0 1 Update on Beam Tests with PS Finemet Cavity H. Damerau, L. - PowerPoint PPT Presentation

0

1 Update on Beam Tests with PS Finemet Cavity H. Damerau, L. Ventura LIU-PS Working Group Meeting June 9, 2015 Many thanks to Matthias Haase and Mauro Paoluzzi

2 Overview • Introduction • Coupled-bunch feedback • Beam-loading compensation feedback • Low intensity and high intensity, triple splitting • Excitation of coupled-bunch oscillations • Mode scans • Excitation rates • Summary and outlook

3 Introduction • LS1: - New Finemet cavity installed in SS02 (M. Paoluzzi et al.) • 2014: - First cavity gap available for test with LLRF system - Tests of beam-loading compensation feedback - 12 harmonics damped with nominal intensity LHC25ns • 2015: - 3 gaps available, 4 gaps after June technical stop - Excitation of coupled-bunch instabilities in open loop - Measurement of excitation amplitudes vs. cavity voltage

4 Overview • Introduction • Coupled-bunch feedback • Beam-loading compensation feedback • Low intensity and high intensity, triple splitting • Excitation of coupled-bunch oscillations • Mode scans • Excitation rates • Summary and outlook

5 Beam-loading compensation (1 harmonic) sin( h FB f rev t + f) sin( h FB f rev t ) Low-pass Cavity Cavity ADC DAC return drive Low-pass cos( h FB f rev t ) cos( h FB f rev t + f) cos(( h RF - h FB ) f rev t + f) f s side- Wall band filter current ADC monitor f s side- band filter sin(( h RF - h FB ) f rev t + f)

6 First measurements with beam in 2014 • Moderate feedback gain Open/closed loop transfer function (very first test!) • Transfer function measurement: ~ 10…12 dB • Spectrum of beam induced voltage in Finemet cavity • 26GeV-test cycle, low intensity single bunch 500 kHz 6 MHz accelerated on h = 8  Impedance reduction observed with beam as expected

7 High-intensity, 6/7 filling, triple splitting • First 12 harmonics simultaneously  1 bunch for nominal LHC25ns beam injected from PSB  6 bunch with nominal intensity for LHC25ns injected  15…20 dB reduction of beam induced voltage, also during triple splitting

8 Overview • Introduction • Coupled-bunch feedback • Beam-loading compensation feedback • Low intensity and high intensity, triple splitting • Excitation of coupled-bunch oscillations • Mode scans • Excitation rates • Summary and outlook

9 Coupled-bunch oscillations, freq. domain F. Pedersen, F. Sacherer, PAC77, pp. 1397-1399  Synchrotron frequency sidebands of the f rev harmonics:  In the case of LHC-type beams in the PS ( h = 21) upper lower

10 Beam excitation with the Finemet cavity sin( h FB f rev t + f) sin( h FB f rev t ) Low-pass Cavity Cavity ADC DAC return drive Low-pass cos( h FB f rev t ) cos( h FB f rev t + f) Amplitude Excitation frequency, D f D f sin Low freq. f DDS h FB f rev cos  Excitation frequency ~ f s Side-band selection away from hf rev Amplitude  ~ 400 Hz at 476 kHz

11 First excitation test using Finemet cavity • Observe beam stability during acceleration after transition No excitation, gaps open Excitation at exactly 20 f rev  No effect with voltage from Finemet cavity at f rev harmonic

12 First excitation test using Finemet cavity • Frequency offset of ~ 300 Hz at start of excitation Excitation, D f = + 300 Hz Excitation, D f = - 300 Hz  Strong excitation with frequency offset with respect to 20 f rev  Beam qualitatively behaves as expected

13 Overview • Introduction • Coupled-bunch feedback • Beam-loading compensation feedback • Low intensity and high intensity, triple splitting • Excitation of coupled-bunch oscillations • Mode scans • Excitation rates • Summary and outlook

14 Mode scan measurements LHC25 ns beam with ~1.3  10 11 ppb equivalent intensity • • 4+2 and 4+3 bunches (full ring) injected from PSB g tr or Start excitation • L. Ventura Two independent mode analysis techniques:

15 Mode scan with 18 bunches in h = 21, cavity 11 Data with old coupled bunch feedback  Some modes can be excited very cleanly, others as a mixture; artefact?

16 Mode scan with 18 bunches in h = 21, Finemet New coupled-bunch feedback LLRF, excitation of each mode  All 18 modes can be excited

17 Mode scan with 21 bunches in h = 21, cavity 11 Excite each mode individually and measure mode spectrum  Clean observation of all possible modes

18 Mode scan with 21 bunches in h = 21, Finemet New coupled-bunch feedback LLRF, excitation of each mode Upper side-band: n = n exc Lower side-band: n = 21 - n exc  Every oscillation mode from n = 1…21 can be excited on both side -bands

19 Overview • Introduction • Coupled-bunch feedback • Beam-loading compensation feedback • Low intensity and high intensity, triple splitting • Excitation of coupled-bunch oscillations • Mode scans • Excitation rates • Summary and outlook

20 Excitation amplitude scan Vary excitation amplitude and check mode spectrum ~20 ms after excitation starts: Upper sideband of f rev Lower sideband of f rev  Oscillation amplitude proportional to excitation  linear regime  Mode amplitudes comparable to excitation with spare cavity C10-11 • Absolute excitation voltages to be analyzed

21 Growth of coupled-bunch oscillation Mode amplitude versus time: two measurement techniques Down-converted sidebands at 20 f rev Oscillation amplitude from mode analysis Signal around 20 f rev L. Ventura ~ 38 ms V RF Finemet cav. 40 ms  Both measurement techniques give very similar results  Growth not exponential • How to derive growth times?

22 Growth of coupled-bunch oscillation Mode amplitude versus time: two measurement techniques Down-converted sidebands at 20 f rev Driven harmonic oscillator model Signal around 20 f rev V RF Finemet cav. 40 ms  Both measurement techniques give very similar results  Growth not exponential • How to derive growth times? • Driven harmonic oscillator model?

23 Overview • Introduction • Coupled-bunch feedback • Beam-loading compensation feedback • Low intensity and high intensity, triple splitting • Excitation of coupled-bunch oscillations • Mode scans • Excitation rates • Summary and outlook

24 Summary • Beam induced voltage reduction tests  12 harmonic damped simultaneously with up to 20 dB gain  Nominal intensity of 25 ns beam, follows RF manipulations • First tests without and with beam successful  Coupled-bunch oscillations excited as expected  Each mode can be excited individually  Confirms measurements with C10-11 in 2013  Qualitatively: Finemet cavity touches beam as expected  Quantitatively: More measurements/analysis needed

25 Outlook • Future MDs to complete excitation measurements  Scan excitation amplitude and frequency offset  Derive growth/damping rates for given cavity voltage  Check with shorter bunches (50 ns-like beam)  Excite multiple modes simultaneously  Excite quadrupolar oscillation modes • Follow-up firmware development  Complete filter design for synchrotron frequency side-bands  Close the loop on one harmonic

26 26 THANK YOU FOR YOUR ATTENTION!

27 Operational PS coupled-bunch feedback div/Meetings/APC/2005/apc050609/JL_Vallet_slides.pdf J.-L. Vallet, https://ab-div.web.cern.ch/ab- • Analogue signal processing, two channels Two accelerating cavities as feedback kickers  limits to modes h – 1 and h - 2 •  New wide-band Finemet cavity as kicker Cover all modes  Digital feedback electronics

28 New cavity (#25) in the PS ring • Wide-band (0.4 – >5.5 MHz, V RF = 5 kV) cavity based on Finemet material • No acceleration, but damping of coupled-bunch oscillations SS02 6-cell cavity unit G F Accelerating gap Power amplifiers M. Paoluzzi (solid state) • Cavity installed in SS02, amplifiers on 2 gaps  First installation of transistor power amplifiers close to beam in PS

29 Damping rate versus gain and intensity Versus gain Measured damping rate with feedback on Corrected for natural damping  Zero damping at zero gain  Natural damping in- dependent from gain Versus intensity  Damping increases with intensity, more signal for given CB oscillation amplitude ‐ Saturation leads to non- zero damping with zero N p ?

Damping rate versus e l and cycle time 30 Versus e l (RMS)  Uncorrected: damping efficiency increases for larger e l  Reduced natural stability for smaller e l  Corrected damping independent from e l During cycle ~26 GeV ~26 GeV  Damping efficiency reduces at higher ~9 GeV energy ~9 GeV  To be checked with simulations

31 Kick voltage versus oscillation amplitude • Excite a coupled-bunch oscillation and measure its amplitude • Observe maximum damping voltage required • Only order of magnitude for kick voltage  Overestimate expected as feedback normally started before oscillations are well developed Basic specifications of kicker Frequency range 0.4 to 5.5 MHz cavity: RF voltage per sideband, V mode ~ 1 kV Maximum total RF voltage, V max ~ 5 kV M. Paoluzzi, H.D., < 200 W Un-damped shunt impedance at n · f rev CERN-ACC-NOTE-2013-0019