Commissioning of the Bunch By Bunch Feedback System
Commissioning of the Bunch By Bunch Feedback System U. Iriso, J. Moldes, A. Olmos (ALBA – CELLS) M. Abbot, G. Rehm (Diamond) U. Iriso ESLS XXIII Nov. 2015
Outline • BBB Layout • BBB Commissioning • System Features U. Iriso ESLS XXIII Nov. 2015
Bb. B Layout Hybrids Libera Front. End IFI 100 W ampli Libera Bb. B Button BPM Stripline Kickers Scope 100 W atten U. Iriso ESLS XXIII Nov. 2015
Bb. B Layout (2015) Timing EVR Clk 1 - 500 MHz ALBA Master Clock 500 MHz Bb. B - X Xout Yout RF FE Hybrids A B C D BPM Bb. B - Y Clk 2 - 500 MHz 100 W FFK - Y Load Scope 100 W Att FFK - X 100 W U. Iriso Four Amplis Configuration 100 W Att ESLS XXIII Nov. 2015
Bb. B Layout (2016) Timing EVR Clk 1 - 500 MHz ALBA Master Clock 500 MHz Bb. B - X Xout Yout RF FE Hybrids A B C D BPM Bb. B - Y Clk 2 - 500 MHz 100 W S 180 -HP U. Iriso FFK - Y Load Scope 100 W Att FFK - X ESLS XXIII Two Amplis Configuration + High Power Splitters Nov. 2015
Bb. B Layout (2016) New Hardware at ALBA – ready to test S 180 -HP LPF [0 - 300]MHz (to protect amplis) 100 W Ampli U. Iriso ESLS XXIII Nov. 2015
Outline • BBB Layout • BBB Commissioning • System Features U. Iriso ESLS XXIII Nov. 2015
Bb. B – Diamond TMBF Commissioning We did a not very successful try with I-Tech Matlab interface Diamond colleagues offered a collaboration to port Diamond TMBF to ALBA system Main steps: * HW checks/modifications on Amplifiers, Hybrids and phase matching of cables * Migrate the full FPGA code from System. Verilog to VHDL language * Adapt it to ALBA machine (harmonic number, revolution frequency, tunes, …) * Install an “EPICS worm” (EPICS Base) inside our Tango control room * Build libca library to transfer data to Matlab * Permanently install Diamond TMBF code in our Liberas Bb. B (“brain” washed) TMBF Commissioning with beam with M. Abbot and f G. Rehm (May 2015) U. Iriso ESLS XXIII Nov. 2015
Bb. B – Diamond TMBF Commissioning Phase Offset Optimization (Mixer) Hor RF Front. End X/Y outputs Ver Need to maximize the signal to have the highest levels at Libera BBB Phase. X=20 deg ; Phase. Y=-20 deg U. Iriso ESLS XXIII Nov. 2015
Bb. B – Diamond TMBF Commissioning Amplifiers characterization One of the amplifiers broke, so we could only install the BBB in the one plane (ver) [0 -200] MHz, Df~10 deg ok Clear degradation occurs >200 MHz 50 d. B gain – quite flat over all interesting freq range [0 -250]MHz U. Iriso ESLS XXIII Nov. 2015
Bb. B – Diamond TMBF Commissioning Single Pulse from Libera BBB + Ampli + Kickers + Attenuators Single Pulse (2 ns) from Libera BBB The amplis distort the signal, produces a ripple of ~6 ns We will not kick a single bunch, but alsoadjacent bunches! In the near future, we plan to exchange our amplis U. Iriso ESLS XXIII Nov. 2015
Bb. B – Diamond TMBF Commissioning Fine phase delay tuning adjusted adding short cables in the loop: 300 ps, 500 ps, 1 ns Synchronize BBB pulse with bunch passage (single bunch) BBB Pulse + delay + 1 ns cable Bunch Passage U. Iriso ESLS XXIII Nov. 2015
Bb. B – Diamond TMBF Commissioning FIR Phase adjustment with stable beam Decrease usual chroms and switch on BBB… It works!! BBB OFF BBB ON Tune Phase ~ +/-180 deg U. Iriso ESLS XXIII Nov. 2015
Bb. B – Diamond TMBF Commissioning DCCT and beam size evolution during first switch On Inject 100 m. A, uniform filling, chroms = (1. 6, 0. 2) 30 min U. Iriso ESLS XXIII Nov. 2015
Outline • BBB Layout • BBB Commissioning • System Features U. Iriso ESLS XXIII Nov. 2015
Bb. B – Diamond TMBF Features Diamond TMBF allows to: • Setup different “Banks”: ways to act on the beam (damping, excitation, sweep…) • Combine them with the “Sequencer” • This can produce interesting features (see next, and also Guenther’s ppt at DEELS 14 ) U. Iriso ESLS XXIII Nov. 2015
Bb. B – Diamond TMBF Features Unstable modes identification Prepare the following sequence: 1. 2. 3. 4. 5. 6. U. Iriso Get a stable beam using TMBF Excite mode m Measure the growth time Switch TMBF back on for that mode Measure damping time Damp again mode m, and repeat for m+1 ESLS XXIII Nov. 2015
Bb. B – Diamond TMBF Features Unstable modes identification Growth rate, 1/turns ? RW No Damping Excitation Mode m U. Iriso ESLS XXIII Nov. 2015
Bb. B – Diamond TMBF Features Bunch Cleaning: • Define a backwards sweep around Qv • Activate DAC only in buckets B 77 & B 80 Killed B 76, Killed B 79, B 80 B 77 Before BC After BC Killed in previous tests U. Iriso ESLS XXIII Nov. 2015
Bb. B – Diamond TMBF Features Bunch Cleaning: Easier to clean a single bunch! Parasitic bunches are produced by the Linac with 18 ns delay!? U. Iriso ESLS XXIII Nov. 2015
Bb. B – Diamond TMBF Features Tune Measurement using PLL: • Measure beam phase response to an input freq (needs to be very close to Qv) • Keep track of frequency (tune) change to match phase • DQ ~ 1 e-5! • Input signal needs to be very small, compatible with normal operation • In this case, one measurement/200 turns • Applied only to single bucket U. Iriso ESLS XXIII Nov. 2015
Bb. B – Diamond TMBF Features Tune Measurement using PLL: Compatible with tune jitter measurements with other means: FFT Amplitude Main tune jitter peaks at 100 and 300 Hz Frequency, Hz Measurement taken with Chrom. V~0 FFT of a BPM buffer of 2. 5 e 5 turns U. Iriso ESLS XXIII Nov. 2015
Bb. B – Conclusions • Diamond TMBF has been commissioned at ALBA • Quite easy and straight-forward (thanks to Guenther and Michael!) • An “EPICS worm” has been installed in our Tango-based Control Room • Diamond TMBF allows several interesting features, already tested at ALBA: o Sequences to identify unstable modes o PLL tune tracking with 1 e-5 precision (compatible with machine operation!) • So far, only commissioned in the vertical plane due to hardware (ampli) failure. U. Iriso ESLS XXIII Nov. 2015
http: //www. ibic 2016. org/ Thank you and see you soon! U. Iriso ESLS XXIII Nov. 2015
BBB – Extra slides U. Iriso ESLS XXIII Nov. 2015
Commissioning procedure * Finding proper RF FE settings (phase offset, clock phase, …) * Problem in Hybrids X signal * Bb. B “brain washing” failed processor board already reflashed at Diamond * Ugly single pulse out of the amplifier. That will affect adjacent bunches * 2 ns jump of the single pulse out of one Bb. B unit. Diamond has reproduced the problem in their lab (DAC initialisation) * Change ALBA timing to TMBF needs * Aligning system to the bunches @ machine SBM * Inject 100 m. A uniform filling beam * Vertical loop closed just by setting the feedback phase * Chrom reduced to ‘ 0’ and beam kept stable * Bunch cleaning test: - no bunch kicking until scrapper closed - 2 bunches kicked out because of ampli * Unstable modes identification (m=-8) * * U. Iriso ESLS XXIII Nov. 2015
Extra slides: 2 ns jumps Scope in persistence mode, to show the 2 ns jump (green). The violet trace is the trigger. U. Iriso ESLS XXIII Nov. 2015
Extra Slides: ADC offsets Fixed w. o. beam, just using the GUI capabilities U. Iriso ESLS XXIII Nov. 2015
- Slides: 28