1. High Efficiency Klystron
I will talk about high efficiency klystron.
K. Watanabe (KEK, Accelerator Lab 3rd division)
eeFACT2016
2016/Oct/26　20min

2. * Introduction of high efficiency klystron
Content
* Introduction of high efficiency klystron
* CPD (Collector Potential Depression) applied to Gyrotron
* How to applied CPD method to the CW klystron
* Current status of R&D of CPD klystron
* Summary
This is a content of my talk.
1st , Introduction of high efficiency klystron.
2nd , about CPD method applies to klystron. This is our study item.
In this talk, it is main part.
And current status of our study also talk, and summay.

3. Introduction (1)
* Example of a high-power system at SuperKEKB
The high power RF system always needs large electric power in the operation. An improvement of efficiency is also always required as a technology component for the energy saving.
The vast majority of the existing commercial high-power rf klystrons in the electronic efficiency range between 40 and 55 %. Only a few klystron available on the market are capable of operating with about 65 % efficiency or above.
Generally, the dissipate power in the collector is emitted to the air as thermal energy through a cooling devices.
600 kW heat loss emits to the air by AFC
Vapor cooling
Collector
600 kW
Output coupler
RF out 1.2 MW
(Saturation)
Solenoid
magnet
As you know, あとはそのまま。
DC gun
90 kV, 20 A
KPS, DC 1.8 MW (Max)
Toshiba　E3732 (CW klystron), MHz
Used for the TRISTAN, the KEKB and the SuperKEKB.

4. Introduction (2)
Two examples to obtain the high-efficiency of the klystron are shown as following,
(1) Toward High-Power Klystrons With RF Power Conversion Efficiency on the Order of 90 %
@ IEEE TRANSACTIONS ON ELECTRON DEVICES, VOL. 62, NO. 10, OCTOBER 2015
Klystron type: L-band MKB (average 150 kW, Peak 20 MW, Pulse length 150 us)
Method: more stronger bunching on the output cavity.
*To obtain the smaller perveance and the weaker space charge effect, the multi-beam klystron chose. (Amount of beam current is covered by the number of the beam)
* Increase of the number of cavities from 5 cavities (existing) to 7 cavities (this simulation).
* By the optimization of the parameters, the 90% of the efficiency was achieved under some conditions by the simulation.
(2) The CPD (Collector Potential Depression) method applied to Klystron.
CPD is an energy-saving scheme that recovers the kinetic energy of the spent electrons after generating rf power. Collector loss is reused to do the energy recovery (pick-up as electrical power from it).
See next slide…… This is a main part of this talk.

5. CPD method applied to Gyrotron
Collector
To improve the efficiency of high-power source, the CPD (Collector Potential Depression) method already was developed to apply the Gyrotron. (Up to 10 ~ 20 %)
prden.php?type=cat&search=
In this case, an ceramics insulator is used to insulate between the collector and the body to be applied high-voltage of Vc (~ 30 kV) for an energy recovery.
The energy spread of spent beam at operation is important for it. At saturation, the spent beam has small energy spread through electromagnetic interaction in the cavity. Therefore, ~ 30 kV of Vc can apply on the gap to pick-up as the electrical power from the spent beam to the outside of gyrotron after generating rf power.
Spent beam
Insulator Vc Vc
0~30 kV
Ref. PHYSICAL REVIEW LETTERS, Volume 74, Number 26, pp , 26 December 1994
K. Sakamoto et al., “Major Improvement of Gyrotron Efficiency with Beam Energy Recovery”

6. How to apply CPD method to klystron
*An insulator needs to insert between the body and the collector to isolate the collector from body.
Schematic diagram of one for CPD
Conventional
Collector
High-voltage applies to Collector-body
Isolate a collector from body to apply high-voltage for energy recovery.
Without insulator
Ground Vc
Insulator and gap
RF out Vk Vk
クライストロン電源は既存の構成を出来るだけ保持する方式を考える。下図はVkとVcを直列にする「Power=(Vk+Vc)*Ib」が、
後述する並列「Power=Vk*(Ik+Ic)」または、6.6kV交流へ変換し既存電源の入力とする方式を考えている(詳細検討未だ)。
Body is a ground.

7. Limitation of the rf output power by backward electrons
In case of the klystron, the spent electron beam has large energy spread through electromagnetic interaction in the cavities at the saturated operation.
Therefore, the collector potential cannot be increased beyond the lower limit of distribution of the spent electron beam, otherwise backward electrons hit the cavities. Therefore, the range to apply CPD for the klystron is restricted to the unsaturate situation.
Saturate : 1.2 MW ~ 65 % <- cannot apply CPD
Unsaturate: kW % < % by CPD
Toshiba, E3732 (508 MHz, 1.2MW CW)
Saturate
Not applicable of CPD
Eo = 90 kV
Eo = 90 kV Vc
0~50 kV
Unsaturate
Applicable region of CPD
Saturate: 1 MW out
Unsaturate: 200 kW out
Drive power
Energy distribution of the spent electron beam after pass through the output cavity.

8. Fabrication of the CPD klystron for the demonstration
Existing klystron was used to fabricate the CPD klystron at 2013.
(Toshiba CPD)
※Recycled components were the electron gun, the Input cavity and the middle cavities in drift tube.
※Newly fabricated components were the output cavity, the output coupler and the collector with CPD gap.
E3786 (T-44A)
Fabricated at 1989
Newly
Concern
*Leak wake field to outside klystron from CPD gap
*Corona discharge around CPD gap
*Radiation shield
Must be clear these items for the operation.
Toshiba CPD
Recycle
June 2014

9. The parameters compared with E3732 and E37703 CPD
Item
E3732　(E3786)
E37703, CPD
Frequency
MHz
Max output
1.2 MW
500 kW
Efficiency
20~65%
(Max efficiency at saturation)
Without CPD　20-60% (Unsaturation)
With CPD　40-80% (10-20 % improved)
Collector size
1 MW
Cooling method of collector
Vapor cooling (130 l/min + AFC)
Water cooling 　(360 l/min)
Cooling item
Klystron body
Output coupler
Focusing coil
Ceramics insulator, Microwave absorber
KPS
B-type　x 1
B-type x 1
PS for CPD x 1 Vk
47-90 kV Va
25-60 kV, Max Ib 20 Adc
25-65 kV, Max Ib 20Adc
Vc (CPD)
none
0　~　-50 kV
We were examined that to demonstrate the CPD method to apply a klystron modification by existing klystron. Moreover, the parameters were determined as a result of restrictions that the KPS also uses the existing thing.

10. Example of the Power balance of klystron within without CPD
Collector loss
Without CPD -> 420 kW
With CPD -> 184 kW
(recovery: 236 kW )
UHF MHz
RF out: 300 kW (CW)
RF in: ~ 2W (CW)
Example of
: Toshiba E37703, CPD
In case of 300 kW RF output (η = 42%) power,
* Without CPD (Vk = 90 kV、Ib = 8 A)
　Required DC power is 720 kW
-> RF out 300 kW,
Dissipate power in Collector 420 kW
* With CPD (Vk = 90 kV、Ib = 8A、Vc = 30 kV)
　η = 42 -> 62%
-> RF out 300 kW
Dissipate power in Collector 184 kW
　Recovery power 236 kW <- Pick-up by cable from collector, then reuse to drive a klystron
After driving CPD, the required DC power is shifted to about 500 kW using by recovery power from the collector.
Power picks by cable
3Φ AC 6.6 kV, ~ 270 A
90 kV, 8 A -> 720kW supply
KPS, DC power supply
(90 kV, Max 20 A)

11. List of tasks
Now ( ),
(0)　Fabrication of a CPD klystron using by the existing klystron (TRISTAN and KEKB)
Design of the RF and radiation shield to optimize it
Future plan,
(1) Commissioning starts to drive a normal klystron　(After Phase2 of SuperKEKB)
　　* DC aging (until collector loss of 300 kW)
　　* RF out (RF out of 300 kW with the collector loss of 420 kW)
(Note, not insulate between the body and the collector to be connecting a cable)
Check the amount of RF leakage and Radiation level
(2)　Check the output signal from collector without Vc
　　* Small collector loss (~ 100 kW ?)
* Check the output signal from the collector
(The output signal will be terminated to a water cooling type high-power dummy load) 　　
(3)　Energy recovery test with Vc for the proof-of-principle
　　* Apply the high-voltage (Vc) to recovery the kinetic-energy of spent beams.
　　* Measures amount of the recovery power dependence on each parameters.
(Another klystron with Marx-circuit will be used for the high-power Dummy load)

12. RF test station at KEK (D2-AT)
Another klystron to use a dummy load (Toshiba E3732, CW, 1.2MW)
KPS (B-type)
Red one is a cable to connect between the body and collector for commissioning.
Work of necessity
* Stages
* Cooling water line
* Interlock system
* Set-up of LLRF system
* Radiation shield
* RF shield
Etc…
Operator control panel of the KPS
Circulator
Dummy load
CPD klystron +
Focusing coil with socket

13. Issues must be addressed for CPD Klystron
The CPD klystron has an insulator to insulate between the body and the collector for the energy recovery. In this case, three issues must be addressed for the operation. That is;
(a) Corona discharge, breakdown;
around the ceramics insulator and the outside of klystron .
(b) RF leakage;
from the physical gap and the dielectric materials.
(c) Structure of the Radiation shield;
to cover the top of the klystron.
Vc :
~50 kV (DC)
Only dielectric materials
Ceramics Insulator
Vacuum
Plastic cover
Orbit of the spent beam
3M Fluorinert to cool the insulator
Physical Gap (apply the Vc)

14. Calculation of the RF leakage from CPD gap
* RF leakage from output cavity (by HFSS)
Check the cutoff performance from fundamental to 3rd order of the output cavity.
Resonant frequency: Fundamental -> 500 MHz
2nd -> 1000 MHz
3rd -> 1500 MHz
Result: The leakage field level of fundamental and 2nd　
order are very small at 1 MW output .
The 3rd Order must be care (-> ~ 100 V/m)
* RF leakage by wake field (by particle studio)
When the modulated electron beam passes through at CPD gap, the strong leakage field excites from klystron.
Frequency range: 500 MHz ~ 6000 MHz
: In case of DC aging -> No problem
: In case of the Vk= 90 kV (Ib=10A) that is a possibility to build the high voltage (~ kV) near outside of klystron.
But, under calculation to obtain accurate result.
E-field
FFT(4)
HFSS
Pout
Fundamental
500 MHz
FFT(3)
CPD gap
RF out
3rd Order
1500 MHz
90 keV
10 A
Pin
RF in

15. Drawing of the Radiation shield and RF shield
* Radiation shield is already designed by Toshiba lately.
* RF shield is under designing using by a result of simulation.
(i) to make the access ports for the cables and cooling pipes
(ii) to set an absorber for reducing the field level in rf shield
Pb shield
Pb shield Vc
Pb shield
Field level must be less than 28 V/m at outside of the rf shield (Integrated value).
Pb plates set on the roof of RF shield to cover top of the klystron.
Punching plate set on the walls.
In case of 2000 MHz
Cutoff frequency check for access port to set the cooling water pipes and the cables.
RF absorber set in the rf shield
Access port

16. Future task: how to design the KPS to optimize the CPD klystrons
Where does the recovery power returned ? (to AC line ?, in KPS ?)
Transformation installation
Collector loss emits to the air
Recovery power
Incoming panel
(three-phase, AC 6.6kV)
Energy recovery system
Go to ???
IVR(induction voltage regulator)
To AC line with DC-AC convertor ?
To the DC line ?
To the DC line with the marx-circuit ?
Must be consider it.
transformer-rectifier assembly
(12 phase full-wave rectifier)
図の構成を用いた実証実験においては、回収電力を別のクライストロンの入力(DC高電圧)と出来る(別のクライストロンをダミーロードとする)。
数十kW以下の場合、水冷の抵抗負荷も準備している。実用の場合には、6.6kV交流へ変換し既存電源入力とすることも想定している（未検討）。
幸い、別の通常のクライストロンをダミーロードとして使用することが、比較的容易に設定できる状態である。
：別の電源システム、クライストロンが近くにある。
実証試験では、できる限り既存システムと回収電力系を切り離すことを考える。
：電圧変動を経由してバンチングの恐れもあるか？
Crowbar-circuit
Smoothing capacitor
DC power supply to the klystron
Cathode and Heater-Anode power supply
KPS for CW klystron at KEKB

17. Summary
* Two examples (the stronger bunching and the energy recovery from collector loss) show to obtain the high-efficiency of the klystrons.
In our study,
* A CPD klystron was fabricated at 2013, to recycle an existing klystron of Toshiba E3786.
(It was used to TRISTAN and KEKB.)
Target is a proof-of-principle of CPD to apply CW klystron in the unsaturated region.
* Now, we are doing the design of RF and radiation shield to optimize the CPD klystron.
* The test to demonstrate the proof-of-principle of the CPD method for klystron is planning to after Phase2 of SuperKEKB.
Note:
This CPD klystron will not be installed in SuperKEKB in the future.
Commercial viability of it is required development of KPS (Klystron Power Supply) to optimize it. At present, this is a future task.