Environmental Noise Studies

(1)

1

Environmental Noise Studies

(May-July’09)

I.Fiori, Virgo Week, Commissioning Meeting, 20 July 2009

by

The VIRGO Commissioning Team and

EGO Vacuum, Electronics, Infrastructure

Teams

(2)

Topics

• Diffused light from End benches

• External Injection Bench

• Turbo pumps

• Cryo-trap

• Coil Drivers crates

• U-pure-water sys (investigation)

• HVAC

2

(3)

Diffused light from WEB, mitigation of bench motion

• Diffused light from WEB was limiting sensitivity before May 2009 (purple)

• WEB motion at legs resonance (20Hz) has been reduced by:

• Slow down HVAC fan (-10%)

• New cover box (more rigid and compact)

• Resonant damper on bench

(4)

Microseism

• VSR2 inauguration, with exceptional microseism (<1% of time)

4

• Possible intervention:

Replacement of WE tower output window: tilted and AR coated (same as was done at NE).

Could reduce noise coupling by a factor 1.5 - 2.

(5)

Noise in 40-50Hz

Contributions not yet fully understood, but progressing:

• TCS noise seems to contribute (see G.Vajente).

• Some cooling fans (turbo pumps, CD crates) found to contribute and are being mitigated (see later in this talk).

• External Injection Bench is suspected to couple seismic noise in this region because of its large motion at 41-47 Hz (bench vertical modes).

5

(6)

EIB noise?

• Coherence changes , and now it is much lower than in May.

• Coupling path is being investigated (indications that it might depend on IMC transmitted power).

• Understand coupling and study possible reduction of EIB motion is also important for V+MS.

6

(7)

EIB vertical resonances:

investigation

• Simulation: 40 Hz resonance is a mechanical mode of the top plate of bench legs.

7

F.Richard and T.Zelenova Ruggi, Paoletti, Fiori

• Measurement with

mini-accelerometers

confirmed prediction.

(8)

EIB: mitigations?

• Stiffening actual legs: attemps made but not successful

• Sorbothane: pads between bench and legs, can move vertical mode down to 15 Hz, and well damped

• BUT, Test in Optics Laboratory: >>

measured large drifts (100m) which cannot be tollerated by BMS.

Turbo pumps, related

• Excess noise at InjectionBrewster

• NI tower motion

• Peaks from Turbo cooling fans

9

(10)

Excess noise at INJ Brewster

¹⁰

• eLogs 23216, 23261, 23267, 23307, 23419

• Evidence: 100-200Hz noise coherent with seismic noise at INJ Brewster, but also EIB and BMS.

• Noise decrease when IB pump off.

• Residual coherence from PR turbo pump. IB and PR pumps ON IB turbo OFF

PR pump OFF Effect of switching off IB pump:

(11)

Excess noise at INJ Brewster

ACTIONS: (

Pasqualetti and Vacuum Team)

• 1) IB pump now kept off, and replaced by MC pump (MC pump looks “safer”, i.e.

no coherence with dark fringe).

• 2) Bellow installed on PR pump: cut all pump noise at Brewster below 100Hz 11

Switching on PR pump after bellow

installation: NO EFFECT! Only noise from PR reaching INJBrew is that of cooling fans.

>>> Role of INJ Brewster to be better studied for V+MS.

(12)

Bellow at PR

• lpaoli, vacuum team - 14:44, Tuesday 30 June 2009 (23412)

12

(13)

NI motion

Ruggi (eLogs 23408, 23436, 23452, 23457)

• ITF was disturbed by repetitive events of NI motion, which caused unlocks or holes in the horizon.

• Tracked to Saturation of one accelerometer on NI top stage.

• The NI top stage seems more coupled to environment than other towers (reasons to be investigated).

13

Meanwhile, ACTIONS:

1)Temporary cure: tower vibration reduced by switching off NI turbo pump and use

Auxiliary NI pump.

2)Installation of bellow on NI turbo (soon, during Tuesday maintenances).

(14)

Turbo cooling fans

• eLogs 23353, 23757, Vacuum team

• Peaks at 45Hz from WI turbo, coherent with TCS bench seism.

• Mitigation attempt: fans detached from turbo and attached to platform.  seismic excitation reduced (tower, TCS bench) by

5-10.

• First action, to test the damping principle.

The method of damping shall be better defined, if needed implemented on all disturbing fans, in the forthcoming maintenance breaks.

14

(15)

Fans of Coil Drivers crates: BS and NE

• eLogs: 23471, 23714; 23684, 23536, 23483, 23477, 23410: Marque, Vajente, Huet, Nocera

• Peaks in dark fringe at (44Hz, 88Hz) and (28Hz, 56Hz, 84Hz)

• present in NE/BS coil current monitoring even with ITF unlocked (if coil driver in low noise)

• Disappear when coil driver fans are switched off.

15

(16)

CD crates, mitigations:

FACTS:

1) Culprit seems magnetic noise emission from crates fans.

Magnetic noise seems local to crate, not present few meters away

( coupling: likely not Env. to mirrors magnets, but via the current sent to the coils).

2) Noisy crates (NE and BS) are of one type.

No noise detected from the CD crate of WE, which is of different type.

ACTIONS:

– Noise path inside crate to be better investigated (mechanical, ground loops, electromagnetic...) to plan proper actions.

– Simple check: if problem related just to fans type (if so, change just fans could be sufficient).

– Ultimatey, evaluate if necessary to purchase crates of the “good”

type.

16

(17)

“Cryo-trap noise”

eLogs: 23722, 23637, 23636, 23634, 23794, Swinkels, Ruggi, Vajente.

• Noise increase in 60-120Hz after routine Cryo-trap refills

• Corresponds to increased vibration of LN tank and pipe.

17

(18)

“Cryo-trap noise”

• Noise (dark fringe and seismometers) decreases with time constant of

1.5day, which is correlated to the average temperature of the tank.

• When “listening” to sismometer on Cryo-trap, it sounds like “bubbles”.

• Vacuum group interested to understand this mechanisms for future installations. Tests to be planned during next maintenances.

18

(19)

“Cryo-trap noise”: mitigation?

• Hypothesis:

since the removal of the suspended baffles (May 15), the wall of the Cryo-trap can now be seen from the detection bench. Any light scattering e.g. from L1 could thus hit the wall of the tank.

Possible ACTIONS:

• Install a new baffle, to cover the “naked” Cryotrap (careful design to avoid upconversion noise effects, very rigid one)

• Reduce back-reflected light from L1, replace with better lens superpolished (J.Marque).

19

(20)

Infrastructures related noises

• Water pump #1: Noisy during VSR1. Replaced. New pump produces no

detectable environmenatal noise in CB hall. However, pump is off, apparently not needed.

• U-pure-water production sys: needs to run in stand-by mode during VSR2.

Noise investigated (eLog 23519, Ciardelli, Fiori).

One pump that starts periodically produces some seismic noise on EIB. No detectable effect in dark fringe. Mitigation tests to do during maintenances.

20

(21)

Environmental Noise Studies