MAX IV magnets

MAX IV magnetsMartin Johansson, 29 Jan. 2016

Outline

● MAX IV overview

● the magnets

● how they were made

● some example results

● Located in Lund, Sweden.

● Facility dedicated to synchrotron radiation research. National Laboratory, hosted by Lund University, with ca 180 employees.

● Old MAX-lab located in Lund University campus (MAX I-III) was shut down Dec. 2015 and is currently being decommissioned.

The MAX IV Laboratory

Full energy linac, in operation since 2014.

1.5 GeV storage ring, installation in progress.

3 GeV storage ring, commissioning in progress since Aug. 2015.

Jan. 9th, 2015 photo: Perry Nordeng

Also, an identical copy ofthe 1.5 GeV ring has beenbuilt in Krakow, Poland. Thisfacility, named Solaris, is Poland’s new national synchrotron radiation lab.

klystron gallery

linac~300 mmax 3.4 GeV

The MAX IV accelerators

1.5 GeV ring96 mDBA lattice

3 GeV ring528 m7BA lattice

• Building construction began 2011

• Linac installation fall 2013 – spring 2014

• 3 GeV ring installation fall 2014 –spring 2015

• 1.5 GeV ring installation fall 2015 - …

11/12 achromats installed, awatingshutdown period for TR1 installation

● Through fall 2015:

3 GeV ring commissioning progress

beam in TR3

Aug 11

first turn

Aug 25

stored beam

0.1 mA

Sep 15

stacking

4 mA

Oct 08

first light

Nov 290 mA

Dec 18

Timeline copied from Pedro Fernandes Tavares’ Dec 2015 staff meeting presentation.

● Then, restart after Christmas/new year-shutdown,– 100 mA by end of 1st week

– Commissioning operation continues to cw5

– Shutdown for installation of 2 ID’s, cw6-10

– Then, resume commissioning…

● MAX IV overview

● the magnets



the MAX IV 3 GeV ring magnets

U1 3d cad view with yokebottom and top halves transparent.

QFm1SFmQFm2

x ySD1

SD2 DIP

Key aspects:

● Relatively small magnet aperture of Ø25 mm.

●magnet block concept → integrated unit containing several consecutive magnet elements.

Magnet blocks in supplier’sassembly hall 2014.

Linac and 1.5 GeV ring magnets

● 1.5 GeV ring magnets = conceptually identical to 3 GeV

● Linac and transfer lines = various conventional magnet types

1.5 GeV magnet block at supplier’s fieldmapping bench 2014.

Linac quad and correctors

Bunch compressordipole

TR3 dipole

3 GeV ring lattice

● Each achromat consists of five unit cells and two matching cells.

● 20 achromats x 7 cells = 140 cells total

● achromat length = 26.4 m

● ring circumference = 26.4 x 20 = 528 m

● 0.33 nmrad bare lattice emittance.

● 1320 magnet elements

3 GeV ring achromat

3d cad assembly:

● Each lattice cell is realized as one mechanical unit containing all magnet elements.

M1

U1

U2

U3

U4

U5

M2

● Each unit consists of a bottom and a top yoke half, machined out of one solid iron block, 2.3-3.4 m long.

a MAX IV magnet block:

M1U1

U2U3

U4U5M2

• a U5 bottom half →

• ↓ an assembled U5

Oct. 10th, 2012 photo: Anna Olsson, Scanditronix Magnet AB

Sep. 27th, 2012 photo: Anna Olsson, Scanditronix Magnet AB

Magnet block features

● Dismountable at horizontal midplane.

● all yoke parts = Armco low carbon steel.

● Quad and corr pole tips mounted over the coil ends.

● 6pole and 8pole magnet halfs mounted into guiding slots in yoke block.

● Electrical and water connections located towards the storage ring inside.

M1 bottom and top halves in supplier’s assembly hall 2013.

Magnet block features

● The main structural parts of the magnet blocks are the yoke bottom and yoke top halves.

● Armco low carbon steel

● Dipole profile, and quad + corr. pole roots machined out of the block.

● Cross section size = 350 x 128 mm

● Lengths = 2299 mm (M1, M2), 2418 mm (U1, U2, U4, U5), 3356 mm (U3)

● Weights = ~450 kg (M1, M2), ~490 kg (U1,…), ~620 kg (U3)

U4/U5 yoke bottom half at subsupplier’s 3D CMM, 2012.

• M1/M2 DIPm soft end reduces thermal load to the long straight.

more magnet block detailsM1 with DIPm bottom coil and pole face strip hidden

• Top half aligned to bottom half by 3 guiding blocks on bottom + top outer reference surfaces.

• Field clamps reduce the dipole fringe field distribution sensitivity to coil shape.

field clamp

soft end

guide blocks

guide block

Inside the 3 GeV ring tunnel

2015 photo: Simon C. Leemann

M1 U1 U2 U3U4

U5

concrete support stands

magnet cabling

magnet cooling water hoses

Inside the 3 GeV ring tunnel

2015 photo: Simon C. Leemann

magnet cooling water hoses

vacuum cooling water hoses

bpm cabling

U1U2U3U4

concrete support stands

3 GeV ring magnet block layout

● The 3 GeV ring lattice has two cell types, unit cells and matching cells.

● -MC is mirror image of MC.

● UC1 and UC2 have identical magnet lengths and distances between elements, but different strengths for quads and one sextupole.

● -UC1 and -UC2 are mirror images of UC1 and UC2.

● UC3 is symmetric around its center, with layout identical to other UC.


… correspondingly, there are two basic magnet block layouts, shown here in 3d cad, bottom halves view from top,

● U2, U3, U4, U5 and M2 are identical/mirror of U1 and M1.

x y OXX OXY OYYQFend QDend DIPm SDendxM1:

QFm1 QFm2SFm y x SD1 SD2DIPU1:

Lattice magnet element lengths in mm:

● U2, U3, U4, U5 and M2 are identical/mirror of U1 and M1.

100100



QFm1 QFm2SFm y x SD1 SD2DIPU1:

100 100250 250

100100100150 150

12 slices, total = 1.5°

12 slices, total = 3°

2d simulations made from cross sections exported from the magnet block 3d cad models:

+ SFm, SD, SDend (Ø25 mm), OYY (Ø36 mm), x/y (g = 25 mm).

3 GeV ring magnet elements

DIP: -0.524 T, 8.607 T/m DIPm: -0.524 T, 8.622 T/mg=28 mm at x=0

QF: -40.34 T/mQFm: -37.79 T/mQDend: 25.10 T/mQFend: -36.59 T/mØ25 mm

SFi: -2069 T/m2

Sfo: -1742 T/m2

Ø25 mm

OXY: -65060 T/m3

OXX: 32394 T/m3

Ø25 mm

design procedure

● 2D simulations were performed using FEMM for all magnet elements.

● 3D simulations were performed using Radia for dipoles and quads as standalone magnets, ie no 3D simulations of the full magnet blocks.

● The dipole was evaluated, and represented in the lattice, as consisting of 12 longitudinal slices.

● Lattice and magnet design were iterated against each other.

21/19

Ø25 mm aperture

• Was required from the lattice design, since…

• the pole aperture has a direct influence on lattice compactness,

o by defining minimum distance between elements

o and by defining minimum lengths for quads, 6poles, ...

• The aperture also has an indirect influence on lattice compactness through coil design, in that the required NI is proportional to g for dipoles, r2 for quads, r3 for 6poles, ... making it easer to fit the coil ends longitudinally.

• For the MAX IV 3 GeV ring design, the enabling factor for Ø25 mm magnet aperture is the choice of NEG-coated copper vacuum chambers throughout the achromats.

Ø25 mm aperture

… by defining minimum distance between elements?

• Rule of thumb used for MAX IV: min. distance ≈ one pole gap

• If shorter, fringe fields overlap, destroying field quality.

• We are at this limit in M1 and M2:

• U1-5 are more relaxed, 75 mm between SF and adjacent QF.

100100


100 100250 250

25 mm 25 mm

Ø25 mm aperture

… by defining minimum lengths for focusing elements?

• Assuming a fixed max pole tip field, Bpt, max quad gradient = Bpt/pole r

• We have chosen max Bpt ≈ 0.5 T, which keeps the whole pole face in the linear region of the iron B(H) curve. Resulting max G ≈ 40 T/m.

• Our quads are at this strength, so with larger pole radius, they would have needed to be longer.

• Our strongest sextupoles have ca 0.3 T at the pole tip…

• max Bpt ≈ 0.5 T is quite conservative. It allows us to have a mechanically simple quad design with upright standing pole roots, and racetrack coil. There are no tapered coils in any of our magnets.

Ø25 mm aperture

● A negative aspect is that with smaller pole gap, the relative strength of random field errors due to manufacturing tolerances increase, since the tolerances are fixed, constituting a larger fraction of the pole gap.

● A positive side effect of small aperture is that the powerconsumption goes down. MAX IV 3 GeV ring magnets total ~300 kW.

The magnet block concept

Does is have an impact on compactness?

• In principle no.

• Practically, maybe in how close you dare to place adjacentcoils. But that can just as well be handled through appropriatespec. and QA.

MAX IV choice was made for,

• Vibration stability.

• Ease of installation.

• The magnet block being an alignment concept which is totallyoutsourced to industry.

● MAX IV overview

● the magnets



Specification and procurement

• Production sourced as build to print-contracts for fully assembled and tested magnet blocks, with MAX-lab providing technical specifications and full sets of manufacturing drawings.

• Suppliers responsible for mechanical tolerances, ±20 µm for the yoke bottom and top blocks (2.3-3.4 m long), and for performing field measurements according to MAX-lab spec.

• MAX-lab responsible for magnetic field properties!

• Contracts signed Sept 2011:• Danfysik A/S, Taastrup, Denmark: M1, M2 and U3 = 60 magnet

block units.

• Scanditronix Magnet AB, Vislanda, Sweden: U1, U2, U4 and U5 = 80 magnet block units.

Specification and procurement

● 3D CMM reports for mechanical tolerances for all 140x2=280 yokebottom/top halves, and for all quad pole tips, 6pole/8pole yokes, etc..

● 3D CMM reports for yoke top halves includes measurement ofalignment target locations.

● Specified field measurements for all 140 magnet blocks,

– Hall probe:• Dipole field map at nom. I

• Dipole field map at nom. I + pole face strips at max I

• Quadrupole single transverse lines at nom. I

– Rotating coil:• Quads/6poles/8poles/corr at I = 0, 100, 0 %

• 6poles/8poles trim coils at max I for each connection mode

– extended with more current levels and repeatability tests for every10th magnet blocks.

Expected challenges

• Meeting the ±20 µm mechanical tolerances.

• Performing Hall mapping with the same level of positioningaccuracy.

• Solving rotating coil meas access while also meeting specifiedaccuracy.

• And doing all this with required production pace!

Outcome

Machining – meeting the tolerances,

• The two magnet suppliers subcontracted machining of the large yoke bottom/top halves to three different CNC machiningcompanies, • Required extra visits to these companies at intervals in the early

stage, and in depth review of 3d CMM tolerance verification.

• But no hands on problem solving in how to perform the machining. Our conclusion: ±20 µm is a tolerance level that canbe met without needing a key person at the accelerator lab togo into the details of how to perform the machining.

• Required continuous monitoring from us to keep pace.

Outcome

Field measurement,

• Both suppliers chose to procureboth new Hall mapping benches, and new rotating coil systems,• Necessitated by meeting the

specified accuracy, and/or keepingtheir previous existing measurement equipment availablefor other customers during the MAX IV production phase.

• Also, rotating coil mechanicallayout necessitated by our magnet block concept.

• Establishing the measurementprocedures according to specrequired both in depth review and hands on problem solving from our side.

Rotating axis

Outcome

• In addition to the expected challenges, the total workload wasalso challenging for the suppliers. Project meetings wererequired at monthly, bi-weekly, or at times even weeklyinterval.

• A positive outcome was that during the project we did not sacrifice any performance requirements!

Production timeline

2011 Sept. Contracts signed

2012 spring start of yoke machining

2012 Sept. 1st yokes approved

2012 fall 1st magnets assembled

2012/2013 fall/winter field measurement systems operational

2013 fall coming into series production phase…

Production statistics from suppliers weekly status reports,

• Yoke machining pace = 1-2 yoke halves/week average

• Field meas pace = 1 Hall and 1 rot. coil /day possible

• Production series completed summer 2014

Series production phase

Danfysik M1, M2 and U3: Scanditronix Magnet U1, U2, U4 and U5:

0

10

20

30

40

50

60

yoke machining

assembled

rotating coil

Hall probe

delivered to MAX IV

0

10

20

30

40

50

60

70

80

yoke machining

assembled

rotating coil

Hall probe

delivered to MAX IV

● MAX IV overview

● the magnets



3 GeV ring magnet field meas. results

U1,2,4,5 quads spread in strengthby rotating coil (160 magnets):

• max-min ≈ 0.7 % total

• or ≈ 0.5-0.6 % amonginner/outer

• Expected from mech. tolerance±25 µm at r = 12.5 mm: ± 0.4 %

0

0,1

0,2

0,3

0,4

0,5

0,6

0,7

0,8

mai

n t

erm

, (ab

s(sc

ale

d)-

no

m)/

no

m [

%]

outer QFm

inner QFm

outer QF

inner QF


U1,2,4,5 SD sextupoles spread in strength by rotating coil (160 magnets):


• or ≈ 1.2 % among inner/outer


-0,2

0

0,2

0,4

0,6

0,8

1

1,2

1,4

1,6

mai

n t

erm

, (ab

s(sc

ale

d)-

no

m)/

no

m [

%]

inner SXD

outer SXD

Examples indicate that spread in strengths agree with expected.


M1,2 OXX/OXY octupoles spreadin strength by rotating coil (80 magnets):


• No difference in averagestrength between OXX and OXY.


0

0,2

0,4

0,6

0,8

1

1,2

1,4

mai

n t

erm

, (ab

s(sc

ale

d)-

no

m)/

no

m [

%]

OXX

OXY


U1,2,4,5 DIP spread in integratedgradient by Hall probe (80 magnets):

• max-min ≈ 0.5 %, similar toquads.

-0,1

0

0,1

0,2

0,3

0,4

0,5

0,6

intB

', (a

bs(

scal

ed

)-n

om

)/n

om

[%

]

DIP pfs=0 U1,2,4,5

Field quality example, rotating coil measured higher order harmonic content for 10 quads:

• Expected from mech. tolerance ±25 µm at r = 12.5 mm: 10-15 1E-4 error terms in 2d simulations of different “worst case” displacements of poles.


-10,0

-8,0

-6,0

-4,0

-2,0

0,0

2,0

4,0

6,0

3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

@r=

10

mm

, Bn

[1

E-4

]

n

U2-19

U2-18

U1-19

U1-18

U1-17

U5-19

U1-20

U2-20

U5-20

U5-16-8,0

-6,0

-4,0

-2,0

0,0

2,0

4,0

6,0

8,0

3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

@r=

10

mm

, An

[1

E-4

]

n

U2-19

U2-18

U1-19

U1-18

U1-17

U5-19

U1-20

U2-20

U5-20

U5-16


Rotating coil meas, magneticcenter offsets to rotating axis by harmonic content feed down:

same, subtracting linear fits, shows relative alignment of 4 magnet elements over 675 mm:

-0,1

-0,08

-0,06

-0,04

-0,02

0

0,02

0,04

0,06

0,08

0,1

0 500 1000 1500 2000 2500dx

[mm

]

z [mm]

M1-16

M1-18

M1-19

M1-20

M1-15

M1-17

QDend

OXY

QFendOXX

-0,1

-0,08

-0,06

-0,04

-0,02

0

0,02

0,04

0,06

0,08

0,1

0 500 1000 1500 2000 2500

rela

tive

dx

[mm

]

z [mm]

M1-16

M1-18

M1-19

M1-20

M1-15

M1-17

Qdend OXY QFend OXX

elements

length[mm]

eval. [pcs]

rel. align

min [µm]

max [µm]

RMS [µm]

M1,2 4 675 39/40 dx -10 12 3

dy -27 20 9

U1,2,4,5 3 400 79/80 dx -16 12 4

dy -24 30 6

U3 4 928 20/20 dx -10 13 5

dy -41 29 18


Relative alignment within magnet blocks, statistics for 39 pcsM1/M2, 79 pcs U1,2,4,5, and 20 pcs U3:

• dy includes rotating shaft sag.

• MAX IV DDR requirement = 25 µm RMS with 2σ cut-off.

• Our conclusion: rotating coil measurements indicate that MAX IV magnet block alignment concept works!

0

5

10

15

20

25

30

-0,04 -0,03 -0,02 -0,01 0,00 0,01 0,02 0,03 0,04

no

of

mag

ne

ts

relative dx [mm]

M1,M2

3 GeV ring commissioning- from the magnet perspective:

lattice correctors

Aug 11 beam in TR3

Aug 25 first turn nominal 0

many turns nominal manually tuned

Sep 15 stored beam 0.1 mA nominal -||-

Oct 8 stacking 4 mA QFE, QDE adj. -||-

meas. bpm-offsets LOCO -> all Q adj, to differentvalues in different achromats.

kicker changed from 3.5 to 1.7 µs

Dec 18 90 mA -||- SOFB

• nominal = current set values calculated from the suppliers field measurements.

• Conclusion, from the magnet perspective, is that alignment and fieldmeasurements accuracy was OK!

Thank you for your attention!

Extra slides …

3 GeV ring magnet block features

● 3d cad view from top of U1 magnet block bottom half with vacuum chamber in place:

● Vacuum chamber bolted to magnet block at bpm and two support brackets.

● Vacuum chamber cooling and signal connections located towards outer side.


● 3d cad view from outer side of M1 magnet block with vacuum chamber in place, on concrete support stand:


● U1 magnet block with plastic cover removed; view from inner side, where cooling and electrical connections are located:

● Water cooling circuits are separate for bottom and top halves.

● Electrical connections across the midplane has to be disconnected when dismounting top half – bus bars/external cables for main coils and plug in contacts for corr, trim and thermoswitches.

top half cooling water inlet

bottom half cooling water inlet

top half cooling water outlet

bottom half cooling water outlet

QFm1 terminals QFm2 terminals

SFm terminals

DIP main coil & pfs terminals

shunt boardSD1 terminals SD2 terminals

trim coils and corrector plug-in contacts

Interlock D-sub contacts

… so, one pole gap distance between consecutive elements, doesit work?

• Compare QFend and QDend harmonic content:

• These two elements see different iron surroundings, differencein n=4 could come from this …

• … but this difference is negligible. Ie, our data data indicatesthat in this case one pole gap distance is sufficient.

-10

-8

-6

-4

-2

0

2

3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

@r=

10

mm

, har

mo

nic

co

nte

nt

[1E

-4]

n

QDend production series average

normal

skew

Distance between magnets?

-10

-8

-6

-4

-2

0

2

3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

@r=

10

mm

, har

mo

nic

co

nte

nt

[1E

-4]

n

QFend production series average

normal

skew

Documents

MAX IV magnets