40m TAC, 10/13/05 1

Report to the 40m Technical Advisory Committee

October 13, 2005

The 40m TeamBen Abbott, Rana Adhikari, Dan Busby, Jay Heefner, Osamu

Miyakawa, Mike Smith, Bob Taylor, Monica Varvella, Steve Vass, Rob Ward, Alan Weinstein

with lots of help from: Matt Evans, Helena Armandula, Rolf Bork, Alex Ivanov, SURF students, etc…

40m TAC, 10/13/05 2

Status as of October 2005

§ Slides 3 through 11: presented at August LSC meeting, included for your review, skip for now.

§ Status of lock acquisition (Rob, Osamu, Matt, Rana)§ e2e simulation (Monica)§ DC readout (Smith, AJW, all)§ Proposal for squeezing-enhanced gravitational wave

interferometer at the 40m (Nergis)

40m TAC, 10/13/05 3

Caltech 40 meter prototype interferometer

Objectives§Develop lock acquisition procedure of detuned Resonant Sideband

Extraction (RSE) interferometer, as close as possible to AdLIGO optical design

BSPRM SRM

X arm

Darkport

Brightport

Y arm

§Characterize noise mechanisms§ Verify optical spring and optical

resonance§Develop DC readout scheme§ Extrapolate to AdLIGO via

simulation

40m TAC, 10/13/05 4

AdLIGO signal extraction scheme

§ Arm cavity signals are extracted from beat between carrier and f1 or f2.§ Central part (Michelson, PRC, SRC) signals are extracted from beat

between f1 and f2, not including arm cavity information.

f1-f1 f2-f2

Carrier (Resonant on arms)

• Single demodulation• Arm information

• Double demodulation• Central part information

§ Mach-Zehnder installed to eliminate sidebands of sidebands.

§ Only + f2 is resonant on SRC.§ Unbalanced sidebands of +/-f2 due

to detuned SRC produce good error signal for Central part.

ETMy

ETMx

ITMy

ITMxBSPRM

SRM

4km

4km

f2

f1

40m TAC, 10/13/05 5

5 DOF for length control

: L+=( Lx+ Ly) / 2: L−= Lx− Ly: l+=( lx+ ly) / 2

=2.257m: l−= lx− ly = 0.451m: ls=( lsx+ lsy) / 2

=2.15m

1-2.3E-24.6E-12.7E-33.6E-3f1 × f2PO

7.1E-217.5E-11.5E-3-6.2E-4f1 × f2AP

-1.0E-1-3.2E-21-3.0E-4-1.7E-3f1 × f2SP

-1.7E-81.3E-31.2E-81-4.8E-9f2AP

-2.3E-6-1.3E-6-1.2E-3-3.8E-91f1SP

l sl−l+L−L+Dem. Freq.

Port

Signal Extraction Matrix (in-lock)

Common of armsDifferential of armsPower recycling cavity

MichelsonSignal recycling cavity

Laser

ETMy

ETMx

ITMy

ITMxBS

PRM

SRM

SPAP

PO

lx

ly

lsx

lsy

Lx =38.55mFinesse=1235

Ly=38.55mFinesse=1235

Phase Modulationf1=33MHzf2=166MHz

T =7%

T =7%

GPR=14.5

40m TAC, 10/13/05 6

Differences betweenAdvLIGO and 40m prototype

§ 100 times shorter cavity length§ Arm cavity finesse at 40m chosen to be = to AdvLIGO ( = 1235 )

» Storage time is x100 shorter.

§ Control RF sidebands are 33/166 MHz instead of 9/180 MHz» Due to shorter PRC length, less signal separation.

§ LIGO-I 10-watt laser, negligible thermal effects» 180W laser will be used in AdvLIGO.

§ Noisier seismic environment in town, smaller stack» ~1x10-6m at 1Hz.

§ LIGO-I single pendulum suspensions are used» AdvLIGO will use triple (MC, BS, PRM, SRM) and quad (ITMs, ETMs)

suspensions.

40m TAC, 10/13/05 7

DRMI lock using double demodulation with unbalanced sideband by detuned cavity

August 2004August 2004§§DRMI locked with carrier resonance (like GEO configuration)DRMI locked with carrier resonance (like GEO configuration)November 2004November 2004§§DRMI locked with sideband resonance (Carrier is anti resonant prDRMI locked with sideband resonance (Carrier is anti resonant preparing for RSE.)eparing for RSE.)

Carrier

33MHz

Unbalanced166MHz

Belongs tonext carrier

Belongs tonext carrier

Carrier33MHz166MHz

ITMy

ITMxBS

PRM

SRM

OSA DDM PD

DDM PD

DDM PD

Typical lock acquisition time : ~10secLongest lock : 2.5hour

40m TAC, 10/13/05 8

Lock acquisition procedure towards detuned RSE

ITMy

ITMxBSPRM

SRM

ETMx

ETMy

Shutter

Shutter

Carrier33MHz166MHz

DRMI using DDM Off-resonantarms using DC lock

(POX/TrX),(POY/TrY)

SP166/PRC,AP166/PRC

(POX/TrX) + (POY/TrY), (POX/TrX) – (POY/TrY)

(TrX + TrY), (TrX – TrY) / (TrX + TrY)

Ideas for armcontrol signal

RSE

Done In progressDone

40m TAC, 10/13/05 9

All 5 degrees of freedom under controlledwith DC offset on L+ loop

Both arms locked with DRMIBoth arms locked with DRMI§§ Lock acquisition time ~1 minLock acquisition time ~1 min§§ Lasts ~ 20 minLasts ~ 20 min§§ Can be switched to Can be switched to

common/differential controlcommon/differential controlLL-- : AP166 with no offset: AP166 with no offsetLL+ : + : Trx+TryTrx+Try with DC offsetwith DC offset

Yarm lockXarm lock

Arm power

Error signal

Ideal lockpoint

Offset lockOffset lock

Have started trying toreduce offset from L+ loop

But…

40m TAC, 10/13/05 10

Progress in last 6 months

§ For the last 6 months , we have been able to control all 5DOF, but with CARM offset.

§ Reducing the CARM offset has been made difficult by technical noise sources. We have spent last 6months reducing them;

»suspension noise, vented to reduce couplings in ITMX»improved diagonalization of all suspensions»improved frequency noise with common mode servo»automation of alignment and lock acquisition procedures»improved DC signals and improved RF signals for lock acquisition.

§ We can now routinely lock all 5 DOFs in a few minutes at night.

40m TAC, 10/13/05 11

-150

-100

-50

0

50

100

150

Pha

se[d

eg]

102 3 4 5 6 7 8

1002 3 4 5 6 7 8

10002 3 4 5 6

Frequency[Hz]

50

40

30

20

10

0

-10

-20

Mag

[dB

]

Measured optical gain Calculated by Thomas's tool

RSE peak

RSE peak!• Optical gain of L- loop

DARM_IN1/DARM_OUT divided by pendulum transfer function

• No offset on L- loop• ~60pm offset on L+ loop• Phase includes time delay of

the digital system.

• Optical resonance of detuned RSE can be seen around the design RSE peak of 4kHz.

• Q of this peak is about 7.• Effectively the same as Full

RSE with GPR=1.4 with 1W input laser.

• Model was calculated by Thomas’s tool.

• We will be looking for optical spring peak.

Design RSEpeak ~ 4kHz

40m TAC, 10/13/05 12

Lock acquisition procedure towards detuned RSE

Start withno DOFscontrolled, all optics aligned.Re-align each 1.5 hours.

ITMy

ITMxBS

PRM

SRMSP DDM

13m MC

33MHz

166MHz

SP33SP166

AP DDM

AP166

PO DDM

40m TAC, 10/13/05 13

Lock acquisition procedure towards detuned RSE

DRMI + 2armswith offset

ITMy

ITMxBS

PRM

SRMSP DDM

13m MC

33MHz

166MHz

SP33 SP166

AP DDM

AP166

PO DDM

Average wait : 1 minute(at night, with tickler)

T =7%

T =7%IQ

1/sqrt(TrY)

1/sqrt(TrX)

40m TAC, 10/13/05 14

Lock acquisition procedure towards detuned RSE

Scripto-matic:

ITMy

ITMxBS

PRM

SRMSP DDM

13m MC

33MHz

166MHz

SP33 SP166

AP DDM

AP166To DARM

PO DDM

AP166 / (TrX+TrY)

CARM

DARM+

-1+

Short DOFs -> DDMDARM -> RF signalCARM -> DC signal

1/sqrt(TrX)+ 1/sqrt( TrY)

Discretionary:CARM -> Digital

CM_MCL servo

40m TAC, 10/13/05 15

Lock acquisition procedure towards detuned RSE

Reduce CARM offset:Go to higher ARM power, switchon AC-coupled analog CM_AO

servo, using REFL DC as error signal.Locks somewhat stable at 85% ofmaximum power.

ITMy

ITMxBS

PRM

SRMSP DDM

13m MC

33MHz

166MHz

SP33

SP166

AP DDM

AP166To DARMREFL

DARM-1

PO DDM

AP166 / (TrX+TrY)

GPR=5

Up next:

RF control of CARM: REFL166 (POX/POY 33 ?)

40m TAC, 10/13/05 16

Ready for Transition to RF CARM

40m TAC, 10/13/05 17

Progress in last 2 months

§ Use CARM DC signal to creep closer to full arm resonance (signalgoes away at resonance)

» combined PRC and arm gain as high as 60, 85% of the way to peak,corresponding to offset of 8 pm, assuming arm losses of 200ppm (inferred).

§ Progress due largely to improvements in loop filtering, to obtain more gain and reduce noise

§ Significant improvements in automation of lock procedures.§ All DOFs diagonalized: 3x3 (SRC/PRC/MICH) + 2x2 (DARM/CARM)§ Lock is lost because of large frequency noise (presumably).§ Effort to reduce frequency noise using CM servo

» digital CM servo to MC works well» analog servo using DC signal seems to work, RF under development» frequency noise may be coming from in-vac PZTs. Not beam jitter noise.

§ Much effort to reduce other noise sources» better diagonalization of MC suspensions, core optics suspensions» continual battle with 60 Hz harmonics

40m TAC, 10/13/05 18

Development of e2e simulation:4Om/AdvLIGO package

§ Monica Varvella (visitor from LAL/Orsay) has developed a 4Om/AdvLIGO package with a DRFPMI optical plant, and is developing the control plant with help from Matt Evans and Hiro Yamamoto

§ Tests include a careful comparison with Twiddle, as well as comparison of error signal sweeps between simulation and 40m data.

§ Can be used to extract the velocity of the mirrors in the 40m under controlled circumstances: lock all degrees of freedom, CARM offset, and then HOLD the CARM servo and watch error signal as mirrors sweep through arm resonance.

40m TAC, 10/13/05 19

Error signal sweeps at 10-9 m/sfor the 40m IFO obtained in

E2E framework and comparedwith TWIDDLE predictions

Example:DARM @ AP 166 MHz

TWIDDLE and E2E comparison

e2e SIMULATION:4Om/AdvLIGO package

TWIDDLE

E2E

40m TAC, 10/13/05 20

e2e SIMULATION: 4Om/AdvLIGO package

Comparison between real data(black) and e2e simulated data(red) of the transmitted light for both the arms (full IFO): the mirror velocities used in

E2E simulation are the values obtained fitting the real data

Real data have been used to estimate relative mirror velocity for

both the arms:

Vxarm= (0.35 ± 0.13) µm/s

Vyarm= (0.26 ± 0.13) µm/sE2E

E2E

real data

real data

Tr X

Tr Y

40m TAC, 10/13/05 21

e2e SIMULATION: 4Om/AdvLIGO package

Comparison between real data , e2e simulated data and the

theoretical prediction V(t) of the SP error signal @ 166 MHz

The t and the velocity v is the value obtained fitting real data

t = 0.7 msv = 0.26 µm/s

V(t) ~ exp(t/t ) sin( a t2)

with a = (k v) / (2 T)

40m TAC, 10/13/05 22

DC Readout at the 40m§ DC Readout eliminates several sources of technical noise (mainly due to the RF

sidebands): – Oscillator phase noise – Effects of unstable recycling cavity. – The arm-filtered carrier light will serve as a heavily stabilized local

oscillator. – Perfect spatial overlap of LO and GW signal at PD.

§ DC Readout has the potential for QND measurements, without majormodifications to the IFO.

§ We may not be able to see shot noise at low frequency, given our noise environment. We may not even see any noise improvements, but we might!

§ The most important thing we will learn is : How to do it » How to lock it?» How best to control the DARM offset?» What are the unforeseen noise sources associated with an in-vacuum OMC?» How do we make a good in-vac photodiode? What unforeseen noise

sources are associated with it?

40m TAC, 10/13/05 23

DC readout equipment

§ Most in-vac optics and opto-mechanics have been ordered» Most mirror mounts and mirrors» Output mode-matching telescope with picomotor focus» two Piezo-Jena steering mirrors» but NOT the PZT drivers. Jay Heefner has committed to designing these (also

maybe needed at sites for RBS system). Design based on IO PZTs. Need strain gauge readback.

§ OMC design from Mike Smith» 4-mirror design» super-mirrors with REO coatings coming from LIGO Lab spares

§ DC PD assembly and electronics under design (Abbott, Adhikari)§ OMC control, DC PD readout, and monitoring electro-optics and

readout, under design (Heefner), mostly using existing infrastructure and equipment.

§ Alignment on bench, in air, and in vacuum seem feasible.

40m TAC, 10/13/05 24

Output Optical Train

=

=

Mike Smith

SRM

1st PZT steering mirror

2nd PZT steering mirror

gets a little tightaround IMMT

OOC IOC

BSC

40m TAC, 10/13/05 25

Output Optic Chamber

Mike Smith

from SRM

to AS RF beamline(roughly 1/3 of AS power)

also a convenient pathfor autocollimator beam,for initial alignment in air

to OMCR beamline

to OMCT beamline

from PSL to IMC

IMCR, IMCT, and SP beamlines

2nd PZT steering mirror

PZT steering mirrors and their controlsare duplicates of a pair that we have

already installed and commissioned for steering from IMC to main IFO (in-vac);

controls are fully implemented in the ASC system (by Rolf). Similar systems

can be used for “LIGO I.V”.

Piezosystem Jena PSH 5/2 SG-V, PZT tilting mirror mount with strain gauge, and associated drivers and

power supplies

Existing in-vacseismically isolated optical table (OOC)

Mike Smith has designed a compact, monolithic MMT, similar to our input MMT, using spherical

mirrors.

4-mirror monolithic OMC.Pair of DC PDswith in-vac electronics

on monolithic base.

40m TAC, 10/13/05 26

OMC, four mirror design

Mike Smith

• Mirrors mounted mechanically, on 3 points (no glue)• curved mirror: off-the-shelf CVI laser mirror with ROC = 1 m ± 0.5%• Fixed spacer should be rigid, vented, offset from table § OMC length signal:

§ Dither-lock? >> Should be simple; we’ll try this first.§ PDH reflection? >> There’s only one sideband, but it will still work.

§ Servo:§ Will proceed with a simple servo, using a signal generator and a lock-in amp.§ Feedback filters can easily be analog or digital.

§ Can use a modified PMC servo board for analog.§ Can use spare ADC/DAC channels in our front end IO processor for digital.

§ PZT actuation

From MMT

to DCPD

PZT mirror

SS fixed spacer

~ 20 cm

reflected beam

mechanical clamps(no glue)

40m TAC, 10/13/05 27

OMC design in SolidWorks§ Small number of pieces§ HV compatible

» a bit of glue on the PZT mirror§ Mirrors mounted mechanically,

on silver washers (no glue)§ ALGOR FEA: lowest mech

resonance at ~770 Hz§ Construct out of well-damped

material, to minimize effect of resonances

» Brass? Copper?» Or just stick with aluminum?

§ All high-quality (REO super-polished and coated) mirrors available from LIGO lab spares (well, the 4th HR mirror, 0o

incidence, may need to come from Newport) Mike Smith

40m TAC, 10/13/05 28

In-vac DC photodiodes

Three views of an in-vacuum DC PD assembly, showing a 50% beamsplitter, two photodiodes, a beam-dump, and a vacuum can to hold electronics. The base will not be made of marble. Electronics under design (Abbott, Adhikari)

Ben Abbott