Download Slides - Indico

Accelerator magnets – construction and
measurements
Neil Marks,
ASTeC, CI,
Daresbury,
Warrington WA4 4AD,
[email protected]
Tel: (44) (0)1925 603191
Fax: (44) (0)1925 603192
Neil Marks; ASTeC, CI.
Magnet Construction and Measurements, Spring term 2012
Philosophy
•To give a pictorial representation of the construction of
accelerator magnets.
•Then to cover the possible methods of measuring flux density
but concentrating on the most frequently used methods.
•Note that magnetic field H is a measure of the excitation
(creation) of the magnetic phenomenon; all measurable effects
are driven by the flux density B.
•Note that measurement ‘accuracy’ involves three different
facets:
resolution;
stability and repeatability;
absolute calibration.
Neil Marks; ASTeC, CI.
Magnet Construction and Measurements, Spring term 2012
Contents:
1. Magnet manufacture and assembly methods.
2. Physical effects available for measurement:
a) force on a current carrying conductor;
b) electromagnetic induction;
c) Hall effect (special case of (a));
d) nuclear magnetic resonance.
3. Practical applications:
a) point-by-point measurements;
b) rotating coil methods;
c) traversing coils.
Neil Marks; ASTeC, CI.
Magnet Construction and Measurements, Spring term 2012
Magnet Manufacture and Assembly
Neil Marks; ASTeC, CI.
Magnet Construction and Measurements, Spring term 2012
Contents
The presentation describes the construction of laminated
magnets and shows examples of the various stages in a
magnet manufacturing factory.
Thanks are given to Tesla Engineering for the
provision of photographs of these processes.
Neil Marks; ASTeC, CI.
Magnet Construction and Measurements, Spring term 2012
Yokes
For ac accelerator magnets, laminated yokes are essential but
many d.c. magnets also have laminated cores.
The use of laminations:
• allows shuffling of steel - randomise properties;
• prevents eddy time constants ~ minutes;
Laminations are 'stamped' using a 'stamping tool'; very high
precision and reproducibility (~20µm) is possible.
The stamping is carried out on a ‘press’ (c 200T!) – see next
slide.
Neil Marks; ASTeC, CI.
Magnet Construction and Measurements, Spring term 2012
Stamping Press
Neil Marks; ASTeC, CI.
Magnet Construction and Measurements, Spring term 2012
Lamination Assembly
The laminations are ‘stacked’ to make ‘blocks’. The assembly is
in a jig, with hydraulic pressing.
Small laminations can be glued together; the glue (epoxy type) is
now usually coated onto the lamination material at by the steel
manufacturer and sets as a thin (c 10 mm) solid layer at room
temperature. After stacking the assembly is put under pressure
and then placed in an oven; the epoxy liquefies and then cures,
producing a solid block.
Large laminations are often welded (leading to distortion?) whilst
held in the stacking jig (but Tesla tend to use glue).
Neil Marks; ASTeC, CI.
Magnet Construction and Measurements, Spring term 2012
Laminations stacked to form
blocks.
Diamond
quadrupole
laminations in
the ‘stacking
fixture’ prior
to bonding.
Neil Marks; ASTeC, CI.
Magnet Construction and Measurements, Spring term 2012
A large GSI dipole
block (1/4 of an
‘H’ magnet)
stacked in a single
stacking fixture.
Neil Marks; ASTeC, CI.
Magnet Construction and Measurements, Spring term 2012
A complete GSI dipole block
Neil Marks; ASTeC, CI.
Magnet Construction and Measurements, Spring term 2012
Diamond quadrupole blocks
Two blocks –
now machined
(see chamfer)
and mated
together (as a
test) to form half
a quadrupole.
Note- coils cannot
be added in this
configuration.
Neil Marks; ASTeC, CI.
Magnet Construction and Measurements, Spring term 2012
Coils
Coils are wound with the glass insulation wrapped onto the
copper conductor.
Neil Marks; ASTeC, CI.
Magnet Construction and Measurements, Spring term 2012
Coils
The coils then receive an 'outer-ground' insulation of (thicker)
glass cloth and wrapped in ‘release tape’ (which will not stick to
epoxy resin). They are then placed into a mould and heated in a
vacuum oven to dry and out-gas.
Neil Marks; ASTeC, CI.
Magnet Construction and Measurements, Spring term 2012
Coils
The mould is then flooded with liquid epoxy resin and the
vacuum tank let up to atmosphere, forcing the resin into the coil.
'Curing' of the resin then occurs at high temperature. Total
cleanliness is essential during all stages of this process!
Vacuum ovens
used to ‘cure’
coils.
Neil Marks; ASTeC, CI.
Magnet Construction and Measurements, Spring term 2012
Coils after impregnation
Neil Marks; ASTeC, CI.
Magnet Construction and Measurements, Spring term 2012
Coil testing
Rigorous testing of coils, including:
•water pressure tests to detect leaks from the cooling circuits;
•water flow;
•thermal cycling (between ambient and operating temperature 50
or 100 times);
•then 'flash' testing at high voltage whilst the body of the coil is
immersed in water (terminals only clear);
•inter-turn voltage tests (using transformer techniques);
is strongly recommended.
Neil Marks; ASTeC, CI.
Magnet Construction and Measurements, Spring term 2012
Completed magnet: Diamond Quad.
Neil Marks; ASTeC, CI.
Magnet Construction and Measurements, Spring term 2012
Magnet Measurements
Neil Marks; ASTeC, CI.
Magnet Construction and Measurements, Spring term 2012
Force on a current carrying conductor
where:
F=BI
F is force per unit length;
B is flux density;
I is current.
Advantages:
integrates along wire;
I can be accurately controlled and measured.
Disadvantages:
not suitable for an absolute measurement;
measurement of F is not very highly accurate;
therefore not suitable for general measurements.
Neil Marks; ASTeC, CI.
Magnet Construction and Measurements, Spring term 2012
Use in spectrometry
specialised trajectory tracing
in experimental magnets:
T
‘Floating wire’ technique wire is kept under constant
tension T and exit point is
measured for different
entry points.
B
I
T
Neil Marks; ASTeC, CI.
Magnet Construction and Measurements, Spring term 2012
Electromagnetic induction
curl E
= - B / t;
V = B An sin wt.
(V is induced voltage; B is flux density; A is coil area; n is coil turns.
Advantages:
V can be accurately measured;
Gives B integrated over the coil area.
Disadvantages:
/ t must be constant (but see later);
absolute accuracy limited by error in value of A;
Can be sufficiently accurate to give absolute measurements but best for
relative measurements.
Used:
standard measurements of accelerator magnets;
transfer standards;
Neil Marks; ASTeC, CI.
Magnet Construction and Measurements, Spring term 2012
Hall effect
Special case of force on a
B
moving charge; a metal
(or semiconductor) with a
V
current flowing at right
angles to the field develops
a voltage in the third plane:
V=-R(JxB)a
where:
V is induced voltage; B is field;
J is current density in material;
a is width in direction of V
R is the 'Hall Coefficient' ( fn of temperature ):
R = fn (a, q);
q is temperature; a is temperature coefficient.
Neil Marks; ASTeC, CI.
Magnet Construction and Measurements, Spring term 2012
J
Hall effect (cont.)
Advantages:
small light probe;
easily portable/moved;
J, V accurately measurable – good resolution, repeatability;
covers a very broad range of B;
works in non-uniform field.
Disadvantages:
q must be controlled measured/compensated;
R and a must be known accurately.
Used:
commercial portable magnetometers;
point-by-point measurements;
Neil Marks; ASTeC, CI.
Magnet Construction and Measurements, Spring term 2012
Nuclear magnetic resonance.
In an external magnetic field, nuclei with a magnetic moment
precess around the field at the Larmor precession frequency:
n  (g /2 p) B;
where:
n is the precession frequency;
g is the gyro-magnetic ratio of the nucleus;
B is external field.
A radio-frequency e-m field applied to the material at this
frequency will produce a change in the orientation of the spin
angular momentum of the nucleus, which will ‘flip’, absorbing
a quantum of energy. This can be detected and the r.f.
frequency measured to give the precession frequency and
hence measure the field.
Neil Marks; ASTeC, CI.
Magnet Construction and Measurements, Spring term 2012
Spin transition.
The ‘spin flip’ in a nucleus:
B
n
n
M
Example:
for the proton (H
nucleus):
with B = 1 T;
n = 42.6 MHz.
Neil Marks; ASTeC, CI.
Magnet Construction and Measurements, Spring term 2012
N.M.R. (cont.)
Advantages:
• only dependent on nuclear
phenomena - not influenced
by external conditions;
• very sharp resonance;
• frequency is measured to very
high accuracy (1:106);
• used at high/very high B.
Disadvantages:
• probe is large size (~ 1cm);
• resonance only detectable in
highly homogeneous B;
• apparatus works over limited
B range, (frequency n is too
low at low B);
• equipment is expensive;
Use:
•most accurate measurement system available;
•commercially available;
•absolute measurement of fields;
•calibration of other equipment.
Neil Marks; ASTeC, CI.
Magnet Construction and Measurements, Spring term 2012
Practical Applications – Point by point
A probe is traversed in 2 or 3 planes with B measured by a Hall
plate at each point to build up a 2/3 dimensional map.
Superseded by
rotating coils for
multi-poles, but
still the method of
x
choice for a small
y
number of high
quality dipoles. (It
is too slow for a
z
production series)
Neil Marks; ASTeC, CI.
Magnet Construction and Measurements, Spring term 2012
Modern Hall-probe Bench used at DL for
insertion magnets.
Hall Probe
Teslameter
Longitudinal Range
Horizontal Range
Vertical Range
Longitudinal Resolution (z)
Horizontal Resolution (x)
Vertical Resolution (y)
Nominal Longitudinal Velocity
Maximum Calibrated Field
Hall Probe Precision
Hall Probe Resolution
Temperature Stability
Neil Marks; ASTeC, CI.
MPT-141-3m
DTM-141-DG
1400
200
100
1
0.5
0.5
1
2.2
± 0.01 %
0.05
± 10
(Group 3);
“
mm
mm
mm
mm
mm
mm
mm/s
T
mT
ppm/°C
Magnet Construction and Measurements, Spring term 2012
Rotating Coil systems.
Magnets can be measured using rotating coil systems; suitable for
straight dipoles and multi-poles (quadrupoles and sextupoles).
This technique provides the capability of measuring:
•amplitude;
•phase;
integrated through the magnet (inc end
fringe fields) of each harmonic present, up to
n ~ 30 or higher;
and:
•magnetic centre (x and y);
•angular alignment (roll, pitch and yaw)
Neil Marks; ASTeC, CI.
Magnet Construction and Measurements, Spring term 2012
The Rotating Coil
A coil continuously rotating (frequency w) would cut
the radial field and generate a voltage the sum of all the
harmonics present in the magnet:
B
-C
 = const.
dipole: V = sin wt
quad: V = sin 2 wt
etc.
Neil Marks; ASTeC, CI.
+C
+
C
-C
-C
+C
sextupole: V = sin 3 wt
Magnet Construction and Measurements, Spring term 2012
Continuous rotation
The coil (as shown) is rotated
rapidly in the magnetic field; the
induced voltage is analysed with
a harmonic analyser.
Induced voltage :
V = / t = N
V
coil A coil B r/

2 n 1

n 1
= N coil coil  n r
w
t;

(A n sin nq + Bn cos n q)(q/ t)
Continuous rotation is now regarded as a primitive method!
Neil Marks; ASTeC, CI.
Magnet Construction and Measurements, Spring term 2012
Problems with continuous rotation
Sliding contacts:
generate noise – obscures small
higher order harmonics;
Irregular rotation:
(wow) generates spurious
harmonic signals;
Transverse oscillation
of coil:
(whip-lash) generates noise and
spurious harmonics.
Solution developed at CERN to measure the LEP multi-pole
magnets.
Neil Marks; ASTeC, CI.
Magnet Construction and Measurements, Spring term 2012
Solution:
Rotation and data processing:
•
•
•
•
•
coil cylinder make < 2 revolutions in total;
windings are hard wired to detection equipment;
an angular encoder is mounted on the rotation shaft;
the output voltage is converted to frequency and integrated
w.r.t. angle, so eliminating any /t effects;
integrated signal is Fourier analysed digitally, giving
harmonic amplitudes and phases.
Specification: relative accuracy of integrated field
angular phase accuracy
lateral positioning of magnet centre
accuracy of multi-pole components ±3x10-4
Neil Marks; ASTeC, CI.
±3x10-4;
±0.2 mrad;
±0.03 mm;
Magnet Construction and Measurements, Spring term 2012
Rotating coil configurations
Multiple windings at different radii (r) and with different
numbers of turns (n) are combined to cancel out harmonics,
providing greater sensitivity to others:
-n
+n
-n
+n
-n +2n -2n
+n
r/4
3r/4
All
harmonics
Neil Marks; ASTeC, CI.
All odd
harmonics,
1,3,5 etc.
Dipole and
quadrupole
rejected.
Magnet Construction and Measurements, Spring term 2012
A rotating coil magnetometer.
Neil Marks; ASTeC, CI.
Magnet Construction and Measurements, Spring term 2012
Test data used to judge Diamond quads
(acknowledgement to Tesla Engineering for spread-sheet developed for Quad measurement)
Validity
Iteration No.
Magnet type identifier
Magnet serial
Date of test
Tester
Comments:
DLS comments:
Dipole+NS007 reference angle
Adjusted dipole reference angle
Field quality data
R(ref) (mm)
Current (A)
Central strength (T/m)
L(eff) (mm)
C3 (4-8)
S3 (6-12)
C4 (4-7)
S4 (1-4)
C6 (2.5-10)
|C10,S10|: (N:3-5, W:6-8)
All other terms up to 20 (2.5-5)
Keys to use
Next shims to use (rounded)
Shimming History
Iteration#
Shims in use
Next shims (measured)
3
4
5
Rounding errors
Warnings
Midplane adjustment
(+ to open)
This template is current
1
WM
WMZ086
Next actions (Refer first):
DLS referral done? (Yes/No/NA)
East (um):
West (um):
Top (um):
Bottom (um:)
C3 switch
S3 switch
C4 switch
S4++ switch
Full switch
dx switch
Post-shim
prediction
12/07/2005
Darren Cox
180A preliminary
Please insert comments here
137.89068 (update fortnightly)
137.90085
240
80
80
0
1
1
1
1
1
1
Alignment data
[good pass/pass]
dx [0.025/0.05]mm
dy [0.025/0.05]mm
dz [2.5/5.0]mm
Roll [0.1/0.2]mrad
-0.49 Yaw [0.15/0.3]mrad
-2.33 Pitch [0.15/0.3]mrad
-2.64
-0.04
yes
Reject/Hold for refer? (S4, C6+)
Adjust vertical split (S3)?
Adjust midplane (C3/C4)?
Full align?
Adjust dx only?
Accept magnet?
Yes
Yes
Value
Outcome
-0.089
-0.059
2.414
0.052
-0.048
-0.085
Fail
Fail
Good pass
Good pass
Good pass
Good pass
35.00
180.00
17.6328
407.253
-0.49
-10.88
6.90
0.80
7.97
5.16
4.98
Pass
Refer, or shim vertical
Refer, or shim horizontal
Pass
Refer to DLS
Pass
Refer to DLS
N key
N/A
S key
N/A
NW foot
N/A
NE foot
N/A
SW foot
N/A
SE foot
N/A
N key
32.010
0.000
0.000
0.000
0.000
0.000
S key
32.012
0.000
0.000
0.000
0.000
0.000
NW foot
19.011
0.000
0.000
0.000
0.000
0.000
NE foot
19.020
0.000
0.000
0.000
0.000
0.000
SW foot
19.004
0.000
0.000
0.000
0.000
0.000
SE foot
19.015
0.000
0.000
0.000
0.000
0.000
Neil Marks; ASTeC, CI.
DLS OK?
?Yes/No?
No
No
No
No
yes
No
yes
Adjust X alone?
Alignment OK?
Magnet Construction and Measurements, Spring term 2012
Traversing coils
Used in curved dipoles -similar method of data
acquisition as used in a rotating coil.
Magnet
on test.
Coil
Reference
magnet
(prototype)
The coil (with multiple radial windings) is traversed from the reference to the
test magnet; voltage from each winding is integrated; variation from zero in the
integrated volts, after the traversal, indicates variations from the reference
magnet total flux vs radius values, which are known.
Neil Marks; ASTeC, CI.
Magnet Construction and Measurements, Spring term 2012