RF Window problems - September 2009

Shortly after we have achieved very good RF processing results in September 2009, we had a serious RF breakdown event at the RF window. This RF window separates the RF waveguide section at SF6 overpressure from the RF cavity at high vacuum.

In short, while running the klystron at about 7.5 MW output power a breakdown event occurred at the vacuum site of the RF window. We installed photodiodes to monitor breakdown light flashes from the RF cavity and from the RF window. At this specific event the light output from the window was evident. After this event, we had to turn the RF power level down to 1 MW and even then small discharges were appearing each pulse. RF processing did not improve this situation and lead to the conclusion that the RF window has been damages and had to be replaced.

We decided to replace the current CPI RF window of type VWX-1053 with a pillbox type RF window. After consulting some manufactures, we have ordered a RF window from CML Engineering of type 3020-02. This window has higher specifications in terms of peak and average RF power level and therefore will be more reliable. At the end of November 2009 this RF Window and adapter piece has been delivered.

The RF window that was installed originally.
The new RF window of CML Engineering installed in the RF section. The connection at the bottom is to the adapter piece for the connection to the RF input coupler of the RF cavity.
Side view of the RF window.

Milestones achieved during RF processing - 10th of September 2009

While continuing RF processing of the RF cavity, the level of RF input power is gradually increasing. For instance, on the 10th of September we could ramp up very fast up to 3.9-MW input power and 3.0-MeV electron energy at 50 Hz without experiencing any breakdown.

By the end of the day the level of 7.1-MW input power and 3.8-MeV electron energy has been reached in case of a 1.0-us short pulse and 40-Hz repetition frequency. The measured peak dark current was 0.17 mA and the charge per pulse was 0.079 nC.

7.0 MW and 3.8 MeV

7.1 MW and 3.8 MeV

First accelerated electrons - 3th of September 2009

During the first week of September we continued the RF processing of the RF cavity.

The second day we also turned on the main solenoid magnet around the RF cavity. As a result we had to repeat the training session of the day before, because the magnet influences all the processes inside the cavity. For instance, field emitted electrons follow different paths when the magnet is turned on.

On the 2^nd of September we soon reached 3.0-MW input power and 2.4-MeV electron energy. We continued the RF processing by replacing the 10-dB attenuator at the input of the preamplifier to 6 dB. Now we sometimes observe breakdown events with a measurable electron signal on the beam dump at the exit of the cavity and at the same time observe a light pulse on the photodiode that looks inside the RF cavity. Surprisingly, not all breakdown events are accompanied by electrons and light. Possibly, these events are taking place outside the RF cavity. By the end of this day, we have reached 3.6-MW input power and 2.9-MeV electron energy operating at 50 Hz.

The 3^th of September was a very exciting day, because we observed for the first time dark current coming out the RF cavity. The power level at that moment was 3.7-MW input power and 3.0-MeV electron energy. We were able to optimize the setting of the main solenoid magnet around the cavity by maximizing the dark current on the beam dump. We also observed single-side electron multipacting at the photocathode in the RF gun (see Han et al. in Phys. Rev. ST AB 11, 013501 (2008)). For instance, we have measured at the setting of 3.9-MW input power and 3.0-MeV electron energy a peak dark current of 52 nA and a total charge of 0.11 nC per pulse.

Below you can find two oscilloscope images displaying with the following signals: Light Blue - Modulator current; Purple - Forward RF Pulse; Green - Backward RF pulse; Dark Blue - Beam dump signal.

The first observation of dark current out of the RF cavity measured at the beam dump at the exit.

Dark current at a higher setting of the acceleration field.

FIrst RF Processing session of RF cavity - 31th of August 2009

On Monday 31^th of August 2009 it was the day of starting the RF Processing procedure.

RF Processing is a neccesary step during the installation of the RF accelerator in which gradually increased RF power is applied to the cavity in order to clean and smoothen the inner surface. As a result the cavity can hold increasingly higher electric field levels without resulting in a breakdown event.

In order to perform a successful RF processing procedure, we have equiped the setup with several detection systems:
1- Breakdown detector: A piece of analog electronics monitors the backward reflected RF waves from the cavity and senses very abrupt changes in RF power that will occur upon breakdown inside the cavity. This is due to the change of impedance of the cavity when a breakdown occurs. When a breakdown is detected the RF pulse is immediantely interrupted.
2- Vacuum pressure monitor with switch: When the vacuum pressure inside the cavity exceeds a certain level the RF power to the cavity will be interupted untill the pressure has droped below the level again.
3- Faraday cup or electron beam dump at the exit of the cavity to monitor the electrons coming out of the cavity. The signal is presented on an oscilloscope.
4- Photodiode that monitors the light output of the cavity to detect the intensity of the breakdown event. The signal is presented on an oscilloscope.
5- Photodiode that monitors the light output near the RF window to detect the intensity of the breakdown event.

We have a few parameters that we can set in the system:
1- RF peak power by changing the RF input power to the Klystron tube by changing a variable discrete attenuator between 0.0 and 15.5 dB in steps of 0.5 dB.
2- RF peak power by changing the high voltage setting of the modulator. This can be seen as a fine adjustment of the RF power as compared to option 1.
3- RF pulse length by changing the modulator pulse length. This can be changed between 0.50 and 2.80 us flat top.
4- RF pulse repetition frequency can be set between 0 - 100 Hz as long as the average power of the modulator does not exceeds 6.5 kW. In general, we do not exceed 50 Hz.

The method that we apply is to start with the setting of the shortest pulse and increase in RF power. After a certain level has been achieved, the pulse length is increased, while keeping the maximum field strength in the cavity constant. Do to this the RF filling time of the cavity has to be taken into account. Once dark current is observed from the cavity, this is easily achieved by keeping the dark current constant. The last step is to increase the pulse repetition frequency. This procedure can be repeated untill the maximum RF input power is reached.

At 18:00 that day we started under supervison of the radiation safety department of TU Delft the RF processing procedure. We replaced the 20 dB attentuator at the input of the preamplifier to a 10 dB. Soon after the start we reached the level of 1 MeV electron energy. By midnight we haved reached 2.9 MW input power and 2.5 MeV electron energy. The estimated electric field inside the cavity is about 60 MV/m.

The accelerator is control from the control room. Left to right: Juleon Schins, Laurens Siebbeles, Koos van Kammen and Walter Knulst.	Walter Knulst is controling RF processing from the accelerator GUI build in LabView.
Walter Knulst together with Martien Vermeulen who is the allround electron accelerator engineer.	Another look at the control room.
A breakdown event is recorded by the oscilloscoop. In green you can see the abrupt change in RF reflection from the cavity and 400 ns later the breakdown detector has switched of the RF power.	Another breakdown event, which has unfortunately not been detected by the breakdown detector due to the fact that the rising edge is less steap than the event to the left.

Low power tests of RF Cavity - 27th of August 2009

On Thursday 27^th of August 2009 the system was ready to be switched on. All subsystems have been tested before. We connected on this day the RF waveguide for the first time to the accelerator cavity.

The first test is to use very low power RF and measure the resonance frequency as a function of cavity temperature. This test is very exciting, because all RF parts have never been tested as one assembly.

The RF cavity has been checked by measuring the resonance frequency using a network analyzer after manufacturing. You can read more about electrical validation in a post from 2006. It has been observed that the resonance frequency is 2997.802 MHz at 20 ^oC in air.

In between the RF waveguide section and the RF cavity, a RF window and input coupler is placed. You can read more about the RF waveguide section in a post from 2006. The input coupler has been tuned in such a way that maximal RF power is coupled from the rectangular waveguide into the coaxial waveguide feeding the RF cavity. However, this has never been tested in combination with the actual cavity.

For the lower power test of the RF cavity, we have modified the RF system so that the maximum output power is reduced by 20 dB, i.e. 0.1 MW instead of 10 MW. This is done by inserting a 20 dB attenuator at the input of the 100 W preamplifier. You can read more about the RF system in a post form 2007. This safety measure protects the RF cavity from severe breakdown events.

We started by putting the RF cavity at 20 ^oC. We achieved a perfect incoupling at a frequency of 2999.080 MHz. The difference of -1.28 MHz with the measurement shortly after manufacturing is more than can be explained by the difference between air and vacuum of -0.95 MHz.

We continued the test by increasing the temperature by steps of 5 ^oC up to 35 ^oC. The resonance frequency shifts to lower frequencies when increasing the temperature. The resulting coefficient of 48 kHz/K is very close to the theoretical value of 48.5 kHz/K.

The conclusion is that the resonance frequency at 30 ^oC of the RF cavity (2998.600) is very close to the center frequency of the PLL. Therefore, we have decided to operate the RF cavity at this temperature setting.

We finished the low power test of the RF cavity by increasing the RF power.

Below you will find several scoop displays showing the reflection of RF power at the resonance frequency of the RF cavity at different temperatures. Dark blue: high tention of modulator; Light blue: current of modulator; Purple: Forward RF pulse; Green: Backward RF Pulse.

RF Cavity at 20 oC and frequency 2999.080 MHz.	RF Cavity at 30 oC and frequency 2998.600 MHz.
The temperature dependence of the RF Cavity.	RF Cavity at 30 oC and frequency 2998.600 MHz. Modulator pulse setting is 0.90 us and higher RF power.