Stefan Wilke, DESY MHF-e 1Dortmund, 2006-09-27

News from RF at PETRA

•• PETRA PETRA todaytoday and in and in futurefuture•• changeschanges in in powerpower distributiondistribution of of waveguideswaveguides•• waveguidewaveguide shuttershutter, , phaseshifterphaseshifter•• Klystron Working Point ControllerKlystron Working Point Controller•• 8 longitudinal 8 longitudinal feedbackfeedback kickercavitieskickercavities•• ELWIS (2nd ELWIS (2nd radiationtestradiationtest))•• klystronprotectionklystronprotection ((crowbarcrowbar, , pearsonpearson probe)probe)•• conditioningconditioning of of inputinput powerpower couplerscouplers (>250 kW)(>250 kW)

Stefan Wilke, DESY MHF-e 2Dortmund, 2006-09-27

PETRA today

positrons/electronspositrons/electrons for HERA:for HERA:beam energy: Ebeam energy: E00 = 7= 7--12 GeV12 GeVbeam current: Ibeam current: I00 = 45 = 45 mAmA

protonsprotons for HERA:for HERA:beam energy: Ebeam energy: E00 = 7.5= 7.5--40 GeV40 GeV

beam current: Ibeam current: I00 = 120 = 120 mAmAsynchrotronlightsynchrotronlight for HASYLAB:for HASYLAB:beam energy: Ebeam energy: E00 = 11.5 GeV= 11.5 GeVbeam current: Ibeam current: I00 = 50 = 50 mAmA

positrons/electrons in HERA:positrons/electrons in HERA:beam energy: Ebeam energy: E00 = 12= 12--27 GeV27 GeV

beam current: Ibeam current: I00 = 40 = 40 mAmAprotons in HERA:protons in HERA:

beam energy: Ebeam energy: E00 = 40= 40--920 GeV920 GeVbeam current: Ibeam current: I00 = 105 = 105 mAmA

Stefan Wilke, DESY MHF-e 3Dortmund, 2006-09-27

and the future...an impression of the new hall

Beam Energy: Beam Energy: EE0 0 = 6 GeV= 6 GeVLength: Length: l = 2304 ml = 2304 mArc radiusArc radius 191.73 m and 22.918 m191.73 m and 22.918 mBeam Current: Beam Current: II00 = 100 = 100 mAmA (200 (200 mAmA))loss per turnloss per turn UUll = 7.590 = 7.590 MeVMeVEmittanceEmittance ((horhor)) ε = 1 ε = 1 nmradnmrad !!Topping upTopping up

FrequencyFrequency:: ffRFRF = 499.66 MHz= 499.66 MHzBeamBeam Power Power (100mA)(100mA):: PPbeambeam = 759 kW = 759 kW ((dipoldipol, , undulatorsundulators, , dampingdamping wigglerswigglers and HOM and HOM losseslosses))CircumferencialCircumferencial Voltage:UVoltage:Ucc = 20 MV = 20 MV (in 12 7(in 12 7--cell cell cavitiescavities, , powerpower per per couplercoupler: 124 kW): 124 kW)rfrf--PowerPower (100mA)(100mA):: PPrfrf = 1573 kW = 1573 kW (2 (2 transmittertransmitter á 786 kW)á 786 kW)

1 meter!

accelerator-tunnel

beamlines

10 m10 m

Temp. regulation in tunnel: 0.1 deg C! In hall: 1 deg C.Temp. regulation in tunnel: 0.1 deg C! In hall: 1 deg C.

Stefan Wilke, DESY MHF-e 4Dortmund, 2006-09-27

current vs future WG-distribution

TX 4 WG parallel

TX

PETRA-II2 TX on 16 cavities

PETRA-III2 TX on 12 cavities

dismounting of 2 cavities on each side,magic Ts exchanged by hybrids

to the other TX

to the other TX

Stefan Wilke, DESY MHF-e 5Dortmund, 2006-09-27

top view in tunnel

Stefan Wilke, DESY MHF-e 6Dortmund, 2006-09-27

rf-system at PETRA III

Waveguide Power Distribution

7-cell cavities

MT

Waveguide Power Distribution

WG-ShutterWG-Shutter

100 kW Load

100 kW Load

400 kW Load

400 kW Load

800 kW Klystron

800 kW Klystron

4-Port

Circulator

Phase-Shifter

Phase-Shifter

7-cell cavities

Interconnection Line

MT

.

WG-ShutterWG-Shutter

100 kW Load

100 kW Load

400 kW Load

400 kW Load

800 kW Klystron

800 kW Klystron

4-Port

Circulator

Phase-Shifter

Phase-Shifter

Transmitter Hall

PETRA-SR

Transmitter Hall

PETRA-SL

PETRA Tunnel

3 dB Hybr. Cplr.

3 dB Hybr. Cplr.

3 dB Hybr. Cplr.

3 dB Hybr. Cplr.

Normally each transmitter drives 'its'

own 6 cavities.

Option to run with only one transmitter (1440 kW) on all cavities,but reduced beamcurrent:

ca. 83 mA

Stefan Wilke, DESY MHF-e 7Dortmund, 2006-09-27

klystron 2

circulator

stairway to tunnel

spare klystronelectronic

klystron 1

switching area

waveguide-power distribution

waveguides to cavities SR and to transmitter SL

building

exten

sion:

klystr

onpow

ersup

plys,

klystr

onmo

dulat

ors,

crowb

arand a

dditio

nal

HV-in

stallatio

ns

Senderhalle SR (Geb.42b)one transmitter hall (SR)

Stefan Wilke, DESY MHF-e 8Dortmund, 2006-09-27

waveguide shutter

Stefan Wilke, DESY MHF-e 9Dortmund, 2006-09-27

Klystron Working Point Controland fast RF-Loop

T=3s

Cath.-PS

T=10ms

Anode-PS

500-MHz Reference KlystronsDriver

CavitiesRF

I/QModulator

I/Q

Demodulator

Iset Qset

Setpoint

Circulator

φ, dBI Q

RF-Att

PID PID

Pout

Pin Pdrv

Set-Vc

Set-Va

Set-Pdrv

KlystronWorking Point

Controller

(PXI-Crate)

Philips YK-1304Ausgangsleistung vs Treiberleistung; Parameter Kathodenspannung

(Kathodenstrom konstant, Ik=Ikn)

0,0

0,1

0,2

0,3

0,4

0,5

0,6

0,7

0,8

0,9

1,0

0,0 0,2 0,4 0,6 0,8 1,0 1,2 1,4 1,6 1,8 2,0

D

P

Ukn, Ikn0,9Ukn, Ikn0,8Ukn, Ikn0,7Ukn, Ikn0,6Ukn, Ikn0,5Ukn, Ikn

outputpower vs. driverpower at different kathodevoltages. kathodecurrent constant

In order to run the klystron over the full output power range close to saturation and at highest efficiency a Klystron Working Point Controller (KWPC) is under development.

The optimum values of klystron voltage, current and drive for any klystron output power are set by the KWPC.

Stefan Wilke, DESY MHF-e 10Dortmund, 2006-09-27

PETRA III longitudinal feedback

Commercially available amplifiers( 8 * 500 W solid state)

&

8 „Daphne“ kicker, tuneableby nose cones (10 MHz/mm)running at 1375 MHz

To achieve the design current (100 mA) in PETRA III powerfulfeedback systems are necessary since the threshold currentsfor coupled bunch instabilities are around 7 mA!Required bandwidth ≥ 62.5 MHz (8 ns bunch distance) !!

A longitudinal feedback kicker is a special cavity used as feedback element for longitudinal multibunch feedback. This picture shows the 1:1 measuring model of the longitudinal feedback kickers for PETRA III. The measuring model serves to verify the measuring technology.

Stefan Wilke, DESY MHF-e 11Dortmund, 2006-09-27

cavity drawings

Stefan Wilke, DESY MHF-e 12Dortmund, 2006-09-27

bandwidthkicker-design for PETRA-3.1 (2 ns bunch distance, B = 250 MHz)reduction of bandwidth for PETRA-3.0 (8 ns bunch distance, B = 125 MHz).

SSBDSB

Stefan Wilke, DESY MHF-e 13Dortmund, 2006-09-27

4 feedback-kicker on girder

<0,05°

100±4 336±4

2 Lambda 1375MHz

436

<0,1

1644

draft

Stefan Wilke, DESY MHF-e 14Dortmund, 2006-09-27

stressing ELWIS

ELWIS : Eier Legende WollmIlch Sau = Eggs laying wool milk sow (pig)

kitkit: PXI : PXI cratecrate, , controllercontroller (PC), (PC), signalsignalconditioningconditioning, FPGA, fast ADC incl. TRC, , FPGA, fast ADC incl. TRC, triggerdistributiontriggerdistribution

TLDTLD

FPGAFPGA

fast ADCfast ADCPCPC ionisationionisationchamberchamber

Stefan Wilke, DESY MHF-e 15Dortmund, 2006-09-27

2nd ELWIS radiation test in DORIS2nd ELWIS radiation test in DORIS

48,6 Gy

damage of FPGA

Stefan Wilke, DESY MHF-e 16Dortmund, 2006-09-27

klystron protection

-75 kV / 36A (nom.)

from outdoorklystron ps

Light (laser) TriggeredThyristor Crowbar

2 independent logics gets theprobe signals and decides

wether the crowbar is fired

M5 checksthe trigger

Pearsoncurrentmonitor

Sensitivity: 10 mV/A

Stefan Wilke, DESY MHF-e 17Dortmund, 2006-09-27

conditioningconditioning of of inputinput powerpower couplerscouplers

Stefan Wilke, DESY MHF-e 18Dortmund, 2006-09-27

test transmitter

Light detector

Cavity withtwo couplers

DeviceUnder Test

Vacuumpump

Stefan Wilke, DESY MHF-e 19Dortmund, 2006-09-27

test cavity with light detector

Stefan Wilke, DESY MHF-e 20Dortmund, 2006-09-27

reaching 250 kW

Vacu

um/

mba

rrf

powe

r/

kW

Parameter: start and end power, power step size, slope, waitingtime

Stefan Wilke, DESY MHF-e 21Dortmund, 2006-09-27

interlockinterlock and and experimentsexperiments•• TheThe light light detectordetector interlockinterlock ((reactionreaction time 2 ms) time 2 ms) isis veryvery

successfulsuccessful. . ThanksThanks to to thisthis interlockinterlock oneone couplercoupler survivedsurvived moremorethanthan 200 200 sparkssparks. . TheThe reactionreaction time has to time has to bebe optimizedoptimized in in thethenextnext teststests..

•• An also An also testedtested interlockinterlock withwith ee-- detectorsdetectors didn‘tdidn‘t workwork so so reliablereliable..•• An An interlockinterlock basedbased onlyonly on on thethe vacuumvacuum pressurepressure isis tootoo slowslow!!•• TheThe slopeslope of of thethe powerpower rampramp of of eacheach powerpower stepstep has to has to bebe

< 60 kW/min. 15 kW/min < 60 kW/min. 15 kW/min seemsseems acceptableacceptable. . ButBut also also thisthis shouldshouldbebe optimizedoptimized..

•• A A longerlonger waitingwaiting time at time at eacheach powersteppowerstep seemsseems to to bebe necessarynecessary. . WeWe havehave to to showshow wetherwether 30 30 minutesminutes areare thethe optimumoptimum..

•• Best Best possiblepossible matchingmatching betweenbetween couplercoupler and test and test cavitycavity and and additional additional frequencyfrequency modulationmodulation will will produceproduce muchmuch betterbetter resultsresultsin in conditioningconditioning..

•• RecentRecent resultsresults: to : to roughenroughen onlyonly thethe inner inner conductorconductor of of thethecouplercoupler byby sandblastingsandblasting reachedreached >250 kW >250 kW withinwithin oneone dayday

Stefan Wilke, DESY MHF-e 22Dortmund, 2006-09-27

the end and thank you!MHF-e at DESY: 25 colleagues