Proposal for experiments with beam loaded field profile
Proposal for experiments with beam loaded field profile in the CLIC accelerating structure. Vasim Khan & Igor Syratchev CERN, BE/RF June 2012
Motivation: The last? piece of the puzzle in our understanding of accelerating structure behaviour at a record high accelerating gradient is how the beam loading affects the breakdown trip rate. TD 26 “Dog-leg” experiment in CTF 3 60 MW from XL 5 klystron CTF 2 Low loss H 01 line (17 m) CTF 3 Drive beam (>1 A) Accelerating structure Currently we are preparing the special experiment (proposed by Walter), where the drive beam current (50 Me. V x 1 -3 A x 3 GHz x 150 ns) will be send through the accelerating structure using old dog-leg beam line in CTF 3 and the same time the RF power from XL 5 klystron will be delivered from CTF 2 area using existing old 30 GHz low loss RF transfer line. We are fabricating all the necessary RF components to modify the transfer line to be used at 12 GHz and prepare the accelerating structure test bench. The installation and experiments are scheduled for the next year. CERN, BE/RF June 2012
Even there is no doubt that such an experiment will give us a lot of information we will be certainly facing a number of challenges: 1. The necessity to install new and expensive components. 2. Operation time needed for experiment will be in a conflict with other CTF 3 programs, as well as the klystron test stand will not be possible to run during that period for other tests in parallel. 3. To do the experiment we have to run rather complex CTF 3 accelerator anticipating all the maintenance issues. 4. The beam losses in the structure (drive beam is not perfect) can lead to the effects that we will not be able to disentangle from the high gradient effects themselves. 5. There will be no access to the test set up during experiment. Even after the experiment one can expect that the structure will not be removed from the test area for a long time due to the radiation restrictions. As an alternative, we have studied possibilities of how to organize (fake) the beam loaded field distribution along the structure by specially prepared artificial Q-factor reduction in each cell. Such a structure can be tested with klystron, providing full control of processing. It appeared, that if the external Q-factor is about equal to the Ohmic one, then the required field distribution can be established. To provide such additional ‘damping’ we propose to modify the positions of HOM damping loads in the cell’s damping waveguides. The details of proposal are presented in the next slides. CERN, BE/RF June 2012
In the CLIC accelerating structure, the HOM loads are radially displaced by about 50 mm. The HFSS simulation showed that if the loads positions changed to ~25 mm, then the loaded Q-factor is twice lower than the original one. Thus the Ohmic and External Q-factors are about the same. First, mid and last cells parameters vs. loads edge position (HFSS). Example: last cell in TD 26 Loaded, External and Ohmic Q-factors in each cell for 25 mm loads offset. Note, that we did not tried to optimize the individual position of the loads in each cell. Q Ext Q Ohmic QL Another important observation is that the other cell parameters at optimal (~25 mm) distance are practically not perturbed. Thus their changes will not contribute to the modified field distribution. Other words, the same structure can be used for the experiments with unloaded and loaded E-field profiles by simply moving the loads racks. Last cell Mid cell First cell CERN, BE/RF June 2012
Calculated field profiles along the structure for the different loads positions (dashed lines). Calculated peak power losses into the single (one of four) load in each cell. unloaded L=30 mm L=25 mm Loaded (target) L=20 mm Average accelerating gradient for the fixed (60. 4 MW) input RF power for the different loads positions. The 100 MV/m corresponds to 24. 7 mm loads offset. The 350 k. W max RF peak (5 W average) power can be considered as a very conservative number for the Si. C loads. But certainly the cooling needs to be introduced. By simply displacing the HOM loads racks in damping waveguides we can organize different field profiles along the structure simulating various beam loading conditions. CERN, BE/RF June 2012
The theoretical beam loaded field profile can be fully reconstructed if the loads positions along the structure can be varied, so that loaded Q-factor in each cell follows: 4 loads racks inserted The new “unloaded” field Losses per single load Following, the calculated loads radial positions: The cells frequencies detuning are in a range from 1. 1 to 1. 7 MHz Constant offset 25 mm 26 mm 24. 2 mm Tapered offset 26 -22 mm 22. 2 mm The frequency dispersion cause some SW pattern in the structure? CERN, BE/RF June 2012
Unloaded Fully loaded Inserting more loads racks The possible experimental scenario can be based on the loads racks with a fixed position and the structure placed in the vacuum tank to minimize the hardware cost. The experiment can be done in stages by inserting the new loads rack (from 0 to 4) and adopting the new input peak power for the fixed average gradient: CERN, BE/RF June 2012
- Slides: 7