Yamamura has studied performance of electrons at rest in various cases of acceleration gradient (E0) and initial RF phase (&phi0) for estimation of the capture threshold by using his program. In this study, he found a bug in his program. This bug has large effects in his results which have been presented at meetings. So, he must re-calculate all the previous results. After the debug, longitudinal position of the electron (Zf) was calculated as a function of initial phase (&phi0) in RF pulse for accelerating period of 10nsec with E0 of 100 and 10MV/m. He found that the electrons could be captured in the angular region centered at &phi0=&pi. Then, the Zf, longitudinal momentum component (pZ,f) and the final phase were calculated at &phi0=&pi as a function of E0. The capture threshold seems to be 60MV/m.
Yokoya commented that the capture threshold can be analytically expressed by; mec2 k /2, where k is 2&pi/&lambda, &lambda=26mm (x-band), so it is calculated to be 61.7MV/m, where sine-like wave is assumed for the accelerating one. He will confirm this analytic formula.
(2) Final Focus System (S. Kuroda)
(transparencies, 1 page, pdf, 26KB)
Kuroda estimated errors of longitudinal positions and rotational angles fo all the quadrupole magnets in the FF system, which blow up beam sizes by 50% in both directions at IP. i.e. 1.5&sigma*x=300nm and 1.5&sigma*y=4nm. The longitudinal position errors can be larger than 1cm expect for the final doublet. The nearest QD0 has the smallest errors comparable to &beta*y. The rotational angle error is the smallest at QF1 and QD0, both of which have 5.68μrad.
(3) Main Linac Simulation by SLEPT - no correction (K. Kubo)
(transparencies, 7 pages, pdf, 157KB)
Energy spread in a bunch is generated by BNS damping at the x-band linear collider (GLC, NLC). The BNS damping must be required in order to control a single bunch blow up and emittance dilution at the LINAC. The magnitude of energy spread can be varied with various schemes of BNS damping, i.e. variations of off-crest phases along the LINAC, as mentioned in the ACFA report of "Particle Physics Experiments at JLC", p20, KEK Report 2001-11, August 2001.
In this study, Kubo made four sets of parameters with the energy spread (rms) of 0.16% and 0.25% at the beam energies of 500GeV (parameter sets of A and B, respectively) and 250GeV (D,E). For a case of 500GeV beam energy with 0.25% energy spread, new optics was also made with lower beta functions, whose parameter set is called as C. The 0.16% spread is nearly the minimum for the initial spread of 2.5%. The parameter set of A (D) is close to that in the GLC roadmap report, KEK report 2003-7. He simulated the emittance dilutions in these 5 cases (A-E) as functions of injection (vertical position) error, random misalignment of Qs, AT of the ATL rule and email@example.comHz of the GM model C. The simulations assumed no correction in this study.
Results are briefly summarized as follows; (1) on the injection error, the parameter sets with smaller energy spread have larger emittance dilution, where lower beta-function optics can reduce the dilution; (2) on the random misalignment, also the smaller energy spread causes larger dilution, where the lower beta-function optics significantly reduce the dilution, (3) on the AT, similar results as (2) were obtained, while the AT effect was larger at higher beam energy than 250GeV, and (4) on the a2@2,5Hz, very similar results as (3) were obtained. For all the cases, the lower beta-function optics could reduce the emittance dilution very significantly. Especially the lower beta optics is very preferable at noisy GM site in order to decouple "the 2.5Hz wave" and the beta-function at high energy of greater than 250GeV.
(4) FEATHER status and Feedback with the beam blow-up (N.Delerue)
(transparencies, FEATHER status, 7 pages, pdf, 450KB and Beam blowup studies, 11 pages, pdf, 570KB )
Nicolas presented results of last beam test before summer shutdown, First, beam positions were measured as a function of bunch number by wire scanner at MW3X. They were surprisingly non-flat, that is, 300μm upward in rear bunches. Also, position fluctuations were visible at several dozen μm. So, it was difficult to see the kick effect in the fast feedback. Second, he failed to calibrate the BPM with 20 bunched beam. No position signals were observed.
He listed issues for improving the kicker as follows; (1) impedance mismatch of the kicker, (2) unable to go to gap lower than 1.2 mm, (3) kicker is too long for operations at 1GHz/357MHz, (4) available RF power limited to 40dBm; The solutions depend on the future choice for FEATHER (and a lot on $$money$$) in terms of following questions: Do we want a very fast correction or is a slower FONTlike correction acceptable? Which amplitudes do we want to be able to correct? To which precision?
Finally he commented for the above issues; (1)Matching the impedance of the kicker is only possible at a given gap. Might require a lot of work (fine tuning). May increase kicker power by 10%(?) (2) Straightening the electrodes would allow smaller gap. We double our kicker power each time we divide the gap by 2. Gap below 0.5mm seem unreasonable. May increase kicker power by 100% (3) Shortening the electrode to a tuned value for 354MHz could allow operations at that frequency. But shorted electrodes have less kicker power. Using 2 parallel electrodes could provide both advantage. May increase kicker power by huge factor@354MHz (but not enough) (4) Our amplifier are the state-of-the-art commercially available. Either we buy TWTA (expensive) or we need to use many amplifier in parallel... All this cost money. We have to discuss our priorities...
Nicolas has investigated feedback performances with the beam blow-up as a function of crossing angle at GLC and CLIC. The beam blow up was estimated by CAIN. The crossing angle ranges from 7mrad to 30mrad, and the L* (distance of the final Q from IP) was set to be 4m. He assumed the fast feedback can be applied after 10 and 20 bunches at GLC and CLIC, respectively, since the bunch separation of CLIC (0.67ns) is half of GLC (1.4ns). At GLC, simulations show that the FEATHER fast feedback delayed model with a higher gain can solve the problem of the beam blowup, while at CLIC increasing the gain of the fast feedback system won't be enough, an extra feendback system is needed. Finally, he concluded that (1) beam-beam interactions can led to beam blowup that would reduce the delivered luminosity; (2) beam blowup at the GLC can be handled by regular fast feedback system with higher gain; (3) at CLIC an extra feedback system might be needed.