EPAC 08, Genoa, Italy RF Reference Signal Distribution System for FAIR M. Bousonville, GSI, Darmstadt, Germany P. Meissner, Technical University Darmstadt, Germany Dipl.-Ing. Michael Bousonville Page 1
Overview Introduction Goal Phase synchronisation Reference Signal System Basic Principle Optical Network Delay Measurement Reference Generation Performance Summary Page 2
Goal 500 m Cavity synchronisation signal generator (DDS) synchronisation central clock Ref Ref signal generator f, ϕ cavity CC Ref Ref Ref Therefore necessary: Distribution of phase synchronous reference signals Ref Ref Problems: Different distances different time delays Time delays not constant τ = f( L,T,... )= f( t) Ref Ref reference generator Page 3
Phase Synchronisation tolerance central clock φ Clock transmission unit reference generator 1 reference generator 2 φ Ref,1 µ 2.5σ j φ Ref,2 φ Crucial: Accuracy between the reference phases Accuracy requirement: 1 at 5.4 MHz Optimisation Parameter: µ and σ Page 4 µ + 5σ j < 514 ps j
Reference Signal Reference 1 T = 20 ns Reference 2 T = 10.24 µs t 0 + n T Ref,2 t f f Clock,1 Clock,2 = 200MHz 97.7kHz f f Ref,1 Ref,2 = 50MHz 97.7kHz Page 5
Basic Principle Assembly of one system branch transmission unit φ Clock φ Clock + φ(τ) φ Ref transmission system reference generator signal generator cavity τ delay measurement unit Any delay variation can be compensated absolute delay drift irrelevant ϕ Ref = f ( ϕ Clock ) f ( τ) Page 6
Basic Principle Star-shaped distribution transmission unit φ Clock φ Clock + φ(τ 1 ) φ Ref transmission reference generator signal generator cavity φ Clock + φ(τ 2 ) φ Ref transmission reference generator signal generator cavity φ Clock + φ(τ N ) φ Ref transmission reference generator signal generator cavity τ n delay measurement unit One instead of N transmission units One instead of N measurement units Page 7 no different time drifts systematic error irrelevant much less effort
Optical Network Configuration of one transmission branch transmission λ Tx 1 1 λ Clock 1 Add/Drop fiber 1 multiplexer FBG λ 1, λ λ Rx 1 1, λ 2, λ M λ M Clock 1 2 λ1, λ 2 demultiplexer λ 2 IN Tx 2 OUT λ Clock 2 2 Rx 2 Clock 2 ADD transmission unit circulator receiver unit λ M λ M Tx Rx I1 measurement unit I2 Page 8
Optical Network Star-shaped distribution transmission unit optical amplifier Gain = M x 3dB splitter 1 x 2 M Add/Drop receiver unit 1 Add/Drop receiver unit 2 distribution Add/Drop receiver unit N optical switch I1 measurement unit reflector (calibration) I2 Page 9
Reference Distribution Optical Network - Prototype laser multiplexer mirrors splitter modulator network analyser switch Page 10
Optical Network - Performance 1. Transmission Channels bandwidth = 1.2 GHz more channels possible 0 dbm opt receiver input dbcele noise < 143 Hz σ jtrans, < 305 fs crosstalk < 70dBele 2. Measurement Channel bandwidth = 10 GHz 0 dbm opt receiver input crosstalk < 130 db total decoupling ele 3. Low Costs only standard components only one transmission unit only one measurement unit Page 11
Delay Measurement Delay determination via phase measurement λ 1, λ 2 IN ADD/ DROP ADD OUT λ 1, λ 2, λ M opt. fibre λ M FBG F M = { } f M 1, f M 2,..., f MN Φ M = { ϕ M 1,ϕ M 2,...,ϕ MN } τ = f ( Φ M ) circulator 1 boundary condition : > 2τ f M,min λ M Tx λ M Rx 1 ϕaccuracy τ accuracy = 2360 f M,max phase measurement f M ϕ accuracy < 0.4 F M = { 50kHz,500kHz,50MHz,6GHz} τ accuracy < 1 6GHz 0.4 = 92,6 fs 2 360 Page 12
Reference Generation delay measurement τ phase correction ϕ cor = f(τ) command data central clock ϕ clock,1 ϕ clock,2 fibre ϕ clock,1 + ϕ(τ) ϕ clock,2 + ϕ(τ) ϕ cor,1 DDS 1 Update ϕ ref,1 signal generator ϕ cavity ϕ cor,2 ϕ ref,2 DDS 2 Update reference generator I1 I2 I3 Page 13
Reference Generation Reference Signal Generation via DDS 1. No phase adjustment limit 2. Resolution of phase adjustment 1.22 ps (Reference Signal 1) 3. Jitter 7.6 ps RMS (Reference Signal 1) 4. Standard components Page 14
Performance - Two reference points - Distance from center clock 1 km each - Average interval 1 s Phase deviation [ps] 20 10 0-10 -20 +0.03 @ 5.4 MHz -0.03 @ 5.4 MHz 0 2 4 6 8 10 12 14 16 Time [h] Current performance µ < 15 ps σ j 7.6 ps ϕ Stability < 1 ps change of DDS-model temperature stabilisation Rx Anticipated performance µ << 15 ps σ j < 250 fs ϕ Stability << 1 ps Page 15
Summary 1. Transmission of two clocks with DWDM noise < 143 dbc ele Hz crosstalk < 70dB σ j,trans < 305 fs more transmission channels possible 2. Separate measurement channel crosstalk < 130dB ele total decoupling measurement accuracy better than 100 fs 3. Reference generation via DDS σ j 7.6ps 4. Phase deviation between two reference points < 0.03 of cavity frequency 5. Only standard components are used Page 16
M.Bousonville@gsi.de Page 17
Additional Slides Page 18
Interfaces Interface 1 ϕ ϕ Clock,1 Clock,2 ϕ ϕ Interface 2 Clock,1 Clock,2 + ϕτ ( ) + ϕτ ( ) τ Interface 3 ϕ = f( ϕ ) f( τ) Ref,1 Clock,1 ϕ = f( ϕ ) f( τ) Ref,2 Clock,2 multiplex central reference source transmission demultiplex delay measurement phase correction RF-DDS asynchronous synchronous Page 19
System Design 1. WDM (Wavelength Division Multiplex) 2. Optimal input power at the optical receiver (0 dbm) This is possible for two reasons a) Low insertion loss of the passive WDM-components b) High transmitter power maximum SNR 3. Crosstalk attenuation >> SNR a) No effect on jitter jitter 1 SNR b) Measurement do not influence the reference signals minimum jitter Page 20
System Design 1. One instead of N transmission units No different time dirfts in different branches 2. One instead of N measurement units Systematic error irrelevant Much less effort Page 21
Realisation WDM-Laser 13 dbm output power RIN < -145 db/hz channel (ITU-norm) opt. frequency ν [THz] opt. wavelength λ [nm] λ 1 32 193,2 1551,72 λ 2 34 193,4 1550,12 λ M 36 193,6 1548,51 Page 22
Realisation External Modulator Page 23
Realisation External Modulator 10 GHz bandwidth under proper conditions no jitter will be added Page 24
Realisation Passive optical components Page 25
Realisation Passive optical components Page 26
Realisation fibre optic cable Low drift velocity dτ ps < 6 L T dt km K min S (calculation: analytical and finite element method) Virtually strain proof dτ L df G ps = 0 km N for F < 4500N G Page 27
Realisation Optical receiver The biggest noise source other than RIN (Relative Intensive Noise) Page 28
Realisation Results Reference Signal Transmission 1. Bandwidth transmitter: 10 GHz 2. Bandwidth receiver: 1.2 GHz System bandwidth: 1.2 GHz 3. Noise dbc ρ N < 143.2 Hz SNR > 52.4dB 4. Crosstalk < -70 db << Noise σ jtrans, s N SNR db 20 9 1 10 5.56 10 305 = < s s 2 N fs Optimisation Parameter J tter: ( σ, ) i σ j f j trans = ok * Sine wave Standard deviation Page 29
Delay Measurement General 1. One separate measurement channel 2. Bandwidth = 10 GHz 3. Measurement channel totally decoupled form the reference channels 4. Nearly all measurement methods are applicable Page 30
Delay Measurement Noise of the measurement channel - Amorphous structure of glass Rayleigh Scattering optical fibre - Noise due Rayleigh Scattering dominates - Spectrum (calculated analytical/numerical) S Norm [dbc/hz] -100-110 -120-130 -140-150 carrier RIN receiver noise shot noise rayleigh backscattering calculation rayleigh backscattering measurement measurement signal -160 0 5 10 15 20 25 f [MHz] 2 2 m 4 2a f 2b f SNorm( f 0) δ ( f ) δ ( f fm ) 1,2 10 = + + + 2 2 2 2 2 π f + f π f + ( f fm ) ( ) SNR B = 10Hz 95dB ϕ < 1 τ < 1ps Optimisation Parameter Mean: Page 31 accuracy ( accuracy ) accuracy µ = f τ ok
Verification Delay Measurement frequency standard 10 MHz 6 GHz 0 dbm switch fiber 1 km network analyser 2 laser CH34 modulator ADD/ DROP 3 db FBG Rx Agilent E8357A 100 m laser CH36 modulator Rx network analyser 1 10 MHz control analysis PC Agilent 8753ES Page 32
Verification of Delay Measurement 60 40 accuracy < 100 ( optical domain) fs 20 deviation [fs] 0-20 Prediction correct! -40-60 0 2 4 6 8 10 12 14 16 18 time [h] Page 33
Test under extreme conditions Fiber length > 1 km; temperature change: -26,5 C 24,5 C Delay [s] Branch 1 5.384E-06 5.382E-06 5.380E-06 5.378E-06 τ > 8 ns 5.376E-06 5.374E-06 5.372E-06 0 5 10 15 Time [h] Branch 2 4.67826E-06 - = Delay [s] 4.67824E-06 4.67822E-06 4.67820E-06 4.67818E-06 4.67816E-06 τ < 0.1 ns 4.67814E-06 0 5 10 15 Time [h] Phase deviation [s] Uncorrected 8.0E-09 6.0E-09 4.0E-09 2.0E-09 φ > 8 ns 0.0E+00-2.0E-09-4.0E-09 0 5 10 15 Time [h] correction Phase deviation [s] Corrected 6E-11 4E-11 2E-11 0-2E-11 φ = 20 ps -4E-11-6E-11 0 5 10 15 Time [h] Page 34