First, Yamamura briefly reviewed previous studies on the dark current simulation. Next, he showed results of the simulation in the LINAC by using his program and SAD. In his program, dark currents was simulated in the first eight accelerating structures ( 400 cells in total). Then, survived electrons in the dark current were input to SAD program for simulation with the nominal LINAC lattice. As a result, there was no survived electron. Therefore, he estimated phase space for survival in the SAD program, where initial conditions of generated electrons are uniform in (xi,yi) and (x'i,y'i). Resultant phase space was very small, which were |xi|,|yi|<10-32m.
Accelerating gradients were set to be 50MV/m and 60MV/m in his program and the SAD, respectively. Yokoya pointed out that there must be a capture threshold which strongly depends on the gradient as well as the frequency. At the capture threshold, electrons at rest can ride on accelerating phase of traveling wave in the accelerating structures. In the x-band case, the capture threshold gradient would be around 60MV/m.
Following discussions, immediate tasks are (1) analytic estimation of the capture threshold and (2) calculation of survival probability as a function of beam line, and (3) dark current profile at SBPM (structure BPM) and QBPM (quadrupole BPM) for cavity-BPMs are sensitive to the asymmetric profile (p209, GLC Project, KEK Report 2003-7).
(2) FEATHER status and feedback with the beam blow-up (N.Delerue)
(transparencies, 12 pages, pdf, 651KB)
Nicolas showed results of the movable BPM calibration by beam. Opening fully the kicker gap, position signals, which are differences between upper and lower electrode signals and is linear to the vertical beam position there, could be seen down to 0.3mm BPM gap. Especially, at 1mm BPM gap, the 1mm was experimentally verified with the beam. However, at smaller gap, the signal seems to be dominated by noise. Closing the kicker gap, the BPM acceptance was observed to decrease as the beam could not scan in the 1mm BPM gap together with 1.2mm kicker gap. He could find a good beam orbit with the 2mm kicker gap and 0.5mm BPM gap although the signal was noisy. So, the nominal values were set to 2mm kicker gap and 1mm BPM gap for the feedback study (dated 1 June 2004), while previous feedback test (dated 20 May 2004) with no detector circuit has used 0.5mm BPM gap. BPM sensitivity was also shown in unit of mV/um as a function of BPM gap, e.g. 0.6mV/um at 2mm gap. Each sensitivity point represent a slope of linear fit in the position signals and the beam positions at a specific gap, while RMS of the signals relative to the linear fit was measured to be 5-20mV.
The feedback circuit is briefly described as follows. Upper and lower electrode BPM signals are input to a hybrid combiner, where their sum (&Delta) and difference(&Sigma) are output. The upper electrode and sum signals are labeled by A and C, respectively, in figures. The &Sigma signal is or is not detected and amplified (~20dB) which is labeled by B. Finally, the B signal is amplified by LZY-1 (RF amplifier with 39dB gain and 20-512MHz band width ), which is labeled by D. A, B and D signals are divided ones by couplers/splitters. D signals were observed with 20-30dB attenuator by a 1GHz digital oscilloscope. "LZY-1 off" means that LZY-1 is shut-off by a remote control, where outputs are attenuated to -50dB.
Results of the feedback test were shown in "With No Detector Circuit, 0.5mm BPM gap and 2mm kicker gap, 20 May, 2004 data " and "With Detector Circuit, 1mm BPM gap and 2mm kicker gap, 1 June 2004 data ". Although we could see different BPM signals (labeled A) between the two results, no feedback effect was observed in both cases.
He found that train to train fluctuation of the BPM signal were dominant the measurement. Also, the beam positions were measured by a wire scanner (MW2X) right in front of the BPM, with LYZ-1 on and off. While the wire scanner can only measure average beam profile/position for 20 bunches with present system, an order of 20 um kick-effect might be observed, which is still comparable with measurement resolution and the train-to-train fluctuation. Therefore, we need to get rid of these fluctuations.
Also, he showed results of feedback simulations with taking account of bean blow up in a case of 7mrad crossing angle. Initial offsets were set to 2 and 10 in unit of IP vertical beam size. With proper gain in the feedback loop, the offset could be corrected as similar to those without the beam blow up.