1. Activities of TPS in-vacuum undulators
Taiwan Photon Source
Activities of TPS in-vacuum undulators
Jui-Che Huang
Magnet Group
Good Afternoon ,
On behalf of magnet group, I am presenting the status and design work on insertion device.
In this presentation, I will focus on the section of in-vacuum undualtor.

2. Phase-I beamline
In the TPS Phase I, the machine will be equipped with ten insertion devices to serve seven beamlines. All undulators are installed in storage ring and just finished part of commissioning in storage ring.
In the TPS Phase I, the machine will be equipped with ten insertion devices to serve seven beamlines.
This means in long straight section, we will installed two undulators for double mini beta Y to enhance light source.
For short straight section , 3 m undualtor will be installed.
Four IUs have been delivered to NSRRC (IU22-2m-1/2 ,IUT22-3m and IU22-3m-1). Three in-vacuum undulators are under construction.

3. Parameters of IDs in Phase-I
EPU48
IU22-2m
IU22-3m
EPU46
photon energy /keV HP
( ) VP --
( )
current [A]
0.5
λu [mm] 48 22 46
Nperiod 67 95
140 82
By [T]
0.83
0.74*
(1.08)
0.76*
(1.13)
0.73
(0.81)
Bx [T]
0.55
0.47
(0.54)
Kymax
3.72
1.52*
1.56*
3.14
(3.48)
Kxmax
2.47
2.02
(2.32)
Lmech [m]
3.436
2.58
3.57
3.89
Gap [mm] 13 7*
(5.4) 15
(13.5)
number of devices 2
5(1 tapered) 1
Total power[kW]
6.31
3.26*
5.08*
7.05
Power density[kW/mr2]
23.93
30.25*
45.78*
28.58
I will start my talk from ID parameters in phase I.
In phase I, Hard X ray beamline, invacuum undulator will be used as light osurce. Two 2-3 undualtor and five 3-meter IU. One taper undulator is prepared to provide pink light.
For Soft X ray beamline, EPUs are served for light source. Two EPU48 are constructed in Taiwan .
Next section, my colleague will give detail presentation.
* Final operation parameter. EPU46: Bx=By=0.454T, . EPU48: Bx=By=0.46T

4. Performance of IU22
Undulator performance data are from factory acceptance test.
FAT test items
2m-1
2m-2
3mTaper
3m-1
3m-2
3m-3
3m-4
Spec.
Peak field
0.74
0.73
0.78
0.75
0.81
0.72T
Phase error K>1
<2.5˚
<2˚
<3˚
<2.8˚
RMS Trajectory variation [μm] @ gap7mm
<±1.5
<±2
on-axis @gap 7mm 11
-4 / -4* 3 8 -3
-13
<100
Iy on-axis @gap 7mm
26 / 18* 17 -1
-16
IIx on-axis @gap 7mm
2200
180
-1000 /800*
1160
-1100
-1200
-8000
<20000
IIy on-axis @gap 7mm
2300
3300
14000 /12000*
3440
1000
2000
-1000
The filed quality of four delivered ID was shown in table
Phase error are less than 3 degree and it allow user to use 11th harmonics and flux density degration less than 20%.
Frist and second integral is within spec. so this will not affect closed orbit
Vacuum test was done by switch on IP and after NEG activation . the pressure reaches to spec. requirement 3x10-8 Pascal without baking after two days.
After two weeks pumping. Final pressure reaches to ~ 1.3x10-8 Pascal , close to SAT test condition.
* Taper mode operation: minimum gap and maximum taper

5. Measurement at vendor site
Conventional Hall probe
bench measurement
SAFALI measurement
Self-Aligned Field Analyzer with Laser Instrumentation, T. Tanaka, FEL2007, Russia. WEPPH052
NEOMAX Engineering

6. Measurement from Vendor
IU22-3m#2
Gap7mm
IU22-3m#1
Gap7mm

7. Magnetic Measurement at NSRRC
In-situ measurement system is developed at NSRRC. 2 1 3
Optical positioning system – Monitor the Hall sensor positions
Hall sensor moving system – Move the Hall sensor along longitudinal (Z) axis
Rail calibration system – Calibrate the Hall sensor positions in vertical (Y)and transverse (X)axes.

8. Overall View

9. In-situ measurement at NSRRC
Acceptance test results show a good agreement with factory test. However, junction position at two magnet array has 40~60 micrometer displacement after re-assembled /shipment.
Field reduction is due to wide gap
After IU delivered at NSRRC, we took magnetic field measurement using in-situ measurement system which is developed at NSRRC.
Acceptance test results show a good agreement between NSRRC and factory measurement. But we can found the junction position of two magnet array has some field difference.
The gap was widen by 30 micrometer after shipmement.
The trajectory and phase erorr still within spec.

10. Multipole error Device Gap Dipole Quadrupole Sextupole Octupole Spec.
≦100 [G.cm]
≦50 [G]
≦100 [G/cm]
≦100 [G/cm2]
IU22-2m-1
Gap7mm
Normal -4 14 32
-38
Skew 10
-27
-53 98
IU22-2m-2 6 24 42
-15 3
-17
-45 11
IU22-3m-1 5 15 -7 4
-24
-12
IU22-3m-2
-10
-40 40 16
-14 28
IU22-3m-3 -6 12
-23
-20
IU22-3m-4 -8
-26
-34 8 65
-32
IUT22-3m
Gap7mm Non-taper 22 -9 44 20
Taper 2.0mm
-18 23 56 -2
-22 43
The multIpole error is examed by Stretch wire both in NSRRC and Factory site.
We have found integral field is offset due to earth field difference at two sites.
All multipole components are within spec.
The integral field in taper mode are( maximum taper ) does not change the multipole componets.

11. Data comparison (IU22-3m#2)
NEOMAX
NSRRC
Field integral measurement from vendor and NSRRC.

12. Data comparison (IU22-3m#3)
NEOMAX
NSRRC

13. Data comparison (IU22-3m#4)
NEOMAX
NSRRC

14. Data comparison (IU22-3m#4)
Phase error
In-situ measurement system can move only small transverse distance, so we do not know what happened at |x|>2mm.

15. Dynamic field integral
Dynamic field integral maybe small for one IU22, but cumulative effect shall be considered when seven IDs is operation. The field inhomogeneity comes from undulator pole saturation .The pole saturation is caused by screw hole and air venting holes.
The simulation results suggest to edge of pole is easily saturated increase height of pole and remove air venting holes can avoid field saturation.
Measured Data
Although the influence on tune shift is small, We try to understand the reason and find a solution for the inhomogeneity issue.
We simply explain the field inhomogeneity is caused by undulator pole saturation. The pole saturation is caused by screw hole and air venting holes.
The simulation results suggest to edge of pole is easily saturated .
increase height of pole and remove air venting holes can avoid field saturation can avoid field saturation.
Simulation by Radia
(Config. I) has vacuum venting holes on the side of the Permedure block (42mm (W) × 10.5mm (H) × 3.6 mm(D)) , (Config. II) same size, no venting holes. (Config III.) (permedure block : 46mm(W) × 16mm (H) × 3.6mm (D))

16. Dynamic field integral
Configuration I II
III
Max. Dynamic Integrals (|x|<15mm)
14G.cm
3G.cm
2 G.cm
Tuning shift due to quadrupole kick
1×10-5
9×10-6
Tuning shift due to octupole kick
4×10-4
7×10-6
No spectral width broadening range
12 mm
>16 mm
>18 mm
The spectral broadening does not occur in the range of +/-11mm. Therefore, even if we suppose misalignment of the undualtor or closed orbit distortion up to +/- 1mm, the broadening can be avoided.
This has no significant effect on beam dynamic aperture and photon energy spectrum.
The field homogeneity measurement does not match the results from magnetic circuit design. And we can see from this graph, if undualtor gap is closed to 5mm , the field homogeneity become bad.
The field roll-off at +/-15mm is 0.8 and 2.2% respectively.
We calculate dynamic field integral , which is due to electron snake trajectory in the in-homogeneity field.
We can see the dynamic field integral is low as 3 and 14 gauss at gap of 7 and 5mm respectively. We estimate the tune shift by transverse gradient of dynamic integral, the tune shift is 8.0E-6. this is very small.
Another aspect we need to be consider is spectral broadening issue, To avoid spectral broadening due to the non-uniformity, Spectral broadening due to non-uniformity of magnetic field distribution < due to the energy spread.
the necessary uniformity We use three sigma rules to obtain a window width where the above uniformity is necessary.
Beam size σ: 118 μm

17. Installation of TPS Insertion Devices
April ~May 2015
10 insertion devices are installed at TPS tunnel. Installation include Undulator levelling and alignment. We spent 10 weeks to install all insertion devices.

18. Ultra–high vacuum condition
Total baking period is 8 weeks for 7 in-vacuum undulators. The current vacuum pressure is <1.2 x10-8 Pascal. After 3weeks, average current vacuum pressure of 7 IU22 was 9.8 ×10-9 Pascal. Each Undulator baking takes 3days and two undulator baked at same time.

19. TPS Insertion Device

20. Beam test on IU22 7- IU22 closed at minimum gap
First test --with Steering coil correction, no FOFB correction
RMS horizontal close orbit distortion :5.1 μm
RMS vertical close orbit distortion :3.4 μm
Tune shift is around 0.01
Injection efficiency changes is small.

21. Beam-based measurmenet
Undulators
Location
Gap
[mm]
C.O.D [μm]
Before correction
After correction
3m#4
7mSS 6
37.5
1.9
5.8
0.5
Taper 3m
10.2
2.0
4.2
1.5
3m#3
51.3
1.10
3.2
1.4
3m#1
12mSS
69.6
1.2
2m#1 7
21.3
1.1
2.2
0.4
3m#2
35.7
2.3
6.3
1.6
2m#2
22.2
1.0
6.1
0.9

22. TPS Double Undulator configurations
TPS long straight section is designed to install collinear double undulators to enhance photon source.
Original design is two identical in-vacuum undulators of length m were intended to be located collinearly and the separation of two undulators is m.
Two collinear IU22
IU22-3m
Q500
Q600 e-
本張投影片主要表達：
中文：
揭示兩種插件磁鐵於Double mini betay section目前之空間圖面配置。另，為避免插件磁鐵於安裝時，與相鄰之元件產生干涉衝突等事宜，所有插件磁鐵皆先行使用3D工程圖檔與相鄰元件進行空間檢視確認。
EN：
The two figures shown two type insertion device installed in the double mini betay mini section. To avoid components confliction during the ID installation, all undulators have been subjected to inspections with a 3D engineering drawing and confirmed there are no interferences between ID and related components nearby.

23. Sketch of SR heating
ID1 receives radiation only from the bending magnet and the upstream part of the undulator. At ID2, the irradiated power on a magnet surface is from a bending magnet, ID1 and the upstream part of itself.

24. Synchrotron Radiation heating
Heat Source
Double-IU (early design)
Double-IU (latest design)
ID1
ID2
ID2+VAb s
Minimum undulator gap (mm) 5 7
IU length (m)
3.124
2.214
Power density (W/m)
Bending magnet
1.17
1.26
0.41
Upstream IU
1.66
(ID1)
(itself)
0.21
(ID1)
(itself)
Image current*
8.50
6.07
Max. total power density (W/m)
11.33
76.45
6.69
7.69

25. SR heating -Small Gap (5mm)
High Risk to have avalanche meltdown.
ID1 receives radiation only from the bending magnet and the upstream part of the undulator. The maximum linear power density is 2.83 W/m at the end of the magnet array.
At ID2, the irradiated power on a magnet surface is from a bending magnet, ID1 and the upstream part of itself. The total linear power density of SR increases from 30 W/m to 68.0 W/m along the longitudinal distance.

26. Vertical photon absorber
Upstream SR heating is very critical, so in downstream undualtor ,we decide to use 2m instead of 3m undulator. A vertical photon absorber is designed to shield radiation from upstream ID and 2m undulator is moved to further downstream. Allowable minimum gap become 7mm. Still, end part of magnet array will irradiated by upstream radiation. Power of 0.17W will irradiated on magnet foil of ID2.
D=3.946m 3m
ID1
ID2
D=6.615m
ID1
ID2
D=4.946m 3m 2m
D=7.615m e- e-
Vertical photon absorber

27. Vacuum pressure during ID commissioning
Downstream in-vacuum undulator
Downstream in-vacuum undulator

28. Summary
The performance of 7 in-vacuum undulators satisfy the requirements of specification , but measurements from NSRRC and vendor have some difference, especially in stretched wire measurements.
7 in-vacuum undulators has been installed in the storage ring and baking completion. The total period is 3 months.
Beam based measurements are tested at last week, and maximum/minimum RMS closed orbit distortion is around 70 μm and 10 μm ,respectively. After C.O.D correction by steering magnets, the RMS horizontal/ vertical closed orbit distortion is 5μm and 3 μm, respectively.
These two extremely case are made in the same patch and so far , we can not find any relationship between closed orbit distortion and lab measurements data.
Heat load problem was found in double conlinear undulator configuration, the vacuum pressure shows the SR generated by ID1 is hit on ID2.

29. Thank you very much.