Transcript Kicking the Can Down the Road (SCU*s)*

D. Arbelaez
for the LBNL LCLS SCU team
Mar. 3, 2016
1
Outline
Magnetic design
– Periodic section
– End design and corrections
– Periodic field corrections
Coil fabrication
Quench protection
Coil testing
Undulator assembly
Measurement results
2
Coil Optimization
Find optimal number of turns per layer and total layers necessary to
meet peak field requirement (1.86 T @ λ = 19.0 mm, g = 8.0 mm)
Wire diameter = 0.6 mm, insulation thickness = 60 μm
5 layers is sufficient to operate at 80 % of load line margin
Modest gains margin can be obtained by adding more layers
Load Line Margin
Load Line Comparison
Number of turns per layer
Critical Current
On-Axis
Field
Peak Conductor
Field
Number of layers
3
End Designs and Non-Ideal Effects
Saturation of the undulator core and poles leads to nonideal effects
– Pole saturation changes the local kick strength
– Pole and core saturation leads to non-ideal global effects
No Core
Saturation
near end
Variation in Pole Saturation
Flux through
the end
 2D calculations are shown to demonstrate the principles
 For accurate results 3D calculation must be used
4
Global Field Effects
Even # of poles, Odd # of coils
cw
On-Axis Field Profile
B
cw
z
ccw
ccw
Odd # of poles, Even # of coils
cw
On-Axis Field Profile
B
ccw
z
ccw
cw
• Number of turns are chosen to cancel these effects in an ideal case
• Global field effects are present due to saturation
5
End Design Principles
2 Independent Correctors
Correction of global field effects
– 1 corrector (coil at each end wired in series) is used for
correction of the global field effect
– corrector produces both a local kick and a global field
– In principle the two ends can be wired independently to
produce both a constant and linearly varying global fields
Correction of local end kick
– 1 corrector at each end wired independently for entrance and
exit kick correction
– This correction is decoupled from the main core and produces
no global fields
– Field clamps are included for this corrector in order to avoid
interference with nearby magnetic components
6
Global Field Correction
Coils are wound in first and last pocket of each core
Produces a global correction + local kick
Strength is chosen to cancel only global field error
Effect of end corrector
global field
local kicks
odd# of poles
7
Local End Kick Correction
Magnetically decoupled from main undulator core
Produces only local kicks
Field clamps are used to minimize stray field
Compact design (fits under splice joint in Nb3Sn device)
Effect of end corrector
no global field
local kicks
odd# of poles
8
Switch-Based Tuning Concept
One superconducting path - with heater
One resistive path (low resistance)
When heater is on the superconducting path becomes resistive (high resistance)
Heaters ON
OFF
Superconducting path
Current
Resistive
solder joint
(low resistance)
9
Current path via lithography on YBCO Tapes
Commercial tape from SuperPower Inc.
Masks designed for photolithography process
Chemical etching used to remove Copper, Silver, and YBCO layers where desired
Solderable thin film heaters were developed for efficient and reliable fabrication
Laser cutting is used to separate joint section
10
Undulator Coil Fabrication: Magnet Core
Mandrel is machined from a single piece of steel
Tight fabrication tolerances were met by the vendor
CMM inspection of the magnet cores was done at various steps to verify
the quality of the parts
Worked with a plasma spray vendor to develop the coating process to
protect against high potential to the magnet core during a quench
11
Undulator Coil Fabrication: Winding
Computer controlled winding machine is used
1.6 km of wire are used for a single core (~ 8000 turns)
Wire braid insulation trials were performed to develop a
thin and robust insulating layer on the wire
Increasing braid angle
(with respect to wire axis)
0.71 mm
Final braid insulation
12
Undulator Coil Fabrication: Heat Treatment
Wind and react process
Large furnace at LBNL can accommodate these coils
Coils are compressed in steel tooling
Heat treatment performed in Ar atmosphere at 650 degrees for 50 hours
(total cycle time is ~ 7 days)
13
Undulator Coil Fabrication: Epoxy Impregnation
Coils are transferred to impregnation tooling
Solder joints are made at the ends to flexible NbTi lead cables
Instrumentation wires are installed in the coils
Impregnation process is carried out
14
Quench Protection
The Nb3Sn undulator operates at extremely high current densities in the
wire (nearly a factor of two greater than NbTi)
Hot spot temperature scales with current squared
Fast detection time and current decay are necessary to protect the Nb3Sn
undulator
Hot Spot Temperature
Magnet Protection Requirement
NbTi
Safe Area
Safe temperature
[A2s]
15
Quech Protection System and Test Results
Quench Protection system uses state of the art electronic components
– Precision isolation amplifiers (Allow high voltage signals)
– IGBT solid state switches (fast switching)
– Dump resistor (fast decay)
Dump resistor value is chosen for fast decay but voltage must also be kept at
reasonable values
Protection system was tested on a four period undulator half
– The coil reached a maximum current of 930 A
– At high currents the rate of decay increases after a short time (suggest quench back behavior)
Current Decay for
short undulator half
16
Coil Testing
Magnet test facility at LBNL used to test individual coils for transport current
Quench protection system was implemented to protect the full length coils
17
Coil Test Results
50
45
40
35
30
25
20
15
10
5
0
7000
6000
5000
4000
3000
2000
1000
0
200
400
600
Peak current (A)
18
800
0
1000
Iits [A2s]
Time Constant [ms]
Quench protection system was successfully applied to the full length coils
550 mΩ resistor was selected (max. voltage = 430 V)
Coil 1 reached 766 A after two quenches
Coil 2 reached 729 A after 80 quenches
Corrector Fabrication
Process to adhere correctors to the vacuum chamber was developed (correctors
and glue take up 0.2 mm of excess thickness)
Short prototypes were fabricated and successfully tested in cryogen free cryostat
19
Undulator Assembly (Cooling System)
Copper plates in contact with undulator poles (Indium interface)
Copper plates are split in order to reduce the effect of thermal
contraction mismatch between materials
Thermal links are used to transfer heat from the plates to the LHe pipe
20
Final Undulator Assembly
Undulator Assembly at LBNL
Lead Connection and LHe Tank Assembly at ANL
21
End Corrections
End corrector response as expected for decoupled end correctors
Observed hysteresis effects for the on-board end correctors
– Higher than expected quadratic component in the magnetic field along the length
– The effect could be minimized with a hysteresis cycle
On-Board End Corrector Response
Decoupled End Corrector Response
22
Magnetic Measurement Results
Hall probe measurements performed at 500 A with coupled end corrector at 2.5 A
De-coupled end corrector is not energized (second field integral corrected numerically)
Later measurements verified that the sensitivity and strength of the de-coupled end
correctors is sufficient to meet the first and second field integral specifications
Magnetic Field
Second Field Integral
First Field Integral
Phase Error
23
YBCO Field Corrector
The YBCO correction network was powered to 50 A
Six of the single-turn correction coils were wired and the functionally of all
six was verified
Two correctors were powered with a main undulator current of 300 A
The two correctors gave a field integral of 30 μTm at 50 A
Effect of Two Correctors on the Second Field Integral of the Undulator
Derived from Hall Probe Measurements
24
Conclusions
Developed basic undulator design and correction methods
– End corrections
– YBCO switch network (central field corrections)
Developed fabrication and processing techniques for Nb3Sn undulators
– Wire and core insulation
– Winding, reaction, epoxy impregnation
Developed quench protection system for high current density undulators
Individual coil test performance was good (~ 95% of design value)
although training behavior between two coils was very different
Good intrinsic field quality was obtained with the Nb3Sn device
Functionality of the YBCO field corrector system was demonstrated
High field in test at ANL was not achieved
– The current limiting is indicative of a local problem (not a fundamental problem with the
technology)
– Will further investigate the cause of the lower than expected current limit
25