Detailed Design Report

Transcription

1 Detailed Design Report Chapter 2 MAX IV 3 GeV Storage Ring 2.6. The Radio Frequency System MAX IV Facility

2 CHAPTER 2.6. THE RADIO FREQUENCY SYSTEM 1(15) 2.6. The Radio Frequency System 2.6. The Radio Frequency System General Discussion Motivating the 100 MHz System The 100 MHz RF System Description of the 100 MHz Cavities RF Station Structure Reliability HOM Suppression Control Loops MHz Higher Harmonic Cavities Description of Prototype Higher Harm. Cavity...14 References

3 CHAPTER 2.6. THE RADIO FREQUENCY SYSTEM 2(15) 2.6. The Radio Frequency System General Discussion Motivating the 100 MHz System The Radio Frequency (RF) system is of vital importance for the performance of the storage rings. It must provide the voltage necessary for the required beam lifetime and it should not induce Higher Order Mode (HOM) instabilities. Other important aspects are the capital and running costs. Super Conducting (SC) cavities offer a high shunt impedance and a very small contribution of HOM. In order to achieve the necessary voltage and power two cavities have to be used. One spare cavity may be needed for operation safety. One straight section with isolation valves between the cavities is needed in this case. The RF amplifiers could be of the newly developed Inductive Output Tube (IOT) type. These tubes can have a very high efficiency (>60%) and a typical power of 80 kw. Four tubes can be combined to a 320 kw amplifier unit. Special cryogenic knowledge and infrastructure is needed to operate these cavities. Higher Order Mode-Damped Cavities (Euro cavities) as developed in the EU project HPRI-CT offer a high degree of HOM damping. These cavities can be housed in the short straight sections and the same type of amplifiers as in the SC case can be used. 100 MHz capacity loaded cavities as developed for MAX II and MAX III are also HOMdamped. These can be housed in the short straight sections. The amplifiers are of the tetrode type which offer quite a high efficiency when operated in the class C mode (appr. 70%). Any of these three systems are able to carry out the function of the RF system needed and the final choice among these will require some additional detailed studies. In Table 1 the characteristics of the three systems are given for reaching an RF-acceptance of 4.5% in a 60% ID equipped ring (energy loss of 756keV/turn) at 500mA. Notable is the fact that to reach 4.5 % energy acceptance, both the SC cavity case and the Euro cavity case requires higher electric power than the 100 MHz Capacity loaded cavity case. Bunch length calculations assume a HHC system, which seems to be more complicated for the 500 MHz cases.

4 CHAPTER 2.6. THE RADIO FREQUENCY SYSTEM 3(15) Table 1: Comparison of different RF systems for the MAX IV storage ring. SC Euro Cap.loaded Operating frequency [MHz] Nr of cavities Shunt impedance [MΩ] 1) K [W] Refrigerator power [kw] Amplifier type IOT IOT Tetrode Voltage [MV] for 4.5% bucket height Total RF power [kw] Electric power [kw] Nr of RF amplifiers Rms bunch length [mm] Cost High Medium Low Ease of operation Resistive wall effect Higher order mode spectrum Good Medium Good In the following we study only the 100 MHz Capacity loaded cavity case.

5 CHAPTER 2.6. THE RADIO FREQUENCY SYSTEM 4(15) The 100 MHz RF System Table 2 presents the data for two alternative 100 MHz configurations. Both alternatives are for 4.5% RF energy acceptance, however alt I is for a 60% ID equipped ring (360 kev+396 kev per turn), and alt II is for a fully ID equipped machine(360kev+660kev per turn). The reason to study the 60 % case is that for a quite long time after commissioning, the ring will not be fully equipped. However, we must keep in mind that a solution for the 100% equipped case should be in reach after this intermediate stage. Table 2: Alternative 100 MHz configurations for reaching 4.5% RF energy acceptance. Alternative I II Energy loss with Ids 756keV 1020keV Circulating current 0.5A 0.5A Total beam power 378kW 510kW Total RF voltage 1.5MV 1.8MV Number of cavities 6 6 Cavity shunt impedance 3.2Mohm 3.2Mohm Cu losses 117kW 169kW Total RF power needed 495kW 679kW Nr of RF stations 6 6 Nr of transmitters Transmitter power 41.5kW 56kW Power to cavity 83kW 113kW Cu losses/cav 20kW 28kW Coupling (beta) Cavity voltage 250kV 300kV Cavity gap 4cm 5cm Kilp. Limit 260kV 310kV Alt I: Represents a solution for a 60% ID equipped ring, with the present MAX II/ MAX III cavities. Alt II: Represents a solution for a fully ID equipped ring, with slightly modified MAX II/MAX III cavities.

6 CHAPTER 2.6. THE RADIO FREQUENCY SYSTEM 5(15) It has been decided to go for an upgrade of the present 100 MHz main cavity design, that is alternative II. By improving the cavity design, we are confident to reach 300 kv in each cavity. This in turn implies that only six cavities are needed to cover the case of a fully ID equipped ring: 1.8 MV equivalent to 4.5 % RF energy acceptance (see table). With the old cavity design we would have needed eight cavities. With the new cavity design five cavities will be sufficient to reach 4.5% RF EA for the intermediate stage of a 60 % ID equipped ring. Two major advantages can be seen: 1) The HOM shunt impedances are reduced 2) the coupling (beta) is reduced. Some more comments: In the chosen alternative II, the natural choice of transmitter power is 60 kw. However, the individual transmitter powers can in principle be halved, by combining four transmitters instead of two, as indicated in the table. There would thus be the possibility to use transmitters of exactly the same type as for MAX II and III. Likewise can the power to the cavity be halved, by splitting up power line somewhere after the circulator (see below) into two branches feeding two coupling loops. In this way the demands on the RF power feed-through to the cavity can be relaxed, if needed. We have also studied the case where the cavities are made double-sided. One reason to go from single- to double-sided RF cavities is the possibility to decrease the number of cavities and in this way lower the shunt impedance of HOMs. However, it turns out that while the number of HOMs is scaled down a factor of two, their strengths are roughly doubled. That is, no dramatic gain can be expected Description of the 100 MHz Cavities The cavity design for the MAX II/MAX III rings (Alt I) is seen in Fig 1. It is of the capacitor loaded type. Tuning of the cavity is accomplished by changing its capacitance by a slight movement of the left (front) sidewall of the cavity. The stepper motor is seen on top of the cavity and the mechanical transmission system is indicated to the left. Three cavities of this type are currently in operation at the MAX II storage ring and another one is placed in the MAX III ring [1]. The cavities have been conditioned to an accelerating voltage of about 250 kv. The inner profile of this cavity is given in Fig. 2. Its theoretical shunt impedance is 3.5 Mohm and the theoretical Q value is However surface roughness and ports on the cavity body lower these values by roughly 7%. The characteristics of this type of cavity have been mapped and the parameter values are presented in Table 3. The fundamental accelerating mode of the electric field is concentrated to the capacitor part at one end of the cavity. This mode has a very small electric field at the other side of the cavity. However, some of the longitudinal HOM:s has a substantial electric field at this back-plate. Damping (capacitive) antennas are introduced at this position to damp the HOM:s while keeping the shunt impedance of the fundamental mode intact. Other HOMs which cannot easily be damped by the capacitive couplers, are planned to be inductively coupled out via one or two loops penetrating from the mantle surface. The input coupler loop was used as such a damping antenna (50 Ohm terminated) when performing the low power HOM measurements presented in Table 3.

7 CHAPTER 2.6. THE RADIO FREQUENCY SYSTEM 6(15) Figure 1: The 100 MHz capacity loaded cavity. 100 MHz Capacity-loaded Cavity for MAX-II and -III, F = MHz C:\LANL\EXAMPLES\RADIOFREQUENCY\MYONEEXAMPLES\PILL100.AF :40:42 0 Figure 2: Superfish geometry of the MAX II & III cavity.

8 CHAPTER 2.6. THE RADIO FREQUENCY SYSTEM 7(15) Measured Frequency [MHz] Qtheory Qmeasured Rtheory [ Ω] Rscaled [ Ω] * * Table 3: MAX II/MAX III type 100 MHz cavity parameters for longitudinal modes. Measured Q-values are for the case of two capacitive and one inductive damping antenna. It is noticeable that the first longitudinal HOM appears at such a high frequency, compared to a usual pillbox cavity where it appears only at a frequency 1.4 times the fundamental. This is achieved by the fact that a capacity loaded cavity is compressed in volume. The price to be paid is a lower shunt impedance compare to a pillbox cavity. The longitudinal HOM:s have been mapped up to 1500 MHz. Modes above this frequency will not be activated because of the poor form factor, less than approx. 0.1, given by the long bunch length. Four of the measured modes are still possible candidates to drive coupled bunch mode instabilities (see instability chapter), even with the present damping scheme. Intended enhanced damping of these modes is described below under HOM suppression. The slightly improved cavities which we are considering in alternative II will have an inner profile shown in Figure 3. They will have an acceleration gap of 5 cm, and dimensions of the inner profile as described in Table 4. The theoretical shunt impedance is 3.5 Mohm and the theretical Q value is However surface roughness and ports on the cavity body will lower these values by roughly 7 %. The cavities should withstand 30 kw of Cu losses, giving a peak acceleration voltage of 300 kv. The deposited power, that has to be cooled away by cooling channels within the copper material, is given in Table 4, for the different surfaces.

9 CHAPTER 2.6. THE RADIO FREQUENCY SYSTEM 8(15) Figure 3: Superfish geometry of the MAX IV cavity. The cavities should have six extra ports (apart from the two rotational axis ports) with Conflat type UHV flanges on the mantle surface for pumping, high order mode (HOM) damping coupler and input power coupler. Two measurement ports should be positioned on the back plate. The ports are specified in Table 5. The tuning mechanism will be of the same type as for our present cavities, with some modifications. The tuning range should be within the reversible deformation region of the copper of the cavity front end plate (minimum ± 1mm).

10 CHAPTER 2.6. THE RADIO FREQUENCY SYSTEM 9(15) Table 4: Inner dimensions and power deposition in the rotationally symmetric 100 MHz cavities. Surface nr Z (along axis) [mm] R (radius) [mm] P (Deposited power) [kw] 1 0 -> > > > > > > > > > > > > > > > Table 5: Specifications for the cavity ports. Port type Diameter [mm] Flange type Total length [mm]from Ri Position Z (or R) [mm] Position φ [ ],φ=0 is downward* Slotted, Pump 180 CF Slotted, Pump 180 CF Open, Power 140 CF Open, Power 90 CF Open, HOM 90 CF Open, HOM 90 CF Open, Beam 50 CF (from Zi) Open, Beam 50 CF Open, Probe 30 CF (from Zi) 200 (R) 30 Open, Probe 30 CF (from Zi) 200 (R) -30 * When looking from the front end side, positive angles are at positions reached counterclockwise.

11 CHAPTER 2.6. THE RADIO FREQUENCY SYSTEM 10(15) RF Station Structure RF Station It has been decided to abandon the principle of pair cavities, which would allow operating without a circulator. However this scheme is not yet fully tested and, most important, it decreases the RF station modularity (a station would consist of two cavities and four transmitters). The plan is now to combine two 60 kw transmitters to a 120 kw transmission line, equipped with a circulator and a load, and couple this power into one (improved) cavity. As mentioned above, only five such stations will suffice for a 60 % ID equipped machine. Even four stations, giving 1.2 MV will suffice for quite a time. The major advantages are: 1). We will finally have six independent RF stations, which means increased modularity compared to the pair cavity case. 2). The advantage of power regulation with phasing can still be utilized, which means higher reliability since the transmitters can run at constant power. The MAX IV storage ring RF system will consist of six 120 kw RF stations, each one feeding one cavity. The available power from the net should be 200 kw per station, in order to allow for an overall efficiency of 60 %. The output power is fed via a circulator to the cavity. A 120 kw load is connected to the circulator. Cavity Circ. 120 kw Load 60 kw 60 kw φ -φ

12 CHAPTER 2.6. THE RADIO FREQUENCY SYSTEM 11(15) Amplitude and phases of the cavity fields must be kept stable during different steps of machine operations. These require different power levels from the tetrode amplifiers and different tuning of the cavities. The low level RF system will include a frequency loop for each of the six cavities, amplitude and phase loop for each RF station. We have chosen class C amplifier because of their high efficiency (70%) but they have a very nonlinear relation between input and output power. This non linearity is biggest at very low output power. This makes it difficult to decrease the power to such low level which is needed when beam loading is very low which the case is when the ring is filled after a shut down. One way to do the power regulation is to let the amplifiers deliver full power but direct some of the output power to the dummy load connected to the fourth port of 3 db hybrid used for combining the power of the two 60 kw transmitters. This is done by changing the phase of the drive signals feeding the two transmitters an equal amount but with opposite sign in order not to change the phase of the output power delivered to the cavities Reliability Few or many RF Stations Our experience from the present MAX facility is that the user community in a better way accept a running machine but with somewhat reduced intensity, compared to a complete shutdown. Many users perform their measurements during short periods like one or two days, and a shutdown of this length of course strikes hard on some unlucky users. Therefore a modular RF system, with several instead of one RF station, seems to be preferable. In this way a failure in one RF station or one cavity might be relatively harmless considering the possibility to tune away or even remove a cavity from the ring Measures to Increase Reliability The anticipated scheme of combining the power from two (or four) tetrode amplifiers by help of a 3 db hybride will make a power regulation possible by phasing, in contrast to a traditional power regulation loop acting on the drive amplifier. The advantage of this is that the tetrode amplifiers can be run with constant output power. This should improve overall phase stabilities between RF stations, and subsequently the reliability of the whole RF system. The construction procedure of the present MAX II/III cavities involved a welding seem in the mantle surface, which turned out to be the weakest link regarding vacuum properties. We are considering an electron beam welding for our new cavities. If such a welding could be performed for the entire cavity, we could also solve the problem of the too soft copper properties, which are the result of a final soldering the cavity. These technical solutions are under investigation. For the probe loops in the cavity we are also considering types that are separated from vacuum via a small ceramic window. This would make it easier to supply calibrated signals to the control system, since the probe coupling could be adjusted without breaking vacuum. Furthermore, we would avoid severe shutdowns due to a probe loop failure.

13 CHAPTER 2.6. THE RADIO FREQUENCY SYSTEM 12(15) HOM Suppression Prototype HOM Couplers The intention is to refine one or two of the capacitive couplers that are situated at the back plane of the cavity. Furthermore, two additional inductive (loop) couplers will be introduced from the mantel surface of the cavity. In these feed-through it will be necessary to apply some kind of stop-band or high pass filter, since the fundamental mode coupling will be too high. We are considering a kind of lambda/4 filter integrated in a double coaxial structure [2]. Even for 100 MHz such a filter would not be too space demanding. Since the fundamental carries 20 kw the filter attenuation of the fundamental mode need to be in the order of 20 to 30 db, in order not to cause thermal problems in the filter. In [3] the measured attenuation of the fundamental mode was 22 db, for the case where a non-filter loop would give beta=6 for the fundamental. The coupling of the fundamental mode to the external load was in the order of -50 db, so the load needed only to be specified for the anticipated HOM power. We should also keep the possibility to use waveguide dampers as developed for the 500 MHz Euro cavities. Those have also been scaled slightly down in cut-off frequency for a scaled cavity version (352 MHz) to be installed in ESRF. We need a cut-off frequency of roughly 400 MHz. HOM damper ports on the cavity should be designed to handle such dampers if this option is chosen Control Loops Resonant Frequency Control Six frequency loops, one for each cavity, will keep the cavities tuned by compensating for both beam loading and temperature effects. Cavity tuning is performed by an elastic deformation of the cavity in the direction of its axial length. This is realised by means of a mechanism, driven by a stepper motor, which changes the width of the gap of the capacitor in the cavity and therefore its frequency. The motor is controlled in closed loop with the output signal from a phase detector, which is proportional to the phase difference between the cavity incident signal from a directional coupler and cavity voltage from monitoring pick-up in the cavity. A phase shifter in the control loop makes it possible to adjust the phase offset Phase Control A phase control loop compensates for phase changes in the amplification chain. The RF signal from the master oscillator and the cavity phase are compared in phase and the error signal drives through a PID regulator an electronic phase shifter. The relative phases between the stations are controlled by a mechanical phase shifter at the input of each of the stations amplifier stages.

14 CHAPTER 2.6. THE RADIO FREQUENCY SYSTEM 13(15) Voltage Control The amplitude control loop adjusts the relative phase difference between the drive of the two transmitters in the RF system. By this the amplifiers can deliver full power but some fraction of the output power will be directed to the dummy load connected to the fourth port of 3 db hybrid MHz Higher Harmonic Cavities A rough estimate of the required voltage for the higher harmonic cavities (HHC) are 500 kv that is one third of the main cavity voltages. However, due to the synchronous phase angle of 30 degrees in the case of 756 kev losses per turn and a main RF voltage of 1500 kv, a more precise number is 424 kv. This includes the assumption of a total shunt impedance for the HHC of 15 Mohm (R=U 2 /P), and that they are passively driven by the beam itself (additional 29 kev loss per turn). The form factor influence on the tuning angle of the Landau cavities has been taken into account in a self-consistent way. This gives an rms bunch length of 68 mm, for a symmetric potential well. The bunch shape is rather non-gaussian, and the FWHM/2.355= 85 mm. The achievable rms bunch length increases slightly (2mm/5 Mohm) with increasing shunt impedance, in the 5-20 Mohm range. Above this range, no considerable gain is made. 15 Mohm seems to be a safe choice to stay Robinson stable regarding the double RF system, even for a current of only 200 ma. The natural rms bunch length for the case studied is 13 mm, and the synchrotron frequency 1.08 khz. The double RF will create a spread in synchrotron frequencies, thus providing for Landau damping. The prototype HHC (see below) is of the same capacity loaded type as the main cavities. This is helping us to push up the HOMs in frequency, and making the HOM suppression easier. The shunt impedance for this type of HHCs is 5 Mohm. By choosing three such cavities, each would pull 4 kw out of the beam. We could also think of a higher number (if the HOMs are sufficiently damped) in order to decrease the power consumption, since our experience of several passive cavities are good (phasing problems are not there, because of the passive nature).

15 CHAPTER 2.6. THE RADIO FREQUENCY SYSTEM 14(15) Description of Prototype Higher Harm. Cavity.

16 CHAPTER 2.6. THE RADIO FREQUENCY SYSTEM 15(15) References [1] Andersson et al, "The 100 MHz RF System for MAX-II and MAX-III", EPAC [2] B. Dwersteg et al, "HOM Couplers for normal conducting Doris 5-cell cavities", IEEE Trans. Nucl. Sci. 32, 1985.