energy stability update - stanford university · energy stability update tim maxwell on behalf of...
Post on 06-Jun-2020
3 Views
Preview:
TRANSCRIPT
Energy StabilityUpdate
Tim Maxwell on behalf of the Stability Performance Program:F.-J. Decker, Andy Benwell, William Colocho, Anatoly Krasnykh, Jeff de Lamare, Jim Lewandowski,Minh Nguyen, Phil Seward, John Sheppard, Mike Stanek, Tao Tang, Jim Turner, Lanfa Wang, …
January 22nd, 2015
MD Physics Meeting
2
Outline
• Jitter evolution• Theory• Application to LCLS (L1S sensitivity & L2/L3 setup)
• Recent stabilization activities• L1S AIP upgrade recovery• Soft x-ray jitter reduction (BC2 E)• Impact on (H/S)XRSS
• In development• Auto-ranging of thyratrons• Auto-diagnosis of known station issues• Auto-suggest linac complement• Improved RF loads
3
Jitter Evolution
We know the phase space ellipse• Each point is an electron in one bunch
4
Jitter Evolution
But also note the centroid (or jitter) ellipse• The same matrix math describes evolution of the time-
energy distribution of the centroids of many shots• Each offset point is the centroid of a different bunch*
* Not the same asdist. of each bunch
5
Coordinates
• Do math using covariance matrix
• Easy to relate to (measurable) rms RF jitters σt , σE andtheir correlation | ρ | < 1
6
Jitter Evolution
• We know linear transformations M
• For some coordinates X / Y (e.g.: z / δ), if
• …then new ellipse σ' around new mean vector μ' :
• Abs. E coordinate is calculated, can normalize to μ
7
Linear Operators
• RF accelerating structure:
- R65 (elsewhere “h”) is the rf chirp- Shear phase space in E direction, add energy offset to mean E
• Magnetic chicane:
- Shears phase space in time
( )rfrfE fsin-µ
• Linac sections each have an intrinsic phase-amp. jitter ellipse σrf
• Seen by centroids of (frozen) beam, locally adds purely uncorrelatedenergy jitter at station:
• Phase jitter factors in if station is chirping (R65)• Increases irreducible total t-E jitter (“centroid emittance”) as• Consecutive RF stations chirp and add their jitter independently
8
Primary Nonlinear Contribution: RF Jitter
66,5522
rfssee ¢+=¢
2rf,
265rf,rf,rf65
2rf,
2beam,66rf, gg
2 ttEEE RR sssrsss ++==¢
Chirp of incoming jitter + Increase of uncorr. E jitter due to station instability
9
L1S jitter: critical station
• If all other sources negligible (zero), L1S contribution transforms as:
• If φL3 = 0, final jitter is:
• R56,BC1 (-0.691 ps/MeV), L2 @ 5 GeV & -32º, R65,L2 (67.2 MeV/ps)
L1XBC1L2BC2L3T MMMMMMMMσ =÷÷
ø
öççè
æ= ,
000
2L1Sbeam,,E
final s
L1Sbeam,,final
L1SBC1,56L2,65finalbeam,,L1Sbeam,,BC1,56L2,65finalbeam,, 1or1 dd ssss
EERRRR EE +=+=
L1Sbeam,,final
finalbeam,,L1Sbeam,,finalbeam,,GeV28.11or48~ dd ssss
EEE ==
10
L1S jitter: critical station
• SXR: 4.7 GeV and < 0.05% jitter at the end, 235 MeV @ L1S:
• HXR: 13.6 GeV and < 0.025% jitter at the end, 235 MeV @ L1S:
%021.0
%05.07.428.11
L1Sbeam,,
finalbeam,,
L1Sbeam,,finalbeam,,
<Þ
<
=
d
d
dd
s
s
ss
%030.0
%025.06.1328.11
L1Sbeam,,
finalbeam,,
L1Sbeam,,finalbeam,,
<Þ
<
=
d
d
dd
s
s
ss
Was 0.0196% summer 2014
Not quite… need more jitter overheadfor HXR to allow for > L3 abs. jit. growth
11
L1S jitter: critical station
Compute L1S jit. limits*:
* w/ L1S SLEDed, ρ = -0.35 typ.
2L1S,
265L1S,L1S,L1S65
2L1S,L1S, gg
2 ttEEE RR sssrss ++=
Goal
L1S SLEDed requires:σA,L1S < 0.04%σφ,L1S < 0.04º
Today
Run 10start
July 2014
12
Low Energy Jitter Reduction
• EBC2 historically maintained at 5 GeV, regardless of EL3
• Reconfig time and collective effects primary concerns
• Compare RF phasors for EBC2 = (5 or 3) GeV, EL3 = 3.4 GeV:
220 MeV 3 GeV 3.4 GeV
5 GeV
13
Low Energy Jitter Reduction
• Each station contributes uncorrelated absolute E jitter in quadrature• +Higher abs. t jit. after BC2 makes L3 jit/station larger
• Higher abs. E jit., same final abs E → Higher rel. E jit.• Quasi-linear model predicts ~40% jitter reduction
220 MeV 3 GeV 3.4 GeV
5 GeV
14
L3 HXR “Decker Phasing” – Jitter ellipse evolution
2) After BC2R56 = -24.7 mm
3) L3 linear transformw/ φL3 = 0º (same)
4) L3 uncor.0.38 MeV growth*
* Competition: chirp can remove incomingcorr. E-jitter, but also increases uncor.jitter from L3
From LiTrack, φL3 = -15º should yield < 1%relative increase of proj. e-spread
1) After L2(5 GeV, ~3.5 kA)
t (ps)
ΔE(M
eV)
t (ps)
ΔE(M
eV)
t (ps)
ΔE(M
eV)
3) L3 linear transformw/ φL3 = -14.6º
(positive R65 fightsnegative BC2 chirp)
t (ps)
ΔE(M
eV)
4) L3 uncor.1.72 MeV growth*
t (ps)
ΔE(M
eV)
t (ps)
ΔE(M
eV)
15
Outline
• Jitter evolution• Theory• Application to LCLS (L1S sensitivity & L2/L3 setup)
• Recent stabilization activities• L1S AIP upgrade recovery• Soft x-ray jitter reduction (BC2 E)• Impact on (H/S)XRSS
• In development• Auto-ranging of thyratrons• Auto-diagnosis of known station issues• Auto-suggest linac complement• Improved RF loads
L1S AIP Upgrade Recovery
• L1S Modulator (AIP upgrade) and Tube changed over fall ’14 down,recovered with phase jit. out of tol (> 0.04º)
• History of L1S work maintained in CATER and athttps://portal.slac.stanford.edu/sites/lclscore_public/Pages/Performance_Stability.aspx
• Restored after efforts from A. Benwell, T. Tang, M. Nguyen, …
16
17
June 18th SXR Jitter Reduction Study
Constants: 1.5 kA, L3 Energy 4 GeV1. 32 L2 Klys., BC2 E = 5 GeV, R56 = -24.7 mm2. 22 L2 Klys., BC2 E = 3 GeV, R56 = -24.7 mm3. 22 L2 Klys., BC2 E = 3 GeV, R56 = -27.2 mm
o Calc.+ Meas.
Exp’t proposed by Lanfa from LiTrack MOGA
18
SXR jitter reduction
• BC2 energy change for Efinal < 6 GeV used so far in Run 11• 16 redundant stations removed @ 540 eV (3.4 GeV)• Consistently 40% less jitter, ex: 0.12% → 0.072% @ 540 eV• No (apparent) FEL loss…
19
SXR jitter reduction – beam impact
Slice energy spread by XTCAV (lasing off)
May 9, 2014 Jan 12, 2014
RMS SES = ~2.4 MeV RMS SES = ~2.4 MeV(horns cut @ BC1)
20
SXR jitter reduction – beam impact
LTU EmittanceDec. 5, 2014 Jan 12, 2014
εx = 1.26 μm / εy = 1.13 μm εx = 1.74 μm / εy = 1.23 μm
21
SXRSS on Jan. 12 @ 540 eV (3.4 GeV)
• Had 0.12% jitter in Dec 2014 test, now 0.078%• “Jitter reduction factor”
• F = <U>seeded,δ-cut / <U>seeded,all
• How much higher avg. brightness is w/o jitter• F = 1.66 w/i 2x FWHM seed bw
Avg. spec. in small δ window U within 2x FWHM bw
<U>seeded,δ-cut
<U>seeded,all
22
SXRSS @ 540 eV
May 2014σδ = 0.11%
F = 2.02
Jan 2015σδ = 0.071%
F = 1.66
“Moving the goal post”: Impact of jitter was reduced despite tuning up a narrower “mustache” in Jan ’15
23
HXRSS
• Delivering 0.025% jit. @ 8.3 keV in Run 9 w/ “L3 Decker phasing”• F ~ 1!
May 2012 @ 8.3 keV June 2014 @ 8.3 keV
σδ> 5x10-4 = σSASE σδ= 2.6x10-4 ≈ σSASE / 2
<U> = 0.15 mJ <U> = 0.33 mJ
24
HXRSS
HXRSS on Jan. 14 2015 @ 5.5 keV• Noted L3 phase trick was less effective• Still some “meh” L2/L3 stations and injector still being squeezed• Avg. 0.47 mJ and 26% fluctuations in 1.2 (x2) eV BW• F = 1.17
25
Outline
• Jitter evolution• Theory• Application to LCLS (L1S sensitivity & L2/L3 setup)
• Recent stabilization activities• L1S AIP upgrade recovery• Soft x-ray jitter reduction (BC2 E)• Impact on (H/S)XRSS
• In development• Auto-ranging of thyratrons• Auto-diagnosis of known station issues• Auto-suggest linac complement• Improved RF loads
26
Auto-diagnosis for modulators
William created/logging PVs monitoring BV stability at timesindicating various known issues:
1
2
34
1. [station]:MKBVTRIGJITT, detectBV jitt due to trigger
2. [station]:MKBVTOPJITT, beam-time BV jitter
3. [station]:MKBVTHYBACKSWING,known to multi-state if thyratronrecovery is unstable
4. [station]:[tbd], well after pulse,monitor station BV recovery
5. [station]:MKBVTSREDUCED,beam-time jitter after removingtime slot jitter
Ops Klystron Display -> Jitter Report -> TS Jitter
Thyratron instability
Tao Tang on Thyratron instability:• Self breakdown
- High gas pressure / reservoir voltage- Impact next a few pulses, much lower BV (beam drop out)- Can not start modulator
• Random recovery- Low gas pressure / reservoir voltage- Impact almost every pulse- High PFN instability
* Stations require periodic thyratron ranging to stay in “sweet spot”
28
Auto-ranging thyratrons
AIP Modulators include EPICS-controllable thyratron tuning
29
Auto-ranging thyratrons
William, Tao, Andy, et al, testing new control + jitter PVs to devise anauto-ranging scheme1. Short term: One-click remote ranging for PEM
2. Long term: Monitor and automate RF maintenance from MCC
Thy. res.V
Backswing jitter
Beam-time BV jitter
30
Auto-suggest compliment
Read all station data, find best compliment for requested config
1. Choose the best stations2. Tune the worst a.s.a.p.
Devising clearer at-a-glancecompliment summary GUIincluding jitter & drop outs
Link directly to jitter prediction& diagnosis tools
31
Improved RF Load
• One of two potential new loads in development to handleincreased RF power
• Modified all-metal SLAC design w/ fins & new coating toimprove attenuation
• Goal: Mitigate multipactoring, stabilize refl. power from load
* Image courtesy Anatoly Krasnykh
32
Improved RF Load
• Sufficient performance of one test unit demonstrated• Working to improve consistency
• Anatoly looking into more local vendor, provide more control over processing
• Stress test of new load(s) under high power forthcoming
* Image courtesy Anatoly Krasnykh
33
Run 11 Stability Program Goals
• Attain/maintain ρ/2 jitter (0.05% SXR, 0.25% HXR)• Demonstrate reduced E-jit impact to SS expt’s to users• Continue rollout of AIP modulators• Develop tools for automation of linac maintenance
• Auto-diagnosis, -ranging, -compliment• RoboDecker™
• Demonstration of performance enhancement from newSLAC RF loads
• Improve ESA pulse stealing transverse stability
The Stability Program meets semi-regularly at 2pm on Fridays. Contact Tim Maxwell to be notified.
top related