1. UK FEL development package WP6:
femtosecond synchronisation and stability
Dr Steven Jamison
Accelerator Science and Technology Centre (ASTeC)
STFC Daresbury National Laboratory
Daresbury, U.K.
S.P. Jamison , UK FEL meeting, STFC Daresbury Laboratory, July 2017

2. Context Pump-probe science at LCLS - science obscured by timing jitter
- can be recovered from time-sorting, but with limitations.
LCLS: ~ 170fs rms jitter
LCLS measurement of Bismuth phonon dynamics, with & without time sorting
S.P. Jamison , UK FEL meeting, STFC Daresbury Laboratory, July 2017

3. WP6: femtosecond synchronisation and stability
Task 6.1 Characterisation of laser oscillator, amplified laser, master and high power RF phase noise sources with CLARA test-bed.
Systematic measurements of relative phase or arrival time in laser and RF systems throughout the CLARA facility.
Measurements will require development of suitable RF-laser and laser-laser cross-correlation diagnostics, while low power RF measurements will utilise precision RF phase noise test equipment.
Beam-based measurements (Schottky scan; energy stability) will complement direct sub-system investigations.
Mitigations and sub-systems stability improvements will be implemented as part of continuous improvement. Where key subsystem solutions or problems are identified, knowledge exchange relationships will be sought with key laser and/or RF commercial providers to prepare these key technologies for a future UKFEL .
S.P. Jamison , UK FEL meeting, STFC Daresbury Laboratory, July 2017

4. Results RF oscillator Phase jitter analysis of RF clocks
S.P. Jamison , UK FEL meeting, STFC Daresbury Laboratory, July 2017

5. Analysis RF oscillator Phase jitter analysis of RF clocks
Clock performance is dependent on selected frequency bandwidth:
Ralab (VELA/CLARA ‘custom’ MO, repaired February 2017): integrated Δτrms = 86 fs
Far out (>1 kHz): +2 fs (best performance)
Close in (<1 kHz): +84 fs (worst performance)
Rohde and Schwartz SMA100A: integrated Δτrms = 27 fs
Far out (>1 kHz): +19 fs (worst performance)
Close in (<1 kHz): +7 fs (best performance)
Agilent (temporary loan): integrated Δτrms = 24 fs
Far out (>1 kHz): +16 fs
Close in (<1 kHz): 8 fs
Spurs in Rohde and Schwartz at ~60, 120 Hz – mains coupling which accounts for +5 fs
From device, or measurement artefact (signal contamination)?
What noise bandwidth is important?
S.P. Jamison , UK FEL meeting, STFC Daresbury Laboratory, July 2017

6. Results Laser oscillator Phase jitter analysis of ‘laser’
Strictly speaking: the 36th harmonic of a signal produced by filtering and amplifying the voltage output produced by a fast photodiode generated by incident laser pulses from the optical cavity at a fundamental repetition rate of ~ MHz
Phase jitter analysis of ‘laser’
Spurs at 60/120 Hz from device – these are transferred to laser through PLL, not present in free-running mode
S.P. Jamison , UK FEL meeting, STFC Daresbury Laboratory, July 2017

7. Analysis Laser oscillator Phase jitter analysis of laser
Laser-RF PLL has finite bandwidth, close in only: 10 Hz – 1 kHz
PLL attempts to follow clock input over this range, so influenced by clock noise – note correlation of spurious
Far-out contribution is limited by laser free running noise
Locked/unlocked are very similar outside of PLL bandwidth
S.P. Jamison , UK FEL meeting, STFC Daresbury Laboratory, July 2017

8. Laser-RF synchronization
Results
Laser-RF synchronization
Data so far has been ‘absolute’ noise – really the relative noise of the device with respect to the internal oscillator of the analyser. Statistics prevent us using absolute noise data to be make quantitative predictions about the relative noise between two devices, despite the common reference…
Laser-RF relative noise
Effect of PLL (direct)
RF better than laser
Laser better than RF
The phase noise analyser has an option to pass the internal oscillator (like SMA?) as an output; by locking the laser to this we can do a relative noise measurement of the Femtolock at MHz.
S.P. Jamison , UK FEL meeting, STFC Daresbury Laboratory, July 2017

9. Laser-RF synchronization
Analysis
Laser-RF synchronization
Laser-RF relative noise
Integrated relative noise (synchronization) between laser and RF clock input (10 Hz – 1 MHz) – ΔτRMS = 13 fs
Manufacturer quotes 7 fs for integrated 3 kHz – 10 MHz; c.f. 8 fs for 3 kHz to 1 MHz measured in-house
Good agreement but also dependent on oscillator choice (manufacturer did not provide that information)
Based on absolute noise spectra and relative measurement, can make reasonable judgement as to factors limiting relative noise:
10 Hz – 1 kHz; 2.4 fs – direct effect of PLL (laser cavity doesn’t quite follow source)
1 kHz – 10 kHz; 4.8 fs – RF oscillator is better than free-running laser
10 kHz fs – free-running laser noise is better than RF oscillator
S.P. Jamison , UK FEL meeting, STFC Daresbury Laboratory, July 2017

10. Results Clock distribution
Relative phase jitter analysis of Libera Sync clock distribution
Integrated (10 Hz – 1 MHz): Δτrms = 8.2 fs
Close in (<1 kHz): 1.1 fs
Far out (> 1 kHz): 7.1 fs
S.P. Jamison , UK FEL meeting, STFC Daresbury Laboratory, July 2017

11. Analysis Clock distribution
Additive phase jitter analysis of Libera Sync clock distribution
Note: relative measurement of distribution performed: compares RF clock before and after distribution
Performance is dependent on selected frequency bandwidth:
Far out: +7.1 fs – but not real? Comment from manufacturer i-tech:
“There is a bump in the MHz range that is not real and is present because of two reasons: 1) there is a narrow filter in the receiver unit that cuts the noise in the MHz range and when comparing the source with the one coming out it seems like the Libera is adding noise…”
“There is another contribution to the bump because of the integrated optics. The optical signal is travelling through 800 m of optical fiber (spool in the transmitter unit) and at the end we have a kind of phase dispersion in the noise. The difference [between the input and output] can be observed as a bump”
“If a very clean RF source is used the contribution can be reduced. Stephan Hunziker did such measurements at PSI and obtained 3.8 fs integrated (10 Hz – 10 MHz)
Far out then may be as low as 2.7 fs…
This is a state-of-the-art device
S.P. Jamison , UK FEL meeting, STFC Daresbury Laboratory, July 2017

12. WP6: femtosecond synchronisation and stability
Task Hz stability
Determine the sub-system causes of shot-shot instabilities in the electron beam properties, using CLARA as test-bed (e.g. electrical mains phase dependence RF phase and amplitude); develop feedforward solutions enabling stable mains-asynchronise pulsing.
S.P. Jamison , UK FEL meeting, STFC Daresbury Laboratory, July 2017

13. WP6: femtosecond synchronisation and stability
Task 6.3 Optical master oscillator
Develop and demonstrate an RF master oscillator derived directly from the photocathode laser oscillator;
evaluate stability performance, confirm (or otherwise) improvements over existing RF and optical master oscillators.
S.P. Jamison , UK FEL meeting, STFC Daresbury Laboratory, July 2017

14. Analysis Laser oscillator Laser as RF clock?
Integrated jitter (10 Hz – 10 MHz) of RF extracted from locked laser is comparable to RF oscillator:
Laser (+ photodiode, low noise amplifier): ΔτRMS = 29 fs
Rohde & Schwartz SMA100: ΔτRMS = 27 fs
Why is locked laser a good clock? Forming hybrid with best of both:
Good close in performance of RF oscillator
Good far out performance of laser
Emphasises importance of relationship between PLL bandwidth and free running noise; for low integrated phase noise
Reference should have low phase noise in PLL range
Locked client should have low free-running phase noise outside PLL range
S.P. Jamison , UK FEL meeting, STFC Daresbury Laboratory, July 2017

15. Optical RF oscillator – CLARA development & testing
RF distribution
LLRF
Linacs
S.P. Jamison , UK FEL meeting, STFC Daresbury Laboratory, July 2017

16. Optical RF oscillator – CLARA development & testing
First working system in hand
….optimisation started
RF distribution
LLRF
Linacs
S.P. Jamison , UK FEL meeting, STFC Daresbury Laboratory, July 2017

17. WP6: femtosecond synchronisation and stability
Task 6.4 sub-5 femtosecond electron beam and photon beam arrival diagnostics
Develop electron, photon and RF arrival-time/phase diagnostics for the above programme, and for use in UKFEL stabilisation and time-stamping systems.
S.P. Jamison , UK FEL meeting, STFC Daresbury Laboratory, July 2017

18. WP6: femtosecond synchronisation and stability
Task 6.5 Start-end timing stability evaluations for ‘model’ UKFEL
Examine accelerator and FEL design with stability optimisation.
S.P. Jamison , UK FEL meeting, STFC Daresbury Laboratory, July 2017

19. Slide title
S.P. Jamison , UK FEL meeting, STFC Daresbury Laboratory, July 2017