Passive isolation: Pre-isolation for FF quads

Passive isolation:Pre-isolation for FF quads

A. Gaddi, H. Gerwig, A. Hervé, N. Siegrist, F. Ramos

Page 2IWLC, October 2010, A. Gaddi, Physics Dept. CERN

Passive isolation: Pre-isolator for FF quads

Retrospective view.In the first CLIC MDI layout (legacy of ILC MDI), the QD0s were supported by the detector, or by a pillar on the detector moving platform. However it became clear, after the measuring campaign at CMS, that the vibrations generated by the detector itself would make impossible to achieve the stability requirements given for the FF magnets at CLIC.

Measurements done at CMScourtesy EN-MME/CERN



Scope of work.

Try to attenuate, at its source, ground motion vertical excitations, in the range 1 – 20 Hz, to make life easier to the following stabilization systems.

Objective.

Stabilize FF magnets to better than 1 nm (rms) at 4 Hz,using an integrated approach of three systems, each onewith its dynamic range and frequency response:

Passive pre-isolator

Active mechanical stabilization

Beam-based stabilization



Stabilization block diagram.

from B. Carron

X – ground motionKi – isolator transfer functionD – external disturbances else than ground-motionW – white noise on beam positionY – beam position signal



Pre-isolator – How does it work ?

Low dynamicstiffness (k) mount

natural frequency around 1 Hz

Acts as a low-pass filterfor the ground motion (w)

Large mass (m)

between 50 and 100 tons

Provides the inertia necessary to withstand the external disturbances (Fa), such as air flow, acoustic pressure, etc.)

+



Where does it fit ?

Ideally located at the end of the machine tunnel,just in front of the detector, on both sides.

drawing by N. Siegrist



How can it be realized ?

QF1

QD0

MassElasticsupport

Walk-on-

floor

QD0 support tube

conceptual design



More in details :

QD0 SD0 MULT QF1 SF1

Pre-isolator slab

Accelerator tunnel

Detector side



FE Model Layout.

Things missing in the model:• Pre-alignment mechanics• Final doublet’s geometries (using, for now, 3-D point masses with estimated inertias)• Final doublet’s supporting structures (girders, etc.)• Pre-isolator’s supports (using , for now, 1-D springs with appropriate stiffness)

Lumical

Beamcal QD0 SD0 MULT

QF1 SF1



Response to excitation in the vertical direction.

6

1 Hz

51.2 Hz

Good performance above the first resonance peak



Response to excitation in the horizontal directions.

7

There is a good decoupling between horizontal & vertical

directions



Random vibration response.

11

Vertical ground motion at CMS.



Random vibration response.

12

Vertical ground motion at CMS.

4 Hz

2.2 nm

0.1 nm

Reduction in r.m.s. displacements by a factor 20 above 4 Hz



Tilting mode Vertical mode

Horizontal mode



Experimental set-up – Why ?

How will it react to different noise sources, else than micro-seism ?

Is the system’s performance amplitude dependent ?

What about energy loss mechanisms (friction...) ?

What performance can we expect ?

We have built a prototype to try to answer to these questions.

Although rather different from the reference design, it will serve to validate the idea of having a low-frequency mechanical filter in front of the FF stabilization chain.

The idea is promising, but...



Experimental set-up – How ?

The prototype needs to be:

Simple to design/build/assembleEasy to “debug” & tune

Cheap

Frictionless pivotal joints

Proposal:

40 ton mass supported by4 structural beams acting as flexural

springs



Photo of the prototype.



Experimental set-up – Expected performance

Performance in ideal conditions (ground motion only).

Transmissibility Displacement P.S.D.

1.55 Hz

CMS floor

@Pre-isolator

Integrated R.M.S. Displacement

2.2nm at CMS floor

0.15 nm at Pre-isolator

15x

plots by F. Ramos



Example of pre-isolator application in industry.

Vibration isolation system at the Centre for Metrology and Accreditation – Helsinki, Finland

4 independent seismic masses (3x70 ton + 1x140 ton)

0.8 Hz pneumatic vibration isolators(“air springs”)



Future plans.

• A passive low-frequency pre-isolator has been proposed as a support for the FF magnets. It can be integrated at the interface between the machine tunnel and the detector cavern.

• This pre-isolator will constitute the first “layer” of the stabilization chain.

• This concept has already been used in other industrial and research facilities.

• Experimental tests will be performed to understand how the system behaves in “real life” conditions (no simulation can accurately take into account all these effects). A simple and relatively inexpensive experimental set-up has been conceived and built. Measurements are under way.

• Further studies are necessary to integrate the pre-isolator with the civil-engineering and the other MDI systems (pre-alignement, FF magnets, forward detector region, beam-pipe, etc…)



Back-up slides.



Further considerations.

Supporting the FF quads (and in particular QD0) from the tunnel sets some parameters that seem to be difficult to be met with a conventional design of the tunnel-experiment cavern interface, specially for what concerns the civil engineering.

What counts to maximize machine luminosity is the relative alignment of the two QD0s, rather than their absolute position. We shall profit of the relatively short distance between QD0s, to try to mechanically link them through a rigid structurethat bridges the gap between the two tunnel ends. This would also increase the coherence of low frequency seismic waves across the UX cavern.



Coherence of seismic vibration important for QD0 relative displacement.

Relative (nm)

3m 4m 5m 7m50 9m 10m Abs (nm)

ATF2 12.3 (/10) 15.6 (/8.2) 24.6 (/5.2) 34.7 (/3.7) 42.0 (/3.1) 37.0 (/3.5) 128.4

LHC 1.0 (/12.5) 1.3 (/9.6) 1.8 (/6.9) 2.6 (/4.8) 2.8 (/4.5) 3.1 (/4.0) 12.5

LAPP 0.4 (/7.5) ========= 0.7 (/4.3) 1.0 (/3.0) 1.0 (/3.0) 0.9 (/3.3) 3

Integrated RMS (nm) of relative and absolute motion above 1Hz with a rigid fixation.B.Bolzon, march 2009

Measurements have been planned at CMS site, where a 2.25m thick reinforced concret slab, long several tens of meters, and supported by pillars with foundations into the rock, would be the design reference. We aim to understand if a similar design, in the CLIC experiment cavern, would be beneficial for the FF stabilization at low frequency.



Conceptual design: the Quads’ bridge.

The design has evolved from a first sketch, aiming to provide the same foundations for the couple of FFQ, to a more complex design that includes a low-frequency vertical pre-isolator.

Main linac

Pre-alignment & stabilization

FFQ stabilization & alignment

Quads-bridge

Tunnel floor



Conceptual design: the Quads’ bridge + the pre-isolator.

Main linac

Pre-alignment & stabilization

FFQ stabilization

FFQ pre-isolation& alignment

Quads-bridge

Tunnel floor

The main linac elements are aligned and stabilized on their girders.The FF quads are mounted inside a supporting tube, the stabilization is located inside the support tube.The FF quads pre-alignment is done by positioning the support tube, via the low frequency pre-isolator.The FF bridge has the function of extending the tunnel concrete slab from one side to the other of theexperimental cavern.

Documents

Passive isolation: Pre-isolation for FF quads