Title LIGO-G1100174 DC Readout in Enhanced LIGO 1 Enhanced LIGO - PowerPoint PPT Presentation

Title LIGO-G1100174 DC Readout in Enhanced LIGO 1

Enhanced LIGO • Try out Advanced LIGO technologies • Bet that increased sensitivity outweighs the downtime exposure = time * (range)^3 New Laser More Power New input optics New Thermal Compensation New Alignment Control Output Mode Cleaner DC Readout 2

Interferometer Gravitational waves produce phase modulation in the arm cavities ? How to detect it? 3

DC Readout: sideband view HETERODYNE (RF) -gw +gw optical frequency -rf +rf HOMODYNE (DC) -gw +gw optical frequency -rf +rf laser carrier 4

DC Readout: fringe view power at AS port 10 picometers DARM length 5

DC Readout Looks pretty simple... DARM_ERR 6

Junk Light 7

DC Readout with OMC Clean up the light at the AS port with an output mode cleaner. DARM_ERR 8

DC Readout promises • fundamental improvement in SNR • technical improvement in SNR - perfect overlap of local oscillator and signal beams - junk light removal by OMC • independence from RF oscillator noises - exploit the amazing filtering ability of the interferometer • Easier platform for squeezed light injection • Easier to handle higher power 9

DC Readout: phasor view optical gain: How do we choose the DARM offset? • Must be much greater than residual DARM displacement • Must overcome contrast defect and electronics noise • But not excessively detrimental to power recycling 10 In practice: turn the knob to get the best sensitivity

OMC design beam from interferometer beam steering mirrors bowtie cavity fast and slow length actuators two DC two quadrant photodiodes photodiodes (QPDs) monolithic, suspended, in-vacuum 11 Sam Waldman et al

Dither Locking no first-order response at maximum 180 degrees out of phase in phase 1. put in small dither sinusoid 2. demodulate output at same freq == error signal! 12

OMC Length Control Cavity length dithered at ~10 kHz via PZT actuator PZT offloaded onto slow, long-range thermal actuator DARM 13

OMC Alignment Control The mode cleaner will clean the modes if you can identify what mode you want to keep. Initial idea: maximize transmission through the OMC 14

Junk light confuses simple servo 00 mode 01 mode transmission 00 + 01 versus beam pointing 01 mode leads the servo astray 15

Drumhead Beacon Dither Idea: Tag the photons in the arm by modulating the ETM Excite the test-mass drumhead mode (9 kHz) Dither the "tip tilt" mirrors at low frequency (~3 Hz) detect power in demodulate at drumhead mode dither frequency 16 M.Evans/Nicolás(LHO)

Optical Gain vs Jitter But if we optimize alignment for optical gain... ...and best optical gain does not correspond to maximum transmission ...then we introduce a beam jitter coupling 17

Beam Jitter Noise Most significant new noise source 130 Hz, one of the more prominent jitter peaks 18

A closer look at 130 Hz L1:OMC−QPD3_Y_OUT_DAQ −12 10 130 Hz arbitrary units −13 10 linear coupling L1:LSC−DARM_ERR −12 10 due to HG 01 mode 130 Hz −14 10 ± 0.875, arbitrary units −13 10 ± 1.6 −15 10 125 126 127 128 129 130 131 132 133 134 135 frequency [Hz] L1:OMC−QPD4_SUM_IN1_DAQ −14 10 5 10 0.87 Hz 4 10 1.6 Hz −15 arbitrary units 10 125 126 127 128 129 130 131 132 133 134 135 frequency [Hz] 3 10 bilinear coupling due to gain modulation 2 10 19 1 10 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 frequency [Hz]

A closer look at 130 Hz Quick check: overlay the spectra −6 10 DARM QPD YAW QPD SUM (shifted) −7 10 arbitrary units −8 10 0 Hz 129.9 Hz 129.9 Hz mirror image of qpd sum spectrum −9 10 125 126 127 128 129 130 131 132 133 134 135 frequency [Hz] 20

Can we predict the timeseries? −7 x 10 4.5 DARM QPD YAW 4 SUM � HP(YAW) QPD sum ASD (arbitrary) 3.5 QPD yaw 3 2.5 2 One of several possible 1.5 bilinear couplings. 1 0.5 0 125 126 127 128 129 130 131 132 133 134 135 1 Yes. Good coherence 0.9 shows this is a real 0.8 coupling. coherence 0.7 0.6 0.5 0.4 0.3 (bilinear wiener filtering?) 0.2 have to be careful to not subtract DARM from DARM 0.1 21 0 125 126 127 128 129 130 131 132 133 134 135 frequency

Beam jitter mitigation • Remove the offending resonance or increase isolation i.e. tip-tilt blade springs, no fixed mirror • Reduce sensitivity to the motion i.e. clever telescope design • Cancel the motion (feedback/forward) i.e. 60 Hz magnetometer FF • Reshape the output beam i.e. use WFS1 to push on the ASC 22

Noise Couplings * Oscillator amplitude * Oscillator phase * Laser intensity * Laser frequency 23

Oscillator Amplitude noise 24

Oscillator Phase noise 25

Anatomy of intensity noise coupling 0 carrier + sidebands carrier only −20 sidebands only carrier Michelson −40 transmission Watts per Watt [dB] carrier AM −60 recycling gain cancellation −80 sideband AM −100 −120 −1 0 1 2 3 4 10 10 10 10 10 10 26 frequency [Hz]

Anatomy of intensity noise coupling II 0 carrier + sidebands carrier only −20 sidebands only −40 Watts per Watt [dB] −60 −80 −100 −120 −1 0 1 2 3 4 10 10 10 10 10 10 27 frequency [Hz]

Laser intensity noise Laser intensity noise to DARM −10 10 −11 10 −12 10 m/RIN MODELS −13 10 −14 10 −15 10 2 3 warm colors = positive DARM offset 10 10 cool colors = negative DARM offset 28 frequency [Hz]

Laser frequency noise Laser frequency noise to DARM −8 10 −10 (DARM meters) per (CARM Hz) 10 −12 10 −14 10 −16 10 seems high - calibration? −18 10 2 3 10 10 29 frequency [Hz]

30 Noise Couplings To-Do • Wrap control loops around Optickle results (Pickle) and around measured results to see effect of cross-coupling. • T une parameters to fit results as well as possible • Be informed by / compare with other methods: - Zach's mode tracking absorption msmts - Mode scans - Arm cavity pole (ringdown) measurements

Digital Gotchas * Not new! * But we keep 'rediscovering' them! 31

Digital Gotchas: Synchronization * synchronization of communication OMC LSC * ADC timestamp synchronization reflected memory duotone delay history 40 Duotone tracking shows LSC OMC 35 nondeterministic ADC startup and drift. 30 25 analog reference microseconds 20 signal (two sines) 15 recorded by 10 independent ADCs 5 OMC LSC 0 −5 phase of timestamped digital signals compared (in plot to the left) Jul2009 Oct2009 Jan2010 Apr2010 32

Digital Gotchas: Other * Quantization noise in DACs -- "dark noise" not enough * Floating point dynamic range/oddities -- don't add small numbers to big numbers -- use double precision -- beware denormalized numbers (slow) * Nondeterministic runtimes -- need to leave headroom in cycle time 33

Commissioning MAY 2008 DECEMBER 2009 34

Shot noise L1 DARM Noise. SensMon 19.8 Mpc. −15 10 DARM (13.7 W) Shot Noise (13.7 W) Shot Noise (25 W) iSRD (~8 W) eSRD (~25 W) −16 10 Hz) −17 Displacemnet (m/ � 10 −18 10 −19 10 May 2010 −20 10 2 3 10 10 Frequency (Hz) 35 V.Frolov

Last slide Thanks for listening! 36