Summary of visit to BNL and LARP workshop High Bandwidth Feedback perspective Preliminary Report by...
-
Upload
charlotte-melton -
Category
Documents
-
view
219 -
download
2
Transcript of Summary of visit to BNL and LARP workshop High Bandwidth Feedback perspective Preliminary Report by...
Summary of visit to BNL and LARP workshopHigh Bandwidth Feedback perspective
Preliminary Report by W. Hofle21.05.2014
Monday: discussions on stochastic cooling with colleagues from BNL, LLRF and transverse feedback;
Tuesday: dedicated Highbandwidth Feedback meeting jointly withM. Brennan and M. Blaskiewicz, LARP team present (J. Fox, O. Turgut, C. Rivetta, S. de Santis) + remote participants (presentations LHC feedback, SPS feedback,stochastic cooling in RHIC)
https://indico.fnal.gov/conferenceDisplay.py?confId=8477
Wednesday Thursday:official LARP part, dedicated session on feedback,
kickers, power amplifiers , reduced models, electronic hardware,and summary presentation by J. Fox in plenary session
Overview
https://indico.bnl.gov/conferenceDisplay.py?confId=730
Phased Schedule
LHC post LS1 Physics Run
Year 1
Phase 1:
20172014201320122011 2015 2016 2018Year 2 Year 3
Demonstrator 1
phase 2 beam tests withBunch trains (ecloud)Scrubbing beamIn ramp and at 450 GeV flat top
Phase 2:
Phase 3: Full implementation
decisionsphase 3:
2019LS1 LS2
Bunch Train Demonstrator R&D and operation
New KickerNew Power amplifiers
Electronics R&D
New Pick-up R&D / prototyping ?as proposedJuly 2013
4
Location of Kickers, Infrastructure• SPS Sextant 3, LSS3• Housed in the second half
of the 321 period of the SPS.
• Approximately 8 m of usable space following the dipole magnet
• 8 new 7/8” cables pulled to kicker location during LS1
• Power amplifiers still needed– Essential for operation of
new kickers!– Critical items for post LS1
operationLARP/HiLumi Collaboration Meeting, CM22
5/7/14
Courtesy of E. Montesinos
15S. De Santis – May 7th , 2014
• The stripline kicker essentially transforms the amplifier output voltage applied at the downstream feedthroughs into a transverse force applied to the travelling particles (integrated deflecting voltage)
Full Assembly - CAD
156 mm 55 mm 92 mm 40 mm
Beam Induced Voltage
17S. De Santis – May 7th , 2014
• The stripline characteristic impedance has been designed to be 50 Ω in the kicking mode (odd TEM), but it is 74 Ω for the even TEM mode.
• Usual remedies (ground planes) are not practicable due to large horizontal stay-clear requirements.
• Beam induced voltage on downstream port (i.e. the amplifier’s) have been estimated by analytical models (time domain, frequency domain) and simulations (Particle Studio)
7
Power Amplifiers Evaluation• Evaluation of 4 power amplifiers in 2013, loaned to us
by manufacturers (work by K. Pollock)– Amplifier Research 250W1000AM4– Amplifier Research KAW-2300 (modified)– MilMega 80RF1000-250– R&K GA020M102-5454R
• Developed testing methods and procedure to characterize amplifiers in both time and frequency domains– Typical frequency domain measurements with network
and spectrum analyzers, measure frequency response, phase linearity
– Time-domain measurements performed with 4 GS/s high speed DAC with the capability of generating arbitrary signals• Burst signals, much like those needed for feedback
• R&K (Japan) seemed to have most desirable characteristics in time-domain, but tests on going
• Specifications:– Frequency = 20 – 1000 MHz– Power = 250 – 500 W– Instantaneous BW– 100% amplitude modulation– Linear phase across band
LARP/HiLumi Collaboration Meeting, CM225/7/14
Time-domain measurements
Time (ns)
Time (ns) Time (ns)
Time (ns)
8
R&K (Japan)• We are working with manufacturers to
improve impulse response and fidelity of amplifiers– Cater to our specifications– Expect modified amplifier to test from R&K
• R&K time-domain test using a 40 MHz burst signal– Able to remove tail of signal– Many manufacturers don’t have the
capability to make the time-domain measurements we do
– Therefore, we need to use our equipment to fully characterize the amplifiers
• Status on R&K– As of April, they’ve informed us that they
need to redesign the 90 degree hybrid combiners
– Seems to be a common issue using several stage amplifiers
– Additional 2.5 months of development R&K’s impulse response measurements
Before
After
LARP/HiLumi Collaboration Meeting, CM225/7/14
9
New Power Amplifier Developments
• We’ve reached out to additional amplifier vendors and have had a very positive response of interest in working with us
• Additional companies– Rhode & Schwarz– AR Pennsylvania– Intertronic Solutions
• Rhode & Schwarz– Most amps are developed for broadcasting
transmission purposes– Berlin development laboratory, a 80 – 1000 MHz,
125 W module, with goal of 1 kW in summer 2015.– Sent us preliminary result, using 200, 400, 800
MHz continuous input signals at 100 W output– Another case where the vendor does not have the
equipment necessary to properly characterize the amplifier for our needs. They have invited us to Berlin to test the amplifier with our high speed DAC
– Very excited to work with us, “…we will do everything humanly possible to make sure we are your amp of choice.”
5/7/14
LARP/HiLumi Collaboration Meeting, CM22
Rhode & Schwarz test of BC125
Source Signal
Amp Output
400 MHz
Good progress !
Q20
Power system
• Strip-line kickers to be installed before start-up with beam• Set of power amplifiers / alternative: use existing power amplifiers, two
amplifier companies seem to qualify, R&K (Japan), R&S Berlin, ramp-up CERN involvement in the selection and purchase
• Re-work phase compensation and signal splitting (new cables, different transfer function)
Expanded demonstrator
• FPGA code for bunch trains (planned was 48 bunches) 16 bunches now considered feasible
• No dedicated FPGA code for bunchlet scrubbing (24x2 bunches)• Control algorithmn for Q20
Pick-up
• keep existing PU system for the moment, possibility to use old or new cablesClocks:
• Revisit synchronization with SPS clocks, tests w/o beam end of August
Plans for immediate future