High Power RF Measurements

Ben Woolley,

Amos Dexter, Igor Syratchev.Jan Kovermann, Joseph Tagg

HG2013, Trieste

June 2013

Outline

Phase Stabilisation of CLIC crab cavities• RF distribution to the cavities.• Accurate phase measurements and stabilisation system.

LLRF for future/current X-Band test stands• Production of vector modulated signals for type II SLED pulse

compressor.• Phase and power measurements of RF signals.• Working example using TWT and SLED II pulse compressor.• Easily transferable to other test stands (e.g. TERA C-band test)

CLIC Synchronisation RequirementCavity to Cavity Phase synchronisation requirement (excluding bunch attraction)

degrees1S

1

c

f7204rmsc

x

Target max. luminosity loss fraction S

f (GHz)

sx (nm) qc (rads)

frms (deg) Dt (fs) Pulse Length (ms)

0.98 12.0 45 0.020 0.0188 4.4 0.156

Estimate RF to beam synchronisation ~ 100 fs (0.43 degrees)

Integration

1)Use over-moded waveguide from klystron to the tee

1)35m of waveguide from the Tee to the cavities,

Scale: 20.3m from cavities to IP

2)Two klystrons with low-level/optical signal distribution.

RF Distribution

Option 1: A single klystron with high level RF distribution to the two cavities.

• Klystron phase jitter gets sent to both cavities for identical path length. Δφ=0.

• Will require RF path lengths to be stabilised to within 1 micron over 40m.

Option 2: A klystron for each cavity synchronised using LLRF/optical distribution.

• Femtosecond level stabilized optical distribution systems have been demonstrated (XFELs).

• Requires klystron output with integrated phase jitter <4.4 fs.

Expected Klystron Stability

• From kinematic model of klystron, phase and amplitude stability depend on gun voltage, V:

• For the Scandinova modulator with measured voltage stability of 10-4 phase and amplitude stability should be 0.12° and 0.013%.

• Active feed-forward or feedback would be needed to gain the required stability of 0.02°.

Waveguide Choice

7

Waveguide type35 meters COPPER

Expansion = 17 ppm/K

Mode Transmission Timing error/0.3°C

Width

Timing error/0.3°C

length

No of modes

WR90(22.86x10.16mm) TE10 45.4% 210.5 fs 498.9 fs 1

Large Rectangular (25x14.5mm)

TE10 57.9% 189.3 fs 507.8 fs 2

Cylindrical r =18mm TE01 66.9% 804.9 fs 315.9 fs 7Cylindrical r =25mm TE01 90.4% 279.6 fs 471.4 fs 17

Copper coated extra pure INVAR 35 meters

Expansion = 0.65 ppm/K

Mode Transmission Timing error/0.3°C

Width

Timing error/0.3°C

length

No of modes

WR90(22.86x10.16mm) TE10 45.4% 8.13 fs 19.04 fs 1

Large Rectangular (25x14.5mm)

TE10 57.9% 6.57 fs 19.69 fs 2

Cylindrical r =18mm TE01 66.9% 30.8 fs 12.1 fs 7Cylindrical r =25mm TE01 90.4% 10.7 fs 18.02 fs 17

Rectangular invar is the best choice as it offers much better temperature stability-> Expands 2.3 microns for 35 m of waveguide per 0.1 °C.

RF path length measurement

48MW 200ns pulsed 11.994 GHz Klystron repetition 50Hz

Vector modulation

Control

Phase Shifter

12 GHz Oscillator

Main beam outward pick up

Main beam outward pick up

From oscillator

Phase shifter trombone

(High power joint has been tested at SLAC)

Magic Tee

Waveguide path length phase and amplitude measurement and control

4kW 5ms pulsed 11.8 GHz Klystron/TWT

repetition 5kHz

LLRF

Phase shifter trombone

LLRF

Cavity coupler 0dB or -40dB

Cavity coupler 0dB or -40dB

Expansion joint

Single moded copper plated Invar waveguide losses over 35m ~ 3dB -30 dB

coupler -30 dB coupler

Forward power

main pulse

12 MW

Reflected power main pulse ~ 600 W

Reflected power main pulse ~ 500 W

Waveguide from high power Klystron to magic tee can be

over moded

Expansion joint

RF path length is continuously measured and adjusted

LLRF Hardware Layout (Low BW)• Fast phase measurements during the pulse (20-30 ns).

• Full scale linear phase measurements to centre mixers and for calibration.

• High accuracy differential phase measurements of RF path length difference (5 μs, 5 kHz).

• DSP control of phase shifters. Linear Phase Detector (1MHz BW)

DSP

ADCADC

MagicTee

To Cavity

To Cavity

Wilkinson splitters

-30 dB coupler

DBM

DBM DBM

10.7GHz Oscillator

-30 dB coupler

Piezoelectric Phase Shifter Piezoelectric phase shifter

DAC1

Amp + LPF

LPFAmp

Power Meter

To DSP DAC2

From DAC2

To Phase Shifter

Calibration Stage

Board Development and CW testsFront end electronics to enable phase to be measure during the short pulses to an accuracy of 2 milli-degrees has been prototyped and dedicated boards are being developed.

PLL controller

MCU

10.7 GHz VCO

Digital phase detector

DBMs

Power Meters

Wilkinson splitter

Inputs

400 ns span:RMS: 1.8 mdeg

Pk-Pk: 8.5 mdeg

90 s span:Drift rate : 8.7 mdeg/10s

Total drift: 80 mdeg

Crab cavity stabilisation: Next steps

LLRF board revision:• Some problems with the PLL and tolerances on the

Wilkinson splitters. (Non equal power splitting).• Investigation into pulsed power operation, effect of

amplitude instabilities and higher order mixing products (not visible in CW tests).

RA joining Lancaster in September to increase/broaden the effort on this project.

Future LLRF Generation and Acquisition for X-band test stands

IF2.4 GHz Oscillator

9.6GHz BPF

Amp

Vector Modulator

RFout

LOin LOout

12GHz vector modulated

signal to DUT

2.9 GHz Oscillator

X4 freq.

RF

LO

LOIF RF

12 GHz BPF

12 GHz BPF

12GHz CW reference

signal

2.4GHz vector modulated

signal

2.4GHz CW reference signal

11.6 GHz BPF

X4 freq.

LO

IF

RF Input 1

400 MHzLPFs

LO

IF

RF Input 2

LO

IF

RF Input 3

LO

IF

RF_Referance12GHz CW reference

signal

3dB hybrid

AmpIF Amps

Oscillators should be phase locked

1.6 GSPS

12-bit ADCs

Digital IQ demodulation

LOADTWT

System Testing: SLED II PC

Phase modulated pulse input

-40 dB coupler

-40 dB coupler

SLED II pulse compressor

RF Input 1RF Input 2 RF Input 3

20 dB attenuator

20 dB attenuator

System Testing: SLED II PCNational Instruments PXI crate containing:

• 2 CW generators for the LOs.

• Vector modulator (up to 6.6GHz) with 200 MSPS I/Q generator

• 5Chs 1.6GSPS 12-bit and 4Chs 250MSPS 14-bit ADC each connected to FPGAs.

• 200 MHz digital I/O board for interlock and triggering signals.

Up/down-mixing components and cabling

TWT: 3kW

10-12GHz

SLED II Pulse compressor

Igor Syratchev

RF Load

Power Meter

LLRF System Test: SLED II PCPhasePower400MHz raw signals

Trans.

Refl.

Inc.

LLRF System Test: Dynamic RangePhasePower400MHz raw signals

40 dB below TWT saturation!

Pulse compressor DetuningPhasePower400MHz raw signals

LLRF system Accuracy IPower accuracy: 0.4%rms Phase accuracy: 0.9°rms

Possible sources of error• Problem with locking between the two CW generators

and the 10MHz system reference caused the signal to beat. This affects acquisition AND vector generation. Proposed solution: Clock ADC’s with 400MHz reference.

• Bit-noise on the ADCs limit accuracy to 0.1% (9.5ENOB).• All harmonics generated by the multipliers and/or mixers

will be mixed down to 400MHz and be indistinguishable from the signal of interest.

400MHz Reference

LLRF system Accuracy II

Comparison with calibrated power meter• Transients at the start of the main pulse

and phase flip are observed (input).• Transients reduced when passed through

the detuned system. Hybrid is narrow band extra filtering of harmonics gives cleaner signal when mixing down.

Input

Output

Calibrated Power Meter

Output:Gain 2.9

Input

TERA C-band Cavity Testing

5.7GHz 4MWMagnetron

Circulator

Directional coupler

Cavity under test

PMT and faraday cup

PXI crate

RF detector diodes, PMT and faraday cup inputs.

Timing board/ trigger generation

Pictures: Alberto Degiovanni

Screenshots from ‘CBOX1’

X-band LLRF: Future Developments

• Characterisation of errors/transients in the system. Both on the generation and acquisition side.

• Miniaturisation of LLRF system components into crate.

• Software: logging, interlock control, timing and triggers etc.

• Integration into XBOX-2 test stand.• Scaling up and integration towards XBOX-3.

22

Thank you for your attention!

Extra Slides

Observed Klystron Stability

• Observed ~5% amplitude jitter on the output of the klystron. This was due to a mismatch in the pulse forming diode in the LLRF network and a triggering error in the ADC firmware.

• Amplitude jitter reduced to ~1-2%.

• Phase measurements will be performed in the coming weeks.

Klystron phase noise• Can use a standard method to measure the phase stability of the klystron.

• Reference source is split such that its phase noise is correlated out by the mixer for both channels.

• Phase shifter adjusted as to bring RF and LO inputs into quadrature.

• Digital scope or ADC and FPGA/DSP preforms FFT analysis, to obtain phase noise curve.

• Experiment can also be repeated for different lengths of waveguide to ascertain the effects of waveguide dispersion.

To Cavity

Wilkinson splitter DBM

12 GHz Oscillator

-57 dB coupler

phase shifter

KlystronTWT

Φ

PIN diodeswitch

Oscilloscope/ADC for FFT analysis

RFLO

IF

Waveguide

Measurement requires good amplitude stability as any AM will be present in the IF.

Waveguide Stability ModelUse ANSYS to find “dangerous” modes of vibration for a 1 m length of waveguide fixed at both ends.Fundamental mode 65.4 Hz

Planned CLIC crab high power tests

Test 1: Middle Cell Testing – Low field coupler, symmetrical cells. Develop UK manufacturing.

Test 2: Coupler and cavity test – Final coupler design, polarised cells, no dampers.Made with CERN to use proven techniques.

Test 3: Damped Cell Testing – Full system prototype

Travelling wave 11.9942 GHz

phase advance 2p/3

TM110h mode

Input power ~ 14 MW

Future

• Continued investigations into the phase stability of the 50 MW X-band klystron.

• Development of feed-forward and/or feedback system to stabilise the klystron’s output.

• Continued characterisation of electronics to obtain stand alone phase measurement/correction system.

• Design/procurement of the waveguide components needed.• Demonstration RF distribution system, with phase stability

measurements. • Perform phase stability measurements during the CTF dog-leg

experiments.

• Measure phase across the prototype cavity during a high power test.

29

LLRF Hardware Layout (High BW)• Fast phase measurements during the pulse (50MHz).

• 400MHz direct sampling to centre mixers and for calibration.

• High accuracy differential phase measurements of RF path length difference (5 μs, 5 kHz).

• DSP control of phase shifters.

1.6 GSPS 12-bit ADC’s

DSP

ADC

MagicTee

To Cavity

To Cavity

Wilkinson splitters

-30 dB coupler

DBM

DBM DBM

11.6 GHz Oscillator

-30 dB coupler

Piezoelectric Phase Shifter Piezoelectric phase shifter

DAC1

Amp + LPF

400MHzLPF

Amp

Power Meter

To DSP DAC2

From DAC2

To Phase Shifter

Calibration Stage