Predictions of LER-HER limits

Transcription

1 Predictions of LER-HER limits PEP-II High Current Performance T. Mastorides, C. Rivetta, J.D. Fox, D. Van Winkle Accelerator Technology Research Div., SLAC 2e 34 Meeting, May 2, 27

2 Contents In this presentation we want to address the expected issues as currents are increased towards the planned 4 A in LER and 2.2 A in the HER. Three kinds of limitations studied: Klystron power limits (individual, non-linear data included) Stability of RF Stations (ability to "tune" the stations) Beam stability through extraction of low mode growth rates and comparison with the available damping rates. Successful operation of the machine requires all three conditions. Trade-offs among them may be necessary. The simulation allows us to predict ultimate limits, understand limitations to the technology or implementation details/imperfections without spending machine time. Extensive description of the simulation, its validations, conclusions and predictions available at our PRST paper:

3 PEP-II System Simulation The simulation uses a reduced model to capture the low order mode dynamics in the system bunches are grouped into Macrobunches. Preserves Beam Low Mode Dynamics. All 2 cavity stations are grouped into one macrostation and all 4 cavity stations are grouped in another. Macro stations include the principal loops used in the control feedback impedance. The simulation uses the same algorithms to optimally configure the RF stations as used in the real machine. The growth/damping rates of the beam are measured by the same tools in the simulation and the machine. The similarity and parameter matching between the real machine and the simulation allows studies of parameter sensitivity and the impacts of technical imperfections on the beam stability.

4 Reduced Model - Simulation Slow dynamic loops define the operation point for the fast model. Associated variables do not change in the time frame set in the simulation. Fast Dynamics are modeled. Not modeled components are also shown (but considered through the comparison of growth and damping rates in the analysis). The direct and comb loops are necessary to reduce the impedance of the cavity as seen by the beam control the growth rates. Station reference Gap Loop Mod. RF reference Band limited kick signal Error + Σ Driver Comb Loop Klys. sat. Loop Klystron Direct Loop Longitudinal Low Group Delay Woofer HVPS RF cav. Beam BPM Tuner Loop

5 Limit definition Not hard limits We are predicting the operational points where it will become increasingly difficult to tune the stations and the number of aborts will significantly increase. An example from run 5 shows how a small increase in current can greatly increase the number of aborts. We are trying to define these points.

6 Tuning the RF Station LLRF stability is sensitive to RF station parameters, set by the optimization tools to satisfy gain and phase margins. The RF station parameters are the gain, phase and delay of the direct and comb loops, the detuning frequency and the Q of the cavity. Since the open loop transfer function of the real machine or the simulation cannot be measured, it is estimated using the closed loop transfer function, through an identification based on a linear model. Based on the estimated open loop transfer function, the optimal controller parameters can be calculated. The resulting operating point defines growth rates and is sensitive to beam current. It is very significant in estimating station stability. Station tuning has a great effect on machine performance.

7 Klystron Saturation Effects on Tuning Tuning severely affected by klystron saturation, which reduces the impedance control (parameter optimization). At normal operation, input power to the klystron is around 15 W. To increase output power, power supply voltage is increased rather than input power (move to higher curve). When the required power goes over 8 kw, an increase in input power is also necessary, which leads into saturation and poor control impedance (reduced station stability). Power out (kw) kv 65 kv 7 kv 75 kv 8 kv 82 kv Marconi #4 Power Curve Power in (W)

8 LER Station Stability Klystron Saturation effects From the klystron power curves, we see that the maximum klystron forward power is operationally unsustainable due to high saturation, which causes distortion (peaking). Station stability sets power limits to 13 kw for SLAC tubes and 93 kw for Phillips/Marconi tubes. These limits are based on tuning experience, and allow small drifts and non-uniform station settings as done in practice. 1 Fit Data Gain (db) Frequency (khz)

9 LER Klystron Forward Power Limits 12 Forward Power vs. Current for LER Forward Power (kw) Heinz 4.5 MV Simulation 4.5 MV Heinz 4.5 MV Simulation 4.5 MV Marconi Limit SLAC Limit Current (A) Forward Power (kw) Forward Power vs. Gap Voltage for LER 3 A 3.6 A 4 A Marconi Limit SLAC Limit Figure: Klystron forward power vs. current estimated by the simulation and H. Schwarz s spreadsheet. Limit crossed at 35 ma with 4.5 MV and 365 ma with 4.5 MV. For SLAC tubes, 375 ma and 39 ma respectively Gap Voltage (MV) Figure: Power vs. Gap Voltage. For 3.6 A need at least 4.4 MV (marginal). For 4 A need SLAC tubes and higher gap voltage!

10 LER Growth Rates (4.5 MV) With existing configuration, the limit due to growth rates is at about 31 ma. Improvements necessary to reach 35 ma (new Amps, Comb rotation, SLAC tubes). 4 Growth Rates for LER at 4.5 MV 1.5 Damping Rates for LER at 4.5 MV ms Existing Configuration Improved Driver Amp Imp. Driver Amp and 1 SLAC tube Imp. Driver Amp and 2 SLAC tubes Comb Rotation + Improved Amp Comb Rotation + Imp. Amp + SLAC tube Net Empirical Damping Rate (ms 1 ) Existing Configuration Improved Driver Amp Imp. Driver Amp and 1 SLAC tube Imp. Driver Amp and 2 SLAC tubes Comb Rotation + Improved Amp Comb Rotation + Imp. Amp + SLAC tube Beam Current (ma) Beam Current (ma)

11 LER Growth Rates (4.5 MV) Growth rates limit at about 325 ma. Approached this limit at the end of run 5 (slide 8). Combination of improvements can lead to 36 ma, with small margin. All possible improvements and/or 5.4 MV (HOMs? Heating?) needed for wider margin. 4 Growth Rates for LER at 4.5 MV 2 Damping Rates for LER at 4.5 MV ms Existing Configuration Improved Driver Amp Imp. Driver Amp and 1 SLAC tube Imp. Driver Amp and 2 SLAC tubes Comb Rotation + Improved Amp Comb Rotation + Imp. Amp + SLAC tube 5.4 MV Gap Voltage Net Empirical Damping Rate (ms 1 ) Existing Configuration Improved Driver Amp Imp. Driver Amp and 1 SLAC tube Imp. Driver Amp and 2 SLAC tubes Comb Rotation + Improved Amp Comb Rotation + Imp. Amp + SLAC tube 5.4 MV Gap Voltage Beam Current (ma) Beam Current (ma)

12 Comb Rotation New Configuration The comb rotation is an example of the new optimal configurations that have been developed and of the trade-offs between limitations. It was determined that a small reduction on the phase margin for the direct and comb loops (reducing station stability) can lead to a substantial reduction in Growth Rates. A rotation of the comb phase by 1 has been implemented in the LER (limited by RFP and amplifier distortion). LER Growth rates (ms 1 ) Oscillation frequencies (Hz) o rotation 1 o rotation Nominal Case Mode Number

13 Amplifier Distortion A discrepancy of the klystron response with the real machine led to further understanding of the significance of the klystron driver amplifier s intermodulation. These amplifiers performed as expected with a single tone input, but their responses are plotted below for a two tone input. This distortion prevents optimal station configuration at high power and leads to a GR increase. The amplifiers are being replaced based on this discovery. Phase (deg) Magnitude (db) 1 P out =1W 8 P out =15W 6 P out =2W 4 P =25W out 2 P out =5W x Frequency (Hz) x 1 6

14 LER Estimation Uncertainty Damping Rates estimated at high currents (uncharted territory). The operational HVPS limit for SLAC tubes is set to 82 kv. In the simulation, the same limit is used for the Phillips/Marconi tubes, even though some of them have higher specifications and may be able to operate at higher voltage. This choice is based on their long filament hours and other problems. The specifications for the new amplifiers include a flat response in the bandwidth of interest, but they have not been installed yet and their effect cannot be verified. How will possible distortions in the RFP module affect our ability to apply the comb rotation? (presently limited to 1 )

15 HER Power Limits Power is dominant HER limitation Balance in power for 2 and 4 cavity stations 18.5 MV Gap Voltage for 2.2 A. Small margin implies sensitivity to drifts, Amp and RFP distortion and station non-idealities. Forward Power (kw) Forward Power vs. Gap Voltage for HER (2.2 A) constant voltage per cavity) 2 Cavity Stations 4 Cavity Stations Estimated Limit GR estimated to 1.2 ms 1 for 2.2 A and 16.4 MV, well within our damping range. Station stability within reasonable margins Gap Voltage (MV)

16 Summary The simulation has been used to study the existing system and planned upgrades. Can be used to study any other possible configurations. The HER should be able to achieve the maximum current planned for run 6 (2.2 A) with right balance of power between 2 and 4 cavity stations. For the LER and the existing configuration: expect 31 ma with 4.5 MV and 325 ma with 4.5 MV. with improvements installation (new amps, SLAC tubes, additional comb rotation): expect 35 ma with 4.5 MV and 36 ma with 4.5 MV. For 4 A, all SLAC tubes will be needed for the LER and a further increase in Gap Voltage (HOM? Heating?)