CERN SPS Upgrade New 200 MHz and 800 MHz amplifiers

CERN SPS Upgrade

New 200 MHz and 800 MHz amplifiers

Eric MontesinosCERN-RF

15th ESLS-RF Workshop 2

CERN Accelerator Complex

Thursday, 6th October 2011

Protons and Lead Ions to maximum acceleration LINAC2 (proton) or Linac 3 (Lead ions) Booster (protons) or Leir (lead ions) PS (Proton Synchrotron) SPS (Super Proton Synchrotron) LHC (Large Hadron Collider)

Several other experiences : n_TOF – The neutron time-of-flight facility;

a neutron source that has been operating at CERN since 2001

AD – The Antiproton Decelerator; manufacturing antimatter providing low-energy antiprotons for studies of antimatter

ISOLDE – Isotope Separator On-Line; source of low-energy beams of radioactive isotopes

CLIC – the Compact Linear Collider Study; an international collaboration working on a concept for a post LHC machine to collide electrons and positrons head on at energies up to several TeV


3

CERN SPS – the Super Proton Synchrotron

Thursday, 6th October 2011 15th ESLS-RF Workshop

The second largest machine in CERN’s accelerator complex, nearly 7 km in circumference. It was switched on in 1976 (CERN Nobel-prize for discovery of W and Z particles in 1983)

Presently, SPS accelerates particles to provide beams for the: FT (Fixed Target) program (North Area) CNGS project LHC (Large Hadron Collider) And Many Machine Developments

North Experimental area

CNGS : CERN Neutrions to Gran Sasso

SPS as LHC injector


4

200 MHz RF in the SPS


The RF-SPS started up in 1976 with two accelerating cavities

Since 1980, for the new role of SPS as proton-antiproton collider, there are four cavities operating @ 200 MHz

We have 4 lines : 2 x Siemens: 20 x RS2004 2 x Philips: 68 x YL1530

201020001976TWC#1 / TX1 TWC#2 / TX2

19901980

1978TWC#3 / TX3

1979TWC#4 / TX4

1980TWC#1 / TX1+TX2TWC#2 / TX3+TX4TWC#3 / TX5+TX6TWC#4 / TX7+TX8

Transmitter (TXB)

mW Dummy load

Coaxial transmission line(feeder line)125 to 160 meters

Accelerating cavity

Terminating loads

Configuration of one of the four200 MHz power plant

Transmitter (TXA)

Powercombiner

1975


5

Two ‘Siemens’ lines = 20 x RS2004


1WSolid state

100WSolid state

1kWYL1440 tube

10kWYL1520 tube

100kWRS2004 tubeTXA

4 x 125kWRS2004 tubes

One line (input cavity ~125/140 m away)

1WSolid state

100WSolid state

1kWYL1440 tube

10kWYL1520 tube

100kWRS2004 tube

Ø G

From Beam Control

TXB4 x 125kW

RS2004 tubes

1WSolid state

100WSolid state

1kWYL1440 tube

10kWYL1520 tube

100kWRS2004 tubeTXA

4 x 125kWRS2004 tubes


1WSolid state

100WSolid state

1kWYL1440 tube

10kWYL1520 tube

100kWRS2004 tube

Ø G

From Beam Control

TXB4 x 125kW

RS2004 tubes


6

Two ‘Philips’ lines = 68 x YL1530



1WSolid state

100WSolid state

1kWYL1440 tube

35kWYL1530 tube

Ø G

From Beam Control

1WSolid state100WSolid state

1kWYL1440 tube

TXA16 x 35kW

YL1530 tubes

TXB16 x 35kW

YL1530 tubes

35kWYL1530 tube


1WSolid state

100WSolid state

1kWYL1440 tube

35kWYL1530 tube

Ø G

From Beam Control

1WSolid state100WSolid state

1kWYL1440 tube

TXA16 x 35kW

YL1530 tubes

TXB16 x 35kW

YL1530 tubes

35kWYL1530 tube


7

Travelling Wave Cavities


One section = 11 drift tubes

2 x 4 sections Siemens plants 2 x 5 sections Philips plants

4 Main Power Couplers 2 input couplers 2 output couplers

2 x 550 kW terminating power loads

One 4 sections cavity

One section: 11 drift tubes


8

200 MHz limitations


With present 4 cavities configuration we will have problems at ultimate LHC current

The increased number of shorter cavities with 2 extra power plants should significantly improve the RF performance for ultimate LHC intensities

The best compromise is 6 cavities: 4 x 3 sections cavities with 1.0 MW 2 x 4 sections cavities with 1.4 MW

Total voltage possible on the flat top vs beam current with :

4 cavities (present situation) with 1.0 MW 5 cavities with 1.0 MW RF 6 cavities with 4 x 1.0 MW + 2 x 1.4 MW Dashed lines are at nominal and ultimate

beam currents.

Courtesy of Elena

Shaposhnikova


9

Cavities redistribution


2011 : 4 cavities

2 x 4 sections

2 x 5 sections

+ 3 spare sections

2018 : 6 cavities

2 x 4 sections

4 x 3 sections

+ 1 spare section


10

First upgrade: Present amplifiers


Ratings Present Future

CW5 seconds 650 kW 700kW

Pulsed43 kHz 900 kW 1100 kW

BW-3dB 2.6 MHz 2.3 MHz

Tubes per year 6 + 16 7 + 18

HVPS need a full re-cabling and an air cooling improvement to allow higher pulsed mode

Tetrodes:Present lifetime statistics, operating ~650 kW cw:

RS2004 : 20’000 hrs: 6 tubes per year

YL1530 : 25’000 hrs: 16 tubes per

year

HVPS need a full re-cabling and air cooling improvement to allow higher pulsed mode


11

New RF power plant


New RF Amplifier

LSS3 Tunnel integration

New RF Building

RF amplifier

1 mW

Coaxial transmission line150 meters

Accelerating cavity

1.7 MW

RF amplifier

1 mW

Accelerating cavity

1.7 MW

New RF Building


12

2018 : two new power amplifiers


Must be reliable: 24/24 hours 300/365 days

(2 months winter Technical stop) 20 years of operation,

with 3 years of operation + 1 year off cycle

Pulse mode: 1.7 MW max

Average: 850 kW (thermal limitation)

2 x 1.7 MW Klystron

2 x 4 x 450 kW Diacrodes

2 x 8 x 225 kW IOTs

2 x 8 x 225 kW tetrodes Equivalent to ‘Siemens’

2 x 16 x 110 kW tetrodes Equivalent to ‘Philips’

2 x 1700 x 1 kW SSA


13

1.7 MW amplifier, i.e 1.4 MW cavity


To have 1.4 MW available at the cavity input, 1.7 MW at the Final output are needed

Taking advantage of the long experience we have with tetrodes and combiners, a possible solution could be a 16 x tetrodes combined through 3 dB combiners

A major improvement to present systems would be to have individual SSA drivers per tetrodes

Four contracts : Drivers (SSA) Finals (SSA or Tetrodes) Combiners (3 dB above 100 kW) Transmission lines (coaxial, 345 mm

outer)

Drivers16 SSAFinal16 Tubesor SSA

3 dBcombinersand power loads

120 m and 180 mCoaxial lines

To cavity input 120 m away

From Beam Control

1/16 splitter

1.7 MW

-0.6 dB total

1.5 MW

-0.2 dB

1.4 MW


15th ESLS-RF Workshop 14

SSA vs Tetrodes

Thursday, 6th October 2011

Overdesign requirements : 14/16 tubes shall provide full power,

i.e. each tube shall deliver up to 138 kW

SSA are more ‘reliable’: 2000/2048 of the total number of devices shall deliver full power

Tetrodes tube costs over 20 years will be added : 20 year * 3/4 * 335 * 24 = 120’000

total hours With 20’000 hours per tube = ~ 200

tubes Reduced by warranty lifetime

SSA obsolescence shall be integrated: i.e. 20% additional transistors, not

module, single chips (still under discussion, need experts inputs)

Wall plug efficiency will be part of the adjudication HVPS included (Tetrodes) Losses in all SSA combiners,

circulators and loads included

Final Tetrodes(gain = 12

dB)

SSA(Gain = 20

dB)

Nominal ratings

16 x 106 kW = 1700 kW

2048 x 830 W = 1700 kW

Maximum ratingsFor 1400 kW at cavity input

Maximum2 faulty tubes14 x 138 kW = 1932 kW

Maximum48 faulty modules2000 x 891 W = 1782 kW

Maximum ratingsDriver

16 x 8.7 kW 16 x 1.1 kW


15

New RF building


Only possible location is between two existing buildings

Maximum ‘RF’ foot floor will be

2 x 450 m2 Whatever the solution, SSA or

Tetrodes, the same building, no impact on the choice

RF workshop

Siemens

Philips

800 MHz

Faraday Cage


16

Draft schedule


Install

Build new hardware

Year 4Year 3Year 2Year 1 Year 5

Building

Services

RF :

Building:

20172014201320122011 2015 2016 2018Year 6 Year 7

Tendering

Commissioning

Authorizations

Tunnel :

Build New hardwareInstallation phase 1

(pickups + dampers + CV + EL + …)

Installation phase 2(cavities)

Cavities re-arrangement within a LS ( > 6

months)

Studies (amplifiers, couplers, cavities)

MS

Studies

Studies


17

200 MHZ upgrade Conclusions


We will have two new 1.7 MW pulsed / 850 kW average RF power amplifiers

Building will be the same, no impact

The less expensive solution beetween SSA and Tetrode will be selected !


18

800 MHz RF in the SPS


The proton beams for the LHC are intense and unless careful precautions are taken they become unstable in the SPS and cannot be accepted by the LHC

One of the most important systems in the SPS used to keep the beams stable and of the highest quality is the 800 MHz RF system acting at the second harmonic of the main accelerating 200 MHz RF system

This 800 MHz system in the SPS is essential for maintaining stability of the LHC beams. It is required at every point in the cycle from injection to extraction. It works by increasing the synchrotron frequency spread in the beam

Stability is problematic above 1/5 nominal without the 800 MHz

By applying RF voltages of ~ 1 MV (about 1/7 of the main RF system) via two cavities in the SPS ring this “Landau Damping” system increases the natural spread of synchrotron frequencies in the individual proton bunches, prevents them acting together, and thus ensures stability

The RF power source and its ancillary equipment for this 800 MHz system must be of the highest reliability to ensure beams are available for the LHC at all times


19

Old 800 MHz system (1)


Since 1980, the system is composed of :

2 Travelling Wave Cavities

2 transmitters of 225 kW each connected via ~ 120m waveguides to the TWC

4 x 56 kW klystrons Valvo YK1198 per transmitter combined using 3 dB hybrids


20

Old 800 MHz system (2)


Unfortunately, that system has not been used for a very long time and has not been properly maintained.

We still only have : 2 simultaneous klystrons available

on 1 cavity 6 operational klystrons 10 broken klystrons (could be

repaired for 100’000 $ each)

We also had major difficulties with power converter transformers : 9/9 burnt 4 repaired


21

Upgrade proposal


Replace Klystron Transmitters with IOT Transmitters and re-use all existing peripherals

Maximum power will be slightly increased up to 240 kW CW

BW-1dB will be increased: 1.0 MHz with Klystrons 6.0 MHz with IOTs

New transmitters will include: RF power amplifiers chain Final amplifier IOT based Individual power converters

Individual Monitoring and control compatible with CERN interface

In total it will be 8 + 1 transmitters

One 800 MHz LineLayout

One 240 kW TransmitterLayout Attenuator

RF Power Amplifier 60 kW cw




Sp

litte

r

Monitoringand

Controls

Amplifiers

Power converters

Monitoringand

Controls

Amplifiers

Power converters

Monitoringand

Controls

Amplifiers

Power converters

Monitoringand

Controls

Amplifiers

Power converters

3d

B C

om

bin

er

3d

B C

om

bin

er

3d

B C

om

bin

er

Power load

Power loadCavity Terminating load

Cavity and TransmitterMonitoring and Controls

(CERN)Power load

Ø shifter

AttenuatorØ shifter



waveguide line(125 meters )


22

Selected supplier : Electrosys


Two companies per member state have been contacted (40 companies)

Six companies have been compliant to our specifications

Electrosys has been selected

‘Quasi’ off the shelves Transmitter

Possibility to have Thales or e2v trolleys and tubes for the same price


23

Factory Acceptance Tests:various operational modes


Continuous operation 24/24 hours CW for 10 months continuously

24 hours made prior to our visit. 4 hours made with us

Very Long Pulses operationFc = 800.888 MHz +/- 0.5 MHz :100% from 0 to 240kW with rise and fall time < 0.5 µs5 seconds ON / 5 seconds OFF

One hour made with us

AM modulation #1Fc = 800.888 MHz +/- 0.5 MHz :100% from 0 to 240kW with rise and fall time < 0.5 µsRepetition time 10 µs (100kHz)


AM modulation #2Fc = 800.888 MHz +/- 0.5 MHz :0 to 240kW with 4 MHz triangle AM 25 % in powerRise and fall time < 0.5 µsFlat top pulse and off pulse length of 11 usRepetition time 23 µs (43kHz)This cycle for 20 second then 1 second OFF.


time

Power

5 s

0.5 us 0.5 us

240 kW

0 kW

10 s

time

Power

10 us

0.5 us0.5 us

240 kW

0 kW

time

Power

11 us

0.5 us 0.5 us

240 kW

0 kW

11 us

23 us

20 s 1 s

4 MHz triangle

25 %


24

Factory Acceptance Tests:Bandwidth


Power Amplifier (Driver + Final)

Operating frequency :

800.888 MHz

Bandwidth at -1dB :

6.0 MHz (+/- 3.0 MHz)

CW output minimum power :

60 kW

Amplifier meets all requirements 790

796.

5

797.

25 798

798.

75

799.

5

800.

25

800.

888

801.

5

802.

25 803

803.

75

804.

5

805.

25 806

0

10

20

30

40

50

60

70

Pout vs frequency

Frequency [MHz]

Pout

[kW

]

Pmax = 61.0 kW

Pmax -1dB = 47.5 kW

BW-1dB = 7.0 MHz (-3/+4 MHz)


25

Factory acceptance tests:Phase stability


Carrier f0 at mid power (30 kW)with additional - 20 dB power sweep

Fully fulfill specification

Measurements to be made with each Power AmplifierAND

with the whole Transmitter (i.e. four Power Amplifiers combined together)

Measurements with a 800.888 MHz carrier at Pmax/2 and a frequency sweep 20dB below carrier :

Non linear phase distortionat +/-3.0 MHz: max. +/- 10°

Passband at -1 dB: 5.0 MHzPassband at -15 dB: 8.0 MHz


26

Factory acceptance tests:Power Sweep


Measurements to be made with each Power Amplifier

AND

with the whole Transmitter (i.e. four Power Amplifiers combined together)

With four PA : Po = 240 kW

With one PA : Po = 60 kW

Pout vs Pin must be monotonic from zero to Po

Small signal differential gain g = dPout/dPin, in the range 0.1 Po to 0.9 Po

Local slope variation max +/-15%

Can vary by 3 dB maximum.

Gain saturation curve

Small signal differential gain in the range 0.1 Po to 0.9 Po:

Local slope variation +/- 5% (+/- 2% averaged)

Vary by 2.0 dB maximum

Non linear phase distortion (CW):

Δ φmax < 10º

Non linear phase distortion curve

Phase distortion < 3°

Pout

0.9 Po

0Pin

0.1 Po

3 dB maximum

g

gmax

gmin

Pin

g = dPout/dPin

Phase shift

Pin

Δ Φ max < 10ºAt Pin for Po

-25.

0-2

3.8-2

2.6-2

1.4-2

0.2-1

9.0-1

7.8-1

6.6-1

5.4-1

4.2-1

3.0

19.0020.0021.0022.0023.0024.0025.00

-80-78-76-74-72-70

Gain Phase

Drive in [dBm]

IoT G

ain

[dB

]

Phase [

deg]

-25.

0-2

3.8-2

2.6-2

1.4-2

0.2-1

9.0-1

7.8-1

6.6-1

5.4-1

4.2-1

3.0

2527293133

f(x) = − 0.002588706 x² + 0.232314847 x + 25.76504335

differential TX gain

Diff

ere

nti

al gain

[d

B]

-6-4-202468

10

Local slope

Local slo

pe

vari

ati

on

[%]

0.1 Pmax 0.9 Pmax


27

Factory Acceptance Tests:conclusion


All factory acceptance tests have shown compliance respect to the specification, and even better : Linearity Monotonous Phase stability Maximum output power

All requirements were fulfilled (we repeated all the tests twice to confirm the results)

We checked modularity of the equipments

We controlled noise level

We checked protections: driver output reflected power operated while making tests (due to over range

power sweep) Water cooling, air temperature, current limits, etc …


28

CERN Acceptance Tests


Pre-series Amplifier has been integrated within CERN operational area

All tests cycles have been done for 4 hours each, no trouble has been discovered


29

Long duration tests:CW mode


When we launch CW long duration tests, difficulties arose

While doing the test over six weeks, we were not able to obtain a stable operation

Maximum time slots were : 115 hours : 1 66 hours: 2 33 hours: 5 < 24 hours : 18

→ Not stable enough in CW mode

41 2 0.50.50.15

50.50.50.52.5

13

0.5

16

4

33 33 33

20

710

32

11

60

115

34

8

0.5

14 16

5

66

19

50.5

23 24

1620

13

26

1512

1512 12 12

25

1410 9

0.020.02

84

31

8

32

47

0

20

40

60

80

100

120

140

Ca

lib

ra

te

dir

ecti

on

el

co

up

leu

r

-7 a

pril

Ne

wch

arg

e 2

50

kW

-4

ap

ril

Ad

juste

me

nt

tub

e b

y p

ho

ne

wit

h M

r.

Be

l -1

5 a

pril

Cle

an

ing

filt

er

-6 M

ay

Ad

juste

me

nt

tub

e w

ith

M.

Be

l -1

2 M

ay

HV iInhibit & HVPS Over

Voltage64%

HV Inhibit18%

Inverter Fault18%

CW Alarm repartition


30

Air temperature sensitivity


Transmitter has shown to be Temperature sensible : Water Temp = 26.4 +/- 0.5 Air Temps = 22.1 +/- 2.6 Driver Gain = 6.7 % Cold IOT = + 7.5 % to - 38 % Hot IOT = +/- 4.9 %

Drivers are inverse temp, while IOT is direct Temp

Restart a cold IOT must be done readjusting the drive within the first three minutes

CERN LLRF will manage these variations

Temp waterTemp air

Temp water Temp airRelative gain IOT


31

Long duration tests:Super Cycle mode


41 2 0.50.50.15

50.50.50.52.5

13

0.5

16

4

33 33 33

20

710

32

11

60

115

34

8

0.5

14 16

5

66

19

50.5

23 24

1620

13

26

1512

1512 12 12

25

1410 9

0.020.02

84

31

8

32

47

0

20

40

60

80

100

120

140

Ca

lib

rate

dir

ecti

on

el

co

up

leu

r -7

ap

ril

Ne

wch

arg

e 2

50

kW

-4

ap

ril

Ad

juste

me

nt

tub

e b

y p

ho

ne

wit

h M

r.

Be

l -1

5 a

pril

Cle

an

ing

filt

er

-6 M

ay

Ad

juste

me

nt

tub

e w

ith

M.

Be

l -1

2 M

ay

010203040506070

0 3000 6000

Pout

[kW

]

Time [ms]

To reduce average power and be closer to machine operation, we launched Super Cycle long duration tests, new difficulties arose

While doing the test over four weeks, we were not able again to obtain a stable operation

Time slots were mainly between 12 to 24 hours

The main fault is always the same ‘IGBT 4 gate D’, even with no amplifier connected !

We are convinced the tube itself is not part of the trouble


32

HVPPS instabilities


We tried a 50% RF signal instead of our super cycle, varying the repetition rate

The HVPPS stability is function of the repetition rate !

010203040506070

0 50 100

Pout

[kW

]

Time ON [%]

0

100

200

300

400

500

600

700

800

900

1000

0.1 0.5 1 5 10 50 100 500 1000 1500 1600 1700 1800 2000 2100 2500

Sta

bil

ity

[se

co

nd

s]

Repetition rate [Hz]

Power converter stability vs Repetition rate

20 hours10 hours

10 hours20 hours

100 hours


33

800 MHZ upgrade Conclusions


First tests were very promising, but…

Long duration tests shown lack of HVPPS stability

We asked for a conventional linear Power Converter (with thyratron crowbar)

Installation is foreseen next week …

Many thanks

For your attention, and for inviting me to your workshop


35

RF Group at CERN


RF group is 170 colleagues operating RF over all machines

A. Cobas, L. DupontGroup Secretaries

a d i o – r e q u e n c y G r o u pE. Jensen

Dpt. E. Ciapala

Beamsand RFCavityServos

& ControlsInterfaceRF Feedbacks

& BeamControlKlystrons

& SC CavitiesLinacsRF Synchrotrons RF

BR FB LR

E. ShaposhnikovaDpt. T.Bohl

A. Butterworth Dpt. L. Arnaudon

W. HöfleDpt. P. Baudrenghien

O. Brunner Dpt. G. Mcmonagle

M. VretenarDpt. F. Gerigk

E. MontesinosDpt. C. Rossi

T. Argyropoulos(DOCT)F. CaspersJ. Esteban Muller (FELL)S. Federmann (DOCT)L.Ficcadenti (FELL)S. HancockR.M. Holz (TECH)H. Timko (FELL)J. Tückmantel

M.E. AngolettaP. Baudrenghien

A. K. Bhattacharyya(FELL)J. Ferreira-BentoG. HagmannT. MastoridisJ. NoirjeanD. Stellfeld

A.BlasA. Bullitt (UPAS)H. DamerauA. FindlayJ. Fox (UPAS)M . Hernandez-Flano (FELL)P. Leinonen(FELL)T. Levens(FELL)J. Lollierou(FELL)R. LouwerseD. PerreletT. TruszcynskiD. ValuchU. Wehrle

L. ArnaudonD. LandreS. Totos

F. DubouchetM.JaussiJ. MolendijkM. Naon(UPAS)A. Pashnin (FELL)A. ReyF. Weierud

D. GlenatP. Martinez-YanezP. Maesen

G. Pechaud

G. McmonagleS. CurtG. Rossat

A. Benoit (Stagiaire) I. Mondino(FELL)C. NicouS. Mikulas(TECH)M .PasiniJ. PradierG. RavidaN. Schwerg(FELL)M. Therasse(FELL)W. Weingarten

J. Chambrillon(FELL)T. Junginger(DOCT)C. Liao (FELL)H. Vennekate(TECH)

J. BroereV. CobhamS. Doebert

A. Andersson W. Farabolini(UPAS)J-W. Kovermann(FELL)R. L. Lillestol (DOCT)S. LivesleyJ. MonteiroS. ReyR. Ruber(UPAS)H.S. Shaker (PJAS)L. TimeoT. Wiszniowski

F. GerigkN. Alharbi(UPAS)M. Schuh(UPAS)P. Ugena-Tirado (FELL)R. Wegner

G. GeschonkeJ.M. GiguetJ. Marques BalulaS. Ramberger K. Schirm

A. Boucherie S. CalvoG. Cipolla C. JulieN. Jurado F. KillingC. Marrelli (FELL) M. PaoluzziC. Renaud C. Vollinger

G. RiddoneA. Acker (UPAS) M.Filippova (PJAS)A. French (PJAS) N. Gazis(DOCT)D. Gudkov(PJAS) I. Kossyvakis(TECH)A.Olyunin(UPAS) P. Piirainen(FELL)F. Rossi (FELL) A. Samoshkin (UPAS)V. Soldatov(UPAS)A. Solodko (UPAS)J. Vainola(UPAS)

W. WuenschM. Dehler (PDAS)N. Shipman (DOCT)I. SyratchevA. Grudiev

A. D’Elia (UPAS) G. De Michele (UPAS)V. Khan (FELL) O. Kononenko(FELL)J. Shi(FELL) K. Sjoebaek(PJAS)H. Zha(UPAS)

C. RossiV. Bretin V. DesquiensM. Haase G. LobeauA. Marmillon M. MorvilloS. TavaresRego(UPAS)

DOCT = Doct. Student FELL = Fellow SUMM=Summer Student TECH = Techn. Student PDAS= Paid Ass. UPAS = Unpaid Ass. PJAS= Project Ass. Underlined = Supervisors

September2011

Documents

CERN SPS Upgrade New 200 MHz and 800 MHz amplifiers