Frank Marhauser Kai Tian. November, Jefferson Laboratory Jefferson Avenue VA, 23606

Transcription

1 JLAB-TN Optimization of the Cryomodule Cold-to-Warm Transitions and the VTA QA Test Configuration for CEBAF Upgrade Cavities with Regard to Critical HOMs above Cutoff Frank Marhauser Kai Tian November, 2009 Jefferson Laboratory Jefferson Avenue VA, Abstract Measuring loaded quality factors of critical Higher Order Modes (HOM) in the Vertical Test Area (VTA) is integral part of the quality assurance (QA) for future upgrade style Low Loss seven-cell cavities. Among field profile ( bead-pull ) measurements such control procedures are obligatory before cavities are installed in a cryomodule. The measurements can determine whether the cavities are obeying specified tolerances after cavity cell trimming, electron beam welding and plastical tuning for field flatness. The aim is to reveal an inadvertent cavity deformation causing potentially tilted HOMs that might be damped insufficiently by the HOM couplers positioned merely on one side of the cavity. Such a scenario has been experienced in an upgrade style cryomodule with a single High Gradient type cavity. It gave rise to surprisingly high Qs (~10 8 level) of a trapped dipole mode resulting in vertical beam-breakup instabilities (BBU) in CEBAF [1]. The cryomodule housing the cavity has been removed recently. Similar occurrences have to be avoided by all means for upcoming LL cavities. This note details the numerical effort to improve the HOM damping for a BBU critical mode pair around 2.9 GHz in cryomodule end cavities. Similarly, we aim to establish a standardized configuration in the VTA for QA procedures.

2

3 Table of Contents 1. Introduction Dipole Impedance Spectrum Critical Mode Pair in Third Dipole Passband Inner Cavities Cryomodule Configuration End Cavities Cryomodule Configuration VTA configuration Summary References

4

5 1. Introduction As part of the QA it is foreseen to measure the loaded quality factors of critical HOMs at cryogenic temperature for each single Low Loss (LL) upgrade type cavity ( C100 ). To be most efficient, these low power measurements can be carried out in the frame of obligatory high power rf tests in the VTA utilizing the same Dewar setup. The QA should provide representative results as obtainable in a horizontal cryomodule with eight cavities assembled in a string. This presupposes that HOM fields are not affected by the setup itself. In the VTA - with the cavity beam tubes electrically shorted - the impact on HOM field pattern is negligible only for trapped cavity modes, for instance for the first (TE111) and second (TM110) dipole passband resonating below the first beam tube cutoff (fc,te11 = 2.51 GHz). However, various BBU critical modes can exist beyond the cutoff frequency. The field amplitudes at the HOM coupler position depend on the beam tube short positions, the HOM frequency and mode type. Consequently, the Qs can differ significantly from those measured in a cryomodule and cannot be seen viable to evaluate BBU critical impedances. To at least account for the most critical dipole mode pair above the cutoff frequency, we try to establish a standardized VTA setup for QA of upcoming C100 cavities. The setup will resemble the cryomodule end cavity configuration, which is numerically optimized such to lower the expected external Q (Qext) at the HOM couplers with a minor impact on the existing design. 1

6 2. Dipole Impedance Spectrum Figure 1 shows the impedance spectrum of beam induced vertically polarized dipole modes calculated with CST Particle Studio [2] up to 10 GHz for a single LL cavity. Figure 1: Vertical dipole impedance spectrum (red curve) as calculated with CST Particle Studio for a seven-cell low loss upgrade cavity. The impedances are normalized by the bunch spectrum (σ = 10mm) (gray curve). Six waveguide modes have been considered for energy absorption at each waveguide port (two beam tubes, one FPC port). The dashed vertical lines indicate the first two cutoff frequencies of the FPC (fc,te10 = 1.1 GHz, fc,te20 = 2.2 GHz) and the beam tubes (fc,te11 = 2.51 GHz). The beam tube cutoff frequencies decide, whether a mode is trapped within the cavity. Similarly to the HOM couplers, the FPC can couple to the evanescent beam tube fields of trapped modes providing additional damping. The trapped horizontally polarized (H) TE111 and TM110 HOMs (not calculated here) can couple to the first FPC TE10 waveguide mode with a cutoff at 1.1 GHz. However, the trapped vertically polarized (V) dipole modes cannot be efficiently conducted out via the TE10 mode but rather the TE20 mode [4], so that the relevant cutoff is at 2.2 GHz. This is still lower than the beam tube cutoff. A TM111 mode exhibits the highest impedance although resonating above the beam tube cutoff. Hereby the impedance extrapolation scheme has been used [3] to resolve all impedance peaks based on a 400m wake potential. However, the impedance amplitudes are not fully representative. One reason is, that the HOM couplers are of rather complex shape and could not be meshed adequately, so that the damping is rather underestimated for trapped TE111 and TM110 passband modes resonating below the beam tube cutoff (symmetrical fields presupposed, i.e. no unfavorable tilted fields induced by cavity deformations). On the other hand, the HOM energy of propagating modes (i.e. TM111 modes) can bounce back between adjacent cavities in a real cryomodule. Numerically all energy is absorbed at the end of the cavity beam tubes, such to rather overestimate the damping in such a case. Moreover, the rectangular fundamental power coupler (FPC) waveguide has been allowed to absorb all out-coupled cavity fields beneficial for HOM damping. In reality this assumption is restricted by the limited bandwidth of the FPC double rf vacuum window configuration and the return loss characteristic of the FPC transmission line, which requires a waveguide filter/absorber to beneficially provide further HOM damping [4]. 2

7 Nevertheless, this calculation reveals the most prominent and dangerous dipole HOMs. Obviously, the most critical modes belong to the TM111 π/7 pair exhibiting the highest R/Q values among all dipole modes (identified by Eigenmode calculations). These modes are resonating around 2.9 GHz far above the first beam tube cutoff but are still confined in the cavity such to produce the highest impedances throughout. The horizontally polarized pendant is not excited here due to the chosen boundary condition. We have experimental evidence, that the damping of the TM111 mode pair relies not only on the HOM couplers but also the FPC and sensitively depends on beam tube boundary conditions, i.e. the location of the cavity in the cryomodule [4]. This will be detailed in the following. 3. Critical Mode Pair in Third Dipole Passband 3.1. Inner Cavities Cryomodule Configuration We first want to provide more insight into the geometrical sensitivity of HOM fields in a string of C100 cavities differing from the single cavity scenario simulated in Figure 1. As shown above high impedance TM111 dipole modes are resonating in the LL cavities around 2.9 GHz, which are of utmost concern with regard to potential BBU instabilities. Apart from the trapped TE111 and TM110 modes, we aim to integrate these modes as part of the VTA QA. With previous cryogenic rf test configurations for two existing LL cavities (C100-1, C100-2) - either tested individually in the VTA or assembled as a cavity-pair in a dedicated upgrade style horizontal test bed (HTB) - the corresponding HOM Qs were deviating by partly more than an order of magnitude due to different beam tube boundary conditions. This has been described extensively in ref. [4]. Beam tubes shorted too closely at the HOM coupler side even produce artificially tilted fields falsifying results. To regain the field symmetry, beam tube extensions are necessary. The dimensions have to be thoroughly chosen to account for a proper QA of the critical modes of interest. In a string of cryomodule cavities - differently from the VTA setup - the propagating HOM fields can leak from one into another cavity. An adjacent cavity acts like a filter. The fields can be reflected back in a rather undefined manner depending on the cavity filter characteristic and consequently on geometrical/fabrication tolerances. As long as the HOM fields can leak through several cavities, multiple HOM couplers will participate to damp the mode as a benefit (see also Figure 4). All damping effects combined typically reduce the Q of propagating HOMs below those achievable for trapped modes. However, very confined HOMs may still exist as in case of the C100 cavity caused by the inherently small apertures. A smaller aperture - leading to a smaller cell-to-cell coupling - is also responsible for a higher sensitivity of HOM fields on fabrication tolerances. The field amplitudes of the BBU critical TM111 π/7-like mode pair is plotted in Figure 2 along the beam axis for a cavity sandwiched in a string of three cavities. This setup resembles the condition of inner cavities in a C100 cryomodule 1. 1 for the second, third, sixth and seventh cavity only. It has to be considered, that the cryomodule exhibits a mirrored cavity configuration with respect to the center, i.e. between cavity four and five. Such a mirrored configuration has been resembled in the HTB, which yielded dipole impedances below BBU impedance thresholds [5]. 3

8 Figure 2: Transverse electric field (top) sampled along the beam axis for the TM111 π/7-like mode pair in a sandwiched cavity (zero position denotes the center of the mid cell). Fields have been sampled horizontally (H) or vertically (V) depending on the polarization. Corresponding electric field contour plots are shown at the bottom. For better visibility of weak field levels in the beam tubes, the amplitudes have been logarithmically scaled and normalized to a tenth of the maximum field amplitude. The dark blue regions indicate field minima. Electric field contours are shown at the bottom for both polarizations. The computations have been performed with either CST Microwave Studio (MWS) [2] or Omega3P [6]. The beam tubes and waveguides of the FPC are shorted electrically at the far ends. The fields reflected from the lateral cavities create standing waves in the cavity interconnecting beam tubes, which in turn determine the efficiency of the HOM damping. The standing wave patterns are affected by 4

9 fabrication tolerances, i.e. individual cell distortions and/or how strong the HOM frequencies are detuned with respect to each other. Measured frequencies of as built LL cavities when assembled in the HTB (C100-1, C100-2) or in the upgrade style cryomodule Renascence [7] (LL001, LL002, LL003) are summarized in Table 2. All cavities exhibited an acceptably flat field profile for the accelerating mode mandatory for proper high power testing. The C100 prototypes - never used for accelerating electrons - were not intended to be trimmed nor tuned in frequency at the time of fabrication, such that the accelerating mode frequency is about 2 MHz too high, whereas the TM111 π/7-like modes resonate at lower frequencies than in the Renascence cavities. Table 2: Accelerating mode and TM111 π/7-like H/V mode frequencies as measured for existing LL cavities in cryomodule assemblies at 2 Kelvin. C100-1 C100-2 LL001 LL002 LL003 f TM010 π mode MHz TM010 π mode field flattened yes yes yes yes yes f TM111 π/7-like mode H/V polarization MHz / / / / / As part of standard in-house fabrication methods, cavity cell dumbbells are trimmed before electron beam welding to achieve the envisaged target frequency (except C100-1 and C100-2). For Renascence cavities an overall shortage of several Millimeters (iris-to-iris) has been observed typically when compared to the untrimmed cavity dimensions per drawing. This has been verified for a recently built large grain LL cavity (J100-1). Consequently - to the best of our knowledge - we may expect TM111 π/7-like modes frequencies to alter within ±3 MHz in future production cavities. It is unlikely, that HOMs will resonate exactly at the same frequency. To resemble this expectation numerically, the lateral cavities in Figure 2 have been detuned with respect to the central cavity. This has been done by scaling the cavities uniformly in size by various amounts - with respect to each other - such to yield different frequencies of the TM111 π/7-like modes in each cavity as listed in Table 3. Scaling the cavities by 1 yields a frequency alteration closely to the expected values as mentioned above. Note, that in MWS the detuned cavities have been scaled transversally only, whereas in Omega3P they were allowed to expand/shrink in longitudinal size as well. Table 3: TM111 π/7-like mode frequencies in H/V polarization for the setup shown in Figure 2. detuned cavity left central cavity detuned cavity right detuning amount* f0 (MHz) / / / ± 1 transversally (MWS) f0 (MHz) / / / ± 2 transversally (MWS) f0 (MHz) / / / ± 1 all directions (Omega3P**) * In all cases, the cavities have been modeled according to the cell dimensions per drawing before detuning. ** In Omega3P all cavities have been scaled once before detuning to obtain the ideal frequency of 1497 MHz for the accelerating mode. 5

10 Figure 3 reveals a closer look to the central fields. Figure 3: Same as in Figure 2 with a closer view of fields for the central cavity. The blue vertical lines indicate two possible zones, where the electric fields vanish, here in the beam tubes on each cavity side either beyond the HOM coupler (left) or the FPC location (right) respectively. It is feasible to closely resemble the cryomodule string assembly with a single cavity in the VTA - for the purpose of a proper HOM Q assessment - by utilizing beam tube extensions shorted within the indicated areas. The uncertainty of these locations however is large (here ~ ±1cm) depending on the detuning amount of adjacent cavities as quantified in Table 2. We also observe variations depending on the mode polarization due to symmetry breaking effects. 6 Table 2: Possible beam tube short positions in the VTA to resemble a sandwiched cavity configuration (locations are given with respect to the center of the mid cell) location H mode V mode detuned adjacent cavities short position FPC side m yes, ± 1 (MWS) short position FPC side m yes, ± 1 (Omega3P) short position FPC side m yes, ± 2 (MWS) short position HOM coupler side m yes, ± 1 (MWS) short position HOM coupler side m yes, ± 1 (Omega3P) short position HOM coupler side m yes, ± 2 (MWS) Moreover, the beam tube guide length depends on the HOM frequency f0 according to c 1 λ 0 guide =, f0 2 f 1 c f 0 where c0 denotes the vacuum speed of light and fc the cutoff frequency of the tube (fc,te11 = 2.51 GHz). Therefore, a change in the HOM frequency from cavity to cavity will alter the outcome. A variation of λguide by 1.39 mm per 10 MHz can be derived from eq.(1). (1)

11 In the more unlikely - but possible case - that all cavities are resonating at similar frequencies, the field pattern in the beam tubes change significantly. A comparison of a resonant and a detuned case calculated with MWS for the V TM111 π/7-like mode is illustrated in Figure 4. As already mentioned above, multiple HOM couplers can participate to damp propagating HOMs as a benefit. Moreover, the FPC transmission line is serving doubly duty to extract a significant amount of HOM energy. The end cavities obviously share less HOM couplers and FPCs than the inner cavities. This impacts the achievable Q. In fact we have recorded significantly higher Qs for the critical TM111 modes in the end cavity (LL001) of the Renascence cryomodule compared to inner cavities (LL002, LL003) [4]. The HOM couplers in LL001 were rather ineffective and the FPC has absorbed most of the HOM energy. The FPCs therefore can play a crucial role to provide the required overall damping. Figure 4: Electric field contour plots of the vertically polarized TM111 π/7-like mode in resonant case (top) and detuned case bottom (±2 geometrical (transverse) detuning of left/right cavity). For better visibility of weak field levels in the beam tubes, the amplitudes have been logarithmically scaled and normalized to a tenth of the maximum field amplitude. The dark blue regions indicate field minima. Due to the rather undetermined short positions, the cryomodule configuration for an inner cavity is hard to resemble in the VTA for QA purpose. We rather concentrate on the peculiar end cavity configuration. The cryomodule end configurations actually incorporate cold-to-warm transitions. This yields a well defined boundary condition, which can be improved with regard to HOM damping independent on the neighboring cavity as detailed in the following. 7

12 3.2. End Cavities Cryomodule Configuration The configurations for cavities at each end of a C100 cryomodule are illustrated in Figure 5. Compared to inner cavities one cavity is replaced by a cold-to-warm transition region at the HOM coupler side. The configurations are mirrored with respect to the centre of the cryomodule, i.e. all rf relevant inner dimensions are similar for both end cavities. Each transition includes a metal gate valve directly flanged to the cavity and a subsequent step down of the tube diameter to reduce the static heat load. Figure 5: Upgrade cryomodule end cavity configurations. To improve the HOM damping, we propose to extend the beam tube beyond the gate valve by a length L as illustrated in Figure 6. The transition beam tube with = 1.37 acts as a rather sharp reflection plane due to the comparably high cutoff frequency (TE11 cutoff ~5 GHz). This forms a well defined standing wave on the HOM coupler side. By varying the extension length L, the standing wave pattern can be manipulated to optimize the external Q (Qext) at the HOM couplers particularly for the critical TM111 π/7-like dipole pair. This design change has no effect on the dipole modes trapped below 2.51 GHz. Figure 6: Numerical model of the cryomodule end cavity configuration. With the setup in Figure 6 we are assuming a priori, that the field amplitudes at the HOM coupler location do not change significantly when varying the boundary conditions on the opposing (FPC) side nominally connected to an adjacent cavity. This in fact has been verified with 8

13 MWS for different configurations as shown in Figure 7. Here the short position at the cavity FPC side has been moved by various amounts (keeping a fixed L = 1.93 ). Figure 7: Electric field amplitudes sampled along the beam axis for the TM111 π/7-like H mode in a cryomodule end cavity configuration. The beam tube has been shorted on the FPC side at different locations beyond the nominal cavity flange, i.e. by 1, 2, 3 and 4 respectively. L is 1.93 throughout. Whereas fields are apparently altered on the FPC side - also influencing the mode pattern within the first two outer cavity cells - the alterations are small on the HOM coupler side 2. This facilitates to optimize L independent on the boundary conditions at the FPC side. The outcome of this optimization 3 is shown in Figure 8. The figure of merit is the HOM coupler Qext for the TM111 π/7-like dipole pair varying in dependence on L. Figure 8: External Q of the HOM couplers for the TM111 π/7-like mode pair as calculated with Omega3P. The Qext takes into account the losses in both couplers. It depends slightly on the mode polarization. This can be explained by the azimuthal orientation of the upper and lower couplers, 2 True also for the vertically polarized HOM. 3 optimization done with Omega3P rather than MWS. With Omega3P a complex Eigenmode solution can be carried out utilizing a finite element meshing of the rather involved HOM coupler details as a major benefit to MWS with only a hedrahedral mesh available. 9

14 which favors the capture of V TE11 modes. The computation reveals a Q variation of more than three orders of magnitude within half the guide length (λguide/2 85 mm at ~2.9 GHz ( = 3 )). With L around 80 mm a local minimum can be achieved yielding Qext = / (V/H) respectively. This is well below the BBU threshold requirement [4]. A second minimum is expected for negative L close to 0 mm according to λguide. In fact the requirement can also be met with the present design ( L = 0mm) yielding Qext = / A (V/H) respectively. However, there is a steep increase in the Qext within L = 0-20 mm. The margin to the BBU threshold requirement can quickly be annihilated by fabrication tolerances. We also have to consider, that non-uniform cell distortions may lead to asymmetrical fields, which is beyond the frame of this work. Typically each cavity is manufactured such to guarantee a proper assembly and alignment of components in the cryomodule aiming for the specified overall cavity length of 1 m from flange to flange. It is facilitated by welding a Niobium tube section - including the cavity flange - on the HOM coupler side. This fabrication step corrects for geometrical tolerances after the final plastical tuning of the cavity has been carried out. Depending on tolerances however, the distance from the HOM couplers to the flange - and therefore L - can vary from cavity to cavity. With the experienced shortage of cavities (several Millimeters) after trimming, electron beam welding and tuning, the beam tube length is usually longer than expected per drawing. E.g. for the Renascence end cavity LL001 the discrepancy measured between the as built and nominal tube length was as high as 21 mm. Taking into consideration fabrication tolerances, it is advisable to choose a length L close to the second Qext minimum. We propose to use L = 72 mm yielding Qext 10 5, i.e. far below BBU requirements, within a margin of ± 13mm. This change however yields thermal implications to be considered. Originally, two bellows with an intermediate heat sink were required in the = 1.37 transition beam tube to limit the heat flow to the cavity to acceptable limits. The location of the innermost bellows coincides with the proposed extension line. To regain an adequate thermal heat path, the bellows needs to be integrated in the enlarged = 3 tube as shown in Figure 9 on the right side. Figure 9: Electric field amplitudes sampled along the beam axis for the TM111 π/7-like H mode in a cryomodule end cavity configuration with the optimized = 3 extensions L = 72 mm either with or without bellows. The latter are required to adequately intercept the heal flow from room temperature to 2K. The figure on the right shows the proposed configuration 10

15 Using an MWS model of the assembly 4 revealed that only a slight impact of field amplitudes and standing mode pattern is expected when incorporating the bellows as illustrated in Figure 9 left by means of the absolute electric field on axis. This implies that the proposed extension around L = 72mm is valid to achieve a similarly low Qext at the HOM couplers as previously calculated. We used Omega3P to verify this statement with the result illustrated in Figure 10 (for the V mode only). The Qext in this case is 3.2e4 close to the former findings. Due to prevalent meshing problems of the rather complex bellows we were not able to perform the full survey of Qext versus L with reasonable numerical constraints. Figure 10: Same as Figure 8 including the Qext result (red square) for an extension L = 72mm including bellows (V TM111 π/7-like mode only). It shall be pointed out, that - although we try to consider fabrication tolerances that may lead to different cavity/tube lengths - we could not account for potential individual (rather arbitrary) cell distortions causing asymmetric HOM field distributions even in presence of a flat/symmetric fundamental mode field profile. Field profile measurements on cavities as part of the QA are advisable to reveal cavities with potential asymmetric, i.e. tilted or more confined HOM fields, which could give rise to unacceptably high Qs.: 4 with a rather course hedrahedral mesh discretization of the bellows 11

16 VTA configuration In the VTA no gate valve and no bellows will be used for the QA rf tests. However, to resemble the cryomodule end cavity assembly, the beam tube extension on the HOM coupler side can be optimized in a similar manner using a straight extension of the tube (see Figure 11). Figure 11: Numerical model of the VTA setup to resemble the cryomodule end cavity configuration. According to the cavity beam tube diameter of = 2.75 we obtain λguide/2 104 mm at ~2.9 GHz, such that a different Qext dependence over LVTA is expected when compared to the cryomodule assembly. This has been verified with Omega3P. The corresponding outcome is plotted in Figure 12. With the cavity flange shorted ( LVTA = 0) the TM111 π/7-like modes couple strongly to the beam tube volume creating artificially tilted fields. This has been verified numerically and experimentally [4]. The use of an extension is mandatory to symmetrize the fields. LVTA = 0 is therefore excluded in the plot. We propose to use an extension with a length of LVTA = 85.5 mm beyond the cavity flange on the HOM coupler side to obtain a Qext 10 5 within a margin of ± 18.5 mm. Again this presupposes symmetrical HOM fields. Figure 12: External Q of the HOM couplers for the TM111 π/7-like mode pair as calculated with Omega3P. 12

17 4. Summary We have numerically optimized the warm-to-cold transition regions in upgrade style cryomodules to improve the HOM damping of a BBU critical dipole HOM mode pair confined in the cavity around 2.9 GHz. The quality factors of these modes have been found to sensitively depend on beam tube boundary conditions, which may cover more than three orders of magnitude. By merely extending the beam tube length on the HOM coupler side, the external Q achievable at the HOM couplers can be minimized theoretically below However, this improvement impacts the damping on cryomodule end cavities only. As experienced with a detailed HOM survey for a string of eight cavities in the Renascence cryomodule, the damping efficiency of inner cryomodule cavities is more efficient by nature, but may have to rely on the FPC transmission line doing double duty in providing additional rf losses to reduce the HOM Qs to safe limits. To resemble the peculiar cryomodule end cavity configuration in the VTA, i.e. to implement a proper Q-measurement of the critical HOM mode pair in the QA procedure, we have established a similar configuration using a specific beam tube extension. The QA will include - as usual - all other critical modes trapped below the cavity beam tube cutoff, which are not affected by the implemented design change. For the most reasonable design we have taken into account - to the best of our knowledge based on previously built LL cavities - possible dimensional cavity and beam tube length fluctuations due to fabrication tolerances. These can be as large as several Millimeters. It has to be pointed out, that the same fabrication tolerances may also lead to asymmetric HOM field distributions if cavity cells are deformed non-uniformly after tuning, even in the presence of a flat fundamental field profile. This could lead to higher Qs than achievable with the optimized settings. In this note - to limit the numerical work to a reasonable time scale - the cavity cells have been modeled either according to technical drawings or have been scaled solely uniformly to study the impact on fields caused by detuning effects. We endorse to perform field profile measurements on BBU critical HOMs to reveal cavities with potentially tilted or more confined HOM field pattern before cavities are installed in cryomodules. 13

18 References [1] R. Kazimi, A. P. Freyberger, C. Hovater, G. A. Krafft, F. Marhauser, T. E. Plawski, C. E. Reece, J. Sekutowicz, C. Tennant, M. G. Tiefenback, H. Wang, Observation and Mitigation of Multipass BBU in CEBAF, Proceedings of the EPAC2008, Genoa, Italy, 2008, WEP087. [2] [3] F. Marhauser, R.A. Rimmer, K. Tian, H. Wang, Enhanced Method for Cavity Impedance Calculations, Proceedings of the PAC2009, Vancouver, British Columbia, Canada 2009, FR5PFF094. [4] F. Marhauser, H. Wang, K. Tian, Critical Higher Order Modes in CEBAF Upgrade Cavities, JLab Technical Note, TN-09-50, [5] F. Marhauser, E. Daly, G.K. Davis, M.A. Drury, C. Grenoble, J. Hogan, R. Manus, J.P. Preble, C.E. Reece, R.A. Rimmer, K. Tian, H. Wang, HOM Survey of the First CEBAF Upgrade Style Cavity Pair, Proceedings of the PAC2009, Vancouver, British Columbia, Canada [6] K. Ko, N. Folwell, L. Ge, A. Guetz, L. Lee, Z. Li, C. Ng, E. Prudencio, G. Schussman, R. Uplenchwar, L. Xiao, Advances in Electromagnetic Modelling Through High Performance Computing, Physica C 441, , [7] C.E. Reece, E.F. Daly, S. Manning, R. Manus, S. Morgan, J.P. Ozelis, L. Turlington, Fabrication and Testing of the SRF Cavities for the CEBAF 12 GeV Upgrade Prototype Cryomodule RENASCENCE, Proceedings of the PAC2005, Knoxville, Tennessee, USA 2005, TPPT