US LHC Accelerator Research Program BNL - FNAL- LBNL - SLAC

Transcription

1 US LHC Accelerator Research Program BNL - FNAL- LBNL - SLAC RF Design Progress and Plans beam beam 10 December 2007 LARP Collimator Video Meeting Gene Anzalone, Eric Doyle, Lew Keller, Steve Lundgren, Tom Markiewicz & Jeff Smith 1

2 Recent Progress RF design presented in Steve s talk MAFIA Simulations and Calculations RF Contact DC Resistance Measurement plans Bench top stretched wire (coil) Impedance Measurement plans 2

3 RF design RF design presented in Steve s talk 3 key issues (as I see it): RF contact from Jaw to end socket Resistance must be very small Perform tests to ensure we meet specs Is geometric impedance appreciable compared to resistive wall? Perform bench-top stretched wire impedance measurements Top and bottom transverse RF seals Probably not as critical. 3

4 MAFIA Simulations With the help of Cho-Kuen Ng and Karl Bane at SLAC, have begun studying geometric and resistive wall impedance in Rotatable Collimators In discussions with Karl Bane it was found that for a round collimator of our dimensions the resistive wall wakefield is about a factor of 10 larger than the geometric. This is true for round collimator but what about our hour-glass shape geometry? MAFIA doesn t appear to be successful at accurately modeling smooth gradual tapers (as seems to be a common problem in such codes) Still learning MAFIA so part of the problem may very well be me! Will probably get better results with more experience. Nevertheless, may not get too accurate of an answer for geometric wake Perhaps GDFidL would work better... Haven t looked into using it yet. MAFIA seems to work better for resistive wall calculations 4

5 Resistive Wall Impedance Using MAFIA, a 1 mm half-gap round collimator 0.93 meters long has a transverse kick of 4.0 x 10^14 V/C.m Compare to 3.3 x 10^14 V/C.m using analytical formula (Chao/Tigner Handbook). Would prefer to have better agreement Now calculating transverse kick for rotatable collimator Beam path 6.9 x 10^13 V/C.m 7.000E-02 ~6 times less transverse kick than round collimator 3.500E-02 If geometric wake is larger for rotatable jaws then geometric and resistive wall may be 0.00 comparable E E E-02 Then proper optimization of taper geometry may be benificial Am I using MAFIA correctly? Want to look into this more... Would like to physically measure this to confirm E E E E E E-03

6 RF Contact Measurements Must have low resistance for RF contacts, especially Jaw/transition piece interface This interface is ~11 mm from beam and must have ~<0.02 mohm total low frequency resistance What kind of electric contacts should be used here? Silver plated? Rhodium? Is copper good enough? (probably not) Cold welding copper? Considering results from Sergio Calatroni et al. How much force needed for good contact? How will resistance increase with wear and tear? Will perform RF contact resistance measurements with HP microohm multimeter. Low resistance RF contact Slot for Spiral RF Spring 6

7 RF Contact Measurements LINEAR FEED-THROUGHS ANVIL SPRING - CONTACT Contact Resistance Test Chamber Two axes: Normal & sliding Existing NLC seismometer vac chamber 7 ELECTRICAL FEED-THROUGHS

8 Stretched Wire Impedance Measurements Currently setting up lab to perform stretched wire impedance measurements as developed by Fritz Caspers, et al. Proceedings of the 2003 Particle Accelerator Conference BENCH MEASUREMENTS OF LOW FREQUENCY TRANSVERSE IMPEDANCE Low Frequency Collimator Measurements Preliminary Results - 14 Nov 2007 F.Caspers, T.Kroyer, E.Metral, F.Roncarolo, B.Salvant 8 1 Abstract A. Mostacci, Univ. di Roma La Sapienza, Rome, Italy, F. Caspers, L. Vos, CERN, Geneva, Switzerland, U. Iriso, BNL, Upton, New York, 11973, USA. For frequencies below 10 MHz the classical two wire transmission line method is subject to difficulties in sensitivity and measurement uncertainties. Thus for evaluation of the low frequency transverse impedance properties of the LHC dump kicker a modified version of the two wire transmission line has been used. It consists, in the present case, of a 10 turn loop of approximately 1 meter length and 2 cm width. The change of input impedance of the loop is measured as a function of the surroundings and by using a proper reference (metallic beam pipe) these changes are converted into a meaningful transverse beam coupling impedance. Measurements of several calibration objects have shown close agreement with theoretical results. INTRODUCTION A beam that oscillates from side to side with amplitude induces differential currents and charges on the walls of the vacuum chamber. These in turn produce a transverse magnetic field and an electric field which further deflects the beam. The threshold for beam instability and the growth rates depends on the so called transverse coupling impedance. Transverse impedances are well known in literature and they can be theoretically calculated for a number of particularly simple structures or in general, they are numerically computed with codes. In this paper we are interested on bench measurements techniques, in particular at low frequencies (below few khz). At those frequencies a better sensitivity can be obtained by using a loop [1], instead of the classical two wire technique [2]. The same wall currents and magnetic (deflecting) field result if the beam is replaced by two parallel wires or more simply by a loop of length, width and current. The magnetic field induces a voltage in the loop which increases its impedance (the current is constant). This additional impedance is simply the variation of the loop impedance when inserted in the Device Under Test (DUT) with respect to the loop impedance. Assuming that the loop is coiled times, the transverse coupling impedance can be obtained from where is the (measured) impedance of the loop when inserted in the DUT and is the (measured) impedance of the loop far away from any perturbing object (i.e. in free space). In general, when measuring a very small impedance (as pointed out in [1]), one should also subtract the radiation resistance from the loop measurements in free space. This is appreciable unless the loop is very short compared with the wavelength. Alternatively one could place the loop in a circular perfectly conducting pipe (a copper or brass one is enough), for which the added impedance is easy to calculate. MEASUREMENT SET-UP The coil used in the measurement was 1.25 m long and 22.5 mm wide. To reduce the signal to noise ratio (particularly important in our case since the measured signals are very small), the loop was coiled 10 times. In this way, one can increase the useful signal with a factor (1)

9 Assembling components for measurements Have lab space and VNA Obtaining test sets and DUTs Attending USPAS course on Microwave Measurements in January Stretched wire impedance measurements bench HP 4195A Vector Network Analyzer, 10Hz - 500MHz 9

10 Bench-top Test Plan 1.Reproduce CERN results for graphite and copper plates Confirm we know what we are doing 2.Measurements on copper pipe and rotatable jaw geometry to measure resistive wall impedance How much smaller is resistive wall impedance for rotatable jaw geometry 3.Perform measurements on 15 degree wedges or azimuthal grooves Is the geometric component of our tapers critical? Will azimuthal grooves improve low frequency transverse impedance? 4.Perform measurements on taper sections with no jaws to remove resistive wall wake from jaws Remove jaws, just look at tapers 10

11 Questions for the experts: Perhaps a different E&M modeling software will work better. Recommendations? GDFidL? Is our RF spring contact between jaw and end socket plausible? Would prefer not to use fingers but have another design using them. Contact only ~11 mm from beam axis. Too close? Recommendations for bench measurements? Precision of coil winding? Alignment of coil within Device Under Test? At high frequency (~GHz) is two parallel wires better? Single displaced wire? Top and bottom transverse RF seals Do we even need them? 11