motivation: why to do magnetic · pdf filetutorial: “ overview of magnetic measurements...
TRANSCRIPT
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 1/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Mot ivat ion: why to do magnet ic measurements ?
Instrument types and how to choose them
Dynamic effects (eddy currents)Non-linear ef fects (saturation and hysteresis)
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 2/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 3/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
When t o measure magnet s ?
dif f erent t rade-of f s bet ween accuracy and resources at dif f erent t imes
design phase: test material samples for permeabilit y, coercivit y etc…test prototypes or models (scaled down versions) to validate computer simulat ionsand specif ic design choices (e.g. chamfers, shims, many other details …)
prot ot ypes/ pre-series: test f ield qualit y to verify the respect of mechanical tolerances (inverse problem),give feedback to designer and manufacturing f irms. Carry out a fully detailed magnet ic characterizat ion (of ten the t ime to do so will not be available during series tests)
accept ance t est s: monitor product ion qualit y,t rap errors, tooling wear …
as early as possible to steer manufacturing.Build up stat ist ics to reduce tests and minimize
total cost . Get all data required for f iducializat ion (installat ion) and beam opt ics.
NB: internal acceptance criteria might be elast ic, but legal acceptance is binary
(and may be even obligatory !)
t hroughout l if et ime: characterize magnets af ter repairs,or to allow use in dif ferent ways than originally intended
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 4/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Mat hemat ical modelsAdvant ages• predict behaviour wit hout having t he physical object (!!!)• f ast and inexpensive for relat ively simple cases; allow parameter space searches, opt imizat ion• virtually unl imit ed resolut ion and precision
Limit at ions• part ial physical model: including all couplings (thermal, mechanical) and phenomena
(magnetost rict ion, magnetoresist ivit y …) that may be relevant is extremely expensive• numerical errors: e.g. singularities in re-entrant corners, boundary location of open regions; these
may spoil results. Special techniques (special corner elements, BEM) require skill and time• high cost of detailed 3D models ∝ ∆x2~3, ∆t-1 (2D simulations not always suff icient …)
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 5/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Impact of model uncert aint ies
10 µm/100 mm gap error→ 10-4 f ield error at high f ield5% µr error → 5⋅10-3 f ield error at low f ield (typical values)
Analytical 1D model (assume no leakage, constant sross-section),typical accuracy 10-1∼10-2
leak
𝐵𝐵 =𝜇𝜇0𝑁𝑁𝑡𝑡𝐼𝐼
𝑔𝑔 1𝜇𝜇𝑟𝑟ℓ𝑔𝑔 + 1
𝑔𝑔𝐵𝐵𝜕𝜕𝐵𝐵𝜕𝜕𝑔𝑔
= −1
1𝜇𝜇𝑟𝑟ℓ𝑔𝑔 + 1
impact of geometrical uncertainty (mechanical tolerances, assembly errors)
5 10 50 100 500 1000Aspect ratio
g0.005
0.010
0.0500.100
0.5001.000 µr= 10000
µr= 1000
µr= 100
µr= 10
µr= 1
impact of material property uncertainties
𝜇𝜇𝑟𝑟𝐵𝐵𝜕𝜕𝐵𝐵𝜕𝜕𝜇𝜇𝑟𝑟
=1
1 + 𝜇𝜇𝑟𝑟𝑔𝑔ℓ
5 10 50 100 500 1000Aspect ratio
g0.001
0.01
0.1
1 µr= 10000µr= 1000µr= 100µr= 10µr= 1
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 6/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
FE/ MM comparison (1/ 2): CER PS Boost er Dipole
2D FE with nominal B(H)(tweaking the curve does not work !)
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
H [A/m] 10 4
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
B [T
]
steel plates to be reinforced toequalize the rings at high f ield
(+110% @ 2 GeV w.r.t. design value !)
Outer rings (higher saturation)2% mismatch FE/MM
Inner rings(high f ield, little saturation)
0.3% mismatch FE/MM
Courtesy A. Newborough, R Chritin
Integralf lux loops
4-ring main bending dipoleof CERN PS Booster
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 7/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
FE/ MM comparison (2/ 2): MedAust ron Bending Dipole• modelling issues more complex for dynamic phenomena (eddy currents)• medical hadrontherapy machine requirements: fast energy changes, high accuracy and stabilit y • set t ling t ime: measured 200 ± 20 ms, computed 150 ms
G. Golluccio, A. Beaumont et al., Overview of the magnetic measurements status for the MedAustron project, IMMW18T. Zickler et al., Design and Optimization of the MedAustron Synchrotron Main Dipoles, IPAC11
Integral PCB f luxmeterMeasured eddy current
decay transient
~ 2M elements, 80 h running time
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 8/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 9/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Induct ion sensors
Fixed coil in a t ime-changing f ield(f luxgate, AC st retched wire)
see J. Di Marco IMMW19−𝑉𝑉𝑐𝑐 =
𝜕𝜕Φ𝜕𝜕𝜕𝜕
=𝑑𝑑𝑑𝑑𝜕𝜕𝒜𝒜𝑩𝑩 𝒏𝒏 𝑑𝑑𝑑𝑑 =
𝒜𝒜
𝜕𝜕𝑩𝑩𝜕𝜕𝜕𝜕
𝒏𝒏 𝑑𝑑𝑑𝑑 + 𝜕𝜕𝒜𝒜𝒗𝒗 × 𝑩𝑩 𝑑𝑑ℓ
𝑑𝑑𝑐𝑐𝜕𝜕𝐵𝐵𝜕𝜕𝜕𝜕
𝑑𝑑𝑐𝑐𝐵𝐵𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕
𝑑𝑑𝑐𝑐𝜕𝜕𝐵𝐵𝜕𝜕𝜕𝜕
𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕
𝑙𝑙𝑤𝑤𝐵𝐵𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕
Wire translating in a DC f ield(classical DC stretched wire)
Coil rotating in a DC f ield
Coil translating in a DC f ield
• Sensitive to the f lux rather than the f ield
• Intrinsically l inear t ransducer(some nonlinearity indirectly due to acquisition electronics e.g. f inite input impedance → small currents circulate in the coil)
• Volt age int egrat ion generally required, with attendant advantages (noise abatement proportional to frequency) and drawbacks (drift, results depend on integration time)
• Direct post-processing of the voltage also possible(relies on accurate speed control and measurement)
see G. CaiafaIMMW20
see Z. WolfIMMW19
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 10/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Lorent z f orce-based sensors
see Prof . Popovic, M. CalviIMMW20
see Z. Wolf , IMMW19
𝑭𝑭𝑞𝑞
= 𝑬𝑬 + 𝑩𝑩 × 𝒗𝒗
𝜕𝜕𝑭𝑭𝜕𝜕𝑙𝑙
= 𝑩𝑩 × 𝑰𝑰
𝑉𝑉𝐻𝐻 = 𝑘𝑘𝐻𝐻𝐵𝐵 𝐼𝐼
∆𝜌𝜌𝜌𝜌
= 𝑘𝑘𝐵𝐵2
Vibrating wire
Hall ef fect sensor
Magneto-resistive sensor(hardly used in our f ield)
• Sensitive to a single f ield component (with higher order correction terms)
• Mechanical/solid state phenomena → stronger non linearity → more dif f icul t cal ibrat ion
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 11/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Magnet ic resonance sensors
𝑩𝑩 =𝟐𝟐𝝅𝝅𝜸𝜸𝒇𝒇
see P. Keller, IMMW19G. Boero, IMMW20
• Resonant absorpt ion/re-emission of RF waves in a sample within a uniform f ield (f ield gradient spreads the resonance, impact depends on sample size and shape)
• Transducer sensit ive to the f ield vect or norm(some impact of temperature, orientat ion of t ransducer, chemical nature of sample: < 10-6 for NMR, much st ronger for EPR)
• Gyromagnet ic rat io depends on fundamental constants → met rological golden st andard
𝛾𝛾 = 𝑔𝑔𝑞𝑞
4𝜋𝜋𝑚𝑚= H+(proton) 42.577
free electron 28 015.737𝑀𝑀𝑀𝑀𝑀𝑀 𝑇𝑇−1
NMR (Nuclear Magnetic Resonance)γ constant known to better than 1 ppm
EPR/ESR (Electron Paramagnetic/Spin Resonance), FMR (FerriMagnetic Resonance)actual value in materials depends on: chemical composition, temperature, direction …
r
µ (magnetic moment)
angular velocity ω=2πfangular momentum l
charge qmass m
precession velocity Ω
field Bmagnetic torqueT=µB=γl
Gyromagnetic ratio
Zeeman (quantum)correction factor
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 12/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Magnet ic measurement dat af low
Physical principle
Met hod Rawdat a
Int ermediat eresult
Fieldrepresent at ion
Finalresult s
Induct ion
𝑉𝑉𝑐𝑐 =𝜕𝜕Φ𝜕𝜕𝜕𝜕
f ixed coil𝑉𝑉𝑐𝑐 = 𝑑𝑑𝑐𝑐𝐵
𝑉𝑉𝑐𝑐(𝜕𝜕) Δ𝛷𝛷(𝜕𝜕)
long. field integral vs. time, transv. pos.
1𝑤𝑤𝑐𝑐ℓ𝑐𝑐
Δ𝐵𝐵 𝜕𝜕, 𝜕𝜕, 𝑠𝑠 𝑑𝑑𝑠𝑠 𝑑𝑑𝜕𝜕
Commonly required results:
• f ield polarit y• integrated/ local f ield st rength
(main harmonic)• f ield direct ion
(phase of main harmonic),• integrated/ local f ield errors
(higher harmonics)• magnet ic axis
(t ransversal posit ion)• magnet ic axis
(pit ch and yaw angles)• magnet ic center
(longitudinal)
vs. t ime, current , excitat ion history, environmental condit ions etc.
rot at ing coil𝑉𝑉𝑐𝑐 = 𝜔𝜔𝑑𝑑𝑐𝑐𝐵𝐵
𝑉𝑉𝑐𝑐(𝜗𝜗) 0
2𝜋𝜋
𝛷𝛷 𝜓𝜓 𝜅𝜅 𝜗𝜗 − 𝜓𝜓 𝑑𝑑𝜓𝜓avg. field expansion coefficients
Cn = Bn + i An
t ranslat ing coil
𝑉𝑉𝑐𝑐 = 𝑠𝑑𝑑𝑐𝑐𝜕𝜕𝐵𝐵𝜕𝜕𝑠𝑠
𝑉𝑉𝑐𝑐(𝑠𝑠) −∞
∞
𝛷𝛷 𝑢𝑢 𝜅𝜅 𝑠𝑠 − 𝑢𝑢 𝑑𝑑𝑢𝑢
long. avg. field profile vs. transv. pos.
1𝑤𝑤𝑐𝑐
𝐵𝐵 𝑠𝑠, 𝜕𝜕 𝑑𝑑𝜕𝜕
moving wire
𝑉𝑉𝑐𝑐 = 𝜕𝑑𝑑𝑐𝑐𝜕𝜕𝐵𝐵𝜕𝜕𝜕𝜕
𝑉𝑉𝑐𝑐(𝜕𝜕) Δ𝛷𝛷 𝜕𝜕
long. avg. field integral
1Δ𝜕𝜕 ℓ𝑤𝑤
Δ𝐵𝐵 𝜕𝜕, 𝑠𝑠 𝑑𝑑𝑠𝑠 𝑑𝑑𝜕𝜕
Lorent z f orce
𝜕𝜕𝜕𝜕𝜕𝜕𝑠𝑠 = 𝐵𝐵 𝑠𝑠 𝐼𝐼(𝜕𝜕)
vibrat ing wire𝛿𝛿𝑥𝑥(𝑠, 𝜕𝜕)𝛿𝛿𝑦𝑦(𝑠, 𝜕𝜕)
eigenmode amplitudes
𝛿𝛿𝑥𝑥𝑥𝑥𝑠𝑠𝑠𝑠𝑠𝑠(2𝜋𝜋𝑠𝑠𝑠𝑠ℓ𝑤𝑤
)
𝛿𝛿𝑦𝑦𝑥𝑥𝑠𝑠𝑠𝑠𝑠𝑠(2𝜋𝜋𝑠𝑠𝑠𝑠ℓ𝑤𝑤
)
1𝛿𝛿𝑦𝑦𝑥𝑥
𝐵𝐵𝑥𝑥𝑥𝑥 𝑠𝑠 𝑠𝑠𝑠𝑠𝑠𝑠(2𝜋𝜋𝑠𝑠𝑠𝑠ℓ𝑤𝑤
)𝑑𝑑𝑠𝑠
1𝛿𝛿𝑥𝑥𝑥𝑥
𝐵𝐵𝑦𝑦𝑥𝑥 𝑠𝑠 𝑠𝑠𝑠𝑠𝑠𝑠(2𝜋𝜋𝑠𝑠𝑠𝑠ℓ𝑤𝑤
)𝑑𝑑𝑠𝑠
Hall ef f ect𝑉𝑉𝐻𝐻 = 𝑘𝑘𝐻𝐻𝐼𝐼𝐻𝐻𝐵𝐵
1D/ 2D/ 3D probes 𝑉𝑉𝐻𝐻(𝜕𝜕) 𝐵𝐵(𝜕𝜕)
Magnet icResonance
𝑓𝑓 = 𝛾𝛾 𝑩𝑩
NMR/ EPRprobes 𝑉𝑉𝑅𝑅𝑅𝑅(𝜕𝜕) 𝑉𝑉𝐿𝐿𝑅𝑅(𝜕𝜕) 𝑩𝑩 =
𝑓𝑓 𝜕𝜕𝑟𝑟𝑟𝑟𝑟𝑟𝛾𝛾
s : longitudinal coordinate; x,y : transversal coordinates. B : magnetic field component normal to the coil, the movement of the wire and to the Hall sensor
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 13/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Overview of magnet ic inst rument sInst rument B
[T]B.W.[Hz]
𝝈𝝈𝑩𝑩𝑩𝑩
Sensorsize Remarks
Rot at ing-coil f luxmet er >10-4 ~DC 10-4 ∅8-350 mm
30 – 1300 m
• full 2 D f ield information (absolute and relative, integral or local): strength, multipoles, axis and direction
• coil bucking → higher multipoles at ppm resolution, decreased sensitivity to mechanical imperfections
• time resolution up to ~0.1 s
Fixed-coilf luxmet er >10-4 >10-2 10-4 < 7 m
• natural (and only) option for very fast pulsed magnets• allows easy dynamics studies (eddy current and history-
dependent effects)• integration constant requires separate measurement
Translat ing-coilf luxmet er >10-4 DC 10-4 ~100 mm • adaptable to curved or very long magnets
• longitudinal f ield profile requires deconvolution
St ret ched wire (moving) >10-3 DC 10-4 ∅ 0.1 mm< 20 m
• calibration reference for integral f ield strength, direction and axis (precision of the XY stages)
• equivalent to 1-turn variable-geometry coil• best geometrical f lexibility (long magnets, narrow gaps)
St ret ched wire (vibrat ing) >10-3 DC 10-4 ∅ 0.1 mm< 20 m
• extremely sensitive for axis (at resonance)• only option for harmonics in small gaps• longitudinal resolution possible via FFT (λ>0.1 m)
Hall probe >10-4 <104 ~10-3 <1 mm2 • widespread, vast range of commercial options• high accuracy requires laborious calibration
NMR probe >0.043 <20 10-6 1 cm3• metrological golden standard• works only in highly uniform f ields• limited bandwidth; provides f ield vector norm
Fluxgat e >10-8
<10-3 <102 10-3 1 cm3 • geomagnetic and environmental f ield applications• fringe f ields, residual f ield, safety
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 14/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Typical t ransversal vs. longit udinal size
1.E-03
1.E-02
1.E-01
1.E+00
1.E+01
1.E+02
1.E+03
1.E-02 1.E-01 1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05
Fixed coils
Hallsensor
Rotatingcoils
FMR
Stretched wires
Longitudinal sensor size (mm)
NMR
Traj
sver
sals
enso
r size
(mm
)
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 15/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Typical accuracy vs. f ield range1.E-07
1.E-06
1.E-05
1.E-04
1.E-03
1.E-02
1.E-011.E-12 1.E-11 1.E-10 1.E-09 1.E-08 1.E-07 1.E-06 1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02
Fixed/rotating coils
Hall sensor
FMR
Stretched wires
NMR
Sensor field range [T]
rela
tive
unce
rtai
nty
[-]
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 16/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Accuracy vs. t est t ime
The st andard uncert aint y of an inst rument is a cert if ied funct ion of the operat ing condit ions (f ield range/ f requency, gradient , temperature etc. … ) can be further improved based on the t ime and ef f ort taken
• Repeat to get rid of random errors: σ 𝜕𝜕 = 𝜎𝜎 𝑥𝑥𝑥𝑥
(diminishing returns for large n)
• Oversample (time/angle) to reduce aliasing(e.g. MHz sample rate for kHz bandwidth → much improved drift correction)
• Flip and repeat to estimate and subtract systematic errorseither the magnet or the instrument, as is more practical
• Reverse polarity to recover ambient or intrinsic offsets(e.g. remanent f ield)
• Redundant takes will give you confidence !
αα1
meas∆α
-α
α2meas
∆α
B
reference axis
⇒
I [A]
x
+I
-I
𝜕𝜕(0) =𝜕𝜕 𝐼𝐼 + 𝜕𝜕(−𝐼𝐼)
2𝑑𝑑𝜕𝜕𝑑𝑑𝜕𝜕
=𝜕𝜕 𝐼𝐼 − 𝜕𝜕(−𝐼𝐼)
2
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 17/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Inst rument select ion crit er ia
1) Compat ibil it y with f ield level/gradients (could not work at all!)2) Transverse size (it must f it , and should reach as wide as possible)
- local ripple close to the pole may degrade the accuracy of harmonics- ext rapolat ion further f rom the axis can be applied, at a cost
3) Bandwidt h- sensit ivit y may drop above cutof f- addit ional errors e.g. f rom induct ive cable loops
4) Longit udinal size- the integral can be computed by scanning longitudinally (t ime-consuming)- de-convolut ion of longitudinal scans done with a longer probe → low-pass f ilter, noise
5) Accuracy- uncertainty can be reduced by repetition, changing orientation, cross-checks …
6) Result f ormat : harmonics vs. map (1D/2D/3D)- can be translated into one another, with caveats
7) Pract ical considerat ions: cost, measurement time,output signal format, cabling length,commensurate size of sensors and magnet,availability of trained personnel …
“hard” criteria
“soft” criteria
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 18/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 19/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
CERN ELENA bending dipole
• Dif f icult case: 60° bending, low-energy pbar ring → low f ield (50 to 420 mT),accelerating & decelerating cycles, 2 min-long long e-cooler plateaux
• 0.5 mm laminations with high dilution 2:1 electrical steel M270-50 A HP/304 L to reduce hysteresis; 13° cut angle for focusing
• Measured with Litz-wire f luxmeter (see O. Dunkel, IMMW19), with 2% coil area uncertainty originally intended as a backup for higher quality PCB unit
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 20/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Mult i-ref erence cross-cal ibrat ion
(1) 2D Hall probe mapon the mid-plane
Courtesy of Lucio Fiscarelli
(2) Cent ral NMR for high-f ield calibrat ion
(3) Integral calculated along st raight linesmatched to St retched Wire
(5) Integral of B’ interpolated on a curvedpath to match f luxmeter coil
𝑤𝑤𝑟𝑟𝑒𝑒𝑒𝑒 =Φ0 + ∫0
𝑡𝑡 𝑉𝑉𝑐𝑐𝑑𝑑𝜕𝜕
∫−∞+∞𝐵𝐵′(𝑠𝑠)𝑑𝑑𝑠𝑠
(4) Hall probe gain and offset calibration
𝐵𝐵′ = ∆𝐵𝐵 + 1 + 𝜀𝜀 𝐵𝐵
independentremanent f ieldmeasurement
pulsed-modefluxmeter coilmeasurement
f inal result(3× 9-coil f luxmeters)
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 21/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Is measuring along a curve t ruly necessary ?
• Simplest example: hard-edge dipole field distribution, uniform along s
• Let bn be the transversal harmonic expansionof the integral along straight lines
• Let βn be the transversal harmonic expansionof the integral along arcs of radius R(approximated by parabolas with negligible loss of accuracy)
• bn and βn are connected by a feed-down-like linear relation:
𝑏𝑏1𝑏𝑏2𝑏𝑏3𝑏𝑏4𝑏𝑏5𝑏𝑏3
=
1𝛼𝛼
12𝛼𝛼2
80𝛼𝛼3
448𝛼𝛼4
2304𝛼𝛼5
11264
0 1𝛼𝛼6
3𝛼𝛼2
80𝛼𝛼3
1125𝛼𝛼4
2304
0 0 1𝛼𝛼6
3𝛼𝛼2
405𝛼𝛼3
224
0 0 0 1𝛼𝛼3
𝛼𝛼2
8
0 0 0 0 15𝛼𝛼12
0 0 0 0 0 1
𝛽𝛽1𝛽𝛽2𝛽𝛽3𝛽𝛽4𝛽𝛽5𝛽𝛽3
𝐵𝐵 𝜕𝜕, 𝑠𝑠 𝑑𝑑𝑠𝑠 = 𝐵𝐵0 1 + 𝑏𝑏2𝜕𝜕𝑎𝑎
+ 𝑏𝑏3𝜕𝜕2
𝑎𝑎2+ …
𝜕𝜕 field integrals along straight and curved pathscontain the same amount of information
𝐵𝐵 𝜕𝜕, 𝜁𝜁 𝑑𝑑𝜁𝜁 = 𝐵𝐵0 1 + 𝛽𝛽2𝜕𝜕𝑎𝑎
+ 𝛽𝛽3𝜕𝜕2
𝑎𝑎2+ …
𝜕𝜕𝜁𝜁
𝑠𝑠
𝐿𝐿
2𝑎𝑎
𝑅𝑅 =𝐿𝐿2
8𝛼𝛼𝑎𝑎
𝜕𝜕𝑎𝑎≈ 𝛼𝛼
2𝑠𝑠𝐿𝐿
2
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 22/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 23/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Measurement of eddy current ef f ect s• Method: current plateaux of durat ion >>
expected τ
• High-speed acquisition of integral coil voltage →detailed prof ile of I*-Im
• The relative amplitude a/Im(t s) and logarithmic decay ratio τ of the exponential starting at the end of the ramp = eddy current effect
• Scaling law for the time constant:τ = 𝐿𝐿𝑒𝑒
𝑅𝑅𝑒𝑒∝ ℓ
ℓ ℓ2= ℓ2
5000
4500
4000
3500
3000
2500
2000
1500
1000
-500
0
500
1000
1500
2000
2500
3000
3500
4000
4500
5000
2 3 4 5 6 7 8 9 10 11 12time (s)
-2.50
-2.25
-2.00
-1.75
-1.50
-1.25
-1.00
-0.75
-0.50
-0.25
0.00
0.25
0.50
0.75
1.00
1.25
1.50
1.75
2.00
2.25
2.50
I[A]
Vcoil[V]
-500
0
500
1000
1500
2000
2500
3000
2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0time (s)
I (A)
B scaled (A)
-20
-10
0
10
20
30
3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0time (s)
2996.0
2996.2
2996.4
2996.6
2996.8
2997.0
2997.2
Im(t)
Vcoil(t)
Im(t)
apparent advance / lagdue to saturation
exponential decay
I*(t)-Im(t)
τ
a
I*(t)
ts
𝐼𝐼∗ 𝜕𝜕 =𝑑𝑑𝑐𝑐𝐿𝐿𝑐𝑐𝑐𝑐
𝐵𝐵 𝜕𝜕 =𝐼𝐼𝑐𝑐 𝜕𝜕𝑟𝑟𝐵𝐵 𝜕𝜕𝑟𝑟
𝐵𝐵 𝜕𝜕
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 24/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Eddy current -cancel ing overshoot
3 τ to complete stabilization
1.5 τ to complete stabilization(theoretical …)
end oframp-up
0 5 10 15
Time (s)
0
0.1
0.2
0.3
0.4
0.5
0.6
Bdl (T
m)
Bdl (Tm)
1 2 3 4 5 6
20
40
60
80
100Im
∆t h
1.8 2.0 2.2 2.4 2.6 2.8 3.0
20
15
10
5
0
I*-Im
• Eddy currents can be part ially, totally or over-canceled by a t riangular current overshoot at the end of ramp-up• Example: stable f lat -top reached at the t ime cost of ∼1.5τ (to be compared with exponent ial decay t ime ∼3τ)• Caveats: - power converter needs high dV/dt ;
- the maximum working point may increase considerably, at the risk of saturat ion- hysteresis → f inal f ield level changes (new limit cycle, st ill OK if stable)
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 25/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Fast -cycled quads
-10 0 10 20 30 40 50 60 70-0.5
0
0.5
1
1.5
2
-10 0 10 20 30 40 50 60 70-0.04
-0.03
-0.02
-0.01
0
0.01
0.02
0.03
• Example: fast capacit ive discharge powering of Linac4 inter-tank EMQs
• current spikes lead to minor hysteresis loops →f ield reproducibility degradation
• oscillations at the end of the ramp-down may provide a beneficial “f ree” degaussing, if symmetrical
• the overshoot at the end of the ramp-up may give a more stable f lat-top, but makes it less reproducible
I (A)
I (A)
GdL [T] – Hysteresis cycle
GdL [T]
Hysteresis cycle (linear part removed)
A
B
0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04-10
0
10
20
30
40
50
60
70
2.9 3 3.1 3.2 3.3 3.4 3.5
x 10-3
-0.5
0
0.5
1
1.5
2
2.5
3
3.5
6 6.5 7 7.5 8
x 10-3
65.8
66
66.2
66.4
66.6
66.8
67
67.2
67.4
67.6
t (s)
I (A)
A
B
AB
(uncont rol led) f lat -bot t om oscil lat ions
initial state (0,0)
remanent GdL = 0.002 T
Courtesy Samira Kasaei```
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 26/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Ot her dynamic ef f ect s• Magnet ic af t er-ef f ect (magnet ic viscosit y): class of
relaxat ion phenomena linked to the magnetoelast icinteract ion between ferromagnet ic domain walls and crystal lat t ice leading to a lag between H and M
• Logarithmic time-dependence is a function of relaxation times distribution and is valid at intermediate time scales
• All ferromagnetic metals are affected• does not depend on the geometry (unlike eddy currents)• weakly correlated with f ield level and initial ramp rate• strongly dependent upon temperature• In soft steels: small effect, large time constant → can
usually be ignored
• Disaccommodat ion: after-effect on the initial permeability
• Magnet ic ageing: irreversible phenomena affecting the metallurgical nature of the steel (precipitation, dif fusion, crystal phase transition) on long time scales
∆𝑀𝑀 ∝ 𝑘𝑘𝐵𝐵𝑇𝑇 log 𝜕𝜕
H(t)
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 27/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Non-l inear f eat ures in iron-dominat ed magnet s (1/ 2)
Courtesy Anthony Beaumont, Giancarlo Golluccio
large fluctuations due to history-dependent residual fieldreproducibility degrades at low field
saturation tends to erase previous magnetic history→ better reproducibility at high field
ℓ𝑐𝑐(𝐼𝐼) =1
𝐵𝐵0(𝐼𝐼)−∞
∞𝐵𝐵 𝐼𝐼, 𝑠𝑠 𝑑𝑑𝑠𝑠
eddy current decay (τ=0.2 s)
m drops due to saturation in the ends
m diverges due toBr at center << integral
central coil replaced by NMR (DC cross-check)
time
I(t)
I stand-by (112 A)
I extraction proton (1072 A)
I extraction ion (2800 A)
linear range(up branch only!)
MedAustron main bending
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 28/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Non-l inear f eat ures in iron-dominat ed magnet s (2/ 2)
• Transfer line bending in the ISOLDE heavy isotope test facilit y
• Minor loops span the whole width of the major hysteresis cycle
• Open loop cont rol: « random » cycling → 0.7% errors
• Missing linear range ?!
Courtesy Guy Deferne, Giancarlo Golluccio
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 29/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Increasing dB/ dt
Hyst eresis + eddy current s
• Aim: ext rapolate dynamic measurements to DC to predict behavior at arbit rary dB/dt• Eddy currents ∝ dB/dt → both f ield lag and dissipation (hysteresis loop area) ∝ dB/dt• Measurement result not so ideal … loops cross each other, more drift on slower cycles• hysteresis/drift ef fects need to be corrected by absolute measurements on the plateaux
ramp-up duration
fixed-coilmeasurement
ELENAbending dipole
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 30/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Non-l inear f eat ures in superconduct ing magnet s (1/ 2)
0.705
0.7075
0.71
0.7125
0.715
0 5000 10000Current (A)
Tran
sfer
func
tion
(T/k
A)
MBP2N1
Superconduct ing f ilament magnet izat ion (persist ent eddy current s)• large hysteresis with relat ive errors of the order of 10-3 at low f ield (inject ion)• hysteresis depends on temperature, current and current history (negligible at high f ield)• main f ield and mult ipoles af fected in dif ferent ways
Iron sat urat ion• af fects only small area in the collar (B>2T)• relat ive errors ∼ 1% at high f ield• additional multipoles generated
Linear regime (geomet ric cont r ibut ion)• f ield is proportional to the current(can be computed with B iot-Savart’s law)• the T.F. depends only on the coil geometry
injection
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 31/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Non-l inear f eat ures in superconduct ing magnet s (2/ 2)
Normal sextupole during ramps
-3
-2
-1
0
1
2
0 2 4 6 8 10field (T)
B3 -B
3 geo
met
ric (G
auss
@ 1
0 m
m)
35 A/s50 A/s
10 A/s20 A/s
MTP1N2
0
1
2
3
4
5
0 500 1000 1500time from beginning of injection (s)
b3 (u
nits
@ 1
7 m
m)
500
700
900
1100
1300
1500
dipo
le c
urre
nt (A
)
Coupl ing current s• f inite inter-f ilament and inter-st rand
resistance (RC) gives rise to loops linked with changing f lux
• mult ipole errors ∝ Ḃ, R C-1
• hysteresis depends upon f ield level, temperature and powering history
Decay and snap-back• superconductor magnetization and coupling
currents interact in a complex way → long-term logarithmic time dependence effects (f ield decay)
• hysteresis branch switching at the end of decay → sudden current redistribution and additional multipole errors (snap-back)
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 32/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Sat urat ion model l ing
• Qualitat ive example: anhysteret ic t ransfer funct ion• Simple analyt ical interpolat ion, too coarse for open-loop f ield cont rol
but adequate for inner-loop power converter cont rol
Magnetization curve of PSB main dipole (current configuration)
Outer rings: k = 0.3465 Tm/kA , n=4, I0 = 12000Inner rings: k = 0.3463 Tm/kA, n=7, I0 = 11500
∫𝐵𝐵𝑑𝑑ℓ𝐼𝐼
= 𝑘𝑘 1 −𝐼𝐼𝐼𝐼𝑜𝑜
𝑥𝑥𝐿𝐿 =
Φ𝐼𝐼
= 𝑁𝑁𝑡𝑡 𝑤𝑤𝑝𝑝∫𝐵𝐵𝑑𝑑ℓ𝐼𝐼
= 𝐿𝐿0 1 −𝐼𝐼𝐼𝐼𝑜𝑜
𝑥𝑥
Regos
PSB main dipole
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 33/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
0.002
0.004
0.006
0.008
0.010
0.012
0 1000 2000 3000 4000 5000
L (
H)
I (A)
Inductance L(I) of SPS reference dipole MBB 004
Ld=(V-RI)/IdotL=Φ/ILwLgap
-0.5
-0.4
-0.3
-0.2
-0.1
0.0
0.1
0
2
4
6
8
10
12
0 1000 2000 3000 4000 5000
BdL (
Tm)
I (A)
Magnetization of SPS MBB 004 (reference dipole)
Induct ance model ing
𝑉𝑉 = 𝑅𝑅𝐼𝐼 +𝑑𝑑Φ𝑑𝑑𝜕𝜕
= RI + 𝐿𝐿𝑑𝑑𝑑𝑑𝐼𝐼𝑑𝑑𝜕𝜕 𝐿𝐿𝑑𝑑 = 𝐿𝐿 + 𝐼𝐼
𝑑𝑑𝐿𝐿𝑑𝑑𝐼𝐼
= 𝐿𝐿0 1 − (1 + 𝑠𝑠)𝐼𝐼𝐼𝐼∗
𝑥𝑥
• A large drop of the dif ferent ial inductance at saturat ion is to be expected even for mildly saturated magnets.E.g. SPSmain dipoles: f ield saturat ion 3.4% dif ferent ial inductance saturat ion of 40%.
• Measurement of the inductance curves can be easily done in parallel with standard magnet ic tests.• If this is not possible, the drop of dif ferential inductance may be estimated from the model of f ield saturation
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 34/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Rapid Cycl ing Synchrot ron bending prot ot ype • Dipole prototype opt imized for a possible future RCS with 100 ms cycle t ime
• 0.3 mm Si-steel laminat ions
• ideal testbed to decouple hysteresis f rom dynamics
integral f ixed coilcent ral Hall probeto compute magnet ic lengthand est imate integrator drif t
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 35/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Bipolar repeat abil it y on RCS
0.21320.21330.21340.21350.21360.21370.21380.2139
-1 0 1 2 3 4 5 6
|BdL
| [Tm
]
Cycle index
Stabilization of integrated field @ ±1.4 GeV
+1.4 → -1.4
+2.0 → -1.4
+2.0 → +1.4
0.2718
0.2719
0.2720
0.2721
0.2722
0.2723
0.2724
-1 0 1 2 3 4 5 6
|BdL
| [Tm
]
Cycle index
Stabilization of integrated field @ ±2.0 GeV
+1.4 → +2.0
+2.0 → -2.0
+1.4 → -2.0
1.07921.07941.07961.07981.08001.08021.08041.08061.0808
-1 0 1 2 3 4 5 6
B0 [T
]
Cycle index
Stabilization of central field @ ±1.4 GeV
+1.4 → -1.4
+2.0 → -1.4
+2.0 → +1.4
1.39981.40001.40021.40041.40061.40081.40101.40121.40141.4016
-1 0 1 2 3 4 5 6
B0 [T
]
Cycle index
Stabilization of central field @ ±2.0 GeV
+1.4 → +2.0
+2.0 → -2.0
+1.4 → -2.0
0.19730.19740.19750.19760.19770.19780.19790.19800.1981
-1 0 1 2 3 4 5 6
Lm [m
]
Cycle index
Stabilization of magnetic length @ ±1.4 GeV
+1.4 → -1.4
+2.0 → -1.4
+2.0 → +1.4
0.19390.19400.19410.19420.19430.19440.19450.1946
-1 0 1 2 3 4 5 6Lm
[m]
Cycle index
Stabilization of magnetic length @ ±2.0 GeV
+1.4 → +2.0
+2.0 → -2.0
+1.4 → -2.0
• cycles simulat ing all possible t ransit ions between ±1.4 and ±2.0 GeV beams in the new PSB extraction switch• f ield errors up to 2⋅10-3 just after a transition• f ield errors down t o 4⋅10-5 af t er t wo repeat ed cycles
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 36/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Bipolar reproducibil it y
• Example: bipolar operat ion of ELENA bending dipole• An addit ional intermediate degaussing cycles improves repeatabilit y by a factor 2
(but const rains and delays operat ion !)
Degaussing, Cycle 1, Cycle 2
Degaussing, Cycle 1, Degaussing, Cycle 2 (not complete)
Repeatabilit y: 2.2⋅10-4
Dif ference at ±Imax: 11.3⋅10-4
Difference at ±Imax: 5.4⋅10-4
asymmetry between asymptotedepends on remanent field only!
hyperbolic asymptote due to B(0)≠0(eddy currents + remanent field)
Integral transfer function in ELENA bending dipole: ∫ 𝑩𝑩𝑩𝑩ℓ𝑰𝑰
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 37/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Magnet ic st abil izat ion • Reproducibilit y of magnet ic f ield improves by reset t ing the magnet ic state with current pre-cycles• The operat ing mode of the magnet should be respected:
Bipolar magnets- steerers- correctors- switching dipoles- experimental magnets
→ degaussing
Unipolar magnets- main ring bending/quads
→ pre-cycles
• Random cycling → minor cycles → unpredictable errors within the envelope of the limit cycle• Enforce monotonic cycling for critical magnets (at the cost of more time spent ramping)• Stay as high as possible above zero to improve reproducibility
Imin Imax
B [T]
O
A
I [A]
F
E
DB
C
-0.80
-0.60
-0.40
-0.20
0.00
0.20
0.40
0.60
0.80
-400 -300 -200 -100 0 100 200 300 400
H (A/m)
B (
T)
5 10 15 20 25 30time s
40
20
20
40
60
IA𝐼𝐼𝑘𝑘+1𝐼𝐼𝑘𝑘
= −23
maximum current injected
power supplyresolution
NB: “real” demagnetization requires T≥Tcurie ≈ 948 °C !!
10 20 30 40 50 60times20
40
60
80
100
120
IAmaximum current in operation
minimum current in operation
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 38/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 39/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Summary
• we must measure because mathemat ical predict ion at the level of precision we need is in many cases more expensive or impossible
• measuring under reproducibil it y condit ions allows est imat ion and correct ion of systemat ic errors → convince your management to take more data points
• there is no universal method ! combine complement ary t oolsto optimize resources
• commercial choices increasing but still limited – in-house or (better) col laborat ive R&D often necessary
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 40/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Ref erences
List of CERN Accelerator School proceedings and other resources covering the fundamentals:
ht t ps:/ / t e-msc-mm.web.cern.ch/
An extensive bibliography by one of our founding fathers, including links to the whole IMMW series: ht t p:/ / henrichsen.ch/ magnet / def ault .ht m
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 41/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Thanks f or your at t ent ion
… and good luck wit h new
discoveries !
Tutorial: “Overview of magnetic measurements”IMMW20, Diamond, UK, 04-09 June 2017
Page 42/[email protected]
MAGNETIC MEASUREMENT LABORATORY cern.ch/mm
Acknowledgement
Inst ruments and measurements discussed are based on the collect ive work of the TE/MSC/MM team at CERN:
P Arpaia, A Beaumont , J L Bardanca Iglesias, R Belt ron Mercadillo, N Brut i, G Caiafa,M E Cervera, D Caiazza, R Chrit in, M Colciago, G Deferne, O Dunkel, P Evangelakos,L Fiscarelli, F N Galli, J Garcia Perez, D Giloteaux, G Golluccio, X Gontero, C Grech,
P La Marca, R Mart inez Estebanez, I D Ould-Saada, A Parrella, C Pet rone,M A Roda, G Severino, E Tournaki, J Weick, T Zickler
Special thanks to present and former team leaders that taught me many of the techniques described:
L Bot tura, D Cornuet , S Russenschuck, L Walckiers