The TESLA RF System S. Choroba for the TESLA Collaboration DESY Notkestr. 85, D-22603 Hamburg, Germany Abstract. The TESLA project proposed by the TESLA collaboration in 2001 is a 500 to 800GeV e+/e- linear collider with integrated free electron laser facility. The accelerator is based on superconducting cavity technology. Approximately 20000 superconducting cavities operated at 1.3GHz with a gradient of 23.4MV/m or 35MV/m will be required to achieve the energy of 500GeV or 800GeV respectively. For 500GeV ~600 RF stations each generating 10MW of RF power at 1.3GHz at a pulse duration of 1.37ms and a repetition rate of 5 or 10Hz are required. The original TESLA design was modified in 2002 and now includes a dedicated 20GeV electron accelerator in a separate tunnel for free electron laser application. The TESLA XFEL will provide XFEL radiation of unprecedented peak brilliance and full transverse coherence in the wavelength range of 0.1 to 6.4nm at a pulse duration of 100fs. The technology of both accelerators, the TESLA linear collider and the XFEL, will be identical, however the number of superconducting cavities and RF stations for the XFEL will be reduced to 936 and 26 respectively. This paper describes the layout of the entire RF system of the TESLA linear collider and the TESLA XFEL and gives an overview of its various subsystems and components. INTRODUCTION The TESLA project proposed by the TESLA collaboration in 2001 is a 33km long 500 to 800GeV e+/e- linear collider with integrated free electron laser facility. In contrast to other linear collider concepts it is based on superconducting RF structures. Details can be found in the TESLA Technical Design Report [1]. In the year 2002 the Technical Design Report Supplement was published describing a dedicated 20GeV electron accelerator for free electron laser application [2]. A detailed paper on the RF system for the TESLA 500GeV collider was presented at the modulator klystron workshop in Geneva in 2001 [3]. The TESLA RF system provides RF power at a frequency of 1.3GHz for the superconducting cavities of the TESLA linear collider and the TESLA XFEL. The 500GeV linear collider requires 572 RF stations. For the 800GeV upgrade the number of stations must to be doubled and in both cases 12 stations for the injectors must be added. The dedicated electron linac for the free electron laser requires 26 stations plus 4 for the injector.
Although the requirements of the different accelerators, the 500GeV accelerator, the 800GeV upgrade and the 20GeV linac for the XFEL, differ in certain points, many of the main parameters of the RF systems are almost identical. Therefore the RF system layout will be presented for the example of the 500GeV accelerator. Major differences will be mentioned. For the 500GeV linear collider an accelerating gradient of 23.4MV/m is required. Since the maximum beam current is 9.5mA, 231kW of RF power will be needed for each 1.038m long superconducting cavity. Each particle beam pulse consists of 2820 bunches with a spacing of 337ns and has a length of 950µs. The filling time of each cavity is 420µs and therefore the total RF pulse length is 1.37ms. The repetition rate is 5Hz. At 800GeV center of mass energy the gradient must be increased to 35MV/m, the filling time increases to 510µs, the beam pulse length and the repetition rate decrease to 860µs and 4Hz respectively. The linac for the XFEL operates at a gradient up to 23.4MV/m at a repetition rate of 10Hz. Since the particle beam current is 5mA up to 122kW are required per cavity. The XFEL beam pulse duration is 800µs and the cavity filling time 570µs. BASIC RF STATION LAYOUT It is possible to use one independent RF source per superconducting cavity or to use a more powerful RF source and connect several cavities to one source instead. The latter results in a smaller number of sources and therefore promises higher reliability. At TESLA 36 cavities, 3 times 12 cavities per cryogenic module, will be connected to one 10MW klystron by a RF waveguide distribution system. For the 500GeV linear collider 8.3MW are required by the cavities. Taking into account 6% losses in the waveguide distribution system and a regulation reserve of 10% for phase and amplitude control the klystron must provide up to 9.7MW. For the 800GeV upgrade the number of klystrons will be doubled, thus 18 cavities will be connected to one klystron. For the XFEL only 5.2MW must be generated by the klystron including the reserve for waveguide losses and for regulation. Each klystron is part of one RF station consisting of the klystron, a HV modulator, which transforms the AC line power to pulsed HV power, the waveguide distribution system, a low level RF system, which controls shape, amplitude and phase of the RF, several auxiliary power supplies for the klystron and the pulse transformer and an interlock and control system, which protects the station and the linac. The RF stations will be installed in the accelerator tunnel with a separation of 50m (25m for the 800GeV upgrade) with the exception of the modulators high voltage power supplies and pulse generating units which will be installed in access halls on top of the accelerator tunnel. The access halls have a separation of about 5km. High voltage pulse cables and additional cables for the interlock system connect the components of each station in and outside the tunnel. The number of modulators per hall will be up to 100 for the 500GeV linear collider, up to 200 for the 800GeV upgrade and 30 for the XFEL.
THE 10MW MULTIBEAM KLYSTRON The RF power source for TESLA is a 10MW multibeam klystron. Investigations of different klystrons have shown that klystrons with low microperveance have higher efficiency than klystrons with high microperveance [4, 5]. Since low microperveance beams have higher voltage and lower current than high microperveance beams of the same beam power, space charge forces are smaller, bunching of the beam becomes easier and hence the efficiency is higher. Unfortunately low microperveance results in higher voltage, which in case of long pulse klystrons might lead to higher breakdown probability and hence reduced reliability. By using many small low perveance beams in parallel in one klystron both the beam voltage and the space charge forces become smaller compared to a single beam klystron of the same power. For a multibeam klystron an efficiency up to 70% or more seems theoretically feasible, whereas a standard single beam klystron of microperveance 2.0 has only an efficiency of 45%. FIGURE 1. The THALES TH1801 Multibeam Klystron. Figure 1 shows the multibeam klystron TH1801 produced by Thomson Tubes Electroniques, now THALES Electron Devices [6]. Seven beams of a microperveance of 0.5 are produced by the cathode and accelerated by the klystron gun. The beams
have independent drift tube sections but common cavities. After RF extraction in the output cavity, the spent electron beams are absorbed in one collector. Two output waveguides are required to handle the RF power of 2 times 5MW in the output windows. The total height of the klystron is 2.5m. The multibeam klystron was successfully tested and operated at the TESLA test facility (TTF) at DESY. It achieved an output power of 10MW with an efficiency of 65%. Table 1 summarizes the design parameters and the parameters achieved with the prototype. The repetition rate and average power during test of the prototype was limited by main power restrictions. Later a series klystron was operated at full repetition rate and full average power. More detailed information can be found in [7]. Since the gain of the klystron is 48dB the drive power is below 160W and solid state amplifiers can be used. They will be installed near to the klystrons inside the collider tunnel. The klystrons will be mounted in the horizontal or the vertical position depending on which accelerator tunnel layout will be finally selected (see figure 4). TABLE 1. Design and Measured Parameters of the THALES TH1801 Multibeam Klystron. Design Measurement Operation Frequency 1300MHz 1300MHz RF Pulse Duration 1.5ms 1.5ms Repetition Rate 10Hz 5Hz Cathode Voltage 110kV 117kV Beam Current 130A 131A HV Pulse Duration 1.7ms 1.7ms No. of Beams 7 7 Microperveance 3.5 3.27 No. of Cavities 6 6 RF Peak Power 10MW 10MW RF Average Power 150kW 75kW Efficiency 70% goal 65% Gain 48dB 48.2dB Solenoid Power 4kW 6kW Other klystron vendors have started to develop multibeam klystrons for TESLA too. The following tables and pictures show the parameter and outline of the CPI and the TOSHIBA multibeam klystron for TESLA respectively. The CPI prototype is being assembled and test is foreseen for autumn 2003. Development of the TOSHIBA klystron has just started in cooperation with KEK, Japan.
TABLE 2. Typical Operating Parameters of the CPI VKL-8301. Ave. Power Output 150 kw (min) Beam Voltage 114 kv (typ) Beam Current 131 A (typ) Efficiency 65 67 % (typ-goal) Frequency 1.3 GHz Pulse Duration 1.5 ms (min) Saturated Gain 47 db (min) Number of Beams 6 Number of Cavities 2+3+1 HOM, fund., harmonic Focusing CFF Cathode loading <2.5 A/cm2 (typ) Solenoid Power 4 kw (typ) Length 2300 mm (typ) Diameter 560 mm (typ, gun) FIGURE 2. The CPI VKL-8301 Multibeam Klystron.
TABLE 3. Design Parameter of the TOSHIBA E-3736 Multibeam Klystron. Peak Output Power 10 MW Average Output Power 150 kw Cathode Voltage 115 kv Beam Current 132 A Efficiency > 65 % Saturated Gain 47 db Number of Beams 6 Number of Cavities 6 Cathode Loading < 3 A/cm Solenoid Power < 4 kw Operation Frequency 1300 MHz RF Pulse Duration 1.5 ms Repetition Rate 10 Hz FIGURE 3. The TOSHIBA E-3736 Multibeam Klystron.
FIGURE 4. Old (top) and new (bottom) accelerator tunnel layout. THE MODULATOR The modulator generates 120kV HV pulses of almost rectangular shape. The pulse duration is 1.57ms and the rise and fall time 200µs. Already during the leading edge of the high voltage pulse when the voltage has reached 80% of the flat top level the klystron can start to generate RF power. Although this part of the RF pulse does not meet the requirements regarding phase and amplitude stability for particle beam acceleration in the accelerator it can be used to fill the superconducting cavities with RF. The phase change of 320 between 80% and 100% of the klystron voltage can be compensated by the low level RF control. The flat top ripple should not exceed ±0.5% in order to limit phase and amplitude variations of the klystron RF output during beam acceleration. The pulse-to-pulse stability must be better than ±0.5%. In case of klystron gun sparking the energy deposited in the spark must be limited to a maximum of 20J. The modulator requirements are summarized in Table 4.
TABLE 4. Modulator Requirements. Typical Maximum Klystron Gun Voltage 115kV 130kV Klystron Gun Current 130A 150A High Voltage Pulse Duration (70% to 70%) <1.7ms 1.7ms High Voltage Rise and Fall Time (0 to 99%) <0.2ms 0.2ms High Voltage Flat Top (99% to 99%) 1.37ms 1.5ms Pulse Flatness during Flat Top < ±0.5% ±0.5% Pulse-to-Pulse Voltage fluctuation < ±0.5% ±0.5% Energy Deposit in Klystron in Case of Gun Spark <20J 20J Pulse Repetition Rate for the Linear Collider 5Hz 5Hz Pulse Repetition Rate for the XFEL 10Hz 10Hz Transformer Ratio 1:12 1:12 Different types of modulators were investigated. Finally the bouncer modulator consisting of a HV power supply, a pulse generating unit and a pulse transformer was selected. It promised low cost and high reliability compared to alternative layouts [8]. The basic idea of the bouncer modulator is to use a capacitor discharge and compensate the voltage droop by the linear part of the oscillation of a bouncer circuit. The circuit diagram of the bouncer modulator is sketched in Figure 5. A HV power supply keeps the 1.4mF main capacitor C1 charged to the 10kV level. By closing the switch S1 the capacitor C1 starts to discharge via the pulse transformer into the klystron. The 1:12 transformer converts the 10kV voltage of the capacitor to the klystron level of 120kV. After 1.57ms switch S1 opens and terminates the pulse. Since the capacitor voltage drops during the pulse no flat top would be achieved unless a huge capacitor bank storing a big amount of energy would be used. For a droop of 1% a capacitor bank of 29mF would be required. In order to overcome the disadvantage of a huge capacitor bank the bouncer circuit was added. A circuit consisting of a choke, a capacitor and another switch is connected to the low end of the primary winding of the pulse transformer. By triggering this switch shortly before the main pulse begins this circuit starts to oscillate. The linear part of the 5ms long oscillation is then used to compensate the droop of the main capacitor C1. By this the 19% droop of the 1.4mF capacitor can be reduced to 1%. A very detailed description of the modulator is given in [9, 10, 11, 12]. Several modulators of the bouncer type were built and are in use at the TESLA test facility. The first modulators were built and assembled by Fermilab, USA, and then shipped completely to DESY. The others were assembled at DESY in cooperation with industrial companies from units manufactured by the companies. The first modulators made by Fermilab use GTOs and IGBTs as main switches in the pulse generating unit, whereas the newer version of the pulse generating unit built by industry use IGCTs instead. The size of the pulse generating unit is 2.8m(L) x 1.6m(W) x 2.0m(H). The pulse transformer transforms the 10kV pulses of the pulse generating unit to 120kV required by the klystron. The transformers leakage inductance determines the rise time of the HV pulse. The first transformers have a leakage inductance of more than 300µH whereas the newer have less than 200µH. This allows to meet the specification of rise time of less than 200µs. The size of the transformer is 3.2m(L) x 1.2m(W) x 1.4m(H), its weight is 6.5t.
FIGURE 5. Circuit diagram of the modulator (schematic). Since the pulse forming units of the modulators will be installed in the access halls on top of the accelerator tunnel whereas the pulse transformers will be installed inside the tunnel, pulse cables with a length of up to 2.8km must be used for the connection. In order to limit the ohmic power loss to 2% on average a current lead of 300mm 2 is required. The wave impedance of the cable must match the impedance of the klystron transformed to the transformers primary side to avoid distortion of the pulse shape. In addition the skin affect must be minimized. Therefore four cables in parallel, each of 75mm 2 cross section, 25Ω impedance and an outer diameter of 30mm, will be used. The cables are of triaxial type to minimize electromagnetic interference. The inner lead is at high potential (12kV), the middle cylindrical lead at the potential of the bouncer circuit (±2kV) and the outer cylindrical lead at ground potential. As insulation material VPE will be used. Additional line matching to the pulse transformer will be done via a RC network. Simulation results and further information on the cable can be found in [13]. A special high voltage power supply for charging the pulse generating units capacitor is required. It has to charge the main capacitor to an accurate value of voltage in order to obtain the same voltage at the klystron from pulse to pulse, and it has to suppress the repetition frequency (up to 10Hz) load of the mains in order to avoid disturbances of the mains. The input connection of each power supply is three phase grid and the output voltage is 12kV. One power supply is foreseen for each modulator. In order to ensure a high reliability the power supplies will be built in modules of switch mode buck converters or alternatively of series resonant converters. The power supply regulation is a digital self-learning regulation of the input power, made possible by the high regulation dynamic of the converter supply. In addition the voltage at the capacitor bank at the firing time of the pulse will be regulated within 0.5% accuracy. The size of a high voltage power supply is 1.2m(L) x 1.6m(W) x 2.0m(H). Further information about the power supplies can be found in [14]. Several additional auxiliary power supplies are required for the operation of the klystron and the modulator, e.g. for the klystron focusing solenoid, for the klystron
filament, the klystron vacuum pump and the pulse transformer core bias. The auxiliary power supplies will be installed together in a rack near to the klystron and the pulse transformer. RF SYSTEM EFFICIENCY Taking into account losses of 6% in the waveguide distribution system the klystron has to provide 8.7MW for the linear collider. This power must be generated below saturation because of the required regulation reserve of 10% for fast phase and amplitude control. Thus the klystron will be operated at a voltage of 117kV, sufficient to generate 10MW of RF power. The efficiency at this voltage level is 65%. It is necessary that the voltage stays constant within 1% during the pulse flat top, when the accelerator particle beam is accelerated. However the power in the leading edge of the applied HV pulse is not wasted completely, most of it is used to fill the superconducting cavities with RF. When the klystron voltage reaches 80% of the flat top voltage, the RF output power is already 4MW. During the last 20% of the voltage rise to the flat top the RF power increases, but the RF phase changes by 320. The phase change can be compensated by the low level RF system. Then the rise time efficiency of the modulator, defined as the ratio of the energy per high voltage pulse used for RF generation to the total energy per high voltage pulse, reaches 96%. Taking into account the modulators electronic efficiency of 90% and the ohmic losses of 2% of the HV pulse cables, the total modulator efficiency is 85%. At 5Hz repetition rate an average AC power from the wall plug of 120kW per RF station is required. In addition 14kW for the auxiliary power supplies must be added. The total average AC power required for 560 active RF stations is therefore 75MW. Table 5 summarizes the power requirements for RF generation in the main linac of the 500GeV linear collider. For the XFEL the particle beam current is decreased to 5mA. Thus the klystron peak RF output power is only 5.2MW. The repetition rate is doubled to 10Hz. This results in an average AC power of approximately 160kW per RF station. The auxiliary power stays unchanged at 14kW. TABLE 5. Efficiency and Power Requirements of the RF System for the Linear Collider. RF peak power per RF station 9.7MW Duty cycle 0.685% Average RF power available per RF station 66kW Klystron efficiency 65% Modulator efficiency 85% Total efficiency 55% AC power per RF station 120kW Auxiliary power per RF station incl. LLRF and 14kW waveguide tuner Total wall plug power per station 134kW Number of active stations 560 Total wall plug power 75MW
INTERLOCK AND CONTROL SYSTEM The interlock system protects the accelerator and the RF stations in case of malfunction and allows a fast localization of faults. In case of klystron gun sparking the modulators IG(B)CT switch must be opened fast and reliable to limit the energy deposited in the klystron spark to less than 20J. The energy stored in the transformer leakage inductance and in the power transmission cable will be dissipated in two passive networks, one at the cable end near the IG(B)CT consisting of a reverse diode and a resistor and a second one made up by an 80Ω resistor across the transformer primary and by a 100µF capacitor which limits the peak inverse voltage at the primary to 800V when the IG(B)CT is opened. In case the IG(B)CT does not open when the klystron arcs, a crowbar is fired. More internal signals controlled by the interlock system are e.g. cooling water flow and temperature, klystron focusing solenoid current, klystron vacuum interlock, sparks in the RF distribution system, reflected RF power or RF leaks. External interlock conditions result from e.g. cavity power coupler or cryogenics faults. The technical interlock, which protects the linac and the RF station under fault conditions, will be based on programmable electronic circuits. Programmable logic devices have the advantage that the interlock logic can be modified easily if required. Today electronic boards based on these devices are fast enough to react within µs. Therefore no additional hard wired systems for fast interlocks are necessary. Besides system protection and providing start up and shut down procedures for the RF stations, the control and interlock system will offer a comprehensive diagnostics of the RF systems. It will allow to measure and to diagnose actual parameters as well as to adjust set points within certain limits for each RF station and its subsystems and to react to different fault conditions in a flexible manner. Communication with the accelerators main control can be accomplished by different protocols. Since each RF station consists of two parts, one in the tunnel and one in the access hall, there will be two interlock units connected to each other by glass fiber cables. Interlock conditions are summarized into categories, each starting a certain action in the other unit, thus only a limited number of fibers is required. RF WAVEGUIDE DISTRIBUTION SYSTEM Instead of a treelike RF waveguide distribution system a linear system is foreseen for TESLA. It is similar to the waveguide distribution system used for the superconducting cavities in the HERA tunnel and similar to the one already in use for the TESLA test facility. From each of the two output arms of the multibeam klystron equal amounts of RF power are branched off by hybrid couplers of different coupling ratio. Circulators between the hybrid couplers and the cavity input couplers protect the klystron from reflected power at the start of the RF pulse during the filling time of the cavity and at the end of the pulse. Their load resistors together with the cavity input coupler and the three stub waveguide transformers between circulator and input coupler determine the loaded cavity impedance as seen by the beam. The waveguide transformers provide an impedance matching range from 1/3Z W to 3.0Z W and the
possibility of ±50 o phase adjustment. Each stub will be equipped with a motor, which will be controlled by the low level RF system. The transformers will also be used to adjust the phase deviations of 6 and 12 resulting from thermal expansion of the distribution system at 5Hz and 10Hz repetition rate respectively. The average power loss in the distribution is 6% including 2% in the circulators. Cryomodule 1 Cryomodule 2 Cryomodule 3 cavity input coupler Hybrid C oupl er RF Distributi on waveguide transformer RF from Kl ystron load circulator DETAIL hybrid c oupler Figure 6. RF waveguide distribution of one RF station. Several waveguide components have been developed at DESY and in industry within the last years, especially to mention, the high power circulators and loads, phase shifters and transformers and the hybrid couplers. Additional information on the design criteria of the waveguide distribution system can be found in [15]. LOW LEVEL RF The low level RF system of each of the RF stations is a digital system which controls amplitude and phase of the vector sum of the fields in the superconducting cavities connected to this station. Perturbations of the accelerating fields result from fluctuations of the beam current and fluctuations of the cavity resonance frequency due to mechanical vibrations and due to Lorentz force detuning. Since the energy spread must be kept below a maximum tolerable level of 5x10-4, amplitude and phase errors of the order of 5% and 20 o respectively as a result of the Lorentz force detuning and mechanical vibrations must be suppressed by a factor of at least 10. Fast variations are counteracted by a fast amplitude and phase modulation of the incident RF power whereas long term variations (on the timescale minutes or longer) are counteracted by the use of cavity frequency tuners. Most of the perturbations can be counteracted by a fast feed forward system because they are repetitive from pulse to pulse. For nonrepetitive pulse-to-pulse and intra-pulse variations a feed back system is required. The RF modulator for the incident wave is designed as an I/Q modulator to control the inphase (I) and quadrature (Q) component of the cavity field. More detailed information of the LLRF system can be found in [16, 17, 18, 19, 20].
XFEL ACCELERATOR PARAMETER In the TESLA Technical Design Report Supplement the XFEL is described in detail. Its technology is the same as for the linear collider. Hence the RF system layout for the XFEL is the same as for the linear collider in principle. Table 6 summarizes the parameters of the XFEL accelerator. The layout of the 26 RF stations will be as for the 500GeV linear collider. TABLE 6. TESLA XFEL Parameter. Final energy 10...15 20 GeV Accelerating gradient 10...17...23.5 MV/m Total length 1380 m Active length 859.4 m Modules 78 Cavities 936 Klystrons 26 Bunch charge 1 nc Bunch spacing 200 ns Bunch train length 800 µs Repetition rate 10 Hz Average beam power 600 kw Photon pulse length Photon energy Wavelength Number of photons per bunch Peak brilliance Peak power 100 fs 0.2 3.0 12.4 kev 6.4 0.4 0.1 nm 430 20 1.2 x 10 12 0.06 1.7 5.4 x 10 33 135 100 24 GW THE FUTURE OF TESLA In February 2003 the German Federal Research Minister announced that the TESLA XFEL is to be built at DESY in Hamburg [21]. Germany will cover half of the investment costs of 673million. Negotiations with European partners will start soon so that the construction of the XFEL could start in two years. The total construction time will be approximately six years. No site for the TESLA linear collider was proposed at that time. Since a linear collider can be built only in an international collaboration further international developments must be taken into account. The TESLA collaboration will continue the research work on the linear collider. REFERENCES 1. TESLA, The Superconducting Electron-Positron Linear Collider with Integrated X-Ray Laser Laboratory, Technical Design Report, DESY 2001-011, ECFA 2001-209, TESLA Report 2001-23, TESLA-FEL 2001-05, March 2001 2. TESLA XFEL, First Stage of the X-Ray Laser Laboratory, Technical Design Report Supplement, DESY 2002-167, TESLA-FEL 2002-09, October 2002
3. S. Choroba, The TESLA RF System, Proceedings of the Modulator Klystron Workshop MDK 2001, CERN, Geneva, April 2001 4. C. Bearzatto, M. Bres, G. Faillon, Advantages of Multiple Beam Klystrons, ITG Garmisch- Partenkirchen, May 4 to 5, 1992. 5. R. Palmer, Introduction to Cluster Klystrons, Proceedings of the International Workshop on Pulsed RF Power Sources For Linear Colliders, RF93, Dubna, Protvino, Russia, July 5-9,1993, p 28. 6. A. Beunas, G. Faillon, 10 MW/1.5 ms, L-band multi-beam klystron, Proc. Conf. Displays and Vacuum Electronics, Garmisch-Partenkirchen, Germany, April 29-30 1998. 7. A. Beunas, G. Faillon, S. Choroba, A. Gamp, A High Efficiency Long Pulse Multi Beam Klystron for the TESLA Linear Collider, TESLA Report 2001-01. 8. W. Bothe, Pulse Generation for TESLA, a Comparison of Various Methods, TESLA Report 94-21, July 1994. 9. H. Pfeffer, C.Jensen, S. Hays, L.Bartelson, The TESLA Modulator, TESLA Report 93-30. 10. The TESLA TEST FACILITY LINAC-Design Report, Ed. D.A. Edwards, Tesla Report 95-01. 11. H. Pfeffer, L. Bartelson, K. Bourkland, C. Jensen, Q. Kerns, P. Prieto, G. Saewert. D. Wolff, A Long Pulse Modulator for Reduced Size and Cost, Fourth European Particle Accelerator Conference, London 1994. 12. H. Pfeffer, L. Bartelson, K. Bourkland, C. Jensen, P. Prieto, G. Saewert. D. Wolff, A Second Long Pulse Modulator For TESLA Using IGBTs, Proceedings of the Fifth European Particle Accelerator Conference, EPAC96, Sitges (Barcelona), 10-14 June 1996, p. 2585. 13. H.-J. Eckoldt, Pulse Cables for TESLA, TESLA Report 2000-35. 14. H.-J. Eckoldt, N. Heidbrook, Constant Power Power Supllies for the TELSA Modulator, TESLA Report 2000-36. 15. Conceptual Design Report of a 500 GeV e+ e- Linear Collider with Integrated X-ray Laser Facility, DESY 1997-048, ECFA 1997-182, edited by R. Brinkmann, G. Materlik, J. Rossbach, A. Wagner (1997). 16. Liepe, S.N. Simrock, "Adaptive Feed Forward for Digital RF Control System for the TESLA Test Facility", European Particle Accelerator Conference, EPAC 98, Stockholm, Sweden, June 22-26, 1998, in print. 17. M. Hüning, S.N. Simrock, "System Identification for the digital RF Control for the TESLA Test Facility", European Particle Accelerator Conference EPAC 98, Stockholm, Sweden, June 22-26, 1998, in print. 18. S.N. Simrock, I. Altmann, K. Rehlich, T. Schilcher, "Design of the Digital RF Control System for the TESLA Test Facility", European Particle Accelerator Conference, EPAC 96, Sitges (Barcelona), Spain, June 10-14, 1996, p. 349. 19. S.N. Simrock, T. Schilcher, "Transient Beam Loading based Calibration of the Vector-Sum for the TESLA Test Facility", European Particle Accelerator Conference, EPAC 96, Sitges (Barcelona), Spain, June 10-14, 1996, p. 1866. 20. B. Aune, et al., THE SUPERCONDUCTING TESLA CAVITIES, Published in Phys. Rev. ST Accel. Beams 3:092001, 2000. 21. Bundesministerium für Bildung und Forschung, Pressemitteilung 16/2003, 05.02.2003