Minutes of 71st FFIR/BDSIM meeting on 4/8/2004

The meeting was held in a room of 425 at KEK, 10:00-12:00, 4/8/2004. We discussed on final focus system, FEATHER and dark current at the LINAC.

(1) Support tube R&D (H. Yamaoka)

(transparencies, 7 pages, pdf, 740KB)

Yamaoka updated the support tube vibrational tests and their evaluations by the ANSYS. Previous measurements shows large discrepancies from the calculations. Since they are considered to be due to weaker H-shaped support than that used in the calculations, removing the support was changed to a plate whose equivalent thickness and width were estimated to be 96mm and 170mm, respectively. As in the previous measurements, results are summarized below;
Structureloads (kg)1st (Hz)2nd (Hz)3rd (Hz)ANSYS:1st (Hz)2nd (Hz)3rd (Hz)
tube (a)0.0123343674104278517
tube (a)0.89+0.898627763094238498
tube (a)1.78+0.898124657276233487
tube (b)0.09312260996123483
, where tube (a) has a central tube of 3mm thickness and 200mm length with relative stiffness of 1/2.6, and two cantilevers in the tube (b) are connected with two thin plate of 1mm thickness, 20mm width and 200mm length with relative stiffness of 1/1000. Therefore, agreements became better.

He also estimated vibrational properties of the real sized support tube ( 80cm-diameter, 16m long ) by ANSYS as a function of thickness of central CFRP tube. In order to increase the resonant frequencies, i.e. from 17.8 Hz to 75.6 Hz at the 1st resonance, additional supports have been introduced at 3.85m from the center as in the ACFA report. The input GM power spectrum has been fitted one from measured spectrum at ATF. Differences of the amplitude in ±2m (nm/Hz1/2) were calculated at the 1st and 2nd resonances as follows;
1st mode2nd mode
In the above calculations, the two "cantilevers" were assumed to be identical. Next, he estimated for the non-identical case, where the two cantilevers have 75 and 70Hz at the 1st mode with slightly different thicknesses. Results are listed below;
1st mode2nd mode
During discussions, a serious question was raised. If the limit with null thickness corresponds to the cantilever system, the central tube can only suppress a little relative displacement at the 1st mode in this case. Therefore, it must be very important to have the same vibrational property even for the tube structure. We should know how must is practical error in the resonant frequencies for the "same" cantilevers.

(2) Final focus system (S. Kuroda)

(transparencies, 7 pages, pdf, 436KB)

Kuroda estimated alignment tolerances at quadrupole magnets (Q's) for the L*=3.5m optics (the Roadmap report) at the first setout (link to the BDS system). So, beam sizes (&sigma*x, &sigma*y ) and horizontal beam position (x*) were calculated as a function of horizontal displacements of each quadrupole magnet, whose dynamic range was within ± 10 μm. He found several sensitive Q's to the beam sizes, which are QD8, QF7, QD6, QF5.{1,2}, QD2B, QF3, QD2A and QF1. The horizontal displacements of Q's have linear relation to x* except for QF7 and QD0. Finally, he listed the displacements corresponding to 50% increase of the beam sizes and x*=1&sigma*x for all the Q's. The most sensitive Q to the beam sizes was QF2 except for final doublets, which was 2μm for 50% increase of vertical beam size, while QF9.{1,2},{3,4} and QF5.{1,2},{3,4} are the most sensitive to x*; i.e. 0.5μm for displacement of 1 &sigma*x. These results assumed no correction. Therefore, actual tolerances shall be estimated with corrections. Also, the tolerance shall be given in terms of emittance growth (e.g. Δ&epsilon/&epsilon=10%) . One major interest is requirements (setting resolution and dynamic range etc.) for movers of final doublets.

(3) Dark current in LINAC (T.Yamamura)

(transparencies, 23 pages, pdf, 14.1MB)

Yamamura refined the dark current simulation by dividing the generation area of dark current into 4 regions in a cell; A, B, C and D are top surface, upstream side, downstream side of a iris and adjacent cavity surface (upstream of the iris). In this simulation, the iris is not rounded; unloaded gradient is 65MV/m; 5/6&pi /cell; mode. In A, electrons can survive only in the half upstream region, where their final phases are clustered around two values of -0.7 and -5.58 radian. In the similar way, dark currents were simulated in B,C and D. The final phases were also clustered as those in A. A and B are the most likely survived regions. He also showed the final energies of dark currents as a function of generated position. The higher energy dark currents are generated in A. In future, he simulate the dark currents at least in two structures in order to use SAD-program for further tracking. A goal is an estimation of dark current probability at exit of the main LINAC.

(4) FEATHER (N.Delerue)

(transparencies, 7 pages, pdf, 510KB)

First, Nicolas briefly reported results of test bench study on the electronics before a beam test. Artificial BPM signals with multi-bunch structure were generated by a BPM pulser. Splitting the signals for the two electrodes, their differences were made by a HH107 hybrid circuit. Then, they were amplified. The attenuated signals were observed by an oscilloscope together with the input signals from the pulser. Fluctuations in the peaks corresponding to multi-bunches were not understood. Diode properties were also measured as expected. Bandpass filter of 357MHz ±10MHz was tested. However, the filtered signals were too distorted to be used for feedback signals. Lowpass filters shall be tested too.

Second, the beam test results were reported on the BPM calibration. At this time, he succeeded to get a linear correlation between the ATF intensity and sum of the BPM signals. However, he failed to get the position signals with 1mm and 2mm gap. The reason is not known. One of basic observation is a measurement of pulse height as a function of gap. Also, beam orbit must be checked within aperture of the BPM button electrodes.

The next meeting will be on 12 May, 2004,10:00 - at 3 gokan, 425 .