2 Power Circuits

Stefan Butzmann

2.1 Introduction

This chapter deals with power circuits and their use for actuator control.
Figure 2.1 shows the general structure of such a power circuit, where one or
several power devices are turned on and off by a driving stage. Protection
circuits are used to prevent the power devices from failure due to overload
conditions.

Single I
Integrated multiple
From driver ,.. To actuator
microcontro power devices

Protection

Fig. 2.1. Structure of a power circuit

As most actuators are of inductive type - more than half of our electrical
energy is consumed only by electric motors in household and factory appli-
cations - the focus of this chapter is oriented on topologies, where the power
device is used as a switch. While linear drive circuits for inductive loads al-
ways result in a significant power dissipation in the power device, switching
topologies have the advantage, that the power dissipation in the drive circuit
can be minimized.
In the following, the most important power device types- the rectifier, the
bipolar transistor, the MOSFET and the IGBT- are discussed. As the plu-
rality of actuators is operated in low to medium power and voltage ranges,
where power-MOSFETs or IGBTs are generally the best choice for power
switches, different circuits for gate drive and protection of these devices are
described. An important aspect is thermal management, which is introduced
in the following section. Different topologies for DC and AC actuator control
are described next. The section "future trends" deals with coming technolo-
gies, gives an insight into intelligent switching devices and an outlook into
future developments.

H. Janocha (ed.), Actuators

© Springer-Verlag Berlin Heidelberg 2004
20 Stefan Butzmann

2.2 Power Switching Devices

2.2.1 Overview

Advances in power circuits are strongly related to the continuous improve-

ment of power semiconductor devices. They are used to control the energy
flow from an electrical source to the load. The ideal power switch is charac-
terized by the following attributes:
- zero resistance during on-state,
- infinite resistance during off-state,
- zero transient time from on- to off-state and vice versa,
- zero driving power during static and transient operation,
- electrical isolation between control function and power flow.
Practical semiconductor switches are imperfect. They have a limited volt-
age and current handling capability, an on-resistance greater than zero and
finite switching times.

<(

"E
::J
()
Thyristor
1000 "F=---:----:--,....,..---..
Bipolar transistor

IGBT
100 1 = - - - - - - - , .

10 MOSFET

1
1 10 100 1000 10000 Voltage I V

Fig. 2.2. Application areas of power semiconductors

The first power devices broadly used in the 1950s were thyristors, which
today are able to switch voltages of more than lOkV and currents of more
than 10 kA. The lack of controlled turn-off-capability however makes thyris-
tors unsuitable for many applications where the turn-off of the power device
also needs to be controlled. Thyristor concepts providing this capability like
the gate-turn-off thyristor (GTO) or the MOB-controlled thyristor (MCT)
are generally only used in applications with power ratings in the megawatt
range. Even though thyristors are very cost-efficient devices, they are rarely
used for the drive of actuators as discussed in this book. For this reason, the
working principle of thyristors will only be briefly explained.
 2 Power Circuits 21

A power device, which was used extensively in power applications up

to the 1980s, is the bipolar transistor. One of the main disadvantages of
bipolar power transistors is the relatively large base current needed to con-
trol the device and the long turn-off time limiting the maximum switch-
ing frequency to approximately 5-10kHz. In general, bipolar transistors are
only used in applications of medium frequencies and high voltages, where
they provide a cheaper solution than power-MOSFETs or IGBTs. However,
their more complex driving circuits in comparison to IGBTs has signifi-
cantly reduced their use in many applications. Today, typical remaining ap-
plications for bipolar transistors are electronic lamps or automotive ignition
circuits.
In contrast to bipolar transistors power-MOSFETs possess a high input
impedance and exhibit low power losses. Additionally, they allow for high
switching frequencies up into the MHz range and are still today the best
choice in low-voltage applications. Due to the increase of their on-state losses
towards higher operating voltages, their performance becomes unsatisfactory
in applications with supply voltages above 600 V.
The Insulated Gate Bipolar Transistor (IGBT) aims at the voltage range
above 600 V, which is no longer suitable for MOSFETs. It can be operated
with switching frequencies of up to 20kHz and currents up into the kA range.
IGBTs are frequently used in motor control as well as in advanced automotive
ignition, competing here with bipolar transistors. Their low-effort driving
circuit has led to an extensive use of IGBTs in many applications, even though
the device itself is more expensive than a bipolar transistor.

Table 2.1. Advantages and disadvantages of power switches

Advantages Disadvantages
BJT low on-resistance driving power
small area slow turn-off
inexpensive
MOSFET low drive effort high on-resistance at high
fast switching breakdown voltages
IGBT low on-resistance at higher medium switching speed
breakdown voltages low drive effort
Thyristor extremely high power no turn-off control via gate
capability

Power-MOSFETs and IGBTs are the most popular power switches used
for actuator control at the current time. As the drive of MOS-gated devices
is simpler than that of bipolar transistors because of the lack of base current,
complete gate drive circuits can meanwhile be integrated monolithically on
a single chip. The addition of protection functions and intelligence has pro-
22 Stefan Butzmann

vided a further step towards smart-power devices, which are discussed in the
last section of this chapter.

2.2.2 Power Diode

Most power applications require fast diodes as free-wheeling elements. The

V-I characteristic of the diode is shown in Fig. 2.3.
For a positive voltage from the anode to the cathode the diode is forward
biased and the current through the diode rises exponentially with the voltage
VF. In reverse bias (negative voltage from anode to cathode) only the leakage
current Is flows through the diode as long as the reverse voltage is smaller
than the breakthrough voltage V8 v. If the reverse voltage becomes larger
than the breakthrough voltage, the reverse current increases rapidly due to
the avalanche effect.
The breakdown voltage of the diode Vsv is usually specified for an am-
bient temperature of 25 °C. V8 v is a function of temperature with a positive
temperature coefficient. Also the leakage current Is increases exponentially
with rising temperature.
Of particular interest for power applications is the switching behavior of
power diodes, which is significantly different from that of small signal diodes.
Figure 2.4b shows the principal turn-on behavior of a power diode in the
configuration according to the circuit in Fig. 2.4a. When a constant current
is impressed into the diode, the device shows quasi-inductive behavior until
the pn-junction has been flooded with carriers (Fig. 2.4b). This yields a peak
voltage during turn-on, which is dependent on the current slope.
The turn-off behavior of the diode is shown in Fig. 2.5. Power pn-diodes
show a "memory" effect. If the voltage across a conducting diode is suddenly
reversed, the diode acts like a capacitor at first and a current back into the
voltage source builds up. Only after the injected minority carriers (holes in
then-region) have been swept out by the reverse current or by recombination,
the diode regains its blocking capability.

'o

Vsv

b
Fig. 2.3. Power diode. a Schematic symbol, b transfer characteristic
 2 Power Circuits 23

.t
tvo Vp

....
a b t
Fig. 2.4. Thrn-on behavior of power diode. a Test circuit, b voltage output for
current step

The main parameter during turn-off is the peak reverse current JRR· For
a greater slope ldJ I dtl of the current during turn-off this reverse current in-
creases and causes current spikes, electromagnetic interference and additional
switching losses. It is therefore an important task in power applications to
limit dJ I dt during turn-off to keep the reverse current peak low.
The reverse recovery time trr describes the time between the point where
the turn-off current crosses zero and the point where the reverse current has
fallen back to 25% of the peak value. The reverse recovery behavior of the
diode is generally described by the softness factor S = tblta. If S « 1 a hard
snap-off is assumed.
In comparison to the described pn-diode a Schottky diode has a faster
switch-off behavior. Instead of an implanted p-region Schottky diodes are re-
alized by the junction between the barrier metal and the semiconductor. Since
both metal and silicon are n-type, the conduction occurs through majority
carriers only. Because of the lack of minority carriers injection, storage and

Fig. 2.5. Turn-off behavior of power diode

24 Stefan Butzmann

recombination, the storage time ts is virtually reduced to zero. This makes

the Schottky diode very suitable for fast switching applications.
The disadvantages of Schottky diodes are the relatively large leakage cur-
rent and the lower breakdown voltage in comparison to pn-diodes. Schottky
diodes are available with voltage ratings up to typically 150-200 V.
For fast-switching applications with breakdown voltages of more than
200V FRED diodes (fast recovery epitaxial diodes) are a good compromise.
These diodes contain gold or platinum atoms diffused into the epi-layer cre-
ating areas where the excess holes can recombine with electrons. This reduces
the minority carrier lifetime and improves the dynamic characteristics of the
diode. Alternatively, the minority carrier lifetime can also be reduced by ion
irradiation.
Table 2.2 summarizes the characteristics of the most important types of
power diodes. The selected examples of the rectifier diode 1N3768, of the
FRED diode 30EPH06 and of the Schottky diode 30CTQ100 (all Interna-
tional Rectifier) show that the forward voltage drop at a nominal current of
IN = 30 A is approximately 0.9 V for the Schottky diode, while the voltage
drop at the line-rectifier diode and at the FRED diode is more than a factor
of two higher. This results in lower static power dissipation of the Schottky
diode in comparison to the two other diode types.

Table 2.2. Comparison of the most important types of power diodes

Rectifier diode Fast recovery Schottky diode

diode (FRED)
Current range [A] 3000 500 100
Reverse voltage [V] 6000 2000 200
Switching losses medium small extremely small

Example 1N3768 30EPH06 30CTQ100

Rating lkV/35A 600V/30A 100V/30A
VF[V]@JN 1.8 2 0.86

2.2.3 Bipolar Power Transistor

Bipolar power transistors offer high breakdown voltages of up to 1500 V

and low saturation voltages VcE,sat· In contrast to MOSFETs, which exhibit
an extremely high gate-source impedance, power bipolar transistors require
a base current. The power control ratio of a bipolar transistor is described
by the current gain B = Ic/IB.
In order to achieve high voltage ratings as needed for power applications,
large base widths have to be realized, resulting in larger saturation voltages
and poor current gains B. Figure 2.6 shows the current gain of the bipolar
 2 Power Circuits 25

--- r-- --....

!'-.....____

""'
10 1

'\
19J=125°C
\
VcE=5V

Fig. 2.6. Current gain of bipolar power transistor BU208A as a function of collector
current [11]

transistor BU208A with a voltage rating of VcE,rnax = 1500 V and a maximum

collector current of Ic,max = 8 A. As shown in the diagram the current gain
is a function of the collector current and decreases below a value of ten for
the nominal current.
Bipolar transistors have no overcurrent capability, i.e. the collector cur-
rent Ic always has to stay below the specified value Ic,max· The maximum
collector--emitter voltage of a bipolar transistor is defined by VcEo and VcBo.
VcEo describes the collector-emitter breakthrough voltage with the base
open, VcBo the collector-base breakthrough voltage. A good approximation
for the relationship between VcEo and VcBo is given by:

VcBo
VcEo ;:::j VB. (2.1)

Figure 2.7a shows a bipolar power transistor used as a switch. Usually

the input base current JB will be chosen higher than necessary in order
to ensure saturation of the transistor. In consequence the collector-emitter
voltage will be reduced to VcE,sat, thus resulting in the power consumption
P =VcE,sat · Ic+VBE · IB of the device.
During the saturation mode with a forward biased base-collector diode
there will be an excess of charge in the base, which has to be extracted during
switch-off. If the base is open-circuited by switching off the base current JB,
the slow recombination of the excess carriers in the base will result in a long
26 Stefan Butzmann

A
---- 18 /
_lc/A

······························-···

---------i

t! us
a

15
---- V IV 1
- Icc' A }_ ___________ ..
10

5 ,!
..........................; ..
:
01------f······························ !1--------l
i
:
I
'
-5 .................. ··········1···························--1···
:I !;
vc! -10 ----------!--- .

-150 10 20 30
t! us
b

Fig. 2. 7. Power bipolar transistor as switch. a Control via base current, b control
via base-emitter voltage

switch-off time toff· In order to reduce the switch-off time the base may be
returned to a negative potential via a low-valued resistor RB (Fig. 2.7b).
The slow recovery characteristic of the bipolar transistor is especially
critical when inductive loads are turned off. As the turn-off does not occur
equally distributed over the chip area, parts of the transistor continue to
carry the complete load current at the full operating voltage before the free-
wheeling diode is activated. This may result in "hot spots" and failure of the
transistor. This effect is known as "secondary breakdown" .
The turn-off time may also be reduced by prevention of saturation of the
device. A suitable anti-saturation circuit is the Baker clamp, which is shown
in Fig. 2.8. In this circuit, the diode D 1 prevents the collector potential from
falling below the base potential, thus avoiding saturation of the transistor.
Clamping the collector-emitter voltage in the described way, however, yields
an increase of VeE and thus in a higher power dissipation during the on-
state.
 2 Power Circuits 27

Fig. 2.8. Baker clamp circuit

The largest disadvantage of bipolar power transistors in DC operation

mode is the low current gain. A load current of Ic = 10 A requires a base
current of IB = 1 A if a current gain of B = 10 is assumed. In order to
minimize the driving current of a power bipolar transistor, the Darlington
stage can be used, where two transistors are connected as shown in Fig. 2.9.
The current gain of the resulting Darlington transistor is the product of
the current gains of the two individual transistors. These advantages, how-
ever, are offset by the higher saturation voltage from the collector to the
emitter, which has been increased by VBE:

VcE,sat,Darl = VcE,sat,Tl + VBE,T2· (2.2)

Apart from this, special measures have to be taken to optimize the switch-
ing behavior. For this reason, power bipolar transistors have enormously lost
importance in favor of MOSFETs and IGBTs.

c·

B'

Fig. 2.9. Darlington stage

28 Stefan Butzmann

2.2.4 Power-MOSFET

Static Behavior

MOSFETs (Metal Oxide Semiconductor Field Effect Transistors) have been

in use for power applications since the early 1980s. The popularity of MOS-
FETs for power applications results from two main advantages over bipolar
transistors:
- MOSFET transistors are simpler to drive as their control electrode is
isolated from the current path. Thus, no gate current is required during
static operation.
- MOSFETs have a resistive nature , i.e. the voltage from drain to source
is proportional to the current flowing through the device. In contrast
to bipolar transistors where the saturation voltage VcE,sat has a negative
temperature coefficient, the on-resistance of MOSFETs exhibits a positive
temperature coefficient of 1 %/K. This behavior is a great advantage
for parallel operation of multiple MOSFETs.

0
Fig. 2.10. Cross-section of power-MOSFET

For power applications the most popular MOSFETs are enhancement

types, which are in the off-state for gate- source voltage VGs of zero and in
the on-state for VGs > V'th (n-channel) or VGs < V'th (p-channel). Due to
the greater mobility of electrons over holes, n-channel types offer a lower
on-resistance per transistor area and are therefore more popular in power
switching applications.
The cross-section of a vertical power-MOSFET is shown in Fig. 2.10. The
actual MOS transistor is located at the surface near the gate and the source
contacts. This construction has proven to be suited for high voltages due to
the great distance of the gate and the drain contact. As a further benefit,
the power is dissipated over a larger volume than is the case with lateral
MOS transistors , which makes these transistors well suited for high-power
applications.
 2 Power Circuits 29

100

<(
VGS

V05 / V

Fig. 2.11. Output characteristic of power-MOSFET

Figure 2.11 shows the output characteristic of a power-MOSFET. As long

as the gate- source voltage is zero, only a small leakage current flows, when
a positive drain- source voltage is applied. For a negative drain- source volt age
the intrinsic reverse diode conducts, so a reverse current will flow even for
Vcs = 0. W hen the gate-source voltage is increased, the transistor shows
a quasi-linear relationship between the drain current / 0 and the drain-source
voltage V08 as long as the MOSFET remains in the ohmic region.
The parasitic bipola r transistor and its base-collector capacitance made
early MOSFETs susceptible to turn-on for large gradients of the drain- source
voltage Vos- The source-metallization prohibits t his parasitic turn-on by

Fig. 2.12. Parasitic components in power-MOSFET

30 Stefan Butzmann

50V
J:::::=5=0=0=V=::r 1OO%
Packaging
+ ......···.. - - - - - - 1

c
r:r.o
.8
c
0
:g
.0
............................. t - - - - - - 1 ·;::
c0
JFET
region
u
+
Expitaxial
layer
·····························t-----+
........______.·············.t::===::i.o
Fig. 2.13. Relative contribution of different components to on-resistance of power-
MOSFETs with voltage ratings of 50 V and 500 V

shorting the base and the emitter, thus creating a pn-diode from the source
to the drain (see Fig. 2.12). This diode shows a long reverse recovery time
and its use is suitable only in low-frequency applications.
In switching applications the on-resistance of the power-MOSFET is one
of the most important parameters as it determines the power dissipation for
a given current. It is mainly contributed to by the following components:
- the resistance of the channel,
- the resistance of the parasitic junction-FET,
- the resistance of the drift region,
- the resistance of the substrate,
- the resistance of the bond wires and the metallization,
- the resistance of the source region.
Figure 2.13 shows the contribution of the different parasitic components
to the overall on-resistance for power-MOSFETs with voltage ratings of 50V
and of 500 V. At high voltages approximately 95% of the on-resistance RDs,on
result from the n- -epitaxial resistance, which is a function of the transistor-
breakdown voltage. For a chip-area of Achip = 1 cm 2 it is given by

Repi = 8 3. 10-9.
0 °
(UBR)
v
2.3
(2.3)

Therefore, the on-resistance of a MOSFET is dominated by the epi-

resistance particularly at high breakdown voltages. At breakdown voltages
of more than 600V the on-state voltage drop of usual power-MOSFETs is
higher than that of bipolar transistors of the same size. This makes the use of
IGBTs (Insulated Gate Bipolar Transistors) more attractive in high-voltage
applications.
 2 Power Circuits 31

Dynamic Characteristics

The dynamic characteristics of a power-MOSFET are mainly determined by

the parasitic capacitances, which are the gate-source, the gate-drain and the
drain-source capacitance (see Fig. 2.14).
The gate-source capacitance Ccs consists of three components. These are
the overlap of the gate electrode with the p-channel region, the overlap with
the n- -source region and the capacitance resulting from the proximity of
the gate electrode to the source metallization. The gate-source capacitance
is widely constant over the operating range of the MOSFET. It has to be
charged to the threshold voltage before the drain current starts to build
up.
The Ccn capacitor is formed by the overlap capacitance of the gate elec-
trode with the n-drift drain region, the Cns capacitor by the junction capac-
itance of the body diode. Both capacitances are nonlinear and functions of
the drain-source voltage.
In datasheets, usually the values Ciss, Coss and Crss are found (see
Fig. 2.15) with the following relations to the depicted capacitances:

Ciss = Ccs + Ccn (2.4)

Crss = Ccn (2.5)
Coss = Cns +Gen. (2.6)

These capacitances have a strong influence on the switching behavior of

the MOSFET. Figure 2.16 shows the turn-on behavior of a power-MOSFET.
At first the input capacitance Ciss is charged and the voltage Vcs rises until
the threshold voltage of the MOSFET is reached. The MOSFET then draws
increasing load current yielding a decrease of the drain-source voltage Vns.
As the drain voltage is coupled with the gate voltage via Ccn, Vcs is kept
on a constant level while Vns is changing (Miller plateau). When V DS has
reached its final value Vcs rises to Vin·
A simple approach to compare the dynamic characteristics of power-
MOSFETs is the use of the total gate charge Qc, which represents the

Fig. 2.14. Parasitic capacitances in power-MOSFET

32 Stefan Butzmann
2400
C/pF
I I I
1800 f = 1 MHz
VGS =0v

1200
"--- C4ss

600 1\
,"- r--- Coss
'- Crss
0
0 10 20 30

Fig. 2.15. Ciss, Coss and Crss of MOSFET IRF640 [12]

current-time product required to turn the MOSFET fully on. The datasheet
of the IRF640 by ST, for example, gives a total gate charge of Qc = 55 nC
for a gate-source voltage of Vcs = 10 V. Additionally, the gate-source charge
(Qcs = lOnC) and the gate-drain charge (Qcn = 21nC) are specified. The
gate-source charge specifies the amount of charge necessary to bring the gate
voltage from zero up to the threshold voltage vth, the gate-drain charge de-
fines the amount of charge needed to overcome the Miller effect, while the
drain voltage falls.
A proper figure of merit to describe the static and dynamic behavior
of MOSFETs is the product of the MOSFET's on-resistance and the gate
charge.
Figure of Merit= Rns,on · Qc. (2.7)

-----------·---· . \ lf-Vas/V
y---V /V 08
40 .............................. , ......... ··········.
I
I

.............................. ............. .li ....... ............................ .

I
I
I
I
; :

0 .................. :................

0.5 1.5
t/ us

Fig. 2.16. Turn-on behavior of MOSFET

 2 Power Circuits 33

2.2.5 IGBT

Up to a drain- source voltage of Vos :::; 600 V MOSFETs are in most cases
the best choice for power switching devices. For higher voltages , however, the
bipolar transistor has a lower on-voltage and it is even less expensive. As
described before, it cannot be driven powerless in static operation due to the
limited current gain though.
A power device combining the high-voltage capability of the bipolar tran-
sistor with the input stage of the MOSFET is the Insulated Gate Bipolar
Transistor (IGBT). IGBTs are generally used in applications where the on-
resistance of MOSFETs becomes impractically high. The main difference be-
tween a MOSFET and an IGBT is the additional p-layer at the drain contact
which injects minority carriers into the n-zone and thus improves the con-
ductivity of the device. Figure 2.17a shows the non-punch-through (NPT)
structure of an IGBT, which consists of an n+pn - p+ -sandwich. With this
structure, the breakdown voltage of the device is determined by the doping
and the thickness of then- -layer, where the electric field can spread.
When the device blocks, the upper pn-transition is exposed to the max-
imum critical field strength. The area under the E-curve is proportional to
the collector- emitter voltage of the IGBT, while the slope is proportional
to the doping of the n- -layer. An alternative structure, the punch-through
(PT) IGBT employs an additional n+ -layer between then- and the p+ -layer
in order to increase the blocking voltage. This allows a thinner n - -layer and
reduces the on-state losses of the device at the expense of higher transient
losses.
This can be seen from Fig. 2.18, which shows the trade-off between the
on-state voltage at current densities of 100 A/ cm 2 and the turn-off-energy for
both punch-through and non-punch-through IGBTs with voltage ratings of
1200V.

i£1

c c
a NPT-IGBT b PT-IGBT

Fig. 2.17. Cross-section of IGBTs. a Non-punch-through, b punch-through

34 Stefan Butzmann

15 2 25 3
On-stage voltage IV

Fig. 2.18. Turn-off losses vs. on-state voltages at 100 A/cm 2 for 1.2 kV PT and
NPT-IGBTs [15]

Additionally, the turn-off losses of PT-IGBTs increase at a higher rate

than those of NPT-IGBTs, when the temperature increases. This behavior
can yield thermal instabilities for PT-IGBTs at high switching frequencies,
where higher switching losses result in increasing temperatures and vice versa.
In the worst case, the resulting thermal runaway can destroy the device.
NPT-IGBTs in comparison do not show this behavior and generally remain
thermally stable.
Figure 2.19 shows the equivalent circuit and the output characteristic of
a power-IGBT. The inherent pnp-transistor of the IGBT increases the on-
state voltage of the power device by the forward voltage of its base--emitter
junction.
The collector-emitter voltage of the IGBT is widely constant for changing
load currents. Thus, at high current levels the conduction losses are lower in
comparison to MOSFETs and at low current levels they are higher.
The saturation voltage VcE,sat of the IGBT is a function of the collector
current Ic, the gate-emitter voltage VeE and the junction temperature TJ.
As the IGBT is generally used as a switch, VcE,sat determines the static power
dissipation. It is contributed to by the on-resistance of the MOSFET stage
and the base-emitter voltage of the pnp-transistor. VcE,sat is temperature-
sensitive and can have both a positive or a negative temperature coefficient in
the area of operation depending on the doping of the device. In order to enable
parallel operation of several IGBTs, IGBTs are usually manufactured with
a slightly positive temperature coefficient of VcE,sat. The reverse blocking
voltage of IGBTs is determined by the breakdown voltage of the pn-junction
on the backside of the wafer and is typically in the order of 20 V.
 2 Power Circuits 35

VcEIV ---+
Fig. 2.19. Output characteristic of an IGBT

While the turn-on behavior of IGBTs is similar to that of MOSFETs,

the turn-off behavior is determined by the tail current, which is due to the
storage of minority carriers in the epi-layer (see Fig. 2.20). As the IGBT's
MOSFET stage is turned off quickly, the turn-off behavior of the bipolar
stage is similar to that shown in Fig. 2.7a for the bipolar transistor as the
minority carriers in the pnp-base cannot be removed externally. The resulting
tail current continues until all minority carriers have recombined. This tail
current has a significant influence on the devices's power dissipation as it
increases the turn-off-time of the device.

50 ........ ;........... ;........... ;.......... .

:. : : '-----':----'
. .. ..
40 .......... : ...........;.......... , ........... ;........... ;..........., ..........., ......... .

30 . . . . . . .
.: :. :. :. :. :. :.
20
:
10 ··········!···········!\·.;.;··· .; ........... : ..........;...........;...........[......... .
:
:
:
:
'' .
:
:
:
:
:
:
:
:
:
0 ........ ) ......... ) ........... ........... ........... : : :
1

- 10

tl

Fig. 2.20. Turn-off behavior of IGBT

36 Stefan Butzmann

2.2.6 Thyristor

The thyristor combines high off-state voltages with high on-state currents
and has therefore been the preferred device in high-power applications for
a long time. Figure 2.21 shows a thyristor and its equivalent circuit.
The thyristor has a pnpn-structure with three terminals. During the off-
state both the npn- and the pup-transistor are nonconductive, and the leakage
current through the transistors and their current gain is so low that the device
remains off even for a positive voltage from the anode to the cathode. When
a proper signal is applied at the gate, the npn-transistor starts conducting and
biases the pup-transistor into conduction. The positive feedback structured
device latches and remains conducting until the anode-to-cathode-current is
reduced below the hold-current necessary to keep the device in the on-state.
The necessary gate current to fire the thyristor is dependent on the anode-
cathode voltage. If this voltage is raised above the breakdown voltage Vsn, the
thyristor will fire by itself. Furthermore, the gate current required to fire the
thyristor is also a function of the temperature. With increasing temperature
thyristors become more sensitive. Undesired firing of the thyristor can also
occur due to high voltage slopes dV/dt at the thyristor terminals, which result
in displacement currents through the internal junctions.
While the "vertical" triggering action in the thyristor occurs at high
speeds, the "horizontal" distribution of charge is only relatively slow. Par-
ticularly for large area elements, there is a danger of high current concen-
tration in the gate region during turn-on, which can cause local overheat-
ing and destroy the power device. For this reason, the current slope dJ /dt
during turn-on has to be limited. This is usually done by external snubber
circuits.
The turn-off of a thyristor is generally initiated by a negative voltage
from the anode to the cathode, as it occurs for example with a sinusoidal
supply voltage. Similar to power diodes, the current flows in the negative
direction during turn-off until recombination has been completed. Only after

,node
A

ga

G cathode
VGl lvs
c
a b c

Fig. 2.21. Thyristor. a Principle, b equivalent circuit, c symbol

 2 Power Circuits 37

a minimum off-time may a positive anode-cathode voltage be applied again

without unintentional retriggering of the thyristor.
The lack of controlled turn-off-capability is a tremendous drawback for the
use of thyristors in actuator drive circuits. in particular, the development of
IGBTs, which achieve voltage ratings in the kV range, has provided enormous
competition for thyristors in the field of actuator drives.
Modified thyristors offering turn-off capability are GTOs and MCTs. As
these devices are generally used in high-power applications only, they will not
be discussed here though.

2.3 Drive Circuits for MOS-gated Devices

2.3.1 General Remarks

The ideal drive circuit for a power switch is able to charge or discharge the
input capacitance of a MOSFET extremely fast and has a low source resis-
tance. As MOSFETs themselves only sink or source current during switch-on
or -off, the drive circuit itself should need only low supply current, otherwise
the efficiency of the system would be degraded. Drive circuits can be classi-
fied into the following groups according to the position of the switch relative
to the load (see Fig. 2.22):

low-side drivers, where the load is connected to the supply voltage VB and
the switch connects the load to ground,
- high-side drivers, where the load is connected to ground and the power
switch connects the load to the supply voltage.

L D

D
______ _j PT
1
1
dovec

a b
Fig. 2.22. Transistor as low- and high-side switch. a Low-side switch, b high-side
switch
38 Stefan Butzmann

Isolated gate drives are used to provide high potential differences and
swings between the control circuit and the power switch. If the power switch
is used with an isolated gate drive, it can be applied both as a high-side and
as a low-side switch.
For each topology, different drive circuits are commonly used, which will
be described in the following. A first overview of the different circuits is given
by Table 2.3.

Table 2.3. Classification of drive circuits for MOS-gated devices

Type of drive Implementation Remarks

circuit
Low-side Complementary Vee-output swing
stage input signal required to cover Vee range
fast transition
inverting behavior
danger of cross-currents in the drive cir-
cuit during turn-on and turn-off

Totem-pole stage inverting behavior

no cross-currents
input signal with only small amplitude
required

Emitter-coupled noninverting behavior

automatic protection of gate against
overvoltages
no cross-currents
delay, as input signal has to change by
2VBE before switching action occurs

High-side p-channel driver only for p-channel MOSFETs, therefore

with level-shifting higher power dissipation in switch
stage simple circuit

Bootstrap driver for n-channel devices

external bootstrap capacitor required
high-side switch has to be turned off
regularly in order to recharge bootstrap
capacitor, therefore only for minimum
operating frequency

Charge pump for n-channel devices

additional supply voltage above battery
voltage created
higher effort
 2 Power Circuits 39

Table 2.3. (continued)

Type of drive Implementation Remarks

circuit
Isolated Opto-coupled medium speed
drivers additional floating power supply needed
to drive power switch

Transformer- fast transition

coupled gate drives - no additional power supply on sec-
ondary side
expensive components

2.3.2 Low-side Drive Circuits

Figure 2.23a shows the complementary drive stage. This stage shows an in-
verting behavior and enables a fast turn-on and turn-off of the power tran-
sistor. This configuration of a driver output stage utilizes the Vee range to
its maximum as both driving transistors are operated with lowest drain-
source voltages. Depending on the implementation of the circuit high peak
currents through T 1 and T 2 can occur when both transistors are conducting
(see Fig. 2.23b), so careful design of the driving stage is necessary.
In integrated circuits additional logic or timing components are used to
prohibit the described cross-currents. The circuit in Fig. 2.24 shows the prin-
ciple of a complementary drive stage with delay logic.
If the input voltage Vin is constant, the output of the second inverter
INV 2 is equal to the input voltage. If Vin is high, then the output of the
NAND gate will become low resulting in T 1 conducting. If Vin is low, then
the output of the NOR gate will be high and T 2 conducts. Every time the

Vs
5
---- 'o.n' A
4 - JD,T2/A

vcJ
ov
0

b -10 20 40 60 80 100
a t/ ns

Fig. 2.23. Complementary drive stage. a Schematic, b simulation of cross-currents

through T1 and T2 during switching
40 Stefan Butzmann

Fig. 2.24. Complementary drive stage with delay logic to prevent cross-currents

input voltage changes, the output of INV2 and Vin will have different logic
levels for a delay time, which is determined by the RC network. During this
time, both transistors T 1 and T 2 will be off thus preventing a cross-current
from flowing through T 1 and T 2 .
A further approach of a low-side drive circuit is the push-pull stage with
coupled emitters (Fig. 2.25). The depicted circuit is a noninverting structure
with a low source resistance and is frequently used to boost current from IC
drivers.
At the beginning of a switching cycle the input voltage has to change by
2 VaE before a change of Vcs of the power transistor occurs. Consequently,
only one of the two driving transistors conducts at one time, which results
in the absence of the current shoot-through observed with the complemen-
tary stage. An advantage of this configuration is the ability to clamp the
gate-source voltage Vcs between Vee + VaE and GND- VaE thus providing
a protection of the power-MOSFET against too high gate voltages.

Vcc=15V

PT

Fig. 2.25. Push-pull stage with coupled emitters

 2 Power Circuits 41

5VJ
ov

Fig. 2.26. Totem-pole stage

The totem-pole configuration in Fig. 2.26 with an inverting behavior also

prevents the current peaks, which are caused by the complementary-stage
architecture. However, for Vin =High a continuous current flows through R 1
and T2, which increases the power dissipation of the drive circuit. The mini-
mum gate voltage Vcs of the power-MOSFET is increased to the sum of the
saturation voltage of T 2 and the forward voltage drop of the diode. In order
to ensure a turn-off of the power-MOSFET when the driver supply voltage is
removed, the resistor R 2 can be added. This, however, yields a constant bias
current also for Vin = Low.

2.3.3 High-side Drive Circuits

Drive Circuits for p-channel MOSFETs

Applications with grounded loads require the switch to be placed between

the supply voltage and the load. The easiest way to implement a high-side
switch is the use of a p-channel MOSFET. The device is turned on by pulling
the gate voltage low.
Figure 2.27a shows the basic drive circuit for a p-channel power-MOSFET.
When the transistor T 1 is turned on the gate voltage is pulled low via the gate
resistor Rc. When T 1 is turned off, the gate-source capacitance is discharged
via R 1 and the power-MOSFET is turned off.
If the supply voltage VB is larger than the maximum allowed gate-
source voltage Vcs of the MOSFET, the circuit can be modified according
to Fig. 2.27b, where the maximum gate-source voltage is limited by a Zener
diode.
The main drawback of both circuits lies in the values of R 1 and R 2, which
should be high in order to minimize the bias current when the MOSFET is
42 Stefan Butzmann

a b

c
Fig. 2.27. Drive circuits for p-channel MOSFETs. a Principle, b with gate-source
voltage limiting diode, c with emitter-follower stage

turned on. This however yields large switching times. A simple solution is
the circuit according to Fig. 2.27c, where an additional transistor stage has
been added to provide a low-impedance driving output.

Bootstrap Drivers for n-channel MOSFETs

P-channel-MOSFET s are built on p-type epitaxial material and the major-

ity carriers are holes. Due to the lower mobility of holes in comparison to
electrons the on-resistance of p-channel devices is 2-3 times higher than that
of n-channel devices with the same chip size. This yields a higher power
dissipation, which is often not affordable.
For this reason, in most power applications n-channel devices are used. In
order to turn these devices fully on, their gate voltage has to be pulled above
 2 Power Circuits 43

Fig. 2.28. High-side driver

the supply voltage VB. Figure 2.28 shows the principle of a high-side driver
with a bootstrap capacitor.
For Vin = High the power-MOSFET is turned off and the bootstrap ca-
pacitor CB is charged via the diode D1 and the load ZL. As soon as Vin
becomes Low the Transistor T 1 is switched off and the potential of the gate
of the power transistor PT is increased. As the source voltage is raised to
approximately VB the bootstrap voltage V* is raised to VB + Vee resulting
in a constant voltage Vcs of the MOSFET. Due to leakage currents in the
driving circuit the capacitor is slowly discharged, which yields an increase of
the on-resistance of the MOSFET and thus in a higher power dissipation.
The bootstrap capacitor has to be recharged by a return of the source volt-
age to zero, which is achieved by turning the power switch off. The described

------------------------------------------------------------'
Vee '
'
'
'
'

Pulse
generator
_n_

------------------------------------------------------------'
'

Fig. 2.29. Integrated level-shifting stage for high-side driver

44 Stefan Butzmann

behavior results in a minimum operating frequency, which is dependent on

the size of the bootstrap capacitor CB.
In integrated circuits, often a level shifting stage as shown in Fig. 2.29 is
employed. The turn-on and turn-off state of the power transistor are entered
by two short pulses generated by driving logic. An RS flip-flop, which is
coupled to the boosted supply voltage, translates these pulses into static
signals at its outputs, thus turning the power-MOSFET on and off. This
method of level shifting has the advantage that the actual driving stage is
decoupled from the level-shifting transistor yielding a faster turn-on of the
power switch.

Charge Pumps

In some applications, e.g. DC motors, the on-time of the high-side switch

may be undefined. As explained above, the previously described bootstrap
circuits suffer from the discharge of the bootstrap capacitor, which results
in a required minimum switching frequency. These concepts would fail in
applications where the on-time is greater than the size the bootstrap capacitor
permits. Charge pump circuits are a frequently used alternative in these cases
in order to create a static voltage above VB.
The inverter in the circuit in Fig. 2.30 is driven by a continuous pulse
train and pulls the capacitor electrode from ground to VB and back. When
the inverter output is low, the capacitor Cpump is charged via the diode D 1
to VB- VD, when it is high, the capacitor Ctank is charged from Cpump via D2
to 2VB- 2VD. If the turn-on time of the power-MOSFET is not critical, the

PT

Fig. 2.30. Charge pump circuit

 2 Power Circuits 45

Vs

Fig. 2.31. Elementary version of charge-pump driver with increased turn-on time

driving circuit consisting of T 1 - T 3 can be saved by replacing the capacitor

Ctank directly by the power-MOSFET (Fig. 2.31).
Some integrated MOSFET drivers combine the bootstrap-technique with
the charge-pump technique in order to provide both fast switching action
at high frequencies and static turn-on conditions of the MOSFET. The en-
ergy for the gate switching action itself is taken from the external bootstrap
capacitor and the energy needed to maintain the gate charge constant is sup-
plied via the charge pump. With this architecture, the necessary charge-pump
capacitor can now be integrated as the power to be transferred is low.

2.3.4 Isolated Gate Drive Circuits

Transformer-coupled Gate Drives

In applications with large potential differences between the control circuit

and the power switch galvanic isolation may be desired. Apart from solutions
with opto-couplers, the use of transformers is the most common approach
for realization of isolated gate drive circuits. Figure 2.32 shows the basic
structure of a transformer-coupled gate drive.

PT

Fig. 2.32. Transformer-coupled gate drive

46 Stefan Butzmann

An important property of transformers is the fact that they can deliver

only AC signals. This results in a limitation of transformer drives to a duty
cycle of 50% or the necessity to realize large voltage swings for duty cycles
other than 50%, otherwise the flux in the core would constantly increase and
finally yield saturation of the core. A common approach to enable varying
duty cycles is the use of a Zener diode as shown in Fig. 2.33.

PT

Fig. 2.33. Transformer-coupled gate drive for varying duty cycles

0
-1
0 2 4 6 8 10

0 2 4 6 8 10
t/ ms

Fig. 2.34. Simulation results for circuit according to Fig. 2.33

 2 Power Circuits 47

During turn-on of T 1 the voltage Vee is coupled to the gate-source elec-

trodes of the power-MOSFET via the transformer. When T 1 is turned off
again, the current is decreased via the Zener diode ZD. The voltage Vz of the
Zener diode has to be chosen to cause a total flux reset in the transformer
for the maximum occurring duty cycle. However, the (negative) coil voltage
of the secondary side of the transformer may not exceed the maximum spec-
ified negative Vcs value of the MOSFET, which is usually in the order of
20 V. Thus, with typical on-voltages of the power-MOSFET of 10-15 V, only
relatively small variations of the duty cycle can be realized.
The simulation in Fig. 2.34 shows an additional problem of the circuit.
When the current through the primary coil has been reduced to zero, the
current flow through the diode is cut off and the drain voltage starts to
oscillate.
A better approach allowing wider duty cycle variations and simultane-
ously preventing core saturation is the use of a decoupling capacitor on the

r- r-r-
>
......
0

>-10

-20L_________L_______
0 0.5 1.5 2

15r- r- r- r- r- r- r- r- r- r- r- r-
::: 10
en
5
0

0.5 1 1.5 2
b t/ms

Fig. 2.35. Transformer-coupled gate drive with decoupling capacitor on primary

and secondary winding for varying duty cycles. a Schematic, b simulation results
for secondary coil voltage and Vas of power-MOSFET PT
48 Stefan Butzmann

primary side of the transformer. In order to damp the resulting LC cir-

cuit formed by the decoupling capacitor and the transformer inductance
a damping resistor R 1 is often added. For duty cycle ratios, which are near
zero or 100%, extreme voltage levels may occur, where the turn-on voltage
has decreased so much that the power device will not safely turn-on. Fig-
ure 2.35 shows a circuit which uses a second capacitor and an additional
diode to shift the driving voltage levels back. When the output voltage
of the transformer is negative, the capacitor C 2 is charged to the trans-
former's output voltage minus the forward voltage of diode D. During the
positive half-wave the capacitor voltage is added to the transformer's output
(Fig. 2.35b).
The circuit shown in Fig. 2.36 uses an additional MOSFET to drive the
power switch. When the voltage across the transformer's secondary winding
is positive, the gate-source capacitance of PT 2 is charged via the intrinsic
backward diode of PT 1 . Even when the voltage across the secondary winding
falls to zero, the power transistor PT 2 will remain conducting. After the
output voltage of the transformer has become negative, PT 1 is turned on
and the gate-source capacitance of PT 2 is discharged. The transformer in
this circuit can have a smaller size as it is operated partly in saturation.

Fig. 2.36. Transformer-coupled gate drive with additional MOSFET

Gate Drives with Opto-couplers

In order to achieve galvanic isolation opto-couplers offer the following advan-

tages:
- high isolation voltages of up to 5 kV,
- signal transfer from DC to several MHz,
- simple interfacing to microcontrollers.
Figure 2.37 shows a common circuit used for opto-coupler gate drive iso-
lation, where the signal from the opto-coupler is amplified by two transistor
 2 Power Circuits 49

.. ________________________,.
f PT

Fig. 2.37. Gate drive circuit with opto-coupler

stages. A disadvantage of opto-couplers is the need of an additional power

supply for the driving circuit on the "secondary" side. Furthermore, the rela-
tively large propagation delay of typically several microseconds can be a fur-
ther disadvantage.

2.3.5 Speed Enhancement and Protection Circuits

Turn-off Speed Enhancement

Apart from dvjdt considerations, the turn-off process of the MOSFET should
take place as fast as possible in order to minimize the switching losses. Ideally,
Rc is tuned for the maximum allowed switching speed of the diode. For
different turn-on and turn-off speeds the circuits in Fig. 2.38 may be used.
In both circuits the turn-on speed is determined by the resistor Rm. The
turn-off speed is controlled by RmiiRG2 in Fig. 2.38a and by RG2 only in
Fig. 2.38b. The main advantage of the turn-off with the pnp-transistor is
given by the fact that the (sometimes significant) gate turn-off current is
held in the smallest possible loop between the gate and the source by the
pnp-transistor, thus reducing EMI problems and also reducing the power
dissipation in the driving stage.
Additional speed-up of the turn-off process can be achieved if a negative
supply voltage for the driving circuit is used. By this measure the charge can
be removed even faster from the gate capacitance of the power-MOSFET.

Overvoltage Protection

Overvoltages occur most commonly when the power switch is turned off while
the load inductance is carrying a current. As stated previously, the diode
starts to conduct only after it has been flooded with minority carriers. Until
this happens, the inductive load keeps a constant current and creates an
50 Stefan Butzmann

15V ____c IVin

-5V :=:r- t

Fig. 2.38. Speed-up circuits for faster turn-off. a Reduction of turn-off time by
use of diode, b use of pnp-transistor for additionally reduction of EMI and power
losses in driving stage

overvoltage at the drain of the MOSFET. Active clamping of MOS-controlled

transistors is a popular way to prevent overvoltages. The principle is shown
in Fig. 2.39.
An additional feedback path consisting of a Zener diode ZD and an ad-
ditional diode D is introduced into the circuit. When the drain voltage rises
above the sum of the breakthrough voltage of the Zener diode and the thresh-
old voltage of the power transistor, the gate voltage of the MOSFET is raised
and the transistor is turned on. The MOSFET in combination with the Zener
element thus acts like a power Zener diode. The diode D in the circuit has
the function to prevent current flow from the driver to the drain when the
power transistor is turned on.
Figure 2.40 shows a simulation of the described clamping circuit, which
was performed without a free-wheeling diode. For the diode ZD a Zener
voltage of Vz = 30V was chosen. When the power-MOSFET is turned off at
t = 5 ms, the current through the inductor increases the drain voltage of the
power-MOSFET. When the clamping voltage
 2 Power Circuits 51
.----.---o Vs

ZD

PT

Fig. 2.39. Active-clamping circuit for overvoltage protection

II- Drain voltage I

> 30 H- - · Gate voltage

r-----------------------
10
\. __________ _

0 2 4 6 8 10

<( 10
_o
c
:::J
5
u
c

0
2 6 8 10
t I ms

Fig. 2.40. Simulation of active clamping circuit without free-wheeling diode

Vclamp = vth,MOSFET + Vz + VF (2.8)

is reached, the power-MOSFET turns on again. In the simulated case without

a free-wheeling diode, the MOSFET is kept in the on-state until the current
through the inductor has been reduced to zero. During this phase, the gate
voltage is held at the threshold level.
52 Stefan Butzmann

Protection Against Gate-Source Overvoltages

Possible causes for gate-source overvoltages are not only error conditions in
the driving circuit, but also the occurrence of a short circuit of the load when
the power-MOSFET is already in the on-state. In this case both the drain
voltage and the drain current rise very quickly. The large resulting voltage
slope dV/ dt in conjunction with the gate-drain capacitance can then yield
a further increase of the gate-source voltage. In order to protect the gate
against such overvoltages, usually two Zener diodes are connected back to
back between gate and source according to Fig. 2.41.

PT

Fig. 2.41. Protection against gate-source overvoltages

Overload Protection

Protecting the power device against short-circuit currents requires additional

circuitry in the driver output stage as the transistor should be turned off fast
after detection of the fault condition.
The most popular method of current protection is undoubtedly the use
of a protection circuit measuring the voltage over a shunt resistor, which is
connected between source and ground (see Fig. 2.42a).
This method is simple, but it requires a low resistance for large currents
in order to keep the shunt's power dissipation low. This again yields small
signal amplitudes, which are sensitive to interference and variations of the
ground level due to parasitic inductances. Measuring a current of 50 A with
a shunt resistor of Rs = 1 m!l gives a sense voltage of Vm = 50 m V and still
yields a power dissipation of P0 = 2.5 W in the shunt resistor.
Because of the small signal amplitudes over the shunt resistor a careful
layout is necessary; the use of differential sense amplifiers can be of further
advantage.
 2 Power Circuits 53

Fig. 2.42. Current measurement with shunt resistor. a Shunt resistor in main
current path, b use of shunt resistor with sense-FET

Current transformers offer much lower susceptibility to interference and

even provide galvanic isolation between the power path and the control cir-
cuit. Their cost is much higher than that of shunt resistors, though.
In sense-MOSFETs a small source area is separated from the main current
path providing a second source contact. The current through this source
contact is proportional to the current in the main path reduced by the area
ratio of the two source contacts. An easy way of evaluating the sense current
is shown in Fig. 2.42b. In this configuration, Rs must be chosen to be small
as the measuring voltage over the shunt resistor reduces the sense current.
A more exact way of evaluating the sense current is shown in Fig. 2.43,
where the sense connection is virtually grounded via the operational ampli-
fier. The output of the op-amp is negative, though, thus the circuit requires
a negative supply voltage.

Fig. 2.43. Current measurement with sense-FET and virtual grounding

54 Stefan Butzmann

Figure 2.44 shows the principle of an integrated overcurrent limitation

circuit. The current source IRef creates a reference voltage VRef over RRef
and thus a constant potential at the base of T 3 . While the voltage over the
shunt resistor R 2 is lower than VRef, T 2 is operated in saturation creating
a voltage of VR2 + VBE,4 + VcE,sat,2 0.8 Vat the base of T5. In that case the
Darlington stage T 5,6 is not conducting and the gate of the power transistor is
maintained high via the current source h. If an overcurrent situation occurs,
then the base voltage of T 4 is raised above VRef and the current through T 2
and T 4 is cut off. This yields a turn-on of the Darlington stage and the gate
of the MOSFET is discharged until the current through the MOSFET has
been reduced to IRef · RRed Rz.

Fig. 2.44. Current limitation circuit

dV/dt Protection

When MOSFETs are exposed to high voltage slew rates during turn-off,
a current Ic = C · dVjdt will flow through the collector-base capacitance of
the parasitic bipolar transistor. In this case, it may happen that the product
of this current and of the resistance RB is large enough to spuriously turn the
transistor on. This uncontrolled action results in higher power dissipation and
undesired interferences. In the worst case, it can even destroy the MOSFET.
dV/dt-limiting has therefore become a central issue for drive circuits, also in
order to reduce the insulation stress in the load.
Figure 2.45 shows a dV/dt-protection circuit, where a capacitor is used
to limit the voltage slew rate during turn-off of the power transistor.
 2 Power Circuits 55

PT

Fig. 2.45. dV/dt-limiting with feedback capacitor

Overtemperature Protection

The maximum allowed power dissipation in a power device is a function

of the ambient temperature, the thermal impedance of the setup and the
maximum allowed junction temperature, which is for typical power devices in
the order of 19j,max = 150-175 °C. If the junction temperature rises above the
allowed limit, the power dissipation in the device has to be reduced. Usually
the power-MOSFET is switched off above a certain junction temperature
19 0 ff and switched on again after the junction temperature has fallen below
a temperature 11on·
A simple way to measure the temperature is the use of the base-
emitter junction of a bipolar transistor, where the VsE voltage degrades with
:::::5 -2mV /K. More sophisticated approaches - restricted to monolithically

integrated smart-power devices- evaluate the current generated by a PTAT

current source (PTAT = proportional to absolute temperature), where the
current is a function of the absolute temperature [1].

2.4 Thermal Management

Temperature is certainly one of the most critical factors contributing to device

failure. The increase of power densities in modern mechatronic systems make
an effective thermal management more and more important.
Temperature-caused failures are either caused by mechanical stress or due
to temperature-dependent electrical failures. To the first group belong, e.g.
bond-wire fatigue- caused by badly matched thermal expansion coefficients
between the wire, the bond pad or the substrate - or die fracture in pack-
ages due to thermal expansion coefficient mismatching of the silicon and the
plastic. Failures of this kind are normally caused by thermal cycling of the
device.
56 Stefan Butzmann

RG,int

Fig. 2.46. Equivalent input circuit of MOSFET

Electrical failures are generated when the junction heats up until the
material reaches its melting temperature. Electromigration resulting from
high current densities can cause voids in the metallization which lead to
open circuit failure.
It is therefore extremely important that the maximum specified junction
temperature is not exceeded during device operation. Power dissipation in
power devices is contributed to by the following factors:

- Conduction losses, which are given by the product of the current during
the on-phase and the voltage drop over the device.
For power-MOSFETs the conduction losses are given by

Pn =Ran· f6 (2.9)

As power-MOSFETs do not possess a second breakdown, these devices

are operated safely as long as their junction temperature does not exceed
the specified value, which usually is in the range from 150 oc to 175 °C.
The limiting values of Vns and In are found in the datasheets of the power
devices in the safe operating area diagram (SOA diagram).
- Switching losses, which result from large voltages and large currents at
the same time during the transient phase between the on- and off-state.
Switching losses are proportional to the switching frequency. They are
generally negligible for low frequencies and may become dominant at high
frequencies. Switching losses highly depend on the drive circuit and the
reverse recovery characteristic of the other power elements in the circuit.
- Gate losses in MOS-driven devices are due to the power dissipation in the
internal gate resistance.
Assuming the equivalent circuit of a MOSFET or IGBT input stage ac-
cording to Fig. 2.46, the power loss at Ra,int can be described by

P.
a=
V?
in·
f · Ciss·R Ra,int
+R (2.10)
G,ext G,int

As the input capacitance varies with the drain-source voltage, the de-
scription via the gate charge provides a more suitable description:
 2 Power Circuits 57

p,G = TT
Vin ·
f · QG ·
Rc,int
(2.11)
Rc,ext + Rc,int
Usually, the internal gate resistance is much smaller than the external one.
If a reasonably sized external gate resistance is chosen, it is safe to assume
that the greatest part of the gate drive losses is dissipated in the external
gate resistance. The internal gate resistance thus does not significantly
contribute to the heating of the device and is therefore negligible.

The maximum allowed power dissipation is determined by the ambient

temperature and by the thermal impedance Zth,ja of the power transistor:

'19j,max -'I?a = Zth,ja · P max (2.12)

with:

'19j,max maximum junction temperature,

'I? a ambient temperature,
Zth,ja thermal impedance from junction to ambient,
Pmax maximum allowed power dissipation.

The thermal impedance Zth,ja depends on the die size, the thermal pa-
rameters of the materials and on the geometry of the heatsink. It consists of
a real part Rth, which determines the static junction temperature for a given
constant power dissipation, and of a thermal capacitance Cth, which describes
the temperature transients.
In the manufacturers' datasheets, Rth,jc (the thermal resistance from
junction to case) is specified and is of the order of 1-2K/W for typical
power packages. The thermal impedance is generally estimated by the use
of finite-element programs. Figure 2.47 shows a power module construction
with a power-MOSFET soldered on a DCB (Direct Copper Bonding) sub-
strate. This setup, which is frequently used in modules designed for high
power dissipation, was simulated with a finite-element program in the exam-
ple shown. In order to calculate the static thermal resistance of the setup,
the bottom of the DCB substrate was set to a constant temperature '!?a. As
the setup is symmetrical only one quarter transistor was simulated. The tem-
perature difference between the hottest spot on the die and the base plate
with the constant temperature gives the thermal impedance in relation to
the dissipated power.
In a similar way the transient thermal behavior can be simulated. Fig-
ure 2.48 shows an electrical circuit used to model the thermal impedances.
The transient behavior of the transistor can be approximated by an RC cir-
cuit (Fig. 2.48) in order to describe the thermal behavior.
As stated earlier, the on-resistance of a power-MOSFET nearly doubles
during a temperature rise of the junction temperature from '19j = 25 oc to
'19j = 125 °C. Assuming a constant current flowing through the MOSFET,
58 Stefan Butzmann

Power
Al,O, module
Glue

Heat sink
(Aluminum)

; - - - - - - - - - -- - - - - - - , ANSYS 5 . Z
JlOV 18 1996
1 2: S7:45
PLOT DO .
NO DAL SOLOTIOII

TIP C• 7Z . l61
S!IX • 1. 716

. 190687
. 38l37t
. S7Z06
. ?62?47
. 953434
1.144
1 . 335
J . SZS
l . 716

Fig. 2 .47. Construction of DCB and finite-element simulation of temperature dis-

tribution [20]

the resulting power dissipation Po = Ros,on · I'5 yields an increase of the

junction temperature which in turn raises the MOSFET's on-resistance. This
again results in an increase of Po. This behavior is called the self-heating
effect (see Fig. 2.49) and can be simulated only by iterative simulations. The
results show higher junction temperatures than can be expected without
consideration of the self-heating effect.
A proper way to simulate the self-heating effect is the extension of the
transistor model by an additional resistor in series to the power-MOSFET,
which accounts for the rise of the on-resistance over temperature. The com-
plete electro-thermal model of a power-MOSFET is shown in Fig. 2.50. In the
model the power dissipation over the complete MOSFET is calculated and
the junction temperature {} J is calculated via the thermal model. The new
junction temperature is used to determine the on-resistance of the power
switch and the additional series resistor is adapted to the correct value
.D.Ros-
 2 Power Circuits 59

Cth,1

P(t)
Cth,2

Cth,3

Fig. 2.48. Equivalent electrical circuit for modelling thermal impedances

Figure 2.51 shows a comparison of the simulations of the transient tem-

perature with and without consideration of the self-heating effect. The sim-
ulation was performed for the described DCB setup with a 200 V power-
MOSFET with a nominal on-resistance of Rns,on = 200 mO at a temperature
of {)a = 25 °C. It was assumed that the ·power-MOSFET conducts a drain
current of In = 10 A from t = 0 on. The simulation results show, that the
increase of the junction temperature is L\{}J 30 °C, which corresponds to
a rise of the on-resistance L\Rns = 60 mO. Without consideration of the self-
heating effect, the junction temperature of the transistor is calculated to be
too low by L\{)J = 9 °C.

Power
dissipation
...,.. Heat
generation

/
Increase of
Roson

Fig. 2.49. Self-heating effect in MOSFETs

60 Stefan Butzmann

,..............................................................................................................................................................................................................................................

f1Ros = f(T) - TJ

D'

_i
i__ _______

Fig. 2.50. Electro-thermal model of a power-MOSFET with self-heating capability

- - with self-heating
55 . - - · without self-heating

---- ------ ---------------

25.. . . . . . . . . . .- ........................ .

tis
Fig. 2.51. Comparison of temperature simulations with and without consideration
of self-heating effect

2.5 DC Switching Converters

2.5.1 Topologies

DC chopper circuits are applied where an inductive load is to be supplied with

a variable current from a fixed DC supply voltage. A common application is
the drive of DC motors from a battery supply (see Sect. 3.2).
Figure 2.52 shows the principle of a DC chopper circuit with an inductive
load L.
The operation of the chopper circuit can be divided into two modes. When
the switch is closed, the current flows from the supply voltage to the load.
 2 Power Circuits 61

Fig. 2.52. DC chopper circuit for one-quadrant operation (Buck-topology)

According to dJ Idt = VI L, the current through the inductance rises. When

the switch is opened, the current continues to flow through the inductance
and seeks its way through the free-wheeling diode D. It decreases slowly due
to the forward voltage drop of the diode and due to parasitic resistances.
The described converter topology is referred to as a single-quadrant chopper
with positive load voltage and positive load current (class-A chopper, buck
converter topology).
Another chopper circuit operates with positive supply voltage, but nega-
tive load current and is referred to as a class-B chopper (boost converter).
Figure 2.53 shows the circuit where the voltage source Vemr is for ex-
ample the back electromotive force of a DC motor. When the switch 8 is
closed, the current through the inductor rises. When 8 is opened again, the
inductor current flows through the diode D and returns energy to the supply
voltage VB.
Combining the two circuits allows operation with positive supply voltage
and bidirectional current (class-C converter; two-quadrant-operation) while
the class-D converter provides a unidirectional current and bipolar operation
(Fig. 2.54). The class E converter (Fig. 2.55) provides both motoring and
regenerating operation in both directions (four-quadrant-operation). Usually
the switches 8 1 and 8 4 are on simultaneously and the switches 8 2 and 83
are also on simultaneously, thus allowing the load to be connected to +VB, 0
and -VB.

!Vemf

Fig. 2.53. DC chopper circuit for one-quadrant operation (boost topology)

62 Stefan Butzmann

Fig. 2.54. Two-quadrant converters. a Class C, b class D

Fig. 2.55. Four-quadrant class-E converter

2.5.2 Electromagnetic Value Actuation with a Class-D Stage

Class-D and class-E converters are the most commonly used topologies for
driving inductive actuators. While class-E converters are mainly used in mo-
tor drives to enable forward and backward rotation (see Sect. 3.5.2, 4.2.1),
class-D converters are frequently applied for the drive of electromagnets,
where the direction of the produced magnetic field is of no further interest.
Examples are lifting magnets or valve actuators [21].
Figure 2.56 shows the structure of an electromagnetic valve actuator used
for camless engines. The actuator has two coils exposing an armature to
a magnetic field when a current flows through them. The resulting magnetic
force yields a movement of the armature and thus the closing or opening
action of the valve. Because of the relatively large valve lift of 8-10 mm
two valve springs are employed in order to support the movement.
A typical driving circuit for each of the two coils is depicted in Fig. 2.57.
The power transistors T 7 and T 8 and the diodes D 2 and D 3 form the actual
switching converter stage. The transistors are driven by an integrated low-
side and high-side driver with the high-side driver being bootstrapped as
shown in Fig. 2.28.
The peak current of the circuit is determined by the output of the DAC
converter which itself is controlled by the microcontroller. The resistors R1
 2 Power Circuits 63

Fig. 2.56. Electromagnetic actuator for variable valve train

and R 2 provide a hysteresis with a constant factor in relation to the DAC

output and thus determine the maximum and minimum current level. When
the lower transistor Ts is turned on, the current signal over the shunt resistor
Rs is amplified by a differential amplifier and compared with the current
levels. Depending on the output of the comparators, the RS flip-flop switches
the upper transistor on and off. When the upper transistor is turned on,
then the full battery voltage VB is applied at the actuator coil, when it is
turned off, the coil voltage is reduced to the negative forward voltage of D 2
in addition with the voltage drop over T 8 and the shunt resistor Rs. In this
case, the current through the coil is slowly reduced.
When both the lower and the upper transistors are turned off then the
current through the actuator coil continues to flow via the diodes D 2 and D 3
until it is reduced to zero. During this time, the coil voltage is inverted to
-VB and the current is rapidly reduced.
Careful consideration in actuator drive circuits should be given to layout.
As shown in Fig. 2.57, the ground was partitioned into an analog ground
AGND for the control circuit and a power ground PGND. The current
through the shunt resistor is measured with a differential amplifier and re-
ferred to AGND for the signal conditioning. As both the low- and high-side
driver have to supply their gate signals with respect to PGND the logic in-
put signals were converted to currents and transformed to PGND via the two
current mirrors T 2 ,3 and T 5 ,6 . These measures provide the desired decoupling
of the signal conditioning from the drive circuit.
For a given battery voltage VB only a limited gradient of the current
can be achieved, which is determined by the quotient VB/ L. The achievable
current gradients may often be too small for effective control of the actuator
64 Stefan Butzmann

0
> (")
z
(.!)
N
'<!"
0 a..
II

II
()

-5?
I 0::: I Cl)

0
z

Fig. 2.57. Complete driving circuit for one actuator coil

 2 Power Circuits 65

Fig. 2.58. Modified power stage with dynamic high voltage boosting [22]

and higher current slopes are often desired. Figure 2.58 shows a modified
power stage, which provides an additional high voltage boost mode without
using an external supply.
The basic idea of the circuit is the feedback of the current via the diode
D 2 into a high-voltage storage capacitor CHv instead of the battery supply
during turn-off of the transistors T 1 and T 2 . When a high current gradient is
needed, T 3 instead of T 1 is turned on and CHv is discharged. As the upper
coil voltage is then pulled to Vs + Vc,Hv, the diode D 3 was inserted into the
circuit in order to prevent the intrinsic body diode of T 1 from conducting.

2.5.3 Stepper-Motor Drive Circuits

Stepper-motors are commonly used to provide a reliable controlled motion

based on digital information (also see Sect. 3.4.3). A desired motor position
can be reached with high precision by counting the number of steps applied
to the motor. Often, stepper-motors are driven "open-loop", which avoids
the need for expensive sensors and feedback-loops. In contrast, closed-loop
drives, where the motor position is controlled by position sensors, offer higher
reliability and a smoother motion due to the adjustment of the switching
sequence. In general, stepper-motors have two coils, which can be driven in
unipolar or in bipolar mode (Fig. 2.59).
66 Stefan Butzmann

E F D

Fig. 2.59. Stepper motor with permanent magnet and two coils

For unipolar drive the center taps of the two coils are connected to the
DC supply voltage, effectively resulting in four (half- )coils. A simplified drive
circuit for unipolar mode is shown in Fig. 2.60.
Here, the direction of the magnetic field produced is determined by the ac-
tivated half-coils. A rotation of the permanent magnet is achieved by switch-
ing the transistors as shown in Fig. 2.61. In the depicted case of a simple
drive circuit, the current through the motor coils is limited by the resistance
of the coil windings and the supply voltage.
While current limiting by the resistance of the coil winding is a convenient
option for low-torque motors, it usually results in low speed rates due to the

Vs

Fig. 2.60. Drive circuit for unipolar mode

 2 Power Circuits 67

L-R time constant of the system. An improvement could be achieved by the

use of a higher supply voltage and an external current limiting resistor. Figure
2.62 shows the simulation of the current through a motor coil with a nominal
current of IN = 1 A, an inductance of 10 mH and an internal coil-resistance
of 50. If the supply voltage is chosen as VB = 5 V, then no external current
limiting resistor is needed; however, the time needed for the current to build
up is determined by the ratio of the coil inductance and the coil resistance .

• t

I
Rotor
position

Fig. 2.61. Transistor switching sequence and rotor position for unipolar drive

. . .
········:····.-·-·---_;...
j , ,*'
....... -:·-I/ .......... .. . .................. .................. :· . . . .
0.8
:
: ;,
.. :: :
:
::
········:··!:············· ·:··················:··················:··················:········
:..
:
:
:
:
:
:
:
:
:

[i :
0.4

0.2
=• :
:
:
:
:
:
:
·_VB= . .
--. VB= 25\/, Rext = 200
. _ .• VB =25 Vand current limitation circuit
-0·2 L___io_ _ _..!:::J2=====4====6====a=::::::!.J
t/ ms

Fig. 2.62. Simulation of current through motor coils

68 Stefan Butzmann

The current rises faster with a higher supply voltage and an external resistor,
but dissipates power in the external resistor.
A further approach is the use of an external current limitation circuit (see
Fig. 2.63), where the coils are exposed to almost the full supply voltage V8
when one of the transistors T2 or T3 is turned on. When the desired current
value is reached and the voltage drop over the shunt resistor equals VRef, the
pup-transistor T 1 limits the current through the coils. This circuit provides
further speed advantages (see Fig. 2.62). However, it also dissipates power in
the pup-transistor.
A common approach therefore is to limit the current in the motor coils
by chopping circuits. Figure 2.64 shows the realization of a chopping circuit
with constant off-time, where a user-adjustable mono-flop turns off the active
power transistor T 1 or T 2 for a constant time when the current through the
shunt resistor exceeds the maximum allowed value (VRef = I· Rs). When
the mono-flop resets, the power switch is turned on and the coil current rises
again. This principle has been applied for example in the SLA7024M (Allegro
Microsystems).
In contrast to the unipolar drive mode, the bipolar drive mode always
uses the full coils. The direction of the resulting magnetic field is determined
here by the direction of the current through the full coils. The bipolar mode
gives a higher torque, but requires two class-E power stages. In bipolar drive
mode, both windings are usually continuously energized, while the direction
of the coil currents determines the motor position (see Fig. 2.65) .

.---..----ovs

Fig. 2.63. Use of current limitation circuit

 2 Power Circuits 69

,--------{) Vs

Fig. 2.64. Chopping circuit with constant off-time

Rotor
position

Fig. 2.65. Currents through motor coils in bipolar mode and the resulting motor
positions

Figure 2.66 shows a typical bipolar stepper-motor drive with the controller
1297 (ST) and the dual full-bridge driver 1298 (ST). In the circuit, a mi-
crocontroller generates the signals for forward or backward stepping, which
are translated by the 1297 into the necessary control signals for the power
stages.
70 Stefan Butzmann
+SV

r--t--OVB

osc
L297

INH1

L1
8

Step-clock

Half/full step

Direction

Enable

Ds

03
c
INH2

L2
D
07
04

+SV

Fig. 2.66. Complete stepper-motor drive circuit

 2 Power Circuits 71

The current through the power stages is supervised via two shunt resistors
and kept at a quasi-constant level for the active coil. For a particular direction
of current, each coil can be operated in 3 modes: fast current increase, fast
current decrease and freewheeling mode. For fast current increase in coil £ 1 ,
for example, output A of the 1297 is pulled high and output B is pulled to
ground. For fast current decrease, the IN HI-signal is pulled to ground and
all transistors in the power stage are switched off. In this case, the current
through the motor coil flows through the diodes D 2 and D3 until it has
decreased to zero. The freewheeling mode is realized by pulling signals A and
B to logic high, while IN Hl is also held at high level. In this case, the
current flows through transistor T 1 , the motor coil £ 1 and the diode D 2
and will decrease only slowly. In comparison to the other possible path for
freewheeling operation D3 - £ 1 - T 4 - Rs 1 , the power dissipation is lower as
the shunt resistor is not included in the current path.
Figure 2.67 shows the current limitation circuit used in the stepper-motor
controller 1297. In this circuit, the peak current through the motor coils is
monitored. Two flip-flops FF 1 and FF 2 are set at the rate of an internal
oscillator. When the voltage at the sense input, which is determined by the
current through the coil and the sense resistor, reaches the programmed peak
value, the flip-flops are reset and the following power stage is set into the
freewheeling mode until the next oscillator edge.

to power stage

Input Logic Output Logic

Fig. 2.67. Current limitation by peak-current detection

72 Stefan Butzrnann

Rotor
position

Fig. 2.68. Current waveforms and motor positions in half-step mode

The switching scheme depicted in Fig. 2.65 is the most popular for full-
step mode and provides the highest torque. A further drive method for
stepper-motors is the half-step mode, where one of the motor phases is with-
out current, when the motor is in the half-step position. This mode of opera-
tion provides an increase in the position resolution by a factor of 2. However,
it results in a great variation of torque between the full-step and the half-step
positions. Figure 2.68 shows the currents through the motor coils in half-step
mode and the corresponding motor positions.

2.5.4 Zero-current Converters

As discussed before, switching losses in the power transistor degrade the

efficiency of switching converters particularly at high frequencies. As these
switching losses are proportional to the product of the voltage over the switch
and the current flowing through the switch at the time of switching, they can
be drastically reduced if either the voltage or the current is zero when the
switch is opened or closed. Quasi-resonant converters provide a possible way
to achieve this goal. They have significant advantages over classical PWM
converters, which are low device stress, low electromagnetic interference and
high efficiency. This allows the application of high switching frequencies and
thus smaller-sized magnetic components.
A popular quasi-resonant converter is the zero-current converter, which
shall be discussed here as a typical representative of this family of converters.
Figure 2.69 shows a full-wave zero-current converter, where the switching
transistor T 1 is connected in series with the inductance Lr.
 2 Power Circuits 73

Fig. 2.69. Zero-current converter of full-wave type

For an understanding of the circuit operation, the load current hL

through the inductance LL is assumed to be constant. The resulting wave-
forms of the voltage Vcr across the capacitor Cr and of the coil current
through Lr are shown in Fig. 2. 70.
The transistor T 1 is turned on when the current through Lr is zero. The
voltage across the capacitor Cr swings to approximately twice the supply
voltage Vs and - when hr is negative- back to almost zero. The basic idea of
the zero-current converter is to turn the transistor T 1 off after one oscillation
period, when the current through the transistor has returned to zero again.
However, the requirement for a precise timing of the turn-off of T 1 is relaxed
by the use of the anti-parallel diode, which takes over the current when the
transistor is turned off during the ring-back period of the coil current. The

Fig. 2. 70. Waveforms of capacitor voltage Vc and of current through inductor Lr

for full-wave zero-current converter
74 Stefan Butzmann

T1 01 L,
/LL

I
Vs! c, 02 LL

Fig. 2.71. Zero-current converter of half-wave type

amount of energy transmitted to the load is thus only determined by the

repetition rate of the turn-ons of the power switch. In order to increase the
output power, the frequency has to be increased. As the amount of energy
transmitted with each event is constant, a drawback of this type of control
is a high ripple, if only a low amount of energy is to be transmitted to the
load.
A modification of the full-wave converter is the half-wave zero-current
converter, which is shown in Fig. 2.71. In this circuit, the diode D 1 is con-
nected in series to the transistor T 1 . Thus, the diode prevents the current
from flowing back to VB and the capacitor is discharged only by the load
current hL·

2.6 Current Development and Future Trends

2.6.1 Overview

The goal of current development activities is the continuous improvement

of existing semiconductor switches. This includes the reduction of the on-
resistance and of the switching losses and for some applications the increase
of the maximum off-state voltage. This trend has led to an optimization of
the cell-designs of MOSFETs and IGBTs including the shrinking of the cell
areas due to self-adjusting production processes, but also fundamentally new
structures of power switches. Typical examples for the latter are trench ele-
ments and CoolMOS-FETs. Table 2.4 gives a comparison of the voltage rat-
ings and the on-resistances of trench-FETs, CoolMOS-FETs and the classical
DMOS-FETs. Both trench-FETs and CoolMOS-FETs have been optimized
for particular voltage ratings: CoolMOS-FETs are available with ratings from
500 V to 800 V, trench-FETs with ratings up to 100 V, thus targeting mainly
at the automotive area. The given examples in Table 2.4 show typical repre-
sentatives of each of the device families.
Simultaneously, a trend towards intelligent switches with integrated pro-
tection functions could be observed during the last few years. So-called
"smart-power" devices even provide the combination of power switches and
functional control on a single chip. Figure 2.72 shows the power ratings of
 2 Power Circuits 75

T a ble 2.4. Overview of power-MOSFET technologies

DMOS Trench-FET CoolMOS

Voltage ratings 20V- 1000V 20V- 100V 500V- 800V
2.3
RoniVBR VBR VsR
Ron I
(m!1· mm 2 ) ::::: 200@ 75 V :::::90@ 75V ::::: 3000@ 600V
(typical values)
Example IRF1407 BUK7213-75B SPP03N60C3
(International (Philips) (Infineon)
Rectifier)
VsR IV 75 75 600
Ron I m!1 8 12 1400
Qg InC 160 41 13
Ron· Qg I pVs 1280 492 18000

discretes and smart-power devices versus their complexity. While monolithic

integration of t h e power switch and t he control logic offers the highest relia-
bility, the power ratings of smart-power devices are limited due to the higher
on-resistance of the integrated power switches. For this reason, there is a gap
between intelligent control, on one hand, and the use of h igh-power semi-
conductors, on the other, which can currently only be closed by the use of
discrete drive circuits.

Current I

10

MOS

10 4 Voltage IV

Complexity

Fig . 2. 72 . Power ratings vs. complexity of discretes and smart-power devices [24}
76 Stefan Butzmann

A higher versatility is given by the concept of an intelligent power con-

troller which provides driving intelligence and protective functions for discrete
power-FETs (see Table 2.5), thus closing the gap between high power and
intelligent drive.

Table 2.5. Classification of smart-power devices

Smart discretes Intelligent drivers Smart-power ICs

integration of intelligent driver IC power switch and
protective functions for discrete control functions on
on transistor die power-FETs one die
vertical current flow contains integrated targets at
integration of two protection and medium-current
or more switches on control functions applications
one die only with bus interface multiple isolated
common drain power outputs
also available in
multi-chip modules

Further integration yields even higher complexity, the so-called super-smart-

power devices, where microcontrollers and on-chip memories (EEPROM/
Flash-ROM) are integrated together with analog and power stages. Such
elements are intended for system integration in automotive and computer
applications, e.g. like printers.

2.6.2 Trench-FETs

Most power-MOSFET technologies available on the market use a horizontal

gate structure and a vertical current flow to the drain at the backside of
the wafer (see Fig. 2.10). Power-MOSFETs consist of a high number of the
described cells connected in parallel in order to reduce the on-resistance of
the complete device.
The most important MOSFET technology today is the DMOS-FET. The
cell-structure itself is different from manufacturer to manufacturer, but the
underlying principle of a lateral channel and a vertical current flow to the
drain contact at the backside of the die is the same. With such DMOS devices,
however, there is a fundamental limit to cell density due to the parasitic
junction-PET, which provides an additional resistance increasing with the
cell-density resulting from the pinch-off effect. This limits the optimization
of power-MOSFETs only by lateral shrinking.
A significant breakthrough was marked by the development of trench-
PETs, where the channel is vertically oriented and the gate structure is built
in a trench into the die volume, instead of being parallel to the surface of the
die (see Fig. 2.73).
 2 Power Circuits 77
s G s

p p

Fig. 2. 73. Cross-section of a trench-FET

In a trench-FET the current flow in the channel is vertical from the source
to the drain contact on the backside of the wafer. As the junction-FET prob-
lem is now avoided, the cell density may be increased again, thus allowing
a further r eduction of the Ros ,on p er chip area (see Table 2.4). Driven mainly
by automotive and computer applications, today most trench-FETs are avail-
able in a voltage range from 20 V to approximately 100 V with specific on-
resistances of less than 200 mfl·mm 2 for a breakdown voltage of 100 V. In
comparison to standard power-MOSFETs with similar voltage ratings the
Ran · Qg-product is reduced by approximately a factor 2. An ongoing trend
is the increase of the number of dies per wafer , which gives additional cost
benefits or at least makes up for the more expensive manufacturing costs.

2.6.3 CoolMOS

While trench-FETs offer their m ain advantages for voltage ratings of up to

100 V , the development of medium-voltage transistors up to 800 V has been
marked by the introduction of CoolMOS.
CoolMOS transistors contain vertical implanted p-tubes as shown in
Fig. 2.74. These stripes introduce extra lateral pn-junctions, which are re-
verse biased, when the device is turned off. The result is a complete depletion
of the drift zone , while the off-state voltage is built up in the horizontal di-
rection. As a consequence, then- -drift area can be significantly reduced and
the reduction in the doping towards high off-state voltages as in conventional
MOSFETs is no longer necessary. The turn-on resistance of CoolMOS-FETs
increases only linearly with the breakdown voltage V BR, which is different
from the exponent of 2.3 as is the case for conventional MOSFETs. The in-
creased conductivity results in a n on-resistance a bout a factor of 5 lower than
that of conventional power-MOSFETs for a rating of 600V. Thus a higher
current capability p er chip area can be achieved and a lower gat e charge in
comparison to similarly rated conventional MOSFETs.
As CoolMOS-FETs with similar ratings as conventional power-MOSFETs
are much smaller in chip size, the parasitic capacitances are also reduced . This
78 Stefan Butzmann

Fig. 2.74. Cross-section of a CoolMOS transistor

yields switching losses about 50 % lower than those of conventional transistors

[15]. One disadvantage of CoolMOS-FETs, however, is the poor dynamic
behavior of the intrinsic body diode.

2.6.4 "Intelligent" Switches

Intelligent switches or "smart discretes" present an intermediate step from

discrete power devices towards more complex smart-power ICs. They com-
bine power-MOSFETs with protective functions like t emperature protection,
overvoltage or overload protection.
The integration of additional functionality into the power switch increases
the reliability of the device and allows the implementation of driving func-
tions into the switch, thus reducing the effort for the external drive circuitry.
Smart-power switches can be realized either by monolithical integration of
the control circuit and the power switch on one die or by multiple chip mount-
ing, where the control IC and the power switch are fabricated separately in
different optimized technologies and interconnected in the package. While the
realization of discrete MOSFETs uses approximately 6- 8 masks, t he mono-

Fig. 2.75. Architecture of a T empFET

 2 Power Circuits 79

Statuso-- - - --------,
Supplyo-- - - - - - -----,
Input

Fig. 2. 76. Block structure of a TOPFET

lithically integrated power switch needs 12-15 masks, thus roughly doubling
the price per chip area.
Examples of a multiple-chip solution are Infineon's Temp-FETs, where the
power-MOSFET can be turned off by a thyristor, which in turn is controlled
by a temperature sensor.
More sophisticated devices, like the HITFET (Infineon) include additional
functions like overvoltage protection, overload protection, current and d V / dt
limitation and even a status feedback to the drive circuit via a digital flag.
Other manufacturers produce smart switches with similar functions by
monolithical integration of the control circuit and the power switch on a single
chip. Examples of these types of switches are ST's OmniFETs or Philips'
TOPFETs.

OL1
SL1
OLI
N.C. OL1
BATT BATT
Gt.D SH1
IH1 7 SHI
STATUS SH2
IH2 SH2
BATT BATT
N.C. OL2
SL2
L2 SL2
ot..2 OL2

Fig. 2.77. Full-bridge with "intelligent" switches (Courtesy of Royal Philips Elec-
tronics)
80 Stefan Butzmann

The latter device family combines the trench-FET technology with pro-
tective functions and is separated into low-side and high-side TOPFETs.
The low-side TOPFETs offer logic-level inputs, voltage clamping, dV/dt lim-
iting, overtemperature-, overload- and ESD-protection. An extended version
allows status reporting via an additional pin; however, it has to be supplied
by a separate 5 V source. The high-side version has an integrated charge
pump in order to provide the required additional supply voltage to turn the
n-channel MOSFET on.
Meanwhile, complete smart-power-MOSFET bridges are available in a sin-
gle package, where both high-side and low-side switches have been integrated
into one single package. Figure 2. 77 shows the lead frame and the output pin-
ning of an intelligent full bridge, where the two low-side switches are separate.
The two high-side switches in the center have a common drain connection
and share the charge pump and some of the diagnosis functions.

2.6.5 Smart Power ICs

While intelligent protection functions are implemented in "intelligent" swit-

ches, a further step towards smart-power ICs is the implementation of intelli-
gent control functions and power switches on one chip. Modern smart-power
technologies combine _6ipolar devices, QMOS logic and DMOS power tran-
sistors (BCD technology). This gives the designers the chance to select the
optimum technology for each circuit function: bipolar for low-offset ampli-
fiers and reference circuits, CMOS for high density circuit parts (digital and
analog) and DMOS for power functions. As a consequence, complex systems
can be built on a single chip.
The advantages of this approach are obvious:
reduced size of power control circuits,
reduced number of connections,
higher reliability.
All of the protection functions realized in intelligent switches like temper-
ature supervision or current limiting can also be integrated in a smart-power
chip. In motor applications, for example, simply protected transistors are
found as well as complete bridge drives with control and diagnosis functions.
While first generations of smart-power devices only featured low inte-
gration density, there is a trend towards the integration of increasing func-
tionality on the chip, including the use of microcontroller cores and DSP
blocks. This also yields significant demand of memory in today's smart-power
ICs.
As this development is driven in large part by the car industry, auto-
motive electronics is one of the main application areas of smart-power sys-
tems. With the increased use of the number of microcontrollers in modern
cars logic-level interfacing has become more and more important. The use
of intelligent power components becomes particularly advantageous, where
 2 Power Circuits 81

transducers for disk drives, printers or motor control can directly be driven
from the microcontroller unit.
Besides automotive syst ems, other typical application fields of smart-
power ICs are power supplies, regulators and even audio amplifiers. In the
field of industrial electronics, integrated circuits in high-voltage technol-
ogy are used to realize off-line applications, e.g. power factor correction
circuits.
For the realization of smart-power devices, mainly three isolation tech-
niques are employed today. With self-isolation the power device is realized as
a vertical transistor. The power stage output is at the backside of the wafer
as is the case in conventional power devices. Examples for this technology
are the VIPower technology by ST or the Smart-SIPMOS technology (Infi-
neon). Thus, the voltage and the current ratings are also similar to those of
conventional discretes. This kind of isolation technique is usually applied for
high-voltage integrated circuits. The complexity of these devices is however
limited, as they share a common collector/ drain contact for multiple power
outputs. Yet, t he low on-resistance makes this kind of device an ideal choice
for high-side switches.
Figure 2.78 shows a sectional drawing of a smart-power device in a junc-
tion isolated smart-power technology. With this technology, quasi-vertical
DMOS transistors are implemented, where the drain current flows back to
the surface via a buried n+ -layer. As the transistors can be separated by p+-
areas, it is possible to realize electrically isolated devices on one chip, which
allows the implementation of low-side and high-side switches. BCD devices
may be more complex than vertical structures, however, their on-resistance
is higher.
The third technology used for smart-power systems is dielectric isolation
with completely isolated areas of silicon. The manufacturing process, however,
is relatively complex and the resulting ICs are quite expensive.
Where both complex control functions and the performance of vertical
transistor technologies are required, the separation of power control and the
power switches can be a good choice simultaneously offering a higher flex-
ibility in the design of the overall system architecture than is feasible with

n+ Buried layer

Fig. 2. 78. Realization of a vertical DMOS tra nsistor in a smart-power process

82 Stefan Butzmann

a complete integrated solution. The concept of "intelligent" power controllers

is heading into this direction.
Figure 2.79 shows the block diagram of an "intelligent" power controller,
which integrates the drive and control circuits for eight power transistors.
Each of the control channels is configurable by software as high- or low-
side drivers. The controller provides extensive protection functions like pro-
grammable current limiting, an SPI watchdog and diagnostic output to the
microcontroller via SPI. Thus, a wide range of MOSFETs for different ap-
plications and voltage ranges can be driven with one single IC controlled via
a serial interface , thus reducing the required number of output pins of the
microcontroller.

Control Logic &

Registers

Fig. 2. 79. Intelligent power controller block structure (Courtesy of Royal Philips
Electronics)

2.6.6 Silicon Carbide

In today's technologies , silicon is mainly used for production of power devices.

With increasing temperatures, however, the leakage currents grow unaccept-
ably large until the intended device operation is replaced by uncontrolled
device behavior.
For this reason , new materials are being explored with higher bandgaps
and increased thermal conductivities. Silicon carbide is such a material. Sil-
icon carbide has a bandgap, a thermal conductivity and a breakdown field
approximately 2-3 times higher than that of silicon (dependent on the chosen
polytype). The high operational junction temperatures , which are possible
with silicon carbide, permit the realization of power devices with extremely
high power densities. Due to the higher breakdown field and the greater
 2 Power Circuits 83

bandgap higher switching frequencies can be realized with silicon carbide

devices thus allowing further shrinking of capacitors and inductors in the
circuit. The improved thermal conductivity allows a further increase of the
ambient temperature as cooling measures can be applied more efficiently.
Even though protoype SiC devices have shown excellent performance,
mass production has for a long time been prevented by the presence of defects
such as micropipes in the crystal structure. Only recently the first commer-
cial silicon carbide based Schottky power diodes have become available from
Infineon. The main features of these devices are the virtual absence of the
reverse recovery current and the virtually missing temperature influence on
the switching behavior in the specified temperature range. The forward volt-
age drop of the devices is significantly higher than that of comparable diodes
based on silicon, though. However, it can be expected that lower forward
drop devices will be available in the near future as the production processes
are continuously improved.

References
1. B. Murari, F. Bertotti, G.A. Vignola: Smart Power !Cs, Springer Verlag
2. Stengl, Tihanyi: Leistungs-MOS-FET-Praxis, Pflaum Verlag
3. Schroder: Elektrische Antriebe, Ed. 3: Leistungselektronische Eauelemente,
Springer Verlag
4. Schroder: Elektrische Antriebe, Ed. 4: Leistungselektronische Schaltungen,
Springer Verlag
5. M. Rashid: Power Electronics, Prentice Hall
6. B.J. Baliga: Modern Power Devices, John Wiley & Sons, 1987
7. Power Semiconductor Applications, Philips Semiconductors, 1994
8. Current Ratings of Power Semiconductors, Application Note AN-949, Inter-
national Rectifier
9. J.M. Peter: Characteristics of Power Semiconductors, Application Note
AN512, ST
10. Selecting and Designing in The Right Schottky, Application Note AN-968, In-
ternational Rectifier
11. Datasheet EU208, ST
12. Datasheet IRF640, ST
13. !GET Characteristics, Application Note AN-983, International Rectifier
14. C. Dorlemann: Geregelte Ansteuerung von Insulated Gate Bipolar Transis-
toren. Dissertation, Ruhr-Universitat Bochum, 2002
15. K. Sheng, F. Udrea et a!: PT and NPT !GETs up to 1.2kV: Which is op-
timum? Proceedings, 9th International Conference and Exhibition on Power
Electronics and Motion Control, EPE-PEMC 2000
16. B. Andreycak: Practical Considerations in High Performance MOSFET, !GET
and MGT Gate Drive Circuits, Application Note U-137, Unitrode
17. A.D. Pathak: MOSFET/!GET-Drivers- Theory and Applications, Application
Note AN0002, Ixys
18. Use Gate Charge to Design the Gate Drive Circuit for Power MOSFETs and
!GETs, Application Note AN-944, International Rectifier
84 Stefan Butzmann

19. HV Floating MOS-Gate Driver ICs, Application Note AN-978, International

Rectifier
20. S. Butzmann, J. Melbert: Elektrisch- Thermische Simulation von Leistungs-
bauelementen, GI/ITG/GMM-Diskussionssitzung ,Heterogene Systeme", 11-
menau, 1997
21. Pischinger et a!: Electromagnetically Operating Actuator, US Patent 4,455,543,
1984
22. S. Butzmann, J. Melbert: Device for Controlling a Regulator, Patent WO
00/09867, 2000
23. S. Butzmann: Sensorlose Regelung elektromagnetischer Aktuatoren fiir den
variablen Ventiltrieb. Dissertation, Ruhr-Universitat Bochum, 2000
24. J. Melbert: Power Semiconductors as a Key Component for System Integration
ISPSD '97, 1997
25. W. Pribyl: Integrated Smart Power Circuits Technology, Design and Applica-
tion Proceedings of the 22nd European Solid-State Circuits Conference, ESS-
CIRC96, Neuchtel, Switzerland, 17-19 September 1996
26. A. Lidow, D. Kinzer et a!: The Semiconductor Roadmap for Power Manage-
ment in the New Millennium. In: Special Issne on Power Electronics Technol-
ogy, Proceedings of the IEEE, Vol. 89 No. 6, June 2001
27. B.J. Baliga: The Future of Power Semiconductor Device Technology. In: Special
Issue on Power Electronics Technology, Proceedings of the IEEE, Vol. 89 No.
6, June 2001
28. B.J. Baliga: An Overview of Smart Power Technology. In: IEEE Transactions
on Electron Devices, Vol. 38 No. 7, July 1991
29. K. Satoh, M. Yamamoto: The Present State of the Art in High-Power Semicon-
ductor Devices. In: Special Issue on Power Electronics Technology, Proceedings
of the IEEE, Vol. 89 No. 6, June 2001
3 Electromagnetic Actuators

Hans-Dieter Stolting

3.1 Introduction

3.1.1 Definition of Electromagnetic Actuators

Electromagnetic actuators transform electrical and mechanical energy into

one another using the electromagnetic-mechanical principle. In the follow-
ing small electric machines, especially motors, are described with a nominal
power in the range about up to 1 kW, so-called fractional-horsepower motors
[8]. In the simplest case a mechanical switch serves as a controller. For open-
loop and closed-loop control of motors more and more electronic components
are used, mechanically switched or triggered by hardware or software. In this
chapter the combination of motor and electronic control is called an actuator
or drive.
As there are many special designs besides the usual function principles
and types of electromagnetic machines, the range of fractional-horsepower
motors is marked by a great variety of different types. This is because of
their unusually great application possibilities (Table 3.1.1):

Most of the motors are used as drives for consumer goods. In order to be
very cost-effective these motors are adapted constructively as extensively
as possible to the driven appliance and to the most economical production
technology.
Besides these, there are high-grade motors for capital goods with special,
often extreme features.

Therefore small electromagnetic actuators are special designs needing the

close cooperation of the manufacturer and the user. Doing so, one is often
faced with the problem that actuator development starts at the moment when
the development of the apparatus is almost finished. But the actuator has to
be available before the manufacturing of the apparatus starts.
The variety of motors and their applications impede standardization.
There are partial agreements on fixing dimensions, cable colors, etc. between
user groups (e.g. oil and gas burner) and motor manufacturers.

H. Janocha (ed.), Actuators

© Springer-Verlag Berlin Heidelberg 2004
86 Hans-Dieter Stolting

Table 3.1. Characterising features of small cheap and high grade motors

Cheap motors High grade motors

Integration in the driven Optimal electrical and
apparatus; motor components constructional adaptation to the
take on apparatus functions. driven apparatus
Largely automatic production Small-sized production: cutting
(> 1 mio pieces p. a.): operation, fitting together by
punch-bend-joint technique, screws
using components with High grade components:
commercial (standard) electrical sheet steel or low-loss
dimensions (magnets, sheet, rare-earth or AlNiCo
bearings, capacitor etc.), magnets, rolling bearing
greater manufacture Special characteristics relating
tolerance band, rough to efficiency (low energy
dimension stages (outer and demand, low temperature rise,
inner diameter, length), small weight), speed (extremely
great airgap, low space high or low, great speed range,
factor, very easy winding, constant speed, non-cogging
staved-enamel wire, shaft or/ and synchronous operation)
steps as few as possible, dynamic (low mechanic or/ and
collar bearing, etc. electric time constant),
Power per weight and efficiency positioning, overload capability,
often insignificant: semifinished live time, robustness, freedom
sheet (often without annealing) from maintenance, low noise,
instead of electrical sheet steel, low vibration, electromagnetic
ferrite magnets. compatibility (EMC),
insensitiveness to environmental
conditions (temperature,
vibrations, accelerations,
pressure, pollution (water-,
dust-, gas-proof), electric or/and
magnetic fields, explosion
protection)

3.1.2 General View

For guidance principles of design possibilities of small electric machines are

given in Figs. 3.1, 3.2 and 3.3. With regard to their basic characteristics they
can be assigned to two groups: self-commutated and externally commutated
motors.
Self-commutated motors are characterized due to the fact that the voltage
applied to their windings is dependent on the rotor position. From that, their
special characteristics are:

The maximal speed depends on the dimensions of the drive and not on the
frequency of the power supply. Due to the very high speed these drives
 3 Electromagnetic Actuators 87

Self commutated motors with mechanical commutator

Motor DC motors AC motors

type Magnet Separate or Compound Series Series
excitation shunt excitation excitation excitation excitation

fJ 8 8 8
v v v

Basic
circuit

Speed-
torque
curve

N(min-1 ) 2000 ...20000

T
< 6000
T
<6000
T
k <6000
T
k_ T
3000...30000
P(W) 0.2... 1000 0.2 ... 1000 20... 1000 20... 1000 10... 1000
7) 0.4...0.8 0.3 ...0.7 0.3...0.7 0.3...0.7 0.3...0.7
T,/TN 4 ... 10 4 ...6 4...6 3... 5 2... 5
TmaxJTN < 10 <6 <6 <5 <5
Untapped
3 brush Untapped winding
Speed winding
chopper chopper controller
control phase angle
controller if required with rectification and DC link
control

@ Commutator motor @ Single-phase or three-phase motor

o..
V 0 _, Permanent-magnet, hysteresis or reluctance motor
f"VV"\ fVVV"\ Series or shunt (parallel, separate) winding 1.1 Permanent magnet

0 Squirrel-cage winding CA Starting capacitor C6 Running capacitator

TN Rated torque Ta Locked-rotor torque Tmax Maximum torque

P Output N Speed TJ Efficiency VE.A Excitation, armature voltage
p Number of pole pairs

Fig. 3.1. Direct mains-operated self-commutated motors

can be built with low volume and weight. Therefore they are the ideal
drives for portable apparatus and tools.
Speed control is easy, cheap and often lossless.
The motors are expensive because they need brushes and a commutator
for switching over the windings.
Motors with a mechanical commutator have a shorter life time, and are
more trouble-bone and noisier than externally commutated motors.
88 Hans-Dieter Sti:ilting

External commutated motors

Asvnchronous motors
Motor AC motors
Three-phase
type motor Three phase Two phase
capacitor Capacitor Resistance Split-pole
squirrel cage auxiliarv windinn auxiliarv windinc auxiliarv windina motor

Basic
circuit
kW
® ®
UtJc
@z
Uu
Speed-
torque
curve
T
p2
T
wit!:P
N----------
aUJ(ihary
._..,.(nding
/ With

T
Ntz T
N(min- 1 ) < 3000 < 3000 < 3000 < 3000 < 3000
P(W) >50 > 500 50 ... 2000 < 500 < 200
T) 0.5...0.8 0.3... 0.7 0.5...0.75 0.3 ...0.6 0.05 ... 0.35
T.ITN 1...3 1. .. 2 .. .4, C8 :1 ... 2 0.2 ... 1 0.2 ... 1
Tmax!TN 1.5... 3 < 1.5 < 1.5 < 1.5 < 1.5
Phase angle
Speed- Frequency control Transformer
control converter transformer
pole changing

External commutated motors

Svnchronous motors
Motor
Three-phase AC motors
type motor
Permanent Permanent
hybrid rotor maanet rotor Reluctance rotor Hysteresis rotor
maanet rotor

Basic
circuit
l4J 1Juc
E)z
Uu Uu
9 I I

Nt2 Nc c
I , _ _..II \._ ..../

Speed-
torque
curve

N(min- 1) :;:; 3000

T
:;:; 3000
T
3000
T
l2 TZ 3000
/ __/
T
3000
T

P(W) > 200 >500 :;:; 50 0.1 ... 500 0.01 ... 100
T) >0.5 :;:; 0.6 .;;0.6 < 0.05 ...0.5 < 0.05 ... 0.4
T.ITN <3 - - 0.5...2 0.2 ... 2
Tmax!TN < 1.5 < 1.5 < 1.5 < 1.5 < 1.5

Fig. 3.2. Direct mains-operated externally commutated motors

 3 Electromagnetic Actuators 89

Self commutated motors

Motor
type Permanent-magnet AC commutator
DC motors Brushless DC motors motor

1-t-8----u-----® !3- * <

:.......................::-:::::}'·----
! 1...4-
1;
rectifier electronic switching using
Hall elements or optical phase angle control
encoders with thyristor

1.
FD sensorless switching
transistor driven in its
linear or active region : : 1...4-, •
without FD : :.------ :
2. transistor driven in the speed control by tacho-
ON-OFF mode (Pulse- generator (T) phase angle control
Control Width-Modulation, with Triac
circuits PWM)
FD: free-wheeling diode speed control by
L_ ______ ___ _j tache-generator

position control using a

pulse generator
PWM servo amplifier (P, encoder)
full, H-bridge circuit
speed control by tacho- ._ _________________ ,!

generator or electric
governor resolver (R) for speed
position control using a control position control
pulse generator (encoder) and switching

External commutated motors

Motor
type
Asynchronous motor Stepper motor

1-tSlEJ--®
3-* ..
frequency converter
phase angle control with acceleration and
Triac capacitor motor deceleration by a given
pulse serie

Control
circuits 3- * 3-

frequency converter
three-phase motor
speed control by tacho-
generator

Fig. 3.3. Electronically driven motors

90 Hans-Dieter Sti:ilting

- Motors with an electronic commutator are as robust and low-noise as ex-

ternally commutated motors. But they always need a rotor-position sensor
and expensive electronics, which make them comparatively expensive.
Externally commutated motors are characterized due to the fact that the
voltage is applied to their windings forcible switched by the power supply in
a definite order. From that, their special characteristics are:
- The maximal possible speed depends on the supply frequency. It is
3000 rpm at a frequency of 50 Hz.
- The speed control is very expensive, so that a variable-speed actuator is
relatively costly.
- The motors are cheap, robust and low-noise.

3.1.3 Basic Construction Possibilities

There are more different layouts of small electric motors than those of larger
motors, especially more rotor-stator configurations (Fig. 3.4):
- Internal-rotor motors predominate because of the ease of production, sim-
ple fitting, good stator-winding cooling and the low moment of inertia
(especially motors with slim rotors).
- External-rotor motors are produced for special applications, e.g. fans with
blades directly mounted at the rotor core (Figs. 3.30, 3.70), winders or
motors with good rotational accuracy (low pulsating torque, Fig. 3.21).
- Intermediate-rotor motors (bell-rotor motors) with moving rotor coil
(coreless rotor) have the lowest moment of inertia of all electric motors,
low mechanical and electrical time constant and no cogging torque. They
are relatively expensive and can be built only for low nominal power be-
cause the self-supporting rotor winding is not sufficiently solid at large
rotor diameter and high speed (Fig. 3.9).
- Intermediate-stator motors with self-supporting stator winding consist of
a permanent-magnet rotor inside the stator winding and an outside yoke
rotating with the rotor. Hence there are no eddy currents and the rotor
has good rotational accuracy (Figs. 3.22, 3.24 and 3.25).
- Disc-type or pancake motors and linear motors are built in the same
above-mentioned stator-rotor configuration (Figs. 3.10, 3.11, 3.21, 3.24
and 3.25).
- Regarding the die set or the magnetic field pattern motors can be classified
in single-axle or two-axle symmetrical types (Fig. 3.5), whereby the flux
lines are radially, diametrically or axially straightened. The first are often
easier to be built in and better to be cooled; the latter are often produced
at low cost.
- Most of the motors are of heteropolar type with alternating poles around
the air gap, but there are homopolar motors too, with alternating poles
along the axis. Heteropolar motors predominate because of their better
 3 Electromagnetic Actuators 91

Stator

*
·-

Rotor

Fig. 3.4. Rotor-stator configurations

power /weight ratio. Homopolar machines with a large number of teeth

around the air gap rotate with a small angle per voltage pulse (Fig. 3.49)
or are suitable if low rotor revolutions are required.
- Windings distributed in slots are on the one hand expensive, and on the
other hand these windings generate a better, sinusoidal field pattern and
therefore have lower losses and parasitic torques.
Concentrated windings on salient poles are cheap to produce, but elec-
tromagnetically unfavorable in asynchronous and synchronous motors, as
they generate a nonsinusoidal magnetic field distribution.
Slotless or air-gap windings generate no cogging torques and prolong the
commutation because the inductance is lower due to the larger air gap
(Figs. 3.9, 3.10, 3.11, 3.20, 3.22, 3.24 and 3.25).

a ....__-_--_-_--_-_-_-_-__, b L.___ _ _ __J

Fig. 3.5. Intersection symmetry. a Two-axle, b single-axle symmetrical cross-

section
92 Hans-Dieter Stolting

In the range of small motors there are sometimes special arrangements of

motors, for instance mirror-image mounted motors with different direction of
rotation for reversing or two smaller motors on one shaft for increasing the
output power.
As every small motor type can be built in nearly all of the above con-
struction variations, the designer of these motors has a lot of possibilities to
fulfill the demands of every special application.

3.2 Self-clocked Motors with Mechanical Commutator

3.2.1 Introduction

Typical features of self-clocked motors are:

- The torque-generating magnetic field is excited by permanent magnets or
coils (excitation or field winding) in the stator.
- The torque-generating current is fed to the rotor (armature) winding by
brushes and commutators.
The commutator works together with the brushes as a switch which
changes over the current in the armature coils, so that the best conditions
for torque generation are always given. That means the current direction
and field direction have to be perpendicular to each other. With increasing
speed the brush noise grows louder and the brush wear is greater because of
switching sparks. Small DC motors have neither commutating winding nor
compensating winding. The brushes are shifted out of the neutral axis against
the direction of rotor rotation to improve the commutation. The cause of this
is that the zero crossing of the magnetic field curve, lying at no-load in the
geometric neutral axis, is shifted against the sense of rotation by armature
reaction. Backward brush shift by about 1-1 slot pitch is recommended if
the rotor always rotates in one direction. To protect the commutator from
damage by worn brushes they sometimes have an automatic disconnection
which lifts off the brushes before their brush-lead or cap drag along the com-
mutator. After changing the brushes the motor can be driven again. The
radio disturbance of the brushes always requires considerable expense of ra-
dio and television interference suppression devices. Because of the pulsating
rotor flux the rotor is made up of axially arranged punched and slotted sheet
steels. Therefore the eddy currents of the rotor caused by the alternating ro-
tor field are less. The pros and cons of the different DC commutator motors
are compared with each other in Table 3.2.

3.2.2 Electrically Excited DC Motors

In the lower power range DC motors with wound fields are also produced, but
only with a small number of pieces because of their higher costs compared
 3 Electromagnetic Actuators 93

Table 3.2. Feature comparison of the of DC motors

Excitation Permanent- Separate or Series Compound

magnet shunt
Speed-torque more load- more load- load-dependent more load-
characteristic independent independent independent
Locked-rotor greater greater smaller greater
current
Speed range smaller greater greater greater
with the same
costs
Speed control similar similar expensive similar
Efficiency with greater lower lower lower
the same costs
Risk of exists not applicable not applicable not applicable
demagnetising
Interference more expensive more expensive similar similar
suppression
Costs lower mean mean greater

with permanent-magnet DC motors (Table 3.2). Therefore these motors are

built if special application demands are required, e.g. very low or very high
temperature or a wide speed range (tapped stator winding).
The types of electrically excited DC motors are given in Fig. 3.1. Shunt-
wound motors have a field winding connected in shunt with the armature
winding. Their speed-torque characteristic is linear decreasing with increas-
ing load. The field and the armature winding of separately excited motors
are supplied by different voltage sources. For instance the field winding is
connected with the AC system by a full-wave uncontrolled rectifier and the
armature winding in addition to such a rectifier by a PWM servo-amplifier
with a transistor driven in on off mode (a chopper like Figs. 3.60 and 3.61).
They have the same speed-torque characteristic as the shunt motors.
Series wound motors have field and armature winding connected in se-
ries and have a varying-speed characteristic which is more load-dependent
than the characteristic of shunt motors. Their locked-rotor current is smaller
in comparison with the locked-rotor torque because the torque depends ap-
proximately quadratically on the current. The current/torque dependence
of motor types with constant flux is linear. Series motors can be connected
to an AC voltage source too. Then they are called universal motors or AC
commutator motors (see Sect. 3.2.4).
Compound-wound motors have both a series and a shunt field, and they
therefore have a speed-torque characteristic intermediate between those of
the above mentioned motors. Because of their high costs they are only seldom
produced. Figure 3.6 shows the operating curves of a compound-wound motor
with its low-decreasing speed torque curve.
94 Hans-Dieter Stolting

I N
A rpm

0
0 5 10 15

T: torque, TN: nominal torque, N: speed, P: output power, I: current, T): efficiency

Fig. 3.6. Performance curves of a compound-wound motor

Ordinarily, fractional-horsepower DC motors built for household or in-

dustrial applications have mostly two poles like universal motors (Fig. 3.37).
However, aircraft DC motors often have four or more poles. For easier and
therefore cheaper manufacturing the stator assembly is laminated too. The
armature winding is a drum winding, usually a lap winding. The brushes are
made of a mixture of copper and graphite. Solid brushes with more copper
have a longer life time, soft brushes with more graphite keep the noise down.
In every case brushes and commutator have to be selected very thoroughly
by testing. The same applies to permanent-magnet DC motors too.

3.2.3 Permanent-magnet DC Motors

For the most part fractional-horsepower DC motors are permanent-magnet

excited motors, as they are simpler and cheaper than motors with wound
fields. They have permanent magnets in the stator structure to set up the
required flux and a commutator winding in the rotor. The different kinds,
shapes and mounting manner of permanent magnets offer much more motor-
design possibilities than windings. Permanent-magnet motors can be adjusted
to the driven mechanism especially well. Therefore they are available in a wide
range of power and revolutions. Predominantly these motors are produced for
extra-low voltage(::; 42 V) and used for battery supply because of their good
efficiency [2]. Here are some examples: auxiliary drives for cars (windscreen
wipers, cooling fans, fuel pumps, sunroof and window lift motors, seat po-
sition motors, door locks, electric steering assistance etc.), toys (e.g. model
 3 Electromagnetic Actuators 95

a b c d

f g
Fig. 3. 7. Low-cost constructions of DC motors. a, b Diametral magnetization, c,
d radial magnetization, e, f, g fiat designs

railways, cars and airplanes) , garden and household appliances (grass shears,
tooth brushes, coffee grinder, tin openers, spits, door pass-keysystems 1 ; heat-
ing valves, ventilation and air conditioning technology, electro bikes, con-
sumer electronics (CD-player, magnetic-tape drives 1 ), machine tools (test
stands for small electrical motors 2 ) , cableless hand tools , auxiliary drives in
grinding and welding machines 1 2 as well as lathes), film and video technology
(cameras, camcorders, projectors) , medicine and laboratory appliances (mas-
sage apparatus, wheelchairs, X-ray units, dental apparatus, stirring appara-
tus, centrifuges, shaking and sieving machines) , vending machines 1 (vending
machines for goods, change machines, gaming and slot machines, jukeboxes),
printers, printing machines 1 , passport scanners 2 , golf cars, pneumatic-post
systems, automatic packaging and label machines .
Most small DC motors are of the internal-rotor type (cylindrical outside-
field type). Cost-effective motors (Fig. 3.7) are usually two-pole and have
a housing made of rolled or deep-drawn sheet. The core of the slotted ro-
tor is laminated. Usually the slot number is even for easy winding produc-
tion (loop winding) . There are especially narrow designs (Figs. 3.7e, 3.7f and
3. 7g), in some appliances also with an axial stator yoke simultaneously used
as a bearing bracket (Fig. 3.7e). Only cheap anisotropic ferrite magnets are
used. This material is temperature-sensitive. It has a comparatively low re-
manent flux density and BH product, but it is difficult to be demagnetize
because of its wide hysteresis loop. Therefore the air gap can be wider. That
reduces production costs. The magnets are thin, thin-walled hollow cylinder,
in dovetail shapes or plats. Since the flux density is low the axial length of
ferrite magnets is often longer than the armature length to get more flux.
The magnets are usually fixed with clamps or springs, seldom by adhesive.

1 often bell-type motors

2 often disc-type motors
96 Hans-Dieter Stolting

Magnet

Pole shoe
a b c
Fig. 3.8. High-grade constructions of DC motors

Sometimes they are encapsulated in plastic. Flexible magnets are also used,
which adapt themselves very well to the housing (Fig. 3.70). Mostly the com-
mutator is cylindrical. Flat commutators are used if the motor length is to
be as short as possible without employing a more expensive disc motor (see
below). Double-pole motors usually carry lap windings, and multipole motors
often wave windings because only two brushes are necessary.
High-grade motors are usually fitted out with rare-earth magnets (SmCo,
but increasingly NdFeB) and with AlNiCo magnets (Fig. 3.8). The latter
material is temperature-stable and provides a high flux density, but a low
coercive force. Because of that it is very endangered by demagnetizing fields
generated by high armature currents. Therefore AlNiCo magnets have to be
long in the direction of magnetization (Fig. 3.8b). To limit the demagnetiza-
tion the currents have to be as low as possible. On the other hand pole shoes of
sheet steel are used to conduct the demagnetizing armature-axis flux, so that
it cannot affect the stator magnet (Figs. 3.8b, 3.8c). Besides them pole shoes
concentrate the exciter field to achieve a higher flux density in the air gap.
Because of that it is disadvantageous that the armature time constant and the
commutation get worse. The air gap should be as small as possible. All of this
makes motors with AlNiCo magnets expensive and has to be considered while
designing them. Today the importance of AlNiCo magnets is decreasing.
Other magnets can be small. SmCo has both high remanent flux density
and high coercive force. So the BH product is very much higher than that of
AlNiCo. It is hard to demagnetize, but it is very expensive. NdFeB magnets
are more cost-effective and have an even higher BH product than SmCo mag-
nets. However, high-grade NdFeB magnets are very sensitive to temperature
and corrosion. Cheaper NdFeB magnets with a lower magnetic remanence,
approximately like SmCo, can be used up to 150 °C and are corrosion-stable.
The development of permanent material is not yet completed. In the future
the BH product should increase further. This will influence the design of
motors with permanent magnets and intensify their importance.
In motors with slotted armature the permeability varies according to the
rotor position. This causes flux oscillations and t hrough that torque oscil-
 3 Electromagnetic Actuators 97

lations. These so-called cogging torques make the true running of the rotor
worse. There are possibilities to reduce cogging: nonconstant air gap, odd
slot number , slot skewing, optimized width and shape of tooth tip, diagonal
magnetized or rhombic shaped magnets et c. These actions make motors more
expensive and do not suppress the cogging completely.
A winding arranged uniform on an unslotted armature core or self-
supporting winding (moving coil) in the air gap hardly creates torque os-
cillations. The latter are carried out as bell or disc shaped winding. The
inductivity of these motors is lower than that of motors with slotted rotors
because of their wider air gap. Therefore the commutation is better and the
electrical time constant is lower. The pot shaped rotor of bell-rotor motors
rotates around the internally placed double-pole magnet . Occasionally there
are four-pole motors. Figure 3.9 shows an exploded view of a bell-rotor mo-
tor with a cut-out housing part. The rotor winding is carried out as a skew
winding (Faulhaber winding) or as a rhombic winding (Maxon winding) . The
solid housing is used as the magnetic return path. These careless motors are
of the intermediate-rotor type (cylindrical-inside-field type). As the ironless
bell rotor has a very low moment of inertia, the mecha nical constant is low.
Therefore motors with such rotors have the best dyna mic properties of all
motors and are especially suited for control and servo drives as well as in ap-
pliances which require good true running. Commutator and brushes are made
of precious metal or are silver plated to reduce the contact resistance. AlNiCo
and now increasingly NdFeB magnets are used . These motors are produced
with an output power up to 100 W , occasionally with 250 W . Larger motors
are not able to be built because of the mechanical stability of the bell rotors .
Other coreless motors have flat rotors . Very small motors have four ferrite
magnetic poles in the stator, three flat self-supporting air-core coils and often
a flat commutator (Fig. 3.10). Motors with a power up to 200W have eight
or ten ferrite poles and a plastic disc with punched or printed conductors
on each site brazed together at the inner and the outer site. The conductors
form a wave winding, so that these motors need only two brushes. They are
made of graphite and axially arranged sliding on the inner part of the wind-
ing (Fig. 3.11a). Larger disc motors which are also known as pancake motors
have flat coils imbedded in plastic and eight or more poles of ferrite, AlNiCo

Rotor winding

Brush

Stator magnet

Fig. 3.9. Exploded view of a bell-rotor motor

98 Hans-Dieter Stolting

or NdFeB magnets in the stator (Fig. 3.11b). Normally the commut ator is
cylindrical, but seldom are there fla t commutators. Because of their large
outer diamet er disc motors with an output power up to 500 W have a m e-
chanical time constant greater t han that of yc lindrical outside-field motors
with slotted rotors.

Magnet

Commutator
Fig. 3.10. Flat DC motor with an outer dia meter of 20 millimeters

a b Commutator

Fig. 3.11. Disc motors

Table 3 .3 g ivessome characteristic dat a of t he above described motors

and Fig. 3.12 shows the most important performance curves of a car auxiliary
drive which a re typ ical of all permanent-magnet mot ors.
The s peed r egulation of a p ermanent-magnet motor i s made byseries
resistance (additional losses, cheap) or by a n electronic cont roller (chopper :
low-loss, expensive). Motors of windscreen wipers somet imes have a t hird
brush, so t hat the armature winding is connected eit her with t he two norm al
brushes or with one of t hese brushes and t he t hird brush . In t he la tter case
 3 Electromagnetic Actuators 99

Table 3.3. Typical catalogue data of permanent-magnet motors

Nominal Nominal Nominal Nominal Effi- Inertia Weight
voltage speed torque output ciency
power
v rpm Nm w % gcm 2 kg
Bell type 12 5000 0.02 6 72 16 0.13
motors 12 5000 0.05 20 78 47 0.27
12 12000 0.001 0.6 73 0.4 0.018
12 12000 0.0025 2 76 0.7 0.02
Disc type 24 3000 0.032 10 50 400 0.45
motors 24 3000 0.16 50 68 2100 1.3
24 3000 0.61 200 68 2900 3.9
24 3000 1.27 400 70 7900 6.5
24 3000 3.2 1000 82 10000 9.7
Servomotor 24 3000 0.064 20 55 120 0.58
24 3000 0.16 50 70 400 1.2
24 3000 0.32 100 76 950 2.1
24 3000 0.64 200 78 2900 3.6
24 3000 1.59 500 83 12000 8.8
Cost- 12 14000 0.0007 1 50
effective 24 14000 0.0034 5 55
motors 30 14000 0.0068 10 65
Car 4000 0.12 50 55 0.8
auxiliary 4000 0.24 100 60 1.8
drives 12 4000 1.19 500 65 4.0
or 4000 2.39 1000 65 4.0
24 12000 0.008 10 35 0.08
12000 0.04 50 55 0.25
12000 0.08 100 60 0.5
Pulsating- 220 7000 0.068 50 55 0.42
current 220 7000 0.136 100 57 0.7
motors
with AC
terminal
and recti-
fication

the commutation is worse (three brushes motor [3]). This method should be
used only for a short time.

3.2.4 Pulsating-current DC Motor

Pulsating-current motors are permanent-magnet DC motors which are op-

erated with rectifiers at the 230 V AC system without special smoothing
elements. The current ripple amounts to more than 10%. Table 3.4 explains
100 Hans-Dieter Stolting

28 7000 70
I _l:L p
A rpm
24 6000

20 5000
"'%
16 4000

12

4 10

0 0
5 TN 10 15 20 25 30_I_35
Ncm
T: torque, TN: nominal torque, N: speed, P: output power, /: current, fJ: efficiency

Fig. 3.12. Performance curves of a permanent-magnet DC motor

N I
rpm .; ....······· A
0.8
5000 I .; .;
..········
...··
..; . ...········
.; ..·· 0.7
IE-
4000
to=········
1\'. ·········· .... .. .;

..·· ..···
...··· = Direct current 0.6

···v. *
.;
Diode
... ...·· 0.5
3000 ' ·, ····· .,· ....······ Uncontrolled
······· bridge
;;-..:, ....······· ....
.; 0.4
\. .
\
\
\ .; ·········· .... .. bridge(so·)
2000 .; ...···
':o.
... ... 0.3
'\.····· ... ··········
... ·········· 0.2
.·\····· ········
.... ,
"""
1000 ... ·..... N ··········
0.1
....··· \ ....
0
...... 0
2 4 6 8 10 12 14 16 18 20 T 24
Ncm
T: torque, N: speed, /: current (root-mean-square-value)

Fig. 3.13. Performance curves of pulsating-current DC motors with different elec-

tronic circuits

why these motors are not employed today to such a degree, as was expected
some years ago. They are used in easy household appliances, especially in
those which have a heating coil as series resistor (hair-drier, foil welding rna-
 3 Electromagnetic Actuators 101

chine). Dependent on improvements in electronics and magnetic materials

it is to be expected that its field of application will extend considerably in
the future. Table 3.3 shows typical data of these motors. The easiest rec-
tification electronics need only one diode. Today two-pulse bridge connec-
tions are used even in simple appliances because they render possible better
motor utilization and they are cost-effective as semiconductor devices. Fig-
ure 3.13 shows speed-torque characteristics with various control methods
compared with DC supply. The curves for the uncontrolled and the con-
trolled bridge mark the region where the motor speed can be regulated by
a phase-angle control (Sect. 3.6.4). The latter curve applies to a timing angle
of a= 90°.

3.2.5 AC Commutator Motor

The AC commutator motor (universal motor) still has great importance in

the range of small-drive engineering. A8 it can be built with high 8peeds (up to
40,000rpm), its power/weight ratio is very favorable. Besides them the speed
can be varied in a wider range and more easily than every other motor type
compared to the costs. It is the most important drive of portable appliances
like household appliances, hand or dish tools: food slicer, mixer, coffee grinder,
washing machines, vacuum cleaner, sewing machines, garden pump, hedge
clippers, drilling and grinding machines, saws, milling and planning machines,
hot-air blower, spray-guns, high-voltage switches.
To easier appliances motors with pulsating current can be used. There-
fore in Table 3.4 typical characteristics of both commutator motors of the

Table 3.4. Comparison of the characteristics of DC and AC commutator motors

of the same size
Pulsating current DC AC commutator motor
motor
Torque identical
Production cost identical
Losses identical
Speed characteristic more load- independent very load-dependent
Speed control easier more expensive
Speed-regulation range at the smaller wider
same cost
Starting torque lower greater
Commutation better, longer brush life worse
Rectification at 230V-mains necessary not applicable
operation
Interference suppression more expensive easier
DC component in a fault case special protective not applicable
Demagnetizing danger actions if need not applicable
102 Hans-Dieter Stolting

same size are compared. The wide speed range of AC commutator motors
is used in washing machines, where the speed difference between wash and
spin-dry amount up to 1:60. Its high starting torque is advantageous for some
appliances like stirring apparatus, drilling machines and switches. In contrast
to large DC series motors small universal motors are not able to reach un-
permissible high revolutions at unloading because of the relative high brush
friction.
AC commutator motors are always double-pole built, mostly with a double-
sided symmetrical section (Fig. 3.14a), seldom with a one-sided symmetrical
section (Fig. 3.14b) with the characteristics referred in Sect. 3.1.3. The two
coils of the field winding are situated before respectively behind the armature
winding. Because of that they act as interference suppression coils. Supplied
by a rectifier they work as a smoothing reactor improving the commutation
and because of that the life time of the motor. Table 3.5 puts together char-
acteristic data of universal motors. Figure 3.15 shows the performance curves
of a 500W-pump motor.
In former times tapped field windings were used for supplying universal
motors either with direct current or with alternating current (Fig. 3.16a).
The speed can be regulated

lossless by tapped field winding (Fig. 3 .16b),

- covered with losses by a series resistance or/ and a resistance parallel to
the armature winding (hardly used today, Figs. 3.16c, 3.16d),
lossless by variation of the terminal voltage.

Table 3.5. Characteristic data of AC commutator motors

Nominal speed Nominal Nominal output Efficiency Weight
torque power
rpm Nm w % kg
4000 0.12 50 40 1.5
4000 0.24 100 55 2.2
4000 0.48 200 65 4.5
6000 0.16 100 63 2.0
6000 0.32 200 66 3.0
6000 0.80 500 80 4.5
8000 0.12 100 65 2.0
8000 0.12 200 67 2.3
8000 0.60 500 80 3.0
10000 0.048 50 60 1.0
10000 0.095 100 67 1.5
10000 0.19 200 70 2.0
15000 0.255 400 75 1.1
15000 0.51 800 80 1.8
20000 0.334 700 75 1.5
20000 0.477 1000 80 2.4
 3 Electromagnetic Actuators 103

Rotor or
armature winding
a b

Fig. 3.14. Electrically excited commutator motors

Tapped windings are used if a wide range of speed regulation is desired

(washing machines). Nowadays variation of the terminal voltage takes place
by phase-angle control (Sect. 3.6.4). In the future, universal motors may be
run by pulse-width modulation (PWM) both rectifying the motor current and
varying the motor voltage. Today this is expensive compared with phase-angle
control.

I p1 .Ji..
w
"
.!L
A rpm %
6 600 12000 k:::::
....... ...::::: /(
......... p1
5 500 10000 100

--
...........
'> P.-"
-
4 400 800 80
......... .z;:::.':JI""
17
r--
3 300 600
v 60

2 200 400
v 40

0
100

0
200
v
0 0.1 0.2 0.3 0.4 0.5 TN 0.7 0.8 0.9 1.0 ..L..
0
20

Ncm
T: torque, TN: nominal torque, 17: efficiency, P1: input power, N: speed

Fig. 3.15. Performance curves of AC commutator motor

104 Hans-Dieter Sti:ilting

a b T

c T d T

Fig. 3.16. Speed regulation of universal motors. a Supplementary DC winding, b

field tapping, c series resistance, d stator parallel resistance

3.3 Self-clocked Motors with Electronic Commutator

3.3.1 Brushless DC Motors

Electronic motors have a polyphase winding in the stator and a permanent-

magnet or reluctance rotor. An electronic control switches the stator phases
in cyclic order dependent on the rotor position. Therefore these motors are
also designated as EC- (electronically commutated) motors or BLDC- (brush-
less direct current) motors [2, 4, 5]. They are robust and low-noise just as
asynchronous motors. They have a better dynamic performance than DC
commutator motors. They are applied not only in the field of capital goods
but above all permanently widening in the field of consumer goods, for in-
stance: tape and video recorders, hard-disc drives, camcorders, digital cam-
eras, plotters, printers, laser reflectors, optical storages, analysis and dosing
apparatus, bar-code scanners, robotics, winding and feed drives, welding ma-
chines, electro-cardiographs, dialysis apparatus, artificial-limb control, small
fans, car auxiliary drives. Meanwhile there are increasing applications in the
range of simpler apparatus. The constant development in the field of electron-
ics and of magnet technique reduces on the one hand the cost and improves
the properties on the other, so that new cost-effective designs, for instance
with stators according to the claw-pole principle, are possible (Fig. 3. 70).
More and more DC commutator motors are being replaced by BLDC motors.
In comparison to DC commutator motors the construction and the tasks
of stator and rotor are changed (Fig. 3.17) but without influence on the fun-
damental method of operation and performance. The stator is the armature;
the rotor excites the magnetic field. Often BLDC motors are drive alterna-
tives to synchronous (stepper- )motors. In contrast to them
they cannot pull out of synchronism because the stator phases are
switched just when the rotor position encoder indicates that the rotor
has reached the corresponding position,
 3 Electromagnetic Actuators 105

their speed is load-dependent and not constant (shunt characteristic) ,

they are expensive because of the electronic commutator, and
their vibration ability depends on the resolution of the rotor position
encoder.

Fig. 3.17. Fundamental constructions of mechanically and electronically commu-

tated motors

For reducing torque oscillating conventional DC motors often have more

than three coils in the rotor, as shown in Fig. 3.17. Twelve up to 32 coils
are typical. Because of the cost at most four coils are used in BLDC motors
since the phase number of the control electronics has to be as great as the
phase number of the winding, essentially determining the cost. The electronic
commutation can happen in two ways. Either the phases are switched on
and switched out or their current direction is changed. In the first case the
current flows alternately in opposite directions in two coils per pole belonging
to one phase (unipolar connection). Here the control electronics is cheaper.
In the second case the alternating current flows only through one coil per
pole energizing alternately in the north and south pole (bipolar connection).
Now the motor efficiency is higher. Therefore most modern BLDC motors
are bipolar driven.
The driving of the winding phases is mostly done by Hall elements. They
are built in, so that they are excited by the rotor magnet itself or by a special
control magnet turning round with the rotor. It is also possible to evaluate
the induced voltage of the winding phase which is not used at this moment
(sensorless control). Further possibilities for determining the rotor position
offer pulse generators which supply the electronics with a certain number
of impulses per turn (magnetic or optical incremental encoder) or a definite
impulse code according to each rotor position (code disc). The latter method
detects the rotor position at standstill, too, so that the electronics are able to
start the motor automatically. In the other case a special starting connection
is required taking over the control, until a reset pulse of the rotating slit disc is
given to the counter which now begins to count. After every turn a reset pulse
is generated. The speed control is mostly made by a tacho-generator and -
with some restrictions relating to the accuracy- by evaluation of the induced
voltage of the just not used phase. Besides them resolvers are used which
106 Hans-Dieter St6lting

register the rotor angle in an inductive way. As with the above-mentioned

sensors, they are not only suited for winding switches but for positioning and
control purposes, too.
BLDC motors are able to reach more than 30,000 rpm, to cover a speed
range up to 1:3000 and to generate torques between 0.02Nm up to 100Nm.
Their life time amounts to several 10,000 hours. Typical data are given in
Table 3.6.

Table 3.6. Typical data of BLDC motors

Nominal Nominal Nominal Nominal Efficiency Inertia Weight
voltage speed torque output
power
v rpm Nm w % gcm 2 kg
24 5000 0.02 10 60 30 0.5
24 5000 0.057 30 65 70 0.9
42 3000 0.032 10 65 20 0.3
42 3000 0.095 30 65 200 1.0
42 3000 0.32 100 70 1400 2.5
230 3000 0.095 30 40 200 1.0
230 3000 0.32 100 45 1200 3.0
230 3000 3.18 1000 65 2000 5.0
230 5000 0.38 200 50 400 2.0
230 5000 1.15 600 60 600 3.0

The simplest BLDC motors have only one winding phase in the stator
which gets one pulse every turn generated by a Hall element (Fig. 3.18). The
useful torque is so low and it pulsates so intensively that it is used only for
simple appliances, for instance as a motherboard fan. The torque becomes
better if the winding gets two pulses per turn, so that the current flows
alternating in both directions (bipolar connection, Fig. 3.18c). This needs
four transistors. Two winding phases flowing through alternatively (unipolar
connection, Fig. 3.18b) are cheaper. In both of these cases the generated
torques do not overlap each other. That means that the resulting torque
disappears at the electrical angles of 0° and at 180°. Stopping at these points
the motor is not able to start self-supporting. Therefore an auxiliary torque
is superposed which is different from null at the above angles. Figure 3.19
illustrates the two possibilities to generate such a torque at examples of four
pole motors: the left side shows a motor with unsymmetrical poles exciting an
additional reluctance torque; the right side has no wound auxiliary poles and
unsymmetrical magnetized rotor. The connections (b) and (c) in Fig. 3.18
need only one Hall element. For one current direction the positive Hall pulse
is used, excited by the north pole of the rotor magnet, and for the other
direction the negative Hall pulse is used, excited by the south pole.
 3 Electromagnetic Actuators 107

,]: W
H
R
T
Phase winding
Hall element
Resistance
Transistor

f g

unipolar bipolar

Fig. 3.18. Connections of BLDC motors. a One-pulse connection, b, c two-pulse

connection , d, e four-pulse connection, f three-pulse connection, g six-pulse con-
nection

0: not magnetized

Fig. 3.19. Two-phase motor with auxiliary torque

108 Hans-Dieter Stolting

Four-pulse connections (Figs. 3.18d, 3.18c) need two Hall elements. As in

the two-pulse cases each winding generates a torque over an electric angle
of 180°. But now there are four torques which superpose each other. The
useful torque increases and the pulsating torque decreases. These motors were
used in former times when the cost of Hall elements and the accompanying
electronic circuit are very expensive. Figure 3.19 shows a four-pulse motor
with an air-gap winding describing the position of the four phases. The phase
numbers correspond to the numbers in Fig. 3.18. The phases are supplied in
the sequence 1, 3, 2, 4. Motor production is expensive. First the rotor is put
into a plastic cylinder. Then the four phases are wound axially round the
cylinder. Finally both parts are put together into the housing which consists
of a laminated core within an aluminium cylinder.
Three-phase motors are supplied with three or six pulses (Figs. 3.18f,
3.18g). The first one is only seldom used for simple actuators. Most important
is the latter type. It has a higher useful torque and an about 13% lower
pulsating torque than that of four-pulse motors. As electronic circuits and
Hall probes are cheaper than in the past, four-pulse motors are hardly used
today. Besides of the star connection in Fig. 3.18g the motor can be delta-
connected. The choice of the connection depends on the supply where it is
to be considered that in delta connections loss may arise generated by ring
currents if there are asymmetries in this circuit.
EC motors are produced with an internal rotor or with an external rotor
if this offers constructive advantages, if a good true-running is demanded or
if the magnetic material needs as much space as possible. Figure 3.21 shows
a motor with salient poles which may generate cogging torques. In slotless
round-frame-type motors a winding with axial coils or winding known as
Faulhaber or Maxon winding (like Fig. 3.8) can be fixed in the stator, divided
into three star-connected parts and directly driven via transistors. These
motors are produced in two types. The first one has a la minated stator core
in which the winding is pasted in like the motor in Fig. 3.20, but without
a plastic cylinder. This construction is expensive (Fig. 3.22 at the top). The
second one has a self-supporting winding. Outside of the winding a massive
magnetic return path rotates with the inside rotor (intermediate-stator type).

Fig. 3.20. Four-pulse motor with a ir-gap winding

 3 Electromagnetic Actuators 109

Magnet pole

Coil

Fig. 3.21. Three-phase motor with external rotor (hard-disc actuator)

Because of that no eddy currents are generated. This construction is much

cheaper, but has a much higher moment of inertia. This is necessary if a drive
with good true running is required (Fig. 3.22 at the bottom).
In Fig. 3.23 there are given construction possibilities of BLDC disc motors
with ironless coils or self-supporting windings. Very small disc motors are
made with photolithographic windings. The stator has three phases with two
flat coils each. The rotor has eight poles. Figure 3.24 shows the so called
"penny-motor", a new micro motor. In this motor the return path rotates
together with the rotor, so that there are no eddy currents (see Fig. 5.64).
With wave-formed concentric conductors embedded in plastic the expen-
sive interconnection of coils is avoided, but on the other hand the manufacture
of this winding is more difficult and expensive (Fig. 3.25). Larger disc mo-
tors have a helically wound stator core often without slots. To reduce the

Stator winding

End shield Laminated core

Flux return path

Fig. 3.22. Exploded view of slotless BLDC motors

llO Hans-Dieter Stolting

EmB Winding
Magnet
c::::J

Fig. 3.23. Constructions of disc motors with radial and axial magnetic fields

Shaft

-- Rotor

Nonmagnetic stator

Fig. 3. 24. Disc motor ( "penny motor" )

Return path Stator wind ing Magnet

Fig. 3.25. Disc motor with waved-formed coil

 3 Electromagnetic Actuators 111

axial force the stator is divided into two parts on both sides of the rotor
(intermediate-rotor type). The rotor has to be constructed to be distortion-
resistant to resist the stress caused by high pulse currents. This and the
essentially handmade manufacturing are the causes of the very high cost.
Because of the cost most small linear actuators are still produced as a ro-
tating motor with a gear, for instance with a gear rack, a toothed belt or
a thread bar with a nut. But today linear direct drives are appearing. The
disadvantages of gears (play, noise, losses) are avoided, but the size of the
drive becomes greater. The stator or the sliding element have to be as long
as the drive way. If the motor is not too small an air-lubricated bearing car-
ries the short slider with the field winding. It is necessary that the air tube
is fixed on the slider just as the electric connecting cable. This causes an un-
desirable torque which has to be taken in consideration if the motor should
be driven controlled. Very small linear BLDC motors have the field winding
in the stator, a permanent-magnet slider and a micro-ball bearing. All these
motors can be built with a two-axis motion (xy-motion), so that they are able
to follow any path. Besides them it is possible to realize three-dimensional
motions (xyz-motion) by combination of several motor systems. In addition,
a rotating element can be added [8]. The advantage of these motors com-
pared to voice-coil motors (Sect. 3.5.3) is that they can be driven in a larger
range.

3.3.2 Servo Motors

In the lower power range there are pilot or servo drives too. Their task is
to position mechanical elements at a given position within a given time and
with a given precision. These tasks are given in feed screws and feed drives
in machine tools 1 , in robot systems 1 , in conveyor belts 1 , smaller antennas,
smaller telescopes, butterfly valves, valves. Often several motors have to run
in synchronism. On servo drives heavy demands are made relating to the
dynamics, the positioning exactness, the peak torque (up to the fourfold
nominal torque), rotational trueness (especially at crawl speed), the speed
regulated range (up to 1:10,000), the efficiency, explosion protection, pro-
tection from environmental influences, etc. They are valid not only for the
motor but also for the other drive components, like power amplifiers, sensors,
control electronics and mechanical transfer elements. To reach an optimal
control dynamic the time constants all of the assemblies have to be as low
as possible. Often these motors are used in four-quadrant operation which
means the motors can be driven and broken in both directions. In the lower
power range electromagnetic actuators have prevailed against hydraulic and
pneumatic drives. The causes are given in Table 3.7 (see Sect. 4.8).
Some years ago only DC commutator motors were used as servo motors
because of their easy adjustability. Even though in the lower power range
1 Mostly motors with an output power greater than 1 kW
112 Hans-Dieter Sti:ilting

Table 3. 7. Comparison of hydraulic and electrical servo drives

Hydraulic drive Electromagnetic drives

higher energy density better control response
smaller mounting space higher efficiency
higher accelerating ability easier maintenance
lower cost easy adaption to different condi-
easy generation of linear motion tions
problem: leakage the same energy type for sensors,
control electronic and actuators

many servo actuators are still DC motors, today three-phase drives with elec-
tronic commutators increasingly prevail. As in DC motors with a constant
exciting field the torque depends linearly on the armature current, so the
regulation and control by the current is very easy. The advantage of a DC
commutator motor is that only one current value has to be measured and
controlled. Therefore the control electronics are simpler. Should the moment
of inertia be low, ironless bell-rotor motors are used up to an output power of
lOOW; at powers more than 20W ironless disc or (should the occasion arise:
slotless) bar-type rotors are used. Ironless rotors have especially low mechani-
cal and electrical time constants (see above). The rotors of three-phase motors
with a permanent magnet or with a squirrel-cage winding have a greater mo-
ment of inertia. In Table 3.8 the characteristic differences between DC motors
with and without brushes are listed. Figure 3.26 illustrates the performance
range of BLDC motors, enlarged relative to DC servo motors.

Table 3.8. Comparison of DC and three-phase servomotors

DC current drives Three-phase drives

- easy supply and control unit by one - control at least by two currents and
current capture of the rotor position
- easy servo-amplifiers -expensive converter, higher switch-
ing capacity
- noise and wear of brushes and com- -maintenance-free, robust, low-noise
mutators
- brush life up to 3000 h - brush life several 10000 h
- dynamic limited by commutation - higher protection grade
- heat dissipation over the shaft - ironless motors: essential lower mo-
ment of inertia
-at voltages higher than 400 V -lower weight, smaller size
matching transformator is necessary
- direct mains-operation
- more expensive
 3 Electromagnetic Actuators 113

6 6 --,
T -Raffi:earth-ma9ne1s------ T
I
\ Rare-earth magnets
rN . . . . . ., TN
I
', Dynamic
I
I
I
4 : extreme 4 I
I
I range \
I

3 Ferrite magnets J
I

2 S3-25% Intermittent
. S3-40% / periodic
Vm77777777.S1 duty
Continuous

0.2 0.4 0.6 N 1 0.2 0.4 0.6 N

a NN b NN
Fig. 3.26. Performance range of servo motors. a BLDC motor, b DC motor; 25%,
40%: cyclic duration factor

In Table 3.9 AC motors (three phase induction motor) and BLDC motors
are compared. Both of these motors are dynamically equivalent. It is not
clear generally which of the two is more favorable. Is a higher speed range
wanted; is the AC motor the better one because of its larger field weakening?
If a higher positioning accuracy i8 required a BLDC motor is the better one.
In three-phase AC motors with squirrel-cage rotor the torque causing cur-
rent is generated inductively in the rotor. They need neither high-grade mag-
nets nor commutators or 8lip rings. Their speed i8 close to the synchronous
speed which depends on the frequency of the pole number and the electronic
control. Therefore the speed can be regulated by the frequency (see Sect.
3.4.1). The worse efficiency and power factor, in comparison with BLCD mo-
tor8, require a more powerful and more expensive supply unit. Therefore in
drive8 lower than 1 kW AC motors are 8eldom employed.
Figure 3.27 show8 the control of a servo commutator motor (m). The cir-
cuit exists of three in one another inserted control loops (cascade-circuit).

Table 3.9. Comparison of BLDC and AC motors

BLDC drive Three-phase drive

Robustness high very high
Overload capability high very high
Field weakening only restricted possible in a large range
Ohmic loss in the stator in the stator and in the ro-
tor
Reactive-power demand not applicable necessary
Size at the same power smaller higher
Motor cost high low
Control easier, cheaper large-scale, expensive
Total cost high very high
114 Hans-Dieter Sti:ilting

Three-phase system
L1 Lz L3

Fig. 3.27. Control circuit for servo drives

The desired position (setpoint position) is given by a system, the actual

value by an encoder (e). The close-loop position control determines, with
a setpoint/actual-value comparison, the desired speed N 8 • Comparing this
with the actual speed value Na measured by a tacho-generator (t) results
in the desired motor current is which is adjusted by the inner control loop.
A six-pulse bridge supplied by a three-phase system yields the direct current
ud dependent on the control signal Uc. The control circuit of an electronic
commutated motor is like this, but there at least two currents have be mea-
sured.

3.3.3 Switched Reluctance Motor

Besides the EC motors with permanent-magnet rotor, described up to now,

there are, however, very rarely, motors with reluctance rotors. Commonly sta-
tor and rotor are laminated because of the heavy flux pulsations. The motor
in Fig. 3.27 has three phase windings which are arranged on two opposite sit-
uated poles. The phases are star- or delta-connected. Dependent on the rotor
position the voltage is applied to them in a cyclic succession 1 . The switching
happens by three Hall elements which are displaced by 120°. The unwound
rotor follows the revolving field trying to put itself in such a position which
minimizes the magnetic resistance (reluctance). This is the case if two rotor
poles are in line with the just excited stator poles. As the two rotor positions,
drawn in Fig. 3.28, illustrate, the rotor rotates with half of the field speed
and opposed to it. Besides three-phase motors, four-phase motors with eight
stator poles and six rotor poles are suggested.
The motor is very robust and frequently used. That makes it interesting
and one hopes that it would prevail against the classic motor types. But there
are serious disadvantages. During the rotor motion the magnetic resistance
varies. Therefore the flux is fluctuating. With it the torque pulsates causing
great noise and vibrations. Many suggestions have been made to minimize
this, for instance by duplication of the number of poles per phase, special
1 If the phases are switched independent of the rotor position it concerns a syn-
chronous motor (see Sect. 3.4.2).
 3 Electromagnetic Actuators 115

Hall
element

Fig. 3.28. Switched reluctance motor

shapes of the poles or by current control. All these methods make the motor
more expensive or reduce the utilization. A further problem is the high in-
ductivity of the winding which makes switching of the phase winding more
difficult. Therefore special measures have to be taken to accelerate the dis-
connection, for instance by current control. Otherwise currents are flowing in
several phases simultaneously. This generates braking torques. These are the
reasons why this motor type has not prevailed up to now and why it is not
suitable as a servo motor. Today it is used only in niches applications.

3.4 Externally Commutated Motors

3.4.1 Asynchronous Motors

Appliances and Design

Small asynchronous motors (AC motors, single-phase (induction) motors)

are mostly manufactured for operation with 230 V alternating voltage (AC
voltage), more seldom with 400 V three-phase voltage [1]. In the following
there are put together exemplary apparatus which use the typical charac-
teristics of these motors (cost-effective, robust, maintenance-free, low-noise,
low-vibration) and no speed regulation or at most a simple one is required:
household and garden apparatus (dishwashers, washing and ironing ma-
chines, clothes dryers, refrigerators, freezers, fan-assisted ovens, air dome 1 ,
garden chaff cutter, lawn mover, oil and gas burners, heating pumps and
other pumps, roller blinds, sun blinds, garage doors, fans 1 , stable fans), of-
fice machines (typewriters, copiers\ fax machines, printers, paper shredders,
apparatus fans 1 , coin counter machines), tool machines (circular saws, high
frequency tools, cement mixers, high pressure cleaners, compressors, grind-
ing machines), laboratory and medicine technique (stirring machines in ther-
mostats and cryostats 1 , analysis apparatus 1 , recording instruments 1 , air-bed
116 Hans-Dieter Stolting

Laminated core

Stator winding

Rotor bar
Fig. 3.29. AC asynchronous motor, capacitor motor

Fan impeller

Fig. 3.30. Asynchronous motor with external rotor, radial blower

pumps 1 , incubators, pumps in electrocardiogram apparatus 1 ), picture and

sound technique (film projectors, cut dishes\ juke boxes, scanners).

1 often external-rotor motors

 3 Electromagnetic Actuators 117

Stator and rotor are laminated, to minimize the eddy current losses gener-
ated by the pulsating magnet fields. Besides the most common internal-rotor
motors (Fig. 3.29) there are external-rotor motors. The latter are employed
among other things for fans because of the smaller axial length and the higher
fan efficiency. The fan blades are mounted directly onto the rotor housing
(Fig. 3.30). The rotor housing of winders can be used to roll up the mate-
rial directly. The stator winding being of two or three phases is distributed
mostly in slots. Only very small motors have concentrated coils on salient
poles (Fig. 3.31). The rotor has slots with one uninsulated bar each. All the
bars are connected at the end sides by rings (squirrel-cage winding). In a few
cases very small motors are equipped with massive iron cylinder or massive
bell rotors, possibly the latter as external rotor. The maximum torque of these
motors is the start torque. Ferraris motors with an aluminium or copper bell
rotor are suitable as positioning actuators, but today are hardly produced
because of the better suitable BLDC motors (see Sect. 3.3.2).

Fig. 3.31. Sections of four-pole double-phase asynchronous motors

Method of Operation and Performance Characteristic

The stator phase currents generate a magnetic field which rotates with the
synchronous speed
f
Ns = -
p
where
f line frequency
p number of pole pairs.
118 Hans-Dieter Stolting

The revolving field of a two-pole motor rotates at f = 50Hz with N 8

3000rpm, the field of a four-pole motor with N 8 = 1500rpm, etc. When the
rotor turns round with a speed different from the synchronous speed, the field
induces voltages in the rotor bars and with it currents. The rotor currents
generate the torque together with the rotary field.
Increasing the speed the power /weight ratio can be improved; that means
the weight can be reduced at the same power. This is used for so-called high-
frequency tools. These are portable hand tools in factories (drilling and grind-
ing machines, etc.). Their motors are driven with frequencies from 200Hz to
400Hz, so that speeds are reached from 12,000rpm to 20,000rpm like uni-
versal motors (Sect. 3.2.5). Advantages related to universal motors are the
special characteristics described in the previous section. A disadvantage is
the worse possibility of speed regulation. In addition, an expensive frequency
converter is necessary.
Three-phase motors possess the best power/weight ratio of all AC motors,
but they are produced economically with an output power more than 50 W.
Their locked-rotor torque and their breakdown torque (maximum torque)
are about twice as great as their nominal torque. With a 400 V line generally
the winding is star-connected (Fig. 3.32b). Apart from the appearance of
parasitics the field rotates with a constant quantity (circle rotating field).
A typical speed-torque characteristic shows the curve (a) in Fig. 3.33. It is
a motor with a relatively high rotor resistance (high-resistance squirrel-cage
motor) to get a start torque as great as possible.
A three-phase induction motor can also be connected to an AC voltage
system, linking two terminals to the system directly and the third in series
with a capacitor. Using an existing, usually star-connected 400 V three-phase
motor it is to be delta-connected to a 230V system (Fig. 3.32c). As in this case
indifferent currents flow in the winding phases in contrast to a three-phase
motor, the torque pulsates during the revolution (elliptical rotating field). It
is pulsating with double the system frequency, so that this oscillating torque
superposes the useful torque which is constant at a certain speed. This os-
cillating torque causes noise which is typical for AC motors. The torque of
a single-phase motor with three windings is much lower than that of a three-
phase induction motor. Disconnecting one terminal of a three-phase motor
(Figs. 3.32e, 3.32f) the motor still produces a torque apart from the start-
ing direction like a one-phase motor. Such hand-started single-phase motors
(Fig. 3.32g) are not manufactured today, but often one-phase operation is
given.
To generate a rotating field two windings with out-of-phase currents are
enough. Mainly the phase displacement of the currents is done by capac-
itors, but also by increasing the ohmic resistance of one of the two phases
(Fig. 3.32h). By appropriate dimensioning of the winding and of the capacitor
a motor can be balanced which means the motor operates as a three-phase
AC induction motor. But this is possible only for one speed value. Should
 3 Electromagnetic Actuators 119

.--------o t Right
• 0
t Left
t Right
• 0

k
______._____. t Left
VN

Fig. 3.32. Asynchronous motor, methods of connection

120 Hans-Dieter Stolting

be that the nominal speed the motor gets a running capacitor C 8 which
is switched on constantly. However, in this case the locked-rotor torque is
very low because the capacitor is too small (Fig. 3.33e). If a high locked-
rotor torque is required the motor gets a starting capacitor CA which has
to be much higher than the running capacitor. Now the starting capac-
itor is badly suited to nominal operation. This causes inadmissible high
losses warming up the motor. Therefore the auxiliary phase with the start-
ing capacitor is to be switched off. Now the motor is single-phase driven
(Figs. 3.32i, 3.33d). Motors with a high locked-rotor torque and a high nomi-
nal torque get, as well as a starting capacitor, a running capacitor (Figs. 3.32j,
3.33£).
Mainly two-phase capacitor AC motors are produced because they need
a smaller capacitor than AC motors with three phase windings. Besides them
they have a better utilization than resistance-start motors. But the latter are
still more robust than capacitor-start motors.Therefore they are employed

Tb
0.8 1---+:::#::...._-+___::'-----t-""8.-----+------1

.................................. ··

................,.......... . . .

0.2 0.4 0.6 0.8 N

Ns
a Three-phase motor
b Single-phase motor, triple phase
c Motor with one phase

}
d Capacitor-start motor
e Running-capacitor motor Douple-phase
f Two-value capacitor motor AC asynchronous
g Resistance-start motor motor

Tb : breakdown torque of the three-phase motor

N5 : synchronous speed

Fig. 3.33. curves of different versions of asynchronous motors

 3 Electromagnetic Actuators 121

N
1.0 rpm 4.0
I
A
0.8 3.2

0.6 2.4

0.4 1.6

0.2 0.8

0 0
200 400 600 800 T 1200
Ncm
T: torque, TN: nominal torque, N: speed, I: current, rr efficiency
Fig. 3.34. Performance curves of a running capacitor motor

where frequent switching is required (e.g. refrigerators, freezers). High ohmic

losses arise because of the high resistance of the auxiliary phase. Therefore
the auxiliary phase is switched on only for the start (Fig. 3.33g). Figure 3.34
shows the operation curves of a double-phase double-pole 175 W motor with
a 10 J.l.F running-capacitor motor. Table 3.10 gives the characteristic data of
single-phase induction motors.

Table 3.10. Typical data of single-phase induction motors

Number Nominal Nominal Nominal Effi- Running Weight
of poles speed continuous output ciency capaci-
torque power tor
rpm Nm w % J-iF kg
Capacity 2 2600 0.184 50 53 3 2.2
motors 2 2750 0.347 100 58 6 3.5
2 2790 1.71 500 67 12 8.0
4 1300 0.367 50 52 3 2.3
4 1370 0.697 100 55 4 4.5
4 1380 3.46 500 65 16 10.5

Shaded- 2 2600 0.037 10 22 1.0

pole 2 2650 0.18 50 32 3.5
motors 4 1350 0.071 10 20 1.2
4 1359 0.354 50 30 3.7
122 Hans-Dieter Sti:ilting

Control Methods
To change the rotation direction of single-phase induction motors either the
capacitor can be switched over from one phase to the other (Fig. 3.32k, 1)
or the current direction of one phase can be changed (Fig. 3.32m). The first
possibility is more cost-effective because only a one-pole switch is necessary.
But both phases have to be alike if the motor is to produce the same torques in
both directions. Often single-phase motors are made with two different phases
because such motors are better to be balanced, so that their performance is
like that of three-phase AC motors. Now the second switching possibility
is required. In contrast to three-phase AC motors the direction of rotation
cannot be changed while running in every case. This depends on the load-
torque type. Sometimes it is possible that the rotor rotates on in the old
direction. Therefore the motor should be stopped before the new direction
is switched on. Besides them the main phase should be switched of too.
Otherwise the motor rotates further.
The speed can be regulated by the following methods:
- The phase windings consist of two parts which can be connected in parallel
or in series with another or with the capacitor in different ways. For that
a complex switch is necessary and high losses are generated. For instance
this method is used in circulating pumps of heating systems. Because of
the constant synchronous speed the speed range is small.
- The terminal voltage can be varied by transformers or by phase-angle
control. Dependent on the square of the voltage the torque drops over-
proportionally with sinking voltage and the losses increase, especially if
phase-angle control is used because the voltage is no longer sinusoidal
and possesses large harmonics. Requiring an overdimensioned motor this
electronic solid-state speed regulator is cost-effective. Also in this case the
speed range is small because of the constant synchronous speed. To widen
the speed range high-resistance rotors, having a flatter speed-torque char-
acteristic, are employed (Fig. 3.35). Simultaneously the starting torques
becomes greater. Due to the intensive cooling and due to the convenient
speed-torque characteristic these methods are used for fans because the
motors are in the air flow, and for pumps with rotors turning in the
pumped liquid. This construction is cost-effective because a special seal
between pump and motor can be dropped. The stator is protected from
moisture by a very thin nonmagnetic tube in the air gap (canned or split-
cage motor).
- The change of synchronous speed can be achieved with two separate stator
windings each with another number of poles. The winding with the lower
number of poles can also be single-phase. The motor is started with the
low-speed winding before it is switched over to the single-phase winding.
Sometimes only one capacitor is enough for both windings.
- The most effective method to vary the synchronous speed can be done
by frequency variation with an electronic power converter. As the torque-
 3 Electromagnetic Actuators 123

Speed N
regulation range

Fig. 3.35. AC motor, speed regulation by voltage variation

generating flux is about proportional to the voltage/frequency ratio, the

terminal voltage has to be varied to the same extent as the frequency.
In this way the torque and the utilization of the motor are constant
(constant-flux operation). Starting at a low frequency and at low volt-
age the motor already supplies its nominal torque. Reaching the nominal
voltage with rising speed, no more voltage improvement is possible. Be-
yond it the flux becomes lower. As a consequence of it the maximal torque
becomes lower too. As long as the nominal current is not exceeded, this
so-called field-weakening operation can be used. In this way, speeds are
possible more than twice as great as the nominal speed at 50 Hz opera-

I
1- Constant-flux NN
range
f<fN, lf;<lf;N
Field-weakening
--range-
f>fN, v; =v;N
IN

Fig. 3.36. AC motor, performance range by frequency variation

124 Hans-Dieter Stolting

tion. The breakdown torque of larger motors is constant; the breakdown

torque of small motors becomes smaller with decreasing speed caused by
the influence of the stator resistance which is not negligible in the case of
small motors. This voltage drop can be compensated by a start voltage
greater than null. Only three-phase motors and no capacitor-start mo-
tors are employed because of their worse characteristics. On account of
the high electronic cost this method is still seldom used for high-grade
applications of lower-power drives.

Shaded-pole Motor

The shade-pole motor (split-phase motor) makes use of a further possibility

to achieve phase displacement. The auxiliary winding is short-circuited. Con-
nected with the AC line the main winding induces in this winding a voltage
like a transformer with a short-circuited secondary winding. The auxiliary
current excites a flux lagging to the main flux. These two fluxes generate an
elliptical field (very small ellipse). Therefore a very high oscillating torque
superposing the mean torque generates high losses. This is the reason why
all shaded-pole motors possess bad efficiency. Figure 3.37 shows two con-
structions of these motors. About a third of each pole is "split off'' and
surrounded by the short circuited winding (one to three turns). The coils of
the main winding are wound first and slid on the yoke (Fig. 3.37 left). Then
the yoke is moved into the stator packet. The rotor has the usual aluminium
squirrel-cage winding.
Because of the small phase of displacement of the currents in the main
and the auxiliary winding and because of the winding misalignment lower
than 7r /2 the start torque is very low. In addition, there is a saddle within

Main winding Short circuited winding Main winding

Squirrel-
cage
rotor
Single-axle Two-axle

Symmetrical cross-section

Fig. 3.37. Designs of shaded-pole motors

 --
3 Electromagnetic Actuators 125

3 1500 r--
\ !]_

----
I N lf;j- %
A
2.5
rpm
: ---....... \

2 1000
) 100
: (J
1.5 / 80

500
I
( 60

)I 40

0.5
----- rr ......
\ 20

0 0
..........
......
-- -- /

0
20 T 60
Ncm
T: torque, TN: nominal torque, N: speed, J: current, r-,: efficiency

Fig. 3.38. Performance curves of a shaded-pole motor

the speed-torque curve at about a third of the synchronous speed (Fig. 3.38).
Sometimes this causes running-up problems. Therefore shaded-pole motors
are especially suited for small fans and pumps. Other applications are juice
presses, clothe driers, grills, simple butterfly control waves, massage appara-
tus, hot-air stoves and cabinet fans. Drives for reversing duties can be built
with two motors assembled homologously. Shaded-pole motors are low-cost
motors. Because of their low efficiency they mostly need intensive cooling. To-
day the importance of these motors is diminishing. They are being replaced
increasingly by permanent-magnet motors like single-phase synchronous mo-
tors (next section) or simple BLDC motors (Sect. 3.3.1).

3.4.2 Synchronous Motors

Introduction

In synchronous motors the rotating field is generated in the same way as in

asynchronous motors. Contrary to them whose rotors rotate slower than the
field, the speed of synchronous motors is equal to the field rotational speed.
Therefore the speed is strongly proportional to the supply frequency. Speed
variation is not possible with the help of voltage variation. Small motors are
used in open-loop control techniques. Because of their simple construction
and relatively high efficiency, if they have permanent-magnet rotors, they are
used in household appliances too: program controllers, clocks, time-delay re-
lays, writing measuring instruments, pulse recorders, studio camcorders, gyro
126 Hans-Dieter Stolting

drives, slot machines, fans, pumps (washing machines, dishwashers, aquari-

ums), lemon squeezers, tin openers, recorder, answering apparatus etc.
Like AC motors the rotating field pulsates during the rotation. This re-
sults in huntings which entail disagreeable pitch pulsations if this is not com-
pensated by a high moment of inertia. Otherwise there are for instance loud
sound vibrations of audio apparatus. The rotor tries to follow the rotating
field. At no-load the pole axis of the rotor and field are in phase. The more
the load rises the more the angle between these axes (load angle) becomes
greater and greater, whereat the rotor is running after the field. If the load
becomes greater than the breakdown torque the rotor stands still. Added to
that the rotor swings at load variations (a special problem of stepper motors:
see Sect. 3.4.3). This property is not so marked with AC motors because
their squirrel-cage rotor has a damping effect. But the main problem of syn-
chronous motors is the start. The rotor does not reach the synchronous speed
immediately after it is switched on. For a certain running up there are several
possibilities which depend on the rotor construction demonstrated in the fol-
lowing sections. Small synchronous motors have no rotor coils and slip rings,

Table 3.11. Typical data of single-phase synchronous motors (230 V)

Number Nominal Nominal Nominal Efficiency Weight
of poles speed continu- output
ous power
torque
rpm Nm w % kg

Reluctance 2 3000 0.16 50 40 2.5

motors 2 3000 0.32 100 55 4.0
4 1500 0.32 50 40 3.0
4 1500 0.62 100 55 4.2

Hysteresis 2 3000 0.003 1 5 1.0

motors 4 1500 0.006 1 5 1.2
6 1000 0.012 1 5 1.5
12 500 0.024 1 5 2.0

Permanent- 8 750 0.013 1 30 0.2

magnet 8 750 0.064 5 35 0.6
motors 8 750 0.127 10 40 1.0
12 500 0.019 1 25 0.25
12 500 0.038 2 30 0.4
12 500 0.095 5 35 0.8
16 375 0.025 1 23 0.25
16 375 0.051 2 30 0.4
16 375 0.127 5 35 0.8
24 250 0.038 1 20 0.25
24 250 0.076 2 30 0.5
24 250 0.191 5 33 1.0
 3 Electromagnetic Actuators 127

but magnet, hysteresis or reluctance rotors. Table 3.11 shows typical data of
the different types of 230 V motors.

Permanent-magnet Rotor

The stator of larger synchronous motors (over 100 W) is constructed like that
of AC motors, i.e. they have slots and are wound double- or three-phase. Con-
nected with the AC line the same elements for phase displacement are taken:
capacitors or resistances. Larger motors have varying rotor constructions.
Figure 3.39 shows two examples with a squirrel-cage winding for running-up.
Coming near to the synchronous speed the rotor pulls into synchronism.
Such motors are called synchronized induction motors or hybrid synchronous
motors (Merril motor). Here the squirrel-cage winding affects damping too.
If the load becomes greater than the nominal torque and finally exceeds
the synchronous breakdown torque, the rotor pulls out of synchronism and
runs further on, until it exceeds the resultant breakdown torque too. The
quasi-steady operation is illustrated in Fig. 3.40. Larger synchronous motors
without a squirrel-cage winding can be run up only by an electronic control
whose frequency arises gradually from near null up to the nominal frequency.
Smaller motors with a small number of poles are constructed like shaded-
pole asynchronous motors with a short circuited coil or without it. They
have a permanent-magnet solid or hollow cylinder as rotor. Therefore they
do not start right off. After switching on the AC voltage the rotor begins
to swing and then it pulls into synchronism within a few periods. Therefore
the moment of inertia has to be low and it is advantageous if the load is
not solid-coupled. Often the gear backlash is sufficient. To support the start
swinging the poles are asymmetrically constructed. Here the poles have steps;
in another case the air gap becomes slowly wider and wider from one pole

Fig. 3.39. Synchronized induction motors

128 Hans-Dieter Stolting

Asyncronous torque
T
,..---.....::- - - Breakdown
torques

.I
N

Syncronous torque

Fig. 3.40. Synchronized induction motors, speed-torque curves

side to the other. The rotor takes different positions, when the current is
switched off and switched on. Figure 3.41 shows a typical design of a double-
pole one-phase motor. The motor in Fig. 3.41 is of great importance because
of its favorable cost and efficiency and has replaced shaded-pole asynchronous
motors as drives for small pumps (see above) and juice-squeezers. At start
the direction of rotation is unknown. In many cases this is not problematic.
To rotate only in one direction a backstop can be considered.
For control purposes often a low speed is desired. In this case the motor
gets a claw-pole stator, as shown in Fig. 3.42. A simple ring coil is mutually
enclosed by metal clips. All metal components are only punched, bent and
welded or clamped together. This is a very cost-effective manufacturing pro-
cess for getting a multipole motor. A disadvantage is that a large part of the
flux does not enter the rotor generating a torque, but as stray flux it goes

Fig. 3.41. Pump driven by a single-phase synchronous canned motor

 3 Electromagnetic Actuators 129

Ring coil

Fig. 3.42. Claw-pole principle

over directly from pole to pole. Besides them eddy currents are caused by the
alternating field in the relatively thick sheet metal generating high losses.
When the alternating field becomes a rotating one, short-circuited rings
are built in like those of shaded-pole motors (Sect. 3.4.1). In Fig. 3.43 only
the stator of an external-rotor motor is presented in this account to give

Fig. 3.43. External-rotor motor, design with claw-pole and shaded-pole principle
130 Hans-Dieter Sti:ilting

System 1

System 2

Fig. 3.44. AC synchronous motor with two claw-pole systems

a better survey. The short-circuited rings influence the flux of the main and
the auxiliary claws to a different degree. Instead of the rings often cupper
plates are arranged at the front side of the motor. To support the start or
the running in a decided direction the claws are formed asymmetrically.
The rotating field of single-phase motors pulsates very considerably, so
that the mean torque is only low and a distinct noise is possibly generated. It
is more favorable if two claw-pole motors are built together. Their stators are
twisted with respect to each other by a half pole pitch and their rotor poles
are in line (Fig. 3.44). One of the stator windings is directly connected with
the AC line, the other in series with a capacitor or resistance (Fig. 3.32h or
3.32i). Should the speed direction be changed the capacitor is changed over
at the other phase (Fig. 3.32e). Likewise three claw-pole motors are built to-
gether with stators distorted like the phases of three-phase motors. They are
directly connected with the three-phase system. Also two smaller motors with
aligned stators can be built together to get a more powerful motor. The essen-
tial advantage of synchronous motors with permanent-magnet rotors is their
multiple greater torques compared with the other synchronous-motor types.
Speed regulation of large motors can only done by pole or frequency changing,
the latter for instance by pulse-width modulation (PWM; see Sect. 3.6.5).

Hysteresis Motor

The stator of hysteresis motors is like that of a magnet-rotor motor. The rotor
is a cylinder with a layer of hysteretic material. It is similar to magnetic ma-
terial, but nonmagnetic in normal conditions and possesses no poles. During
running-up the magnetization of the hysteretic layer is permanently reversed.
By that a torque is produced which accelerates the rotor up to synchronism.
In this state poles arise. Now the hysteresis motor works like a magnet mo-
 3 Electromagnetic Actuators 131

tor, but with an about 20-30 times smaller torque. Switching off the motor
the magnetizing disappears. Its advantage is its self-supporting start and the
soft change-over from asynchronism to synchronism (Fig. 3.34). The eddy
currents in the rotor produce a torque additionally to the hysteresis torque.
Hysteresis motors are only seldom used today because of their low torque.

Reluctance Motor

The magnetic resistance (reluctivity) of reluctance motors varies along the

circumference corresponding to the pole number (anisotropic rotor). The ro-
tor tries to align itself with the rotating field, so that the reluctance of the
magnetic circuit becomes a minimum. That happens in synchronism. For
that the rotor of larger motors has a squirrel-cage winding like that of AC
induction motors. Figure 3.46 shows two construction possibilities.
The left rotor is an AC motor rotor in which four poles are milled. The
right rotor possesses non-magnetic spacers which force the flux on decided
paths. The slots, pole gaps and spacers are filled with aluminium. The pro-
duction of the left motor is cheaper, the right motor generates about double
the torque.
Very small multi-pole motors have a sheet-cylinder in which as many
windows as desired poles are punched. So an extremely low rotational speed
is given. The stator is of the claw-pole type like those of magnet motors.
Other rotors are formed like gear wheels with a tooth number like the
stator pole number. The stator has two or three phase windings. These
small motors generate a start torque by eddy currents. Contrary to hys-
teresis motors which pull in synchronism softly, this takes jerkily in re-
luctance motors. Pulling out of synchronism they continue running asyn-
chronously if the load torque is lower than the asynchronous breakdown
torque (Fig. 3.47). Today applications of reluctance motors are unimpor-
tant.

d ·······:r.·:r······
E ••••••••••••• •••• .....,.---Ideal
··········r························::::., motor
TH i Real

Ns-;r
TE: eddy-current torque, TH: hysteresis torque

Fig. 3.45. Hysteresis motor, speed-torque curve

132 Hans-Dieter Sti:ilting

Flux path

Fig. 3.46. Reluctance motors with squirrel-cage windings

---------Asynchronous
t
T
breakdown torque
-Synchronous
-Pull-in torque
I
I
I
I
I

N---
Fig. 3.47. Reluctance motor, speed-torque curve

3.4.3 Stepper Motors

For many positioning tasks step by step rotating motors are required instead
of motors with a continual motion [6]. Contrary to BLDC motors with rotor
position encoder, stepper motors are the cheaper option because they are
driven in open-loop control. There is no check-back signal whether the rotor
has really reached the given position. Therefore it must be sure that un-
der every admissible condition the motor always runs the desired number of
steps. So for a stepper motor it is characteristic that on every electrical pulse
always exactly one mechanical pulse follows. As BLDC motors, the current of
stepping motors is electronically commutated too, but with a frequency given
by the electronic control. Therefore stepping motors belong to the class of
synchronous motors and possess all the characteristics which are typical for
this motor type, like starting problems, pull out at overload and oscillations
under sudden load variations. These characteristics authoritatively influence
the design of a stepper drive because they evidently disturb the observance of
the above-mentioned condition "mechanical equal electronic pulse number"
or even prevent it. In the latter case BLDC motors have to be used.
 3 Electromagnetic Actuators 133

Fig. 3.48. Fundamental constructions of stepper motors

Stepper motors are constructed in principle like BLDC motors. The stator
has a winding made of concentrated coils on distinct poles (Fig. 3.48) or has
a ring coil in a claw-pole system (Fig. 3.44). The rotor is a permanent-magnet
cylinder (permanent-magnet (PM) stepper motor), a gear wheel (variable
reluctance (VR) stepper motor) or a combination of both types (hybrid
(HY) stepper motor). In order to rotate in the desired direction, clockwise
or counter clockwise, the motor consists of two submotors (PM, HY) or of
three submotors (VR), in the following, called "systems". The systems can
be arranged in a plane side by side (Fig. 3.49) or axially one after another

Laminated stator core

Axially magnetized
permanent magnet

System 1

South pole ring gear

Fig. 3.49. Hybrid stepper motor

134 Hans-Dieter Stolting

(Fig. 3.44). Either the system-stator::; are twisted towards each other or the
rotors by half a pole pitch. The other elements are in line.
Contrary to the VR motor which has to be controlled three-phase, PM and
HY motors need only a double-pha::;e control. The latter both have a higher
torque and the advantage that they produce a currentless torque if the sta-
tionary rotor is displaced by the load. This torque is similar to the cogging
torque of BLDC motors. There it is not desired, but here it is advantageous
because the rotor is fixed, when the stator winding::; are unexcited. This de-
tent torque should be not greater than 15% of the maximum steady torque,
when the windings are excited (holding torque). Otherwise large pulse o::;cil-
lations arise. The disadvantage of the PM motor contrary to the HY motor is
that not so many rotor poles can be magnetized along the circumference as
the latter has teeth. Table 3.12 gives characteristic data for the most essential
designs.
The ampere-turns of the systems change, alternating with every pulse.
If in Figs. 3.50a, 3.50b voltage is applied to the systems one after another,
the rotor moves on at 90° and makes four steps every turn. After a step
is finished it stands in the pole axis of the just-excited system. The torque
can be enlarged if two systems are excited simultaneously (Figs. 3.50c, 3.50d).
The number of steps at one turn is four too, but the rotor stops in the middle
of the pole axis of the two excited windings. The rotor makes a full step in
every case; it is working in "full-step mode". If alternately one, respectively,
two sy::;tems are excited in the sequence a, c, b, d, etc., the step angle is
halved and the number of steps doubled. In this case the motor is working in
"half-step mode". The number of steps is given by

2pm
Z=--
ks
where
f line frequency
2p number of poles of the aligning motor part
m number of systems
ks operation-mode factor (full-step operation: k 8 1·
half-step operation: ks = 0.5). '
The number of steps yields the step angle

360°
a=--.
z
Figures 3.50e and 3.50f show the alternating control of the ::;ystems of a VR
motor; Figs. 3.50g and 3.50h the alternating control of a HY motor.
Half-step operation is a possibility to diminish the step angle, but it has
the disadvantage that the torque is different step by step. To avoid this,
expensive current control is required. Therefore mostly full-step operation is
used like Figs. 3.50c, 3.50d, where the systems are reversed alternating. In
 3 Electromagnetic Actuators 135

Table 3.12. Characteristic data of stepper motors

Number Step Holding Moment of Electrical Weight
of steps angle torque inertia time
(rotor) constant
deg Nm gcm 2 ms kg
Claw pole 24 15.0 0.05 3 0.7 0.09
motor 24 15.0 0.11 30 1.6 0.3
24 15.0 0.26 60 1.6 0.55
32 11.25 0.011 3 0.7 0.08
32 11.25 0.115 30 1.6 0.3
32 11.25 0.28 100 2.0 0.43
40 9.0 0.012 3 0.7 0.08
40 9.0 0.125 30 1.6 0.3
40 9.0 0.23 60 1.6 0.55
48 7.5 0.013 3 0.7 0.08
48 7.5 0.13 36 1.4 0.3
48 7.5 0.4 120 2.0 0.43

Hybrid 100 3.6 0.114 10 0.2

motor 100 3.6 0.26 18 0.21
200 1.8 0.18 52 0.34
200 1.8 0.48 120 0.53
200 1.8 1.0 330 1.0
200 1.8 5.5 2800 4.6
200 1.8 9.5 5500 7.25
200 1.8 15.0 8300 10.5

Hybrid 500 0.72 0.25 100 3.0 0.5

motor or or 0.85 400 8.0 1.3
with 5 1000 0.36 3.8 1800 10.0 3.5
systems 11.0 11500 20.0 12.5

PM motor 40 9.0 0.011 1 0.6 0.04

with disc 100 3.6 0.19 12 2.5 0.25
rotor 200 1.8 0.57 100 3.7 0.6

this case the motor is best utilized. With a still more expensive control the
ampere-turns of the systems can be changed in very small steps to move the
rotor in very small steps too. This mode is called "micro-step operation" .
This method has the additional advantage that the rotor reaches the next
latched position without swinging.
The step-number equation points at two further possibilities for dimin-
ishing the step angle:
An increase of the system number m follows a corresponding increase of
control-phase number raising the cost. Therefore mostly motors with two
systems are built. Seldom are there motors with up to five systems.
136 Hans-Dieter Stolting

b d f h
Fig. 3.50. Control principles of stepper motors

- It is more cost-effective to increase the number of poles or teeth. In this

case more exact manufacturing is required. Otherwise the step-angle error
becomes too great. This error means the deviation of the desired position.
Because of the cost, stepper motors are often wired up unipolar, more
seldom bipolar. In order to follow the field during the performance of several
steps the control frequency is not allowed to exceed the so-called maximum
start-stop stepping rate. The corresponding curve is given dependent on the
load-moment of inertia in data sheets of stepper motors. To realize high
stepping rates, e.g. to minimize the positioning time the following methods
are given:

A resistance is connected (constant-voltage drive) in series with each

phase winding. Disadvantages are additional losses and a more power-
ful output amplifier.
- The windings are applied to a high voltage which is modulated so that
the mean current becomes the permissible value (constant current drive
or chopper drive).
- An additional voltage is superposed on the nominal voltage as long as the
nominal current is reached. Then the voltage becomes the nominal value
(bilevel drive).

The dying out of the rotor, after the motor is switched off, lengthens the
positioning time, so that dumping is indispensable (apart from the micro-step
drive). Today this happens generally by specificly placed pulses.
The most cost-effective stepper motor is the PM motor built as a claw-pole
type. It is constructed with two systems like the corresponding synchronous
motor (Fig. 3.44). Both stator systems are displaced to one another by about
half a pole pitch. Mostly each system has two ring coils where the currents
 3 Electromagnetic Actuators 137

flow, alternating in one or other direction (unipolar connection). This connec-

tion needs a simple, i.e. cheaper, electronic control, but the motor is larger
because of the double number of windings. Occasionally each system has
only one coil, where the current direction changes alternately (bipolar con-
nection). In this case the electronics are more expensive, but the utilization
of the motor is better. The polarization of the ferrite-magnet rotor is con-
stant along the axis. Claw-pole stepper motors are produced with step angle
greater than or equal to 7.5 o. The great eddy-current losses and the stray
fluxes are the reasons for the low utilization of these motors. Drives with an
exceptionally low moment of inertia are constructed as PM disc motors, but
they are expensive especially because of the complicated rotor construction.
If small step angles are demanded without using gearing mostly HY mo-
tors come into question (Fig. 3.49). They have the advantage of the VR
motors, the small step angle, with the advantage of PM motors, the large
torque, but they are more expensive than claw-pole motors. The rotor has
two ring gears which are displaced to one another by about half of a tooth
pitch. Between the rings there is a rare-earth magnet which is magnetized
axially. The stator is laminated and generally has two systems in a plane.
Two opposite lying poles of a double pole motor belong to one system. The
poles are toothed too with the same tooth pitch as the rotor, but the teeth
of the both systems are displaced to each other by about half a tooth pitch.
Today bipolar connection is used mostly. In exceptional cases motors of this
type can have up to five systems with step angles down to 0.36°. But these
are very expensive actuators.
All of these motors can be designed as linear motors. Drives with such
motors are very expensive. Therefore mostly they are produced with a rotat-
ing motor whereby the revolving motion is formed into a linear motion by
gearing (feed screws, gear racks, toothed belts). The advantage of linear mo-
tors is that they directly actuate the working element and the disadvantages
of gearings (losses, play, noise) are avoided. But either the stator or slider has
to be as long as the working range and the advantage of gearing (enlarging
the driving force respectively minimizing the load-moment of inertia) is not
given.
PM-stepper motors with only one system like the one-winding-phase syn-
chronous motor (Sect. 3.4.2, Fig. 3.41) are used as clock drives (Fig. 3.66).
Figure 3.51 shows the principle of construction with unsymmetrical poles. If
the motor is switched off the pole axis may be situated in the position of
Fig. 3.51a. If the winding is supplied to voltage the current should generate
a flux pulse in the here drawn direction. On that the rotor rotates clockwise
and stops at the painted position in figure (b) after the current pulse dis-
appears. The next opposite directed pulse causes a further step of 180°. If
the rotor were to start in position (a) by a flux pulse like figure (b) it would
move about a small angle counter-clockwise and then fall back to the start
position. The next pulse is opposite to the first one and therefore the right
 -
138 Hans-Dieter StOlting

-
I I

a • +/ b -I

Fig. 3.51. Single-phase stepper motor

starting pulse is given. All clockwork motors are 180°-stepper motors rotat-
ing in only one direction. Because of that they are hardly used for any other
purpose than clock drives. After a new battery was put into a quartz clock,
it is possible that first a false pulse appears, so that the motor loses one step.
However, the next pulse is the right one. This trouble makes of the exactness
of the clock no worse.

3.5 Limited Motion Drives

3.5.1 Introduction

Many apparatus move only with a short unique or oscillating rotational or

linear motion. Then rotating motors with eccentric movements, rod drives
camshafts and tappets can come into question in individual cases. But today
they are used by way of exception because of the noise, the wear and the
cost. Actuators principally possessing the required motion are better suited.
They use the following electromagnetic principles of force generation:
- Forces at boundary layers of materials with different permeability (Max-
well force).
- Force on current-carrying conductors in a magnetic field (Lorentz force).

3.5.2 Electromagnets

Introduction

Electromagnets operate with Maxwell forces [10]. They have a fixed element
(magnetic core, body or yoke) with a coil (excitation winding) and a movable
element (armature). The magnetic core and the armature are made of ferro-
magnetic material for flux carrying. Switching on the current the armature is
attracted, switching off it falls back to its free position. There are character-
izing differences between direct and alternating current magnets influencing
essentially their applications. Many applications use these drives whereby
three magnet types are distinguished, solenoid and hammering actuators,
turning and swinging magnets:
 3 Electromagnetic Actuators 139

Solenoid actuators: flaps, pneumatic and hydraulic valves, slide valves,

interlocks, brakes.
Hammering actuators: riveting, punching, stumping, chiselling machines.
Thrning magnets: throttle valves, control valves (hydraulic, pneumatic),
material support (for instance web of cloth, paper), turnout in transport
plants.
Swinging magnets, vibrators: electric razors, massage apparatus, piston
pumps, swing pumps, diaphragm pumps, small compressors, swing saws,
vibrating hauling plants, oscillating sieves, oscillating tables, swinging and
helical conveyors.

DC Magnets

The magnet body and the armature are made of solid iron and are very ro-
bust. Both components can be formed so that an appliance-matched torque-
per-way characteristic F(s) results. Figure 3.52 shows fundamental design
possibilities of solenoid and drag magnets serving for adjusting and position-
ing with their associated F(s)-curves. The construction (c) is a combination
of (a) and (b). Hammering magnets are designed similarly. But there the
electric energy is converted into kinetic energy generating mechanical pulses
for hammering, riveting, etc. The armature is reset at de-excitation by grav-
ity or springs. The magnetic forces are between lOmN and lOkN and the lift
amounts to some millimeters (small magnets) up to 20 centimeters (greater
magnets).

E A
a: magnetic core, b: armature, c: winding, d: nonmagnetic limit stop, A: start
position, E: end position
F: force, s: way
Fig. 3.52. Principles of DC magnets and their force-way curves
140 Hans-Dieter Stolting

Besides translatory operating magnets, often rotational working magnets

are demanded. Their rotation can be made in different ways:
Moving magnets with axial air gaps are fundamentally constructed as
illustrated in Fig. 3.52. The lifting movement can be changed in a rotating
movement in several ways, for instance by an inclined ball race (Fig. 3.53),
so the shaft makes a small axial motion. The compressive load per unit
area of the balls causes greater wear. An advantage is the large angle of
rotation. Often an axial motion is undesired. In this case the armature
can be formed as a locked female screw with steep-lead-angle thread (the
shaft acts as a screw fixed for rotation). With this construction essential
large angles of rotation can be reached, but with very great wear. Another
type with a low rate of wear and without a linear motion is built of an
armature locked against longitudinal motion which is fixed with the shaft
and a core within the coil. Both of the components have helix shaped faces.
If the coil is currentless they form a large air gap. As soon as a magnetic
field is excited the armature tries to diminish the air gap by rotation.
Figure 3.54 shows this function with the aid of the unwound armature-
core principle. The production of these magnets is expensive because of
the need for milling the faces.
Solenoids with a radial air gap are rare because they generate only low
start and end forces and the angle of rotation is limited up to a maximal
65°. Figure 3.55 shows the general designs. The magnetic circuit consists
of the housing, one- or double-sided pole systems with the convenient
armature and the internal tube serving as magnetic return path. Pole-
systems and armature can be milled or pressed (porous metal). The reset
ensues by springs whose torque diminishes the working torque. For re-

Fig. 3.53. Rotating DC magnet with an inclined ball race

Start position End position

Core

Fig. 3.54. Rotating DC magnet with helix shaped faces of core and armature
 3 Electromagnetic Actuators 141

Magnetic core

One-sided

pole system

Double sided

Coil Return path

Fig. 3.55. Rotating DC magnet with radial air gap

versing duty two solenoids are mounted, mirrored and excited according
to the motion direction. DC solenoids reach torques of about 1 Nm.
Generally the magnets are connected by rectifiers to an AC system. Excit-
ing the magnet coil is possible by a capacitor discharge. Should the magnet
react especially fast after it is switched on, the electric time constant can be
diminished for instance by a series resistance. But through that, additional
losses arise which require a more powerful supply. A further possibility to
diminish the pickup time of the armature is to switch it on at a higher volt-
age (over-excitation). But it must be dropped down to the nominal voltage
at once because the current increases to an unpermissible value. The voltage
change-over can be done by the magnet it selves time- or current-dependent
(bilevel operation).

AC Magnets

Because of the pulsating field in the magnet core and in the armature iron
losses arise, all flux-carrying parts of AC magnets have to be laminated. This
is done only for larger magnets, but not for small magnets because of the
cost. In the first case the switching time is shorter than that of DC magnets.
The magnetic force pulses with the double the system frequency entailing
a humming noise. To get a small retention force in the end position AC
magnets always have a short-circuited winding in which a current is generated
inductively. This results in a second displaced force so that the resulting force
is greater than null every time.
142 Hans-Dieter Sti:ilting

Magnetic
core

Winding

Short circuited winding

a b
t
F

\
Armature Start postion
c
F: force, s: way
Fig. 3.56. Principles of AC magnets and their force-way curves

There are lots of design possibilities to realize different force-per-lift char-

acteristics. Figure 3.56 shows three fundamental designs with the correspond-
ing F(s)-curves. Construction (c) yields the best utilization because of the
very low stray flux. AC magnets reach forces between 1 and 150 N with a lift
between a few millimeters and 10 centimeters. Larger magnet forces are gen-
erally associated with small lifts. Three-phase magnets are built with the
principle of Fig. 3.56c whereby the windings are fitted on three limbs. These
magnets produce forces between 50 and 1000 N with lifts of 20-60 mm. There
are few AC rotating magnets. Sometimes the rotary motion of a clapper-
armature magnet is used (Fig. 3.56b).

Swinging-armature Magnets, Vibrators

A to-and fro-motion of an armature is possible through when the magnet-

force generating current is switched on and off at the favorable moment or
its direction is changed. Drives with only one coil need a spring for falling-
back into the start position. If the armature swings beyond the position of
minimum energy a restoring force arises. The spring and the armature force
of inertia must be tuned so that the current switch-frequency is equal to the
natural frequency of the system. In drives with two coils the armature-swing
affects the alternating switch-on of both coils. By corresponding tuning the
resonance-effect can be obtained here too. This is the assumption for optimal
efficiency.
 +
3 Electromagnetic Actuators 143

.........

Fig. 3.57. Swinging-armature magnets

Just as in the case of rotating motors, swinging-armature motors can

be subdivided in self-commutated and externally commutated motors with
similar properties as rotating motors (Sect. 3.1.2). In self-commutated drives
the armature switches themselves the one winding on and off respectively
to-and-fro between the two windings. The swing frequency is free eligible
because it depends on the motor design and the load. The switch arc is the
reason for switch wear and the limited life time. Also, these drives are loud.
Electronic commutation is quiet and robust, but more expensive because it
needs an armature position sensor and electronic control.
External commutated drives can be driven with alternating or direct cur-
rent. The first method is the cheapest one, the latter current must be switched
by more expensive electronics. These actuators are especially robust and
quiet. In every case the motor operates according to the synchronous-motor
principle, i.e. the switching rate is fixed. If the armature consists only of
iron the armature swings with double the system frequency (reluctance ar-
mature). If it is premagnetized it swings with the single frequency (magnet
armature).
Swinging-armature motors have magnet bodies and armatures which are
coupled together not mechanically, but magnetically. Only the armature
oscillates, whilst the magnet body stands still. All designs represented in
Figs. 3.52, 3.55 and 3.56 can be built as swinging drives. Figure 3.57 shows
widespread constructions. For instance they are used as drives in electric
razors.

Fig. 3.58. Vibrator

144 Hans-Dieter Stolting

Vibrators are built similarly, but the magnet body and armature are con-
nected to one another by tuned springs, so that both elements are able to
swing (Fig. 3.58). If vibrators are fixed to conveying systems, so that the
swinging motion is directed upward-sloping (throw vibrator) goods can be
transported upwards. The same principle is used in winding-conveying appa-
ratus for the supply of components in automatic production lines.

3.5.3 Voice-coil Motors

Voice-coil motors (plunger electromagnet) are so called because they oper-

ate like electrodynamicalloudspeakers [8]. Here the force on current-carrying
conductors in a magnetic field is used. This force is proportional to the mag-
netic flux and the conductor current. Therefore the deflection of an elastic
fixed coil can be controlled by the current. By changing the current direction
the motion direction can be reversed.
Figure 3.59 shows two fundamental construction variants. In (a) a coil
is arranged coaxially to the magnet body. Instead of arranging within the
magnet the coil can lie outside of the magnet. Cylinder or box coils are used.
Part (b) shows a fiat building construction with a disc coil. As the design of
the coil's bearing and guide is difficulty to produce, the coil is often fixed and
the magnet movable arranged. But this construction often possesses a greater
force of inertia. In every case it is to be decided whether the magnet or the coil
should have the motion length. Both constructions have special advantages.

Fig. 3.59. Principle of construction of voice-coil motors

3.6 Closed and Open Loop Control Methods

3.6.1 Introduction

An overview of regulating and controlling methods of fractional horsepower

motors is given in Fig. 3.3 [9]. These methods are similar to those which are
known from larger drives. However, the type and the expense of the circuit
determine the cost. Therefore power semiconductors should be employed as
little as possible and no regenerating braking is used. Moreover, there are
circuits which are unsuitable in the higher power range. Fractional horsepower
 3 Electromagnetic Actuators 145

motors are often driven in only one direction, so that the drive control must be
suitable only for one-quadrant operation. For reversible operation the current
direction has to be changed. In the range of low power mechanical switches
are often used. Should the rotor be stopped as fast as possible a mechanic
brake is chosen (e.g. spin drier). For instance in sunblind drives this is done
as follows. After the current is switched off a spring presses the rotor a bit
out of the stator against a disc brake. By switching on the current the rotor
is pulled into the stator and so the brake is lifted. Compared to electronic
brake circuits the advantage of a mechanic brake is that it works in the de-
energized state. If the motor is to work and brake in both rotational directions
a four-quadrant circuit is required.

3.6.2 Line-commutated Converter

Line-commutated converters are of use for DC commutator motors with

permanent-magnet stator. But they are also suitable for DC series motors
which are seldom used in the lower power range. Apart from the latter motors
the torque-generating current is practically constant. Therefore the torque is
proportional to the current. With the change of the load-torque the motor
current changes too if the speed is to be constant. The DC motor is driven
mostly at a half-controlled bridge. If it is supplied by AC voltage a recti-
fier is connected in series. For speed control the actual value is measured by
a tacho-generator and compared with the setpoint value. The analysis de-
livers the current setpoint. This value is compared with the instantaneous
current value. This yields the signal for the corresponding output voltage of
the converter (current-controlled speed limiting system).

3.6.3 DC Chopper Converter

At AC voltage supply DC chopper converter (DC chopper) own a line-

commutated rectifier which feeds a self-commutated converter. Figure 3.60
shows a simple converter for an electric hair-dryer with a voltage divider
(heating coil) for voltage matching. The transistor works as a switch. The

<>----"-..(_

220V-

0
h fd
H: heating coil, r: rectifier, fd: free-wheeling diode, c: control,
pm: permanent magnet

Fig. 3.60. DC chopper for a hair-drier (single-quadrant operation)

146 Hans-Dieter Stolting

Fig. 3.61. H-bridge (four-quadrant operation)

free-wheeling diode (FD) carries the current in the switch break. With the
H-bridge in Fig. 3.61 motors can be driven in four-quadrant mode. The cur-
rent course is marked by a solid line for one direction of rotation and the
additional course by a dotted line if the transistor T 1 is switched off. If the
other direction is desired the transistors T 2 and T 3 are switched on. The
speed control ensues in the same way, as mentioned in the case of a line-
commutated converter (see Sect. 2.5).

3.6.4 AC Chopper Converter

By AC chopper converters (phase-angle control) above all universal (AC com-

mutator) motors are driven, but seldom AC induction motors (see Sect. 3.4.1).
As a rectification of the alternating voltage is not necessary, this circuit is
simple and cheap. This is the reason why universal motors became the most
important drives of household appliances and handhold tools. Figure 3.62

Fig. 3.62. Phase-angle control of a universal motor

 3 Electromagnetic Actuators 147

shows the fundamental circuit and the curves of the motor voltage UM and
of the motor current iM. Dependent on the resistance-position the capacitor
charges itself. If its voltage reaches the gate-trigger voltage of the Diac the
capacitor discharges over the Diac and the Triac, so that the Triac is trig-
gered and the motor voltage is switched on. By variation of the resistance the
firing angle a is varied. P is the phase angle between the voltage and the cur-
rent. Today an IC regulates the firing angle a. The speed-control of motors
is done with help of tacho-generators. Often they are very simple. A ferrite-
magnet rotating with the motor shaft induces voltage pulses in a coil. The
pulse frequency is raised with help of a claw-pole system (see Sect. 3.4.2).
It improves the analysis in an evaluation-logic for tuning the firing angle. It
must be protected against disturbing pulses generated by the commutation
which are especially distinct in phase-angle control.

3.6.5 Converter for Polyphase Motors

For small asynchronous and synchronous motors only frequency converters

with constant link voltage (PWM) come into question today and in the future
if the cost accounts for this. With the mark-to-space ratio the amplitude
can be changed as the frequency of the output voltage. By this frequency-
proportional voltage matching is possible, the assumption for constant-flux
operation (Sect. 3.4.1). If it is desired the motor can start with the maximum
torque. Is really the speed regulation of rotating-field motors expensive, the
more it is the speed control. Therefore alternative actuator solutions with
BLDC motors may be more advantageous. This is to be decided in every
case.

3. 7 Applications

The radial blower in Fig. 3.63 is driven by s permanent-magnet-excited 12 V

DC motor. To get an overall length as small as possible the motor is built
within the blower using an internal rotor. This construction is mechanically
more advantageous for a commutator motor and offers more space for the fer-
rite magnets used here. If BLDC motors are used for blowers or fans external
rotors are more effective. In this case the fan baffies are mounted directly on
the rotor housing.
Many new actuators have been developed for car technology. Figure 3.64
shows a compact actuator for power-assisted steering. The driver's hand-force
is measured by a torque sensor. The electronics control the motor current,
so that the on steering-gear acting torque, detected by a second sensor, is
a multiple of the hand-force torque. The ratio of both torques can be reg-
ulated depending on the driver's wishes and on the driving speed among
other things. There are advantages compared with the hitherto used hydraulic
148 Hans-Dieter Sti:ilting
Blade Blower wheel

Fig. 3.63. Radial blower

Torque sensors

Fig. 3.64. Drive for a power-assisted steering (Bosch)

power-assisted steering relative to driving behavior , built-in and working ex-

penses.
As increasing car-electrification leads to a thicker and thicker wiring har-
ness, new techniques for carrying energy and signal fluxes are being devel-
oped. Bus systems should replace the supply cables, for instance, to control
electronic motors. Among other motors the wind-screen wiper BLDC motor
in Fig. 3.65 should replace the currently used DC motor if the cost accounts
for this. In this way, the speed regulation possibilities are much more ex-
tensive, e.g. to regulate the wiper speed corresponding to the rain intensity.
This proves right for many other auxiliary actuators in cars, e.g. for the
 3 Electromagnetic Actuators 149

Diaphragm
for pressure
compensation

Stator
core

Fig. 3.65. Drive for a wind-screen wiper (Bosch)

cooler-blower motor for the same advantages. In the future the starter motor
and the dynamo will be combined into one machine (starter-generator). This
will happen together with the introduction of a second vehicle-network of
42 V besides of the 12 V network. All the large current-users, e.g. the greater
motors, will be connected to the 42 V system. But today there are many
problems, especially a la rger and therefore much more expensive battery and
electronics.
In the clockwork of Fig. 3.66 similar single-phase synchronous machines
are built in, namely in a double manner as stepper motor and as generator.
These stepper motors are typical for quartz-controlled drives of analogous
clocks (see Sect. 3.4.3, Fig. 3.51). The same construction is chosen for the

Electronic

Generator

Fig. 3.66. Clockwork with stepper motor and generator (Seiko)

150 Hans-Dieter Stolting

generator. It supplies a battery driven by an imbalance. It rotates up to

lOO,OOOrpm.
In Fig. 3.67 two hybrid stepper motors drive the slide of a coordinate
table with a spindle gearing in the x - and y-direction. Today such drives are
driven more and more by linear drives because they are more compact and
robust avoid the disadvantages of gearing (Sect. 3.4.3).
The light-gantry robot in Fig. 3.68 is suitable for positioning, mounting
and handling tasks in laboratories and of precision mechanics. Its head can
pick different tools (soldering apparatus, grabs, etc.). It is driven by three
five-phase stepper motors three-dimensionally with a resolution of 0.1 mm
and with a maximum speed of 0.7mjs. The programming of the head- and
tool-control is made menu-assisted on the screen of a personal computer.

Screw spindle

Fig. 3.67. Coordinate table (Phytron)

Fig. 3.68. Light-gantry robot (Berger)

 3 Electromagnetic Actuators 151

Figure 3.69 shows a typical example of a voice-coil drive. It produces the

rotary motion of the read head in diskette units dependent on the coil exci-
tation. Most write-read heads are fitted out with such actuators. Mounting-
and measuring tables have two-dimensional voice-coil drives, e.g. machines
for laser-processing of printed boards.
The motor in Fig. 3. 70 is an example that very old inventions are still used
today. More than llO years ago the rotor of the first widely used three-phase
generator possessed a claw-pole system.

Fig. 3.69. Voice-coil drive of a read-head

Fig. 3.70. Apparatus fan . BLDC motor with claw-pole stator, external magnet
rotor and built-in electronics
152 Ha ns-Dieter Stolting

It was an ingenious idea but in large machines the disadvantages predom-

inate (see Sect. 3.4.2, Fig. 3.42). Employed in small motors t he advantage
is to get a polypole motor with only one coil. Claw-pole constructions are
to be found for instance in stepper motors , in bicycle dynamos , in motorca r
generators and in the BLDC motor of the fan in Fig. 3.70. Often the rotor
magnet is of elastic material. It is not so brittle and difficult to handle as
hard magnet material and adapts better to the rotor housing. Sometimes
the magnet m aterial is molded into the rotor housing. In the field of small
electrical machines there are often such constructions to be found which were
invented in former times for large machines, but because of their disadvan-
tages are no longer used and forgotten. Later they were rediscovered because
for small motors their disadvantages are not so serious compared to their
advantages 7[ ]. Other examples are the bell rotor and the skew winding used
in DC commutator motors and in BLDC motors too (Figs. 3.9, 3.2).
A h eart- lung machine can b e rep laced by two micro pumps as shown in
Fig. 3. 71. They are placed within a catheter in the left and the right cardiac
valve. They pump the blood during a heart operation. A flexible staff (spiral)
carries a ball on the top. At the bottom of the spiral there is the pump a nd
the motor within the pale plastic case. This is necessary caused b y a sep tic

Fig. 3. 71. Micro blood-pump and its placing in the left a nd right cardiac valve
 3 Electromagnetic Actuators 153

demands because the blood flows along this case. By it the blood is warmed
up only about one degree. The three-phase two-pole synchronous motor has
an outer diameter of 6 mm and an output power of 3 W. It looks like the
motor in the top of Fig. 3.22. It is directly coupled with the pump rotating
at 35,000rpm. By this new method the opening of the breast can be made
very small so that the convalescence of the patient is faster and with fewer
problems.

References

1. Stepina, J. (1982): Die Einphasenmotoren, Springer-Verlag, Berlin/Heidelberg

2. Kenjo, T.; Nagamori, S. (1985): Permanent-magnet and brushless DC motors,
Clarendon Press, Oxford
3. Beisse, A.; Lebsanft, L. (1985): Betriebsverhalten permanenterregter Gleich-
strommotoren bei Verschiebung einer der Kohlebiirsten, Etz-Archiv H. 12, pp
389-394
4. T. J. E. Miller, T. J. E. (1989): Brushless permanent-magnet and reluctance
motor drives, Clarendon Press, Oxford
5. Nasar, S. A.; Boldea, 1.; Unnewher, L. E. (1993): Permanent magnet, reluc-
tance and self-synchronous motors, CRC Press, Inc., Boca Ration/ Ann Ar-
bor /London/Tokyo
6. Kenjo, T.; Sugawara, A. (1994): Stepping motors and their Microprocessor
Controls, 2nd edition, Clarendon Press, Oxford
7. StOlting, H.-D. (1998): Alles bewegt sich. Geschichte der Elektrotechnik, Bd.
16, Geschichte elektrischer Kleinmaschinen, VDE-Verlag, Berlin
8. StOlting, H.-D.; Kallenbach, E. (2002): Handbuch Elektrische Kleinantriebe,
2nd edition, Carl Hanser Verlag, Miinchen/Wien
9. Brosch, P. F. (2002): Moderne Stromrichterantriebe, 4th edition, Vogel
Buchverlag
10. Kallenbach, E.; Eick, R.; Quendt, P. (2003): Elektromagnete, 2nd edition, Teub-
ner, Stuttgart
4 Fluidpower Actuators

Wolfgang Backe and Andreas Klein

4.1 Introduction

In fluidpower actuators, the power of a liquid or gas flow is adjusted by

means of suitable devices. These devices are controlled by an electric input
signal, and transform the applied power into a specified mechanical variable
such as stroke (angle of rotation), speed (rotatory speed), or force (torque)
of a longitudinal or rotatory motion. The thermal energy of the pressure
medium is not utilized in any of these cases.
In hydraulics, relatively high pressures (up to 420 bar) and relatively low
flow velocities within the lines are used for energy transmission. However,
high flow velocities do occur, for example whenever there are great pressure
differentials at flow resistors or leakage gaps. These high flow velocities may
have undesirable consequences; e.g. throttle losses at leaks or impulse forces
that affect control elements.
Far lower pressures are usually used in pneumatics (6 to 10 bar). However,
much higher flow velocities within the lines are permissible.
In order to increase the accuracy with which fluidpower actuators adhere
to reference variables and compensate for any errors caused by disturbances,
the principle of feeding back the control variable (and, if necessary, further
auxiliary control variables) and constantly comparing it to nominal values is
employed.
The development of suitable controller concepts - often relying on digital
signal processing - has led to a significant increase in the speed and accuracy
of fluidpower actuators.
Fluidpower actuators are employed wherever large forces or torques have
to be brought to bear; e.g. according to a speed programme. By contrast,
pneumatic actuators are suitable for transporting low masses at high veloci-
ties, for example in the realm of transportation technology or in loading and
unloading production machinery.
Hydraulic energy can be controlled either through resistance (i.e. valve) or
through displacement control. One furthermore distinguishes between feeding
from a source with an impressed pressure or an impressed flow. Figure 4.1
shows the most important circuits of hydraulic drives.
Resistance control usually relies on an impressed pressure. The electrohy-
draulic valves used exhibit good dynamic performance, because low masses

H. Janocha (ed.), Actuators

© Springer-Verlag Berlin Heidelberg 2004