INTRODUCTION

The machines manufactured by S2M comprise a set of fixed parts called the "STATOR"
and at least one moving part, known as the "ROTOR".

The "BEARINGS" position the rotor in relation to the stator and enable the rotation of
the rotor with a minimum amount of friction.

The bearings are characterized by their diameter, their length, the maximum force they
can withstand, their maximum rotation speed, their spring rate K: (F = - K X) and their
damping A: (F = - A s X), X being the displacement of the rotor in the stator.

In general, for one rotor there must be 2 radial bearings and one axial bearing called the
"THRUST BEARING".

PLAIN BEARINGS
These are rudimentary, low-cost bearings for slow rotors, that do not permanently rotate,
such as in toys, the wheel axle of a wheelbarrow, the wheel axles of lawn mowers, etc.

BEARING

STATOR
ROTOR STATOR
RR
The bearings must be lubricated WHEEL
to reduce friction
BALL BEARINGS

Ball bearings are probably the most commonly used type of bearing. There is a wide
variety of this kind of bearing, with diameters ranging from a few mm to several meters.
A traditional bearing comprises two raceways (internal and external), fitted with a
groove, inside for the external ring, outside for the internal ring, balls (made of steel or
ceramic) and sometimes a cage to distribute the balls around the circumference of the
bearing.

EXTERNAL RING
STATOR
RR
BALLS
AND
GREASE
ROTOR

INTERNAL RACEWAY

The bearing is lubricated to reduce friction.

The rotation of the Rotor causes the rotation of the interior raceway and of the balls, on
the other hand the external raceway is fixed to the stator.

The spring rate of a ball bearing is high.

Damping is low.
Characteristics: maximum D.N product (diameter, speed), maximum static and dynamic
load.
OIL BEARINGS

 Oil bearings consist of a very smooth cylindrical surface in the rotor and oscillating
pads in the stator. The distance between the rotor and the pads is very small.

 These bearings equip large machines, such as compressors, alternators etc.

 This type of bearing only functions during rotation. A very fine film of oil (a few
microns thick) is then formed between the rotor and the radial pads of the stator and
the axial thrust bearing.

 The rotor is then centered.

 This type of bearing functions with a lubrication system to circulate the oil in the
bearing.

 The system, comprising a tank, a pump, a temperature regulator, filters, pressure

gauges, valves etc., is the most important part of the bearing.

The spring rate is average.

Damping is average.

Pad LUBRICATION SYSTEM

Rotor Rotor
Tank, filter, temperature control,
valves, pumps, pressure gauges etc.
Oil
 S2M MAGNETIC BEARINGS
Standard S2M magnetic bearings are active and consist of two radial bearings (2 control
loops per radial bearing) and an axial thrust bearing (1 control loop). 5 axes control 5
degrees of freedom of the rotor (a moving object has 6 degrees of freedom) thereby
enabling the rotor to turn.
The magnetic bearing is a mechanical unit together with its control electronics.

There is no contact between the stator and the rotor.

Each axis comprises the same structure:

 Position sensors (including the oscillator and the demodulator),

 “P.I.D”-type (proportional - integral - derivative) structural electronic circuits
 Filters
 Nonlinear circuits
Two power amplifiers (power outputs) supply the electromagnets which “at the rotor"
create the forces that oppose the forces applied to the rotor (gravity, process).

The electromagnetic windings and position sensors are placed at the stator on the pole
tips of the packets of precut, stacked magnetic sheets.
Electromagnets

ELECTROMAGNETIC
FORCES
The dynamic spring rate is approximately 3 times weaker than that of an oil bearing. On
the other hand, the static spring rate is much greater.
Damping is greater than that of an oil bearing and above all is independent of speed.
 BASIC ELEMENTS OF A
MAGNETIC BEARING ASSEMBLY

AXES
As previously indicated, a magnetic bearing assembly is based on the control of the 5
axes of freedom of a moving object.

Axis 3 Y Axis 4

Axis 5

X2 X’2
Y
Axis 1 Axis 2 V
4 Y
’

X’1 X1

Y
’
In the radial axis, the axes are set at 45° from vertical YY'. side 1 and side 2.

By convention, S2M has defined the axes as follows: Y V2

W2
W2 V
2 Z2
Z2
Radial
PALIER bearing 1
raRADIAL 1 X2 X’2
Y
W1 V1 V
V4 WW4
W1 V 4 Y 4
11 ’
1 Radial
PALIER bearing 2
RADIAL 2
X’1 X1
Z1 Z1
V W
3 Y 3
’
V3 W3
The axes are designated as follows:

 Axis V1 to V3: Axis V1.3

 Axis W1 to W3: Axis W1.3
 Axis V2 to V4: Axis V2.4
 Axis W2 to W4: Axis W2.4
 Axis Z1 to Z2: Axis Z1.2

In the case of a vertical machine, the axes will be:

Z1

Y1 W1
BEARING 1

X1 V1

AXIS Z

Y2 W2
BEARING 2

X2 V2
Z2

Electrical convention:

When the rotor is in the home position (no current in the coils) the measurements made
on the position sensors indicate a negative voltage. When the rotor is above the nominal
position, the voltage read value will be positive.
Electromagnets
In general, an axis comprises 2 electromagnets

AXIS 2

Electromagnets

AXIS 1

Reminder: in the radial axis, the axes are set at 45° in relation to the vertical axis (to take
into account the Earth’s gravity).

F
A radial bearing consists of 2 independent axes made up of windings on the pole tips of a
magnetic circuit.
This magnetic circuit consists of fine precut sheets and assembled as packets.
Four electromagnets are therefore necessary for each radial bearing

Radial magnetic circuits have four electromagnets made up of 3 or 4 coils per

electromagnet.
The coils are wired so that the pole tips have north, south, north polarities etc. for one
electromagnet, then north, south, north for the adjacent electromagnet, and so on.
 Electromagnet V1

Electromagnet V3
The axial thrust bearing comprises two electromagnets Z1 and Z2. The winding is a flat
coil. The magnetic circuit is a flat disc with one or two grooves (two flat coils in this
case). On the rotor there is a disc of the same diameter.

N N
STATOR   Z1 electromagnet

S S
D CROSS-SECTION

ROTOR I
S Z
K

S S
Z2 electromagnet  
N N

Note: two electromagnets face-to-face must have the same polarities to avoid a direct
transfer of the field through the disc of the thrust bearing.
Characteristics:

An electromagnet is defined by:

 The current input,

 The polarity (North or South) of the pole tips, especially in Z and induction
with maximum current.

 The number of turns per pole tip and the value of its self-
inducting coil,

 The maximum induction at the maximum current.

This provides an attractive force:

F = k.i²/ ².
I = the current in the coil,
 = air-gap (the air-gap "" is the air distance between the packets of sheets of the stator
and the packets of sheets of the rotor).

K is a form factor of the pole tips and depends on the number of turns in the coils

The air-gap "" is the air distance between the packet of sheets of the stator and the
packets of sheets of the rotor.

Where:

F = B².S/2.µo
B = induction in the air-gap "", S = surface area of the pole tips, and µo = 4.10–7
(permeability of the vacuum).

For five axes, in general we need a minimum of 10 electromagnets.

SENSORS

It must be possible to control each axis separately. There must therefore be at least one
sensor per axis.

The sensors are placed in the same angular position as the electromagnetic unit.

SENSORS

AXIS 2

ELECTROMAGNETS

AXIS 1
These sensors are of S2M design and are electrically powered.
The sensors have the same designations as the axes they control (V13, W13, V24, W24,
Z12).

They consist of coils assembled on the pole tips of a packet of sheets at the stator.

s t at o
r V1
E-
L1

E+ ROTOR

L
L2
V3
These sensors, assembled in "bridge" connections are differential. They do not measure
the absolute position of the rotor but its displacement "X" in relation to the center of
the stator.
As for the electromagnets, the sensors in a given plane are located, most of the time, on
the same packet of sheets.

axial

radial
AUXILIARY BEARINGS

 The ball bearings in an Active Magnetic Bearing (AMB) are said to be auxiliary
bearings because they are not used when the AMB is in operation, because
there is a space (AIR-GAP) between the rotor and the internal raceway (0.15 to
0.5 mm depending on the machine in question).
During operation, the rotor does not make the ball bearings turn because in
general they are prestressed. They are not in contact with rotating parts.

 The external raceway is assembled on the stator with damping tape (Borelly), a
kind of small corrugated sheet between the external raceway and the stator, 0.25
mm thick, which decreases to 0.15 mm after assembly.

 The ball bearing is used to land the rotor when the AMB stops and as a support
bearing in the case of power cut or breakdown.

 The ball bearing (raceways and balls) is lubricated or covered with a special
treatment depending on the application, such as molybdenum disulfide (MOS2),
Neversez etc.

 "Deep groove" or "oblique contact" ball bearings are used.

Cross-section of ball bearings:

STATOR

ROTOR
CONTROL ELECTRONICS

The control electronics provide:

- the power supply for the position sensors,

- the conformation and treatment of the return position signal
- the power supply for the electromagnets.
- the management of the various operating parameters (temperature, vibration, current)

Amplifier
Electromagnets
Signal
treatment Sensors

Reference
signal
 PRINCIPLE OF OPERATION
OF A MAGNETIC BEARING

CONTROL-LOOP SYSTEM

The signal sent by a position sensor is compared with the reference signal, which defines
the nominal position of the rotor. If the reference signal is equal to zero, the nominal
position is in the center of the stator. The error signal is proportional to the difference
between the nominal position and the real position of the rotor at a given moment in time.
This signal is transmitted to the processor, which in turn sends a correction signal to the
power amplifier.

Demonstration:

First case: the rotor is not levitated (i.e. is at rest on the ball bearings)
In this case, the sensor signal will be negative, the control signal at the output of the
“switchpoint” is sent to amplifier A1, which inputs current to coil V1.

A1

Reference
signal

A2
Second case: the rotor is levitated above the nominal value
In this case, the sensor signal will be positive, the control signal at the output of the
"switchpoint" is sent to amplifier A2, which inputs current to coil V2.

A1

Reference
signal

A2

This operation is common to all the axes (reminder: 5 axes are controlled).
In this way each control electronics system comprises the following elements:

- 5 control-loops
- 10 amplifiers (generally 2 per axis)
- 5 detection treatment systems (position sensor signal)
 DETECTION

Sensor operation

The sensor is powered by an oscillator outputting 2 sinusoidal signals in opposite phase

of 20 kHz with an amplitude of 30 volts.
These two signals are designated E+ and E -

s t at o
r V1
E-
L1
L1

E+ ROTOR

L
L2
L2

V3

When the rotor moves from x (for example towards V1) the air-gap of self-inducting coil
L1 decreases (OX) and the air-gap of self-inducting coil L2 increases (o+x). Self-
inducting coil L1 increases and self-inducting coil L2 decreases because the values of the
self-inducting coils are inversely proportional to their air-gap. Signal "L" changes in
relation to the power supply of the smallest self-inducting coil, and therefore to E+.

To make things clearer, it is possible to compare this assembly with a traditional

Wheatstone bridge, in which the resistance levels vary according to the air-gap
E+
E+
R1
 ROTOR reading reading

R2
E-
E-
There are 3 possibilities:

E+

E-

Case in which R1=R2

Read signal

Case in which R1R2

Read signal

Case in which R1R2

Read signal
We therefore have an alternative Read signal comprising the position information and a
carrier signal at 20 kHz.

There must be a positive or negative dc signal (see the control-loop system) to compare
with the reference signal. Demodulation therefore must be carried out.

In the case of an axial sensor:

There are several versions of axial sensor:

A self-inducting coil-inductive sensor similar to a radial sensor: this has a coil at the two
ends of the rotor.

E+ E-
ROTOR

L = measurement of
axial displacement

An axial-radial sensor with three packets of sheets, the central packet being the radial
sensors. On the side packets four coils are wired in series. The inputs are connected to E+
and E-, and the two other ends of the coils, connected together constitute the Read signal
of the axial sensor.
Note: this axial type, known as a slip sensor, generates harmonic four of the rotation
speed.

Circumferential sensor: comprises two flat coils side by side, in an insulating plastic
frame and therefore with no iron in the stator. The rotor comprises three solid-core
raceways: a central raceway made slightly magnetic steel and two side raceways made of
aluminium.

L(x)
STATOR
E+ E-
+

AL ALU
U
STEEL ROTO
R
In all three cases, the reading system is identical to the radial system
Demodulation

The demodulator is an electronic circuit which performs synchronous rectification of the

Read signal, i.e. it eliminates the carrier signal at 20 kHz and outputs a voltage
representing the displacement "x" of the rotor, whatever the form of displacement.

REF +
L A C
GAIN
R V(x)

S
REF -

D
GAIN -1 B R
RESPONSE
C E

0V

The reference signals are output by the oscillator. These signals have the same frequency
as the excitations E+ and E -.
This make it possible to control the analog type switches.

E+

REF+

REF -
 REF+

REF -

Reading

Signal at "C" after throughput via REF + switch

Signal at "D" after throughput via REF - switch

Signal at "S" after summation

 Signal after filtering

THE OSCILLATOR AND SIGNAL DEMODULATION ARE ON THE

SAME PCB
AMPLIFIER

The amplifier can be defined as a voltage-current converter

+150 Volts

Electro-
magnet
Positive dc voltage AMPLIFIER
control signal from
switch

+150 Volts

The amplifier provides a current to the electromagnet when its control signal is
positive. A negative control signal does not produce any current

Present-day amplifiers are all of the switch type (13 kHz to 50 kHz according to the
amplifier). The maximum range of current is from 2 to 60 A.

Their designation is based on the supply voltage and the maximum current that they
output,
e.g. an amplifier that outputs a maximum of 30 A with a supply voltage of 300 V is an
A300/30.

The amplifiers generate a function that depends on the control voltage and also a voltage,
since their load is a coil (self-inducting coil). The voltage necessary is equal to L.2..f.i
in which L is the value of the self-inducting coil of the electromagnet, f is the frequency
of the control signal and i the current output by the amplifier.