Top-Up ?? J. L. Revol On behalf of the Accelerator & Source - - PowerPoint PPT Presentation
ESRF Friday Lecture Top-Up ?? J. L. Revol On behalf of the Accelerator & Source Division With the contribution of: JF Bouteille, L Farvacque, T Perron, E Plouviez, J Jacob, L Hardy, M Hahn, Lin Zang, P Berkvens A Chumakov P Clotens, G
On behalf of the Accelerator & Source Division With the contribution of:
JF Bouteille, L Farvacque, T Perron, E Plouviez, J Jacob, L Hardy, M Hahn, Lin Zang, P Berkvens A Chumakov P Clotens, G Monaco, T Mairs, O Mathon S Pascarelli, A Rogalev L Emery (APS) , T Nakamura(SPring8), P Kuske (Bessy), E Karentzoulis (Elettra), V Kempson (Diamond), M Bieler (Petra), A Luedeke (SLS), A Loulergue (Soleil), G Benedetti (ALBA)
ISDD Friday Lecture -- Top Up ??-- Revol JL, March 23rd, 2012 2
induced by the finite lifetime originating from the different loss processes (vacuum, Touschek, physical aperture, ...) By
Which Top Up do we need at the ESRF? Which Top Up could we provide at the ESRF?
ISDD Friday Lecture -- Top Up ??-- Revol JL, March 23rd, 2012 3
Top-Up Top-Up
Bessy II
Top-Up
Petra
Top-Up Top-Up
SLS
Top-Up Top-Up Top-Up Top-Up
APS
Top-Up every 12h
Decay mode
ISDD Friday Lecture -- Top Up ??-- Revol JL, March 23rd, 2012 4
Survey of Top Up operation at ESRF/ APS/ SPRing8/ PETRA/ SOLEIL/ DIAMOND/ BESSY/ ELETTRA/ ALBA /SLS Top Up is multi parameters with discrepancies between institutes!
ISDD Friday Lecture -- Top Up ??-- Revol JL, March 23rd, 2012 5
Stability in current at SPring8: 0.03 % "If this level of stability is impossible to achieve, constant decay in ~ a few minutes is preferable.“ said spring8 operation physicist.
ISDD Friday Lecture -- Top Up ??-- Revol JL, March 23rd, 2012 6
Current (mA) Delta I (mA) % SPring 8 100 0.03 0.03 Bessy 300 0.4 0.13 Elettra 300 1 0.3
ALBA 400 1 0.25
APS 100 0.5 0.5 SLS 400 2 0.5 Soleil 400 2 0.5 Diamond 250 2 0.8 Petra 100 1 1 ESRF 200 40 20 Lower variation is the best, but what is really needed for optics stability ??
ISDD Friday Lecture -- Top Up ??-- Revol JL, March 23rd, 2012 7
Photon energy, slit aperture,... and optics and beamline design
the layout of beamlines to manage heat load variation, especially for the UPBLs, which should be less sensitive to refills.
make a refill transparent? Difficult to get a single answer from the ESRF experts !! Beamlines have already made a lot of efforts to be less sensitive to heat load variation, solution optimized for the different techniques........ Only experimental data would really give precise information ..
ISDD Friday Lecture -- Top Up ??-- Revol JL, March 23rd, 2012 8
Heat-load on the Si(111) monochromator UPBL06 Inelastic scattering- TDR
Thermal slope error on the first crystal
Displacement field at the beam spot of the first crystal (7 keV) Power on the crystal: 5 mA →10 W 100 mA →200 W 200 mA →400 W 240 mA →475 W
Change the beam divergence corrected by vertically deflecting white beam mirror mounted on a bender
ISDD Friday Lecture -- Top Up ??-- Revol JL, March 23rd, 2012 9
50 100 150 200 250 300 5 10 15
slits, ΔHxΔV (mm
2)
2.0 x 1.0
Rocking curve FWHM (μrad) Storage ring current (mA) Si(333) at 43.236 keV
1 2 3 4 5 10 15
10 mA 50 mA 100 mA 150 mA 200 mA 250 mA 300 mA
Slit 2 x 1 mm
2
Horizontal position (mm) Surface height (nm)
Bragg angle narrow vertical slit at various vertical positions
Setup:
ID18 Measurement of crystal curvature
Optics optimized for 200 mA Ideal collimation between 160 and 200 mA, crystal flat due to the power deposit But crystal too cold to be ideally flat between 60 and 90 mA More disturbance in 16bunch than in 7/8+1
16 Bunch 7/8+1 16 Bunch 7/8+1
ISDD Friday Lecture -- Top Up ??-- Revol JL, March 23rd, 2012 10
Sensitive to short term and long term thermal effects. Optimization of the design and feedback on the slits as a function of the beam current to maintain a constant thermal load
10000 20000 30000 40000 50000
10 20 30 40 50 60 70 80 90 100 110 120 130
Stability (μm) time (s) Stability on a long timescale 2008 Horizontal 2008 Vertical 2011 Horizontal 2011 Vertical
ID24/ UPBL11 Time resolved and Extreme conditions X-ray Absorption Spectroscopy
But the stability of the
issue with the present ΔI, especially with the use of the 4 undulators.
ISDD Friday Lecture -- Top Up ??-- Revol JL, March 23rd, 2012 11
Stability Number of shots Time between injection SPring 8 0.03 % One ≈ 2 to 3 mn
Hybrid 10 to 30 sec
Bessy 0.13 % One ≈ 20 to 30 sec Soleil 0.5 % One ≈ 2 to 5 mn Elettra 0.3 % 20 shots ≈ 6 mn at 2 GeV
25 mn at 2.5 GeV
SLS 0.5 % 10 shots ≈ 1 to 2 mn Petra 1 % 40-50 low
charge shots
≈ 8 mn
1 mn (timing mode)
APS ≈ 0.5 % One 1 and 2 mn Diamond ≈ 0.8 % Few 10 mn ESRF ≈ 20% 400 12 hours
Injection at fixed Delta I Injection at fixed Delta Time
Two Operating Modes
ISDD Friday Lecture -- Top Up ??-- Revol JL, March 23rd, 2012 12
Injection kickers Depends on the closure of the bump and on the lattice Injection septum Depends on the magnet design and shielding Booster magnets field Depends on the leaking magnetic field, the shielding, the compensation field and the fast orbit feedback performances Cleaning process Depends on the process
ISDD Friday Lecture -- Top Up ??-- Revol JL, March 23rd, 2012 13
K1 K2 K3 K4 Septum
Stored beam
SR axis
Shielding
Ideal situation: Bump is perfectly closed, even during transient Septum is perfectly shielded Stored beam is moved close to the injected beam via K1,2,3,4 Injected beam is moved close to the stored beam via septum and K3,4 For the duration of the injected beam pulse (= booster length= 1μs at ESRF) Never perfectly achieved in any of the machines For each shot, the stored beam is shacked and damped with the transverse damping time given by the lattice (typically a few ms)
ISDD Friday Lecture -- Top Up ??-- Revol JL, March 23rd, 2012 14
K1 K2 K3 K4 Septum
Stored beam
SR axis
Shielding
“The issue is we have to actively maintain the tuning of the injection system kicker magnets to keep it at a minimum, then because we are pretty sure that our users can’t see the kick we have been a bit more relaxed about keeping it at a minimum and retune kickers less often leading to a higher value. We have also found the injection efficiency is not necessarily optimum at these minimum levels of stray kick”. Said Diamond Operation Manager...
ISDD Friday Lecture -- Top Up ??-- Revol JL, March 23rd, 2012 15
H Stability perturbation V Stability perturbation Sextupoles in the bump
μm
μm
SLS Negligible 40 No Soleil 100 (peak) 70 (rms) 50 No Elettra 100 (peak) 20 No Petra 100 10 No Diamond 250 (peak) 150 No Bessy 500 (peak) 1000 No SPring 8 400 (peak);(<200 rms) 1 Yes (strength optimized) ESRF 1700 (peak) 100 Yes (4) APS 3000 (peak) 200 Yes (14) All machines are in the μm stability standards, injection disturbs all machines for ≈ a few ms time ! Need gating for the most demanding experiments
ISDD Friday Lecture -- Top Up ??-- Revol JL, March 23rd, 2012 16
Usage of gating trigger SLS Trigger provided, used by some Soleil Trigger provided, not used Elettra Trigger provided to 2 beamlines Petra Trigger provided, not used Diamond Trigger provided, very few use Bessy Trigger provided SPring 8 Trigger provided, not used ESRF Software trigger provided, used by a few APS Trigger provided, not used Need gating for the most demanding experiments Very little use of gating !!
ISDD Friday Lecture -- Top Up ??-- Revol JL, March 23rd, 2012 17
Booster power supply type Injection Every Number of Kicks SPring 8 Ramping 1 Hz 2 mn One APS Ramping 2Hz 1 mn One Elettra Ramping 2Hz 6 mn Few Soleil Ramping 3Hz 2 mn One SLS Ramping 3Hz 1 mn Few Diamond Ramping 5Hz 10 mn Few Bessy Resonant 10Hz 20 sec One ESRF Resonant 10Hz 12 hours Few Petra Resonant 12Hz 8 mn Few Resonant power supply is more efficient. Ramping power supply is more flexible Is few shots injection mode worse? Main issues for Top-Up: Reproducibility & Switching ON sequence
ISDD Friday Lecture -- Top Up ??-- Revol JL, March 23rd, 2012 18
UPBL 4: NINA Nano-Imaging and Nano-Analysis NINA will not be sensitive to heat load variation
(Beam horizontally cut by the primary slits)
…But doing acquisition with 1ms acquisition time Sensitive to fast disturbance during image scan Sensitive to vertical emittance variation ID12: Circular Polarisation Beamline Not sensitive to heat load variation But making a difference between scans with different polarization Sensitive to very small disturbance during acquisition scan
No disturbance for at least 20 mn would be the best.
ISDD Friday Lecture -- Top Up ??-- Revol JL, March 23rd, 2012 19
Reliability of the injector Soleil Close to 100 % Bessy Very good SLS 99.4 Petra 98 % APS 97.6 % Elettra 97 % Diamond 44 missed injections SPring 8 > 8 hours MTBF ESRF
Bad for Top-Up, but presently not an issue , could be improved
Ideal issue: Sustain a few missed injections at the beamline
281 hours
ISDD Friday Lecture -- Top Up ??-- Revol JL, March 23rd, 2012 20
Multibunch Current in 7/8+1 mA 196 + 4 Horizontal emittance nm
4
Vertical emittance
pm 3.5
Lifetime at 200 mA h 45 Reduction of the lifetime in multibunch limited to less than 10 hours despite a reduction of the coupling by one order of magnitude For a few weeks at the beginning of 2011, to compensate for the reduction in lifetime, the number of refills was increased from 2 to 3 per day, with a long decay during the night and two shorter decays during day time. Current variation / 2 After investigation, beamline scientists report that they prefer 2 long
now back to 2 refill/day
Current variation still too high to be beneficial Or Large current variation already sustained
21
91 % of Beam time available for Timing Experiments
ISDD Friday Lecture -- Top Up ??-- Revol JL, March 23rd, 2012
Need cleaning at each injection
ISDD Friday Lecture -- Top Up ??-- Revol JL, March 23rd, 2012 22
Cleaning process SPring 8
In the booster
APS
In the particle accumulator ring (upstream booster) Envisaged in the Booster
ESRF
In the SR Tested in the booster
Soleil
No, Envisaged in the booster
Bessy
No, Envisaged in the SR
SLS
No
Petra
No
Elettra
No
Diamond
No
Operational at ESRF
Robust, but generates a vertical blow up Tested at ESRF, but need more work: Not efficient at the gun of the Linac due to dark current produced in the accelerating sections Not reproducible enough in the booster Operational at SPring8 Flat top during the ramping of the booster
In the storage ring In the injector
ISDD Friday Lecture -- Top Up ??-- Revol JL, March 23rd, 2012 23
Vertical emittance 6 pm 4 pm 160 mA 200 mA
9h00 9h05
Injection ≈< 1mn [600 kicks] Cleaning = 20 sec
3 Dec 2011 11 Sep 2011
Vertical emittance 8 pm 6 pm 57 mA 91 mA
9h00 9h05
Injection ≈> 1mn [600 kicks] Cleaning = 2 *20 sec 27 pm 42 pm
3000 pm !!
Refill in 16 Bunch Refill in 7/8+1 During Injection [1mn] beam is shacked every 100 ms for a few ms
Which results in a small vertical emittance blow up measured by our averaging diagnostic
During Cleaning [20 sec] In 7/8+1: small blow up: Only single bunch (4mA) excited
(gated cleaning)
In 16 bunch: strong blow up: Beam is excited for 20 sec
Which results in an increase of the vertical emittance to 3000 pm !! Extremely high perturbation !! A new method using the bunch by bunch feedback electronics is now available with a the blow -up <50 pm Users are often confused between blow up by injection and blow up by cleaning!
ISDD Friday Lecture -- Top Up ??-- Revol JL, March 23rd, 2012 24
Delivery at low vertical emittance 5 pm (instead of 40-30 pm, with excitation being f(I)) Lifetime of 6 hours instead of 12 hours Reduction of the current variation: Today 40% with 4 refills /day and a vertical emittance *10 Proposed : Refill every 30mn (preserving long decay, for stable data acquisition) Max current 90mA, (decaying to ≈ 82 mA) Improved cleaning in the SR (50 pm blow up),
continue the development in the injector.
Refill time reduced to less than 2mn (with automation) Proposal: Could be tested during MDT time on 29 May and 5 June
Or other MDTs
ISDD Friday Lecture -- Top Up ??-- Revol JL, March 23rd, 2012 25
Reliability of injector (....including control..), Strongly depends on the injection frequency Opening of the In-Vacuum during injection Electrical consumption
Strongly depends on Δt, (for ΔT=30mn: < 0.5 MWh/h, < 40€/h, 1000€/day) improved with, SSA RF on the booster, Faster switching ON sequence on BPSS
Improve present BPSS control. New booster Power Supply ? New injection scheme using no linear kickers Improved septum Injection efficiency Filling pattern diagnostics Improved purity diagnostics in the SR and in the Booster Cleaning in the booster Stability and reproducibility of the Linac and of the Injection/extraction elements ............................. But one of the most important issues is the required minimum stability To be worked out with the beamlines during MDT A lot of additional issues have to be worked on or investigated.......... To determine the optimum ΔI and Time between Injection
ISDD Friday Lecture -- Top Up ??-- Revol JL, March 23rd, 2012 26
Mode multibunch 200 mA,45h, 5pm vertical emittance, no cleaning :
Injector multibunch 5ma, efficiency ≈70%, ΔI/shot= 1 mA 1 shot Δt >≈ 15 mn
mA
% Δt= 30 mn
shots ΔI<≈ 2 mA
% Δt= 300 mn
shots ΔI<≈20 mA
%
Mode 7/8 200 mA, 45h, 5pm vertical emittance, cleaning:
Injector single 0.1mA, efficiency ≈70%, ΔI/shot= 0.02 mA
1 shot Δt >≈ 20 sec
% Δt= 3 mn
shots ΔI<≈ 0.2 mA
% Δt= 30 mn
shots ΔI<≈ 2 mA
% Δt= 300 mn
shots ΔI<≈ 20 mA
%
Mode 16 bunch 90 mA, 6h, 6pm vertical emittance, cleaning:
Injector single 0.1 ma *5 bunch, efficiency ≈70%, ΔI/shot= 0.1 mA
1 shot Δt >≈ 20 sec
% Δt= 1 mn
shots ΔI<≈ 0.3 mA
% Δt= 3 mn
shots ΔI<≈ 1 mA
% Δt= 30 mn
shots ΔI<≈ 10 mA
%
Single shot or few shots for ESRF??
(Figures are order of magnitude)
Standard decay 2mA 30mn
ISDD Friday Lecture -- Top Up ??-- Revol JL, March 23rd, 2012 27