1. L4-PSB working group meeting 25/06/2009
PS TFB
Linac 4 Timing
Summary of discussions held with
M.E. Angoletta, P. Baudrenghien, C. Carli, M. Chanel, A. Findlay
Alfred Blas, Flemming Pedersen
L4-PSB working group meeting 25/06/2009

2. Justification for Longitudinal Painting in PSB 1/3
PS TFB
Justification for Longitudinal Painting in PSB 1/3
During it’s 37 years long history, transverse space charge at low energy has been a dominant cause of intensity (Isolde, CNGS, AD) and brightness (LHC) performance limitations in the PSB.
One very important parameter is the bunching factor at low energy:
Make the bucket long (reduce fs), add a second harmonic RF
Make bunches as long as possible: fill the bucket as much as possible (but without losses)
Avoid bunch shape oscillations and associated oscillations in bunching factor
Modify the potential well to make bunches flatter (= add second harmonic RF)
Modify the injected phase space distribution in the Linac or prior to capture in PSB.
Modified potential well.
During the h = 5 era, it was first studied with a h = 10 prototype in and all 4 C16 systems were added in
During the LHC injector conversion project in 97/983, the dual harmonic operation was retained as we entered the h = 1 & 2 era in the PSB to maximize the brightness of the LHC beams in the PSB
1) Shaping of Proton Distribution for Raising the Space-Charge Limit of the CERN PS Booster, XIth Int. Conf. on High Energy Accelerators, CERN, Geneva, 1980 or CERH/PS/BR 80-14
2) A Second Harmonic (6-16 MHz) RF System with Feedback-Reduced Gap Impedance for Accelerating Flat-topped Bunches in the CERN PS Booster
3) Conversion of the PS Complex as LHC Proton Pre-Injector,
Alfred Blas, Flemming Pedersen
L4-PSB working group meeting 25/06/2009

3. Justification for Longitudinal Painting in PSB 2/3
PS TFB
Justification for Longitudinal Painting in PSB 2/3
Modified phase space distribution (hollow distribution, depressed central density)
The concept was proposed analytically as a sum of two elliptical momentum distributions in 19794
Machine studies were done in 79/801 by using the 400 MHz debuncher of Linac2 with 180 degree phase change as well as empty bucket deposition using PSB 3 MHz rf.
Instabilities associated with closing of the phase loop prevented this from being used operationally; mechanism not well understood at that time.
In the h = 1&2 era, this method was further pursued with machine studies in 995 using the 16 MHz C16 system to inject an empty bucket into the coasting beam prior to RF capture (requires dB/dt = 0 at injection).
Numerical computations of longitudinal bunched beam transfer functions of bunches with depressed central density by Shane Koscielniak (TRIUMF) contributed significantly to the understanding of the mechanism of the instability.
4) Bunches with Local Elliptic Energy Distributions
5) New Technique for Bunch Shape Flattening
Alfred Blas, Flemming Pedersen
L4-PSB working group meeting 25/06/2009

4. Justification for Longitudinal Painting in PSB 3/3
PS TFB
Justification for Longitudinal Painting in PSB 3/3
Longitudinal painting in the Linac4 era
The increased injection energy (160 MeV) reduces the transverse space charge by a factor of 2
The 3 MeV chopper and H- charge exchange injection permits loss-free injection into a moving bucket at full dB/dt which means reduced time at low energy.
Turn by turn energy modulation and turn-by-turn timing control of the 3 MeV chopper timing permits to inject a longitudinal phase space distribution very close to being the stationary distribution such that harmful bunch shape oscillations and losses are avoided and obtain a good bunching factor at low energy which improves both brightness and intensity
The digital chopper timing control system with associated application and controls software layers must be conceived and planned to have turn-by-turn timing control ready from day one of operation of PSB with Linac4
This permits the LHC (brightness, reduced filling time) and other CERN users (intensity) to rapidly obtain maximum profit from the Linac4 upgrade
To fully profit from this longitudinal painting (depressed central density, intensity control) several chopper on/off events must be foreseen within a single turn
The PSB will then finally have the longitudinal phase space density control tool we have always wished we had for the last 37 years!!
5) Linac4, a New Injector for the CERN PS Booster
Alfred Blas, Flemming Pedersen
L4-PSB working group meeting 25/06/2009

5. L4-PSB working group meeting 25/06/2009
PS TFB
Linac 4 - PSB
 2μs
Pre-chopper
+LEBT
Debuncher
RF Feedforward
Distri
4*rf
From F. Gerigk
180 m
Source
45 keV
Chopper
Amplitude modulated
for energy modulation (painting)
from P. Baudrenghien
RF feed-forward, energy modulator, debuncher, distributor and rf have to be in phase with the Chopper
Alfred Blas, Flemming Pedersen
L4-PSB working group meeting 25/06/2009

6. Chopper + Feed-Forward timing
PS TFB
Chopper + Feed-Forward timing
PSB beam demand = some number of Linac 4 bunches (352.2 MHz).
Intensity precision: No jitter, 3 bunches doesn’t mean 3 +/- 1 bunch(es).
Time precision: One L4 bunch at time t doesn’t mean 1 bunch at time t + TL4_RF
This one Linac 4 rf period uncertainty is only accepted as a global shift of the one-revolution bunch train.
Absolute Time reference: Could be time zero of the entire injection process (start injection re-synchronized with the revolution reference)
Could be “start injection” + each revolution reference tic
Reference clock: These constraints imply that the chopper should use the L4_RF or a (sub) multiple.
Alfred Blas, Flemming Pedersen
L4-PSB working group meeting 25/06/2009

7. L4-PSB working group meeting 25/06/2009
PS TFB
Chopper + Feed-Forward timing
Single absolute time reference = start injection
Pros: No intensity jitter
No time jitter (just a single jitter of one L4 rf period for the global filling pattern)
Cons: any error is integrated over the entire injection process duration (≠ FECs to be well synchronized when changing L4 fRF or PSB fINJ REF )
The entire injection process lasts < 400 μs = 140k TL4-RF = 18 bit value
Time Value
Event Number
Alfred Blas, Flemming Pedersen
L4-PSB working group meeting 25/06/2009

8. L4-PSB working group meeting 25/06/2009
PS TFB
Chopper + Feed-Forward timing
Dual absolute time reference = start injection + revolution reference
=> L4 rf re-synchronized at each revolution
Pros: No intensity jitter
Any timing program error is effective only within one pair of revolutions
Cons: One L4 rf period jitter for the global filling pattern within one revolution
Need to treat a pair of revolution periods for each revolution filling pattern as the actual rf overlaps the reference during acceleration
Alfred Blas, Flemming Pedersen
L4-PSB working group meeting 25/06/2009

9. L4-PSB working group meeting 25/06/2009
PS TFB
Chopper + Feed-Forward timing
Triple absolute time reference = start injection + revolution reference + multiple of the reference
=> L4 rf re-synchronized at each Beam ON-OFF
Pros: Any timing program error is effective only within one revolution
Cons: One L4 rf period jitter for any Beam ON-OFF event (intensity jitter)
Most complicated
Alfred Blas, Flemming Pedersen
L4-PSB working group meeting 25/06/2009

10. Functions for the Modulator + debuncher
PS TFB
Functions for the Modulator + debuncher
The function generator will provide the set point value to the energy modulator (Voltage amplitude modulation) and to the debuncher (phase and ? amplitude modulation). The typical signal is a triangle with a >20 μs period.
Requirements : New value every revolution. Both the Modulator and the debuncher need to have the same final response. Staircase signal acceptable when both modulator and debuncher have a high BW (staircase response also = flat line painting). When one or both of the response times are long (>10ns) the homogeneity of the painting will be affected by parasitic exponentials. This would require additional interpolated values.
Alfred Blas, Flemming Pedersen
L4-PSB working group meeting 25/06/2009

11. L4-PSB working group meeting 25/06/2009
PS TFB
Functions for the Modulator + debuncher
Requirements : New value every revolution. Both the Modulator and the debuncher need to have the same final response. Staircase signal acceptable when both modulator and debuncher have a high BW (staircase response also = flat line painting). When one or both of the response times are long (>10ns) the homogeneity of the painting will be affected by parasitic exponentials. This would require an addition of interpolated values to obtain a smooth ramp.
Time precision: one L4 rf period (although not critical)
Absolute Time reference: Time zero of the entire injection process (start injection re-synchronized with the revolution reference)
Reference clock: Injection revolution reference
Alfred Blas, Flemming Pedersen
L4-PSB working group meeting 25/06/2009

12. L4-PSB working group meeting 25/06/2009
PS TFB
Block diagram
 2μs
Pre-chopper
+LEBT
Energy modulation
Debuncher
Distri
4*rf
Source
Chopper
Linac 4
20 m
160 m
Linac rf
feed-forward
Function
Generator
FMC3 (CO)
45 keV
Chopper Control
Application
BIXi.RF_PHASE + dF/dT
TON-OFF Chopper
L4 rf
BIXi.SDIS
CTRV
Timing CO
BIX.SInjChop
BIXi.SInjRF
Rev
Inj. Ref
Source h2
1/2 h1
SP2T
Inj. rf reference (h1 or h2)
BIXi.SDIS
Alfred Blas, Flemming Pedersen
L4-PSB working group meeting 25/06/2009

13. Block diagram Linac4 PSB Application Low level RF, timings ….
PS TFB
Block diagram
 2μs
Pre-chopper
+LEBT
Energy modulation
Source
Debuncher
Distri
4*rf
Chopper
Linac 4
20 m
160 m
45 keV
Linac4
Low level RF, timings ….
PSB
Low level RF and standard timing modules
Trigger(s), possible RF train(s) and E-modulation
Application
Provides data for chopper & E-modulation well in advance (different data for different USER’s for PPM)
Slide from C. Carli
Aim of these discussions: discuss and agree on an interface between these three parties
Alfred Blas, Flemming Pedersen
L4-PSB working group meeting 25/06/2009

14. L4-PSB working group meeting 25/06/2009
PS TFB
Security issues
Requirements : The Linac 4 beam should not reach high (Beam-OFF at the chopper level) energy when:
the Distributor targets the Head Dump
the Distributor is in a transition state
the Distributor targets the Tail Dump
The Tail Dump position of the distributor should be removed, at the kicker level, as the tail Dump is unlikely to withstand a long L4 pulse.
Alfred Blas, Flemming Pedersen
L4-PSB working group meeting 25/06/2009