November 6, 2002
Proposed RF
Gun pulse shaping experiments
John Schmerge and Ronald Akre
The LCLS gun will need to be operated at 120 Hz and at peak field on axis of 120 MV/m. Assuming a 3 ms long klystron pulse and 10 MW rf power to achieve the necessary 120 MV/m field, the gun must dissipate 3.8 kW of power. Preliminary thermal analysis has determined that the gun will undergo significant thermal distortions leading to the resonant frequencies shifts of each individual cell and subsequent field redistribution resulting in an undesirable gun cell field ratio. Thus it is necessary to limit the amount of heat dissipated on the gun walls to something less than 1 kW. One method to achieve this is to inject high power (something greater than required to achieve the desired steady state field) into the gun for a short period of time and then rapidly decrease the incident power to the necessary stead state power value once the desired field value in the gun has been reached [1]. Since this technique has not been used previously the pulse shaping concept should be demonstrated experimentally. The necessary equipment for the experiment is listed below.
All of the required equipment exists at the GTF except for the fast attenuator and linear amplifier. However, the attenuator can be borrowed from SLAC for the duration of the experiment and the Class A linear solid state sub-booster (SSSB) is currently under construction. The SSSB will be built around a module built by Microwave Amplifiers Ltd. The 1kW Linear, Class A, amplifier is due to ship from England in 2 weeks. Testing of the amplifier module and design of the support electronics will take place during November and December 2002. Fabrication will be done during January and the first part of February. The amplifier will be tested mid February 2003. The testing will include bench tests of the following:
After completing the bench testing, the amplifier will be set up to run a 5045 klystron in the klystron gallery. The output of the SLEDed klystron phase and amplitude will be history buffered by the control system. This data will be compared with other klystrons in the sector. At this point the SSSB will be ready for use including an experiment on rf pulse shaping.
The first part of the pulse shaping experiment would be to test the ability to rapidly decrease the rf power exiting a 5045 SLAC klystron. According to simulation the desired pulse shape is 30 MW (or higher) for 800 ns followed by 200 ns of 10 MW rf power. The pulse should be terminated after a complete duration of 1000 ns. The maximum desired rise time for each amplitude change is 100 ns. This test should be conducted using a 3 M SLAC linac section to avoid reflections and potential damage to the klystron. The test could be conducted at either the GTF or the SLAC linac. A fast low power attenuator would be programmed to obtain the desired klystron output pulse shape. The test would consist of programming the low power attenuator and measuring the klystron output pulse shape using the waveguide directional coupler and rf detector.
After completion of this test, an identical experiment should be conducted using a LCLS prototype gun instead of a linac section. However, the only prototype LCLS rf gun available is at the GTF and the klystron connected to it is an XK-5 instead of an LCLS type 5045. Nonetheless the rise time of the klystrons should be comparable and can be measured. The maximum possible klystron output power is approximately 25 MW so the klystron will be switched from 25 MW to 10 MW and the field in both gun cells measured to determine the effectiveness of the rf pulse shaping. The fast attenuator can be re-programmed as necessary to alter the gun fields as desired. A successful experiment would demonstrate the ability to reach a field of at least 100 MV/m (preferably 120 MV/m) in less than two time constants instead of the typical 4-5 time constants. The time constant for the GTF gun is currently 700 ns but may be modified slightly with the installation of a new Mg cathode.
Once the SSSB is available the entire experiment should last no longer than two weeks with one day to set up, and four days for generating, interpreting and modifying the klystron driving the linac. The second week would be spent with the klystron driving the gun. The two experiments can be conducted separately and do not need to be consecutive depending on equipment and personnel availability.
[1] J.F. Schmerge,
Reducing the heat load on the LCLS 120 Hz RF gun with RF pulse shaping,
LCLS-TN-02-07 (Nov. 2002).