plots of all possible costs built with different combinations of blending filters. The orders of magnitude are indicated by the low and high pass indices l and h of the binomial filter in \ref{binomial} and the are going between l =[1,4] and h = [1,4]. The plateau on both plots is given by the fact that the SC filter is dominating over those frequencies, which induces issues in the choice if the blending frequency. This makes a new way to evaluate blending filters in presence of the SC filter necessary.}
\label{blend}
\end{figure}
...
...
@@ -374,7 +374,7 @@ The plot in Fig. \ref{diffham} shows the differential motion of HAM2 and HAM3 in
\begin{figure}[h!]
\centering
\includegraphics[scale=0.3]{images/diffham.png}
\caption[HAM chambers in isolation]{Ham chambers in isolation: motion of HAM2 as a reference. The purple trace is the differential motion between HAM2 and HAM3 that we are interested in reducing.}
\caption[HAM chambers in isolation]{Ham chambers in isolation: motion of HAM2 as a reference. The purple trace is the differential motion between HAM2 and HAM3 that we are interested in reducing below $\sim$ 0.1 Hz.}
\label{diffham}
\end{figure}
...
...
@@ -511,15 +511,20 @@ After every simulation which could possibly work for the system, we locked the i
\caption[PRCL-ISI crossover filter]{Open loop gain (OLG) crossover filter implemented at LHO for a measurement of PRCL signal in offloading conditions.}
\label{prclfilter}
\end{figure}
\section{A follow-up test at LHO}
The new configuration proposed and tested has changed a crucial section of the structure of LIGO. This change might have consequences on other sides of the instrument, for example affecting other noise sources. A test about the effect of this configuration has been performed at LHO in 2020 by the LHO, LLO and Seismic teams. This test studied the impact of CPS differential controls on scattered light glitches on O3b run. The teams investigated the reason of an increase in the rate of glitches. The synchronized motion of the chambers with the ground could in principle make the instrument more sensitive to glitches (and other noise sources) which were hidden by the seismic and sensor noises. The study is showing that the configuration is not responsible for this increase, but there is an effect on the sensitivity of LIGO to glitches, when the configuration is activated. The complete study is exposed in details in the LHO logbook post \cite{lhotest}.\\
This is an example of the impact of the CPS differential control on LIGO: it has been used to test the effect of wind in microseismic regions \cite{lhotest2} and further tests might help to understand the impact of less seismic motion on other noise sources.
\newpage
\section*{Conclusions}
\addcontentsline{toc}{section}{Conclusions}
This study is promising to provide a significant contribution to the improvement of LIGO LSC signals and the detector stability when it is running in observing mode. The tests at LHO demonstrated that the experiment succeeded in lowering the seismic motion of the platforms by a factor of 3 at low frequencies and that also the DARM signal benefited from it. The simulations have shown that it is possible to reduce the differential motion of the chambers by a factor of 3 in order of magnitude below 0.1 Hz. The test on the Power Recycling Cavity Length highlighted that the signal can be controlled by the ISI according with the software simulations.\\
As we saw, the implications go straight to the basics of the instrument: a more stable detector produces a less noisy signal which can last longer into the cavities, assuring a longer observing time and giving the possibility to detect more gravitational waves and in lower ranges of frequency \cite{lantztalk}\cite{jenne}.
LIGO Livingston site has also actuated a similar process, following the progression at LHO during the work on site in 2019 \cite{llo}. Due to the limited time of the commissioning break, it was not possible to take further measurements of ISI motion and LSC signals, especially with an accurate study of the blending filters. However, since the software skeleton of the new configuration has been built and installed on both LIGO CDSs, further studies and tests were due in 2020 to complete the last steps and test it fully on the interferometer. The results make the experiment worthy of future developments and we are confident that these tests could be carried out in the near future.
LIGO Livingston site has also actuated a similar process, following the progression at LHO during the work on site in 2019 \cite{llo}. Due to the limited time of the commissioning break, it was not possible to take further measurements of ISI motion and LSC signals, especially with an accurate study of the blending filters. However, since the software skeleton of the new configuration has been built and installed on both LIGO CDSs, further studies and tests were due in 2020 to complete the last steps and test it fully on the interferometer. Meanwhile, the impact of the new configuration is under investigation on both sites: the reduced seismic motion could affect other noise sources that were previously hidden by the seismic and sensor noises, like glitches and wind. The results make then the idea worthy of future developments and improvements, and we are confident that these tests could be carried out in the near future.
@@ -124,9 +124,9 @@ This thesis focuses on the improvement of the seismic isolation system, which no
\label{ligosei}
Every optic needs to be stable with respect to seismic motion, because movements in the mirrors will cause unwanted displacement of the laser beam on the optical surface, resulting in noise during the laser travel into the cavities and then at the output. The main mirrors (test masses and beam splitter) are suspended from a stabilized bench and every suspension chain is placed in vacuum chambers called \textit{Basic Symmetric Chamber} (BSC). The auxiliary optics are placed on optical benches enclosed in the \textit{Horizontal Access Module} (HAM) chambers.
\begin{figure}[h!]
\begin{figure}[H]
\centering
\includegraphics[scale=0.8]{images/chambers.png}
\includegraphics[scale=0.75]{images/chambers.png}
\caption[Advanced LIGO vacuum system]{Schematic view of the vacuum chambers enclosing the optics \cite{mat}. There are 5 BSCs and 6 HAMs, for a total of 11 vacuum chambers for each LIGO. Each chamber provides a mixture of passive-active isolation from seismic motion, using pendulums, inertial sensors and hydraulic systems.}
\end{figure}
...
...
@@ -141,7 +141,7 @@ The HAMs provide five levels of isolation, among which there is the Internal Sei
\end{figure}
\noindent
The BSCs have a similar design as the HAMs, but they have two stages of ISI to support the suspensions isolating the test masses (Fig. \ref{bsc}).
The BSCs have a similar design as the HAMs, but they have two stages of ISI to support the suspensions isolating the test masses (Fig. \ref{bsc}).\\
\begin{figure}
\centering
...
...
@@ -151,7 +151,7 @@ The BSCs have a similar design as the HAMs, but they have two stages of ISI to s
\end{figure}
\paragraph{The sensors on the chambers}
The devices dedicated to monitoring the seismic motion are inertial and displacement sensors, which are horizontal and vertical, according to the different motion they need to sense. Currently, no sensors for tilt motion are installed on the platforms. Actuators are paired to each sensors, for active isolation of the sensed noise. The vertical displacement sensors are called Capacitive Position Sensors and are placed between every stages of every chamber: they measure the relative motion between the platforms. These are the sensors we will use in Chapter \ref{CPSdiff}. The vertical and horizontal inertial sensors with the dedicated actuators are placed on the platforms, underneath the optical tables, measuring the seismic motion in the horizontal and vertical directions. The position and the use of these sensors are different for HAM and BSC chambers, depending on the number of stages and the presence of the suspensions. The calibration and the specific role of each sensor into the seismic isolation system can be found in \cite{kisselthesis}, with references to the covered range of frequencies in \cite{kisseltalk3}.
The devices dedicated to monitoring the seismic motion are inertial and displacement sensors, which are horizontal and vertical, according to the different motion they need to sense. Currently, no sensors for tilt motion are installed on the platforms. Actuators are paired to each sensors, for active isolation of the sensed noise. The vertical displacement sensors are called Capacitive Position Sensors and are placed between every stages of every chamber: they measure the relative motion between the platforms. These are the sensors we will use in Chapter \ref{CPSdiff}. The vertical and horizontal inertial sensors with the dedicated actuators are placed on the platforms, underneath the optical tables, measuring the seismic motion in the horizontal and vertical directions. The position and the use of these sensors are different for HAM and BSC chambers, depending on the number of stages and the presence of the suspensions. The calibration and the specific role of each sensor into the seismic isolation system can be found in \cite{kisselthesis}, with references to the covered range of frequencies in \cite{kisseltalk3}.\\
\begin{figure}[h!]
\centering
...
...
@@ -168,9 +168,9 @@ Active isolation implies a sensing system of the noise to reduce and a control s
The control loop of a generic ISI stage on the X degree of freedom is simplified in the block diagram in Fig. \ref{control}. The platform motion is the sum of the input disturbance and the contribution from the control signal and it is measured by relative position and inertial sensors. This motion is then low- and high-passed via filters suitably built to fit the requirements and tuned to obtain the best performances combining the best results of both filters. This technique is called \textit{blending}, and the frequency where the relative and the inertial sensors contribute at their best is called \textit{blend frequency}. The result of this blend is called \textit{super sensor}. The output of the super sensor feeds the feedback loop, where the actuators close the loop \footnote{A general overview of control loops theory is exposed in Appendix B}.\\
The sensor correction loop takes the ground motion signal from an inertial instrument, filtering it before adding it to the relative sensor signal. This filter is needed because the sum of the motions from the ground inertial and the relative sensors can in principle provide a measurement of the absolute motion of the platform. However, the ground sensors are affected by low frequency noise and need to be suitably filtered.
\begin{figure}[h!]
\begin{figure}[H]
\centering
\includegraphics[scale=0.7]{images/control.png}
\includegraphics[scale=0.65]{images/control.png}
\caption[Control loop for a generic HAM-ISI]{Control loop of a generic HAM-ISI platform. Similar block diagrams can be applied for BSC-ISI platforms, including relative position sensors between the two stages of ISIs. \textbf{Green:} there is an inertial sensor measuring the ground motion along the x axis (GNDx), a Capacitive Position Sensor (CPS) measuring relative motions between the platform and the ground. Rotational sensors take care of tilt motion and GS13 are seismometers measuring seismic motion. Tilt and GS13 sensors are both placed on the platform. \textbf{Blue:} the Sensor Correction (SC) filter is typically a Finite Impulse Response (FIR) designed to provide required magnitude and phase match at 100 mHz (where isolation is needed). High- and low-pass filters (LP and HP) manipulate the signals from the low and high frequency sensors and are blended to form the super sensor, which output is sent to the control loop in \textbf{pink}. The overall corrected signal is then sent to the plant (\textbf{yellow}), which represents the processing phase for platform motion actuation.}
@@ -288,6 +288,10 @@ Beginning of Gravitational Wave Astronomy}
\bibitem{llo} A. Pele' et al., \textit{ECR: Differential CPS and cavity offload}, proposal, 2019, DCC E1900330-v1
\bibitem{lhotest} LHO and SEI team \textit{The impact of CPS differential controls on scattered light glitches in O3b}, alog.ligo-wa.caltech.edu/aLOG/index.php?callRep=56293
\bibitem{lhotest2} LHO team \textit{Turned off CPS DIFF} alog.ligo-wa.caltech.edu/aLOG/index.php?callRep=53926
\bibitem{lantztech} B. Lantz et al., \textit{Estimates of HAM-ISI motion for A+}, technical note, 2018, DCC T1800066-v2
\bibitem{hammodel} S. Cooper et al., \textit{Ham ISI Model}, technical note, 2018
...
...
@@ -338,7 +342,7 @@ Beginning of Gravitational Wave Astronomy}
%
%\noindent
%I really need to thank all the people on LIGO Hanford, who made me feel included and part of a great family. \textit{Grazie} Jenny, Jim, Rahul and Jeff B. for your infinite patience and for teaching me so much of LIGO; and I was very lucky to meet Dripta, an amazing housemate and friend, thank you for all the time you chose to share with me.\\
%\textit{Grazie} to my family in Seattle, zia Claudia, zio Carlo, Rossana \& Jim, Carla \& Wayne, who helped me to feel at home during my stay in the USA.
%\textit{Grazie} to my family in Seattle, zia Claudia, zio Carlo, Rossana \& Jim, Carla \& Wayne, who helped me to feel at home during my stay in the USA.\\
%
%\noindent
%To my friends in Italy: Vero \& Ludo, Alice \& Anna Paola, Martina, Francesca, Edo \& Nani, \textit{grazie} for your incredible will to stay close to me despite the 2000 km-distance.\\
@@ -624,6 +624,7 @@ Noise measurements of CDS with unplugged electronics have been taken, to check i
\newpage
\section*{Conclusions}
\addcontentsline{toc}{section}{Conclusions}
The analysis of feasibility of this experiment showed that the optical lever can be in principle a good device to sense tilt motion over long lever arms. However, the noise budget indicated a small frequency window of good operation, while below 0.1 Hz the levers are limited by the ground motion along the z axis. It is anyway a good device to be tested.\\
During the test of the prototype, the measurements have shown that we had issues when calibrating the device due to problems highly related to electronics from UoB, since the tests with the AEI electronics showed that the optical setup was well built and aligned. The very short time of the visit did not allow to take more in-depth tests.\\
Other possible reasons to investigate for better performances might lie in the structure of the prototype: further tests might be useful to understand if the device can be improved by changing the position of the lens with respect to the QPD, and let the diode sit at the focus on the lens. This solution will concentrate the power and decrease the size of the beam.\\
