LHC Inject
orUpgrade
SPS 800 MHzRF Power upgradeProgress reportDecember 2013
LHC Inject
orUpgrade
LIU
-SPS
mee
ting,
12t
h D
ecem
ber 2
013
Contents
• IOTs• CERN SPS 800 MHz system
& why an upgrade to IOTs• Major difficulties
while commissioning the pre-series• Series delivery
• Cavities• New Antennae• New Controls
Eric
Mon
tesi
nos
2
LHC Inject
orUpgrade
LIU
-SPS
mee
ting,
12t
h D
ecem
ber 2
013
CERN SPS 800 MHz system & why an upgrade to IOTs
800 MHz power plant• Since 1980 the system is
composed of 2 transmitters of 225 kW
• One transmitter has4 x 58 kW Valvo klystrons with 3 dB combiners
• Each transmitter is connected through~ 120 m waveguides to its Travelling Wave Cavity
Eric
Mon
tesi
nos
3
2 x 225 kWtransmitters
YL1198Klystron
800 MHz TWC
LHC Inject
orUpgrade
LIU
-SPS
mee
ting,
12t
h D
ecem
ber 2
013
CERN SPS 800 MHz system & why an upgrade to IOTs
Obsolescence of the system
• This RF power system is getting very old
• We had major difficulties with klystron ceramic failures and with HV transformers
Eric
Mon
tesi
nos
4Effect of age on HV transformers
Broken ceramic window
of a klystron
LHC Inject
orUpgrade
LIU
-SPS
mee
ting,
12t
h D
ecem
ber 2
013
CERN SPS 800 MHz system & why an upgrade to IOTs
Upgrade proposal• keep all existing
ancillaries and replaceKlystron Transmitters with new IOT Transmitters
• First and only IOT at CERN
• We wanted to get experience with this tube as it could be used for
• LHC upgrade (crab cavities)• new CERN accelerator RF
system (SPL ?)
Eric
Mon
tesi
nos
5
New IOT transmitter
LHC Inject
orUpgrade
LIU
-SPS
mee
ting,
12t
h D
ecem
ber 2
013
CERN SPS 800 MHz system & why an upgrade to IOTs
Long purchasing process• 2009, 40 member states
companies have been contacted
• 2010, only six companies have been compliant to our specifications
• 2010, Electrosys has been selected
• 2011, we received the pre-serie transmitter at cern
Eric
Mon
tesi
nos
6
Two companies per member state have been contacted
LHC Inject
orUpgrade
LIU
-SPS
mee
ting,
12t
h D
ecem
ber 2
013
Main difficulties while commissioning the pre-series
Major troubles
• All Factory Acceptance Tests have shown compliance
• Pre-series Amplifier at CERN was ok with short duration tests
• Long duration tests, we started to experience difficulties• HVSMPS• Conventional HVPS• Erratic faults from 3 to 6 am & from 4 to 7 pm
Eric
Mon
tesi
nos
7
LHC Inject
orUpgrade
LIU
-SPS
mee
ting,
12t
h D
ecem
ber 2
013
Main difficulties while commissioning the pre-series
HVSMPS instabilities 2011
Eric
Mon
tesi
nos
8
IOT
HVSMPS
AM square
Repetition rate
RF Power load
600 Hz 17 kHz
HV DC(+ harmonics)
0.1 1100
15001700
20002500
1
10
100 Time Between Failure
[Hz]
[hours]
RF
LHC Inject
orUpgrade
LIU
-SPS
mee
ting,
12t
h D
ecem
ber 2
013
Main difficulties while commissioning the pre-series
HVPS troubles 1st half 2012
• HV monitoring transformers
• Ferrites in the IOT input circuit
Eric
Mon
tesi
nos
9
IOT
HVPS
RF
AM square
Repetition rate
RF Power load
Burnt HV monitoring transformers for di/dt
protection
HV transformer
MainsDiodes rectifier
Capacitors stack Thyratron
HV DC
LHC Inject
orUpgrade
LIU
-SPS
mee
ting,
12t
h D
ecem
ber 2
013
Main difficulties while commissioning the pre-series
HVPS troubles 2nd half 2012
• Repetitive faults from• 3:00 to 6:00 am• 4:00 to 7:00 pm
Eric
Mon
tesi
nos
10
IOT
HVPS
RF
AM square
Repetition rate
RF Power load
0:00 1:59 3:59 5:59 7:59 9:59 11:59 13:59 15:59 17:59 19:59 21:59 23:590
200
400
600Noise immunity
Time of the day the fault occurs [h]Ti
me
betw
een
faul
ts [h
]
LHC Inject
orUpgrade
LIU
-SPS
mee
ting,
12t
h D
ecem
ber 2
013
Main difficulties while commissioning the pre-series
HVPS troubles 2nd half 2012
• We were not able to link it with any external ‘perturbation’• CERN workers, security people or firemen checking the building• Automatic lighting of the building• GSM or Wifi antenna in the building• Traffic lights close to the building• Public lighting day & night light detection• Airport traffic: no planes before 7 am at Geneva airport• Train: old story with TGV disturbing LEP beams, not able to link with any
train timetable• Etc…
Eric
Mon
tesi
nos
11
0:00 1:59 3:59 5:59 7:59 9:59 11:59 13:59 15:59 17:59 19:59 21:59 23:590
200
400
600Noise immunity
Tim
e be
twee
n fa
ults
[h]
Time of the day the fault occurs [h]
LHC Inject
orUpgrade
LIU
-SPS
mee
ting,
12t
h D
ecem
ber 2
013
Main difficulties while commissioning the pre-series
HVPS troubles 2nd half 2012
• Not always detected by the Driver overdrive protection
• Always triggering the di/dt tube protection
Eric
Mon
tesi
nos
12
IOT
HVPS
RF
AM square
Repetition rate
RF Power load
• Faulty RF generator• Glitch on top of the RF pulse• Equivalent to 10 x power
LHC Inject
orUpgrade
LIU
-SPS
mee
ting,
12t
h D
ecem
ber 2
013
Main difficulties while commissioning the pre-series
HVPS troubles 2nd half 2012
• We solve the problem simply by replacing the RF generator
• We were able to perform2,800 hours without a fault
• We had to stop the test as we started LS1
Eric
Mon
tesi
nos
13
IOT
HVPS
RF
AM square
Repetition rate
RF Power load
12/13/2011 3/12/2012 6/10/2012 9/8/2012 12/7/2012 3/7/20130
500
1000
1500
2000
2500
3000Time between Failure
Date
hours
LHC Inject
orUpgrade
LIU
-SPS
mee
ting,
12t
h D
ecem
ber 2
013
6/1/2011 8/30/2011 11/28/2011 2/26/2012 5/26/2012 8/24/2012 11/22/2012 2/20/20130
500
1000
1500
2000
2500
3000Time between Failure
Date
hours
Main difficulties while commissioning the pre-series
Time spent per problem
Eric
Mon
tesi
nos
14
6 months 5 months 4 months
HVSMPS HV monitoring transformers
&Input circuit
ferrites
RFGenerator
&Faulty
Overdrive protection
20 months before launching the series production
24.06.12 green light for Series production
LHC Inject
orUpgrade
LIU
-SPS
mee
ting,
12t
h D
ecem
ber 2
013
Series delivery
Series production• We asked for several
modifications
• Driver overdrive protection
• Additional air cooling for coaxial lines
• Direct access to hardware interlocks through a monitoring panel with ‘blinking first fault’
Eric
Mon
tesi
nos
15
Hardware Electronic PLC
Remote GUI
Hardware Electronic PLC
Local GUI &
First faultRemote
GUI
PLC clock, minimum ms
Real time
LHC Inject
orUpgrade
LIU
-SPS
mee
ting,
12t
h D
ecem
ber 2
013
Series delivery
Series production
Eric
Mon
tesi
nos
16
Remote GUI
External probes with risky possible false contacts
In addition we have a panel for quick
identification of faults (LEDs)
& with direct access to hardware interlocks
(BNC) for real time monitoring
LHC Inject
orUpgrade
LIU
-SPS
mee
ting,
12t
h D
ecem
ber 2
013
Series delivery
Series tests at CERN• Due to LS1, lack of
water has limited tests up to 60 kW total maximum, i.e. one transmitter at a time
• Between June and December 2013 we received all eight transmitters and commissioned them individually
Eric
Mon
tesi
nos
17
• Transmitter #1 connected to main combiners
• Transmitter #2 connected to RF load for CERN Acceptance Test
• Transmitter #3 not yet connected
#1#2
#3
LHC Inject
orUpgrade
LIU
-SPS
mee
ting,
12t
h D
ecem
ber 2
013
Series delivery
Series tests• Seven transmitters
have been fully commissioned
• We miss the tube for the last transmitter, should be delivered before end 2013
• It has already been validated with the tube of transmitter 7
Eric
Mon
tesi
nos
18
Gino testing one of the transmitter
All eight transmitters in BB3
LHC Inject
orUpgrade
LIU
-SPS
mee
ting,
12t
h D
ecem
ber 2
013
Series delivery
What is a commissioning ?• Long and detailed reception
process
• Static tests• Mechanical inspection• Water pressure test• Ancillaries checks
• Electrical tests• Emergency stop test• HV test with an external HVPS by
steps of 5 kV per minutes• Crowbar without tube and with
tube (thin wire in short circuit)
• RF Low Power tests• Tuning the amplifier
Eric
Mon
tesi
nos
19
BW-1dB must be > 6 MHz (+/- 3 MHz)
with Ibeam = 0.45 A
Phase variation < +/- 30°over 6 MHz (+/- 3 MHz)
LHC Inject
orUpgrade
LIU
-SPS
mee
ting,
12t
h D
ecem
ber 2
013
Series delivery
What is a commissioning ?• RF High Power tests
• Increase the power by step of 10 kW up to 60 kW
• Check all values, Vbeam, Ibeam, Pfwd, Prev, Pdriver, Ibody, Vfocus, Ifocus
• At mid power, retune the cavity for Pmax
• Measure • Thermal power load• Efficiency = RFout / HVin
• Long duration tests with several cycles
• Minimum four hours per cycle• Observe any characteristic variation
Eric
Mon
tesi
nos
20Long duration test description
Measurements to be made with each Power Amplifier AND
with the whole Transmitter (i.e. four Power Amplifiers combined together) Continuous operation 24/24 hours CW Very Long Pulses operation Fc = 800.888 MHz +/ - 0.5 MHz : 100% from 0 to 60kW with rise and fall time < 0.5 µs 5 seconds ON / 5 seconds OFF
time
Power
5 s
0.5 us 0.5 us
240 kW
0 kW
10 s
AM modulation #1 Fc = 800.888 MHz +/ - 0.5 MHz : 100% from 0 to 60kW with rise and fall time < 0.5 µs Repetition time 10 µs (100kHz)
time
Power
10 us
0.5 us0.5 us
240 kW
0 kW
AM modulation #2 Fc = 800.888 MHz +/ - 0.5 MHz : 0 to 60kW with 4 MHz triangle AM 25 % in power Rise and fall time < 0.5 µs Flat top pulse and off pulse length of 11 us Repetition time 23 µs (100kHz) This cycle for 20 second then 1 second OFF.
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 %
LHC Inject
orUpgrade
LIU
-SPS
mee
ting,
12t
h D
ecem
ber 2
013
Series delivery
Circulator• Thales asked us to protect
the tubes with a circulator against full reflection that could occur while conditioning the couplers
• We fully approved it regarding other applications difficulties (ESRF, ALBA, …)
• We kept our Waveguide switch to allow quick test onto a power load
Eric
Mon
tesi
nos
21
Cavity
Load
Short
Circulator – WG switch (cavity or short) – Power load
Circulator
TCU
Short
To cavityWG switch
Load
WG switch
LHC Inject
orUpgrade
LIU
-SPS
mee
ting,
12t
h D
ecem
ber 2
013
Series delivery
Still to be done• Test of the four
amplifiers together with both systems
• Onto the load• Onto the cavity
• Beginning 2014 with a low total power of60 kW (lack of water)
• From April 2014 with full total power of240 kW (water back)
Eric
Mon
tesi
nos
22
Cavity #2Load
Short
Cavity #1
Load
Short
~
2014
LHC Inject
orUpgrade
LIU
-SPS
mee
ting,
12t
h D
ecem
ber 2
013
Cavities
New Antennae• A new antenna has been re-
designed
• It is fully compatible with the existing model and with the TWC 200 MHz
• There were two antennae per cavity, LLRF asked for one per iris
• 100 antennae have been produced
• 37 cells per cavity have been equipped
Eric
Mon
tesi
nos
23
New antenna design
Sébastien adding a new antenna on all 66 cells
LHC Inject
orUpgrade
LIU
-SPS
mee
ting,
12t
h D
ecem
ber 2
013
Cavities
New Controls
• With the obsolescence of the ECAs and controls of the TWC 200 MHz systems, as we were upgrading Power Amplifiers, we decide to also upgrade TWC 800 MHz controls
• This is underway, and we will have a new system ready beginning 2014
Eric
Mon
tesi
nos
24
New controls unit front panel
New controls unit PLC
LHC Inject
orUpgrade
LIU
-SPS
mee
ting,
12t
h D
ecem
ber 2
013
Schedule
Eric
Mon
tesi
nos
25
2014
Jan Feb Mar Apr May Jun Jul AugDec
8 Individual transmitters
TWC 160 kW total
TWC 260 kW total TWC 1
220 kW total
TWC 2220 kW total
TWC 1 & 22 x 220 kW total
+Cavity
conditioning
LLRF commissioning
SPS setup with beam
SPS closed
LLRF antennae calibration
LHC Inject
orUpgrade
LIU
-SPS
mee
ting,
12t
h D
ecem
ber 2
013
Conclusion
All 8 systems are expected to be operational before the end of this year
Even if it has not been easy, High Power systems are pretty well as scheduled
Eric
Mon
tesi
nos
26