Note: Descriptions are shown in the official language in which they were submitted.
CA 02792602 2012-10-19
1
TITLE
Stabilized High-Voltage Power Supply
TECHNICAL FIELD
The present invention relates to a high-voltage power supply comprising a
plurality of DC
power modules connected in series, each module providing a DC link voltage
that may be
selectively switched. This general setup is often called a "pulse-step
modulator" (PSM).
PRIOR ART
Pulse-step modulators are widely used as voltage supplies for high-power
vacuum tubes in
various applications, in particular, as modulation amplifiers in AM
transmitters or radar
systems. An early example of a pulse-step modulator is disclosed, e.g., in EP
0 066 904.
A typical prior-art pulse-step modulator is illustrated in Fig. 1. The
modulator comprises a
plurality of identical DC power modules 10 connected to the secondaries of a
multi-
secondary transformer 14. Each module comprises a rectifier circuit 11 and a
smoothing
capacitance 12 for providing a DC link voltage, and a switching circuit 13 for
selectively
providing the DC link voltage to the output of the module. A free-wheeling
diode in each
switching circuit ensures that an unidirectional current may flow through each
module
even when the output switch of the module is open.
The total output voltage Vta of such a PSM-type power supply can be modulated
by
switching the DC linlc voltages to the outputs of the modules 10 in an on/off
fashion. This
provides a rather coarse, step-wise modulation of the total output voltage,
the steps being
determined by the individual output voltages of the modules ("coarse-step
modulation",
CSM). For more accurate control, pulse-width modulation (PWM) at the outputs
of the
modules is often additionally employed. Intricate switching schemes have been
suggested
CA 02792602 2012-10-19
2
to ensure that the power load is distributed equally over the modules, and to
increase the
effective PWM frequency without increasing the actual switching frequency of
each
module. Since PWM occurs at much higher frequencies than the mains frequency,
PWM is
also usually employed to remove voltage ripple at multiples of the mains
frequency.
Whereas PWM allows for a very accurate control of the total output voltage,
any PWM
switching scheme requires a PWM output filter to eliminate voltage ripple at
the PWM
frequency.
One potential application of a PSM-type power supply is its use as the cathode
power
supply (main power supply, MPS) of a gyrotron. A gyrotron is a particular type
of vacuum
tube, which emits millimeter-wave electromagnetic radiation. Typical beam
output powers
range from some tens of kilowatts well into the megawatt range. Gyrotrons are
used, inter
alia, in nuclear fusion research to heat plasmas. =
A gyrotron typically comprises an electron gun, an acceleration chamber, a
resonance
cavity immersed in a strong magnetic field, and an electron collector. An
electron beam is
accelerated to relativistic energies and subjected to the magnetic field. The
electrons gyrate
around the magnetic field lines and emit electromagnetic radiation. By
interaction of the
relativistic electrons with the radiation field, amplification of the
electromagnetic radiation
occurs. Gyrotrons are as such well known and are available commercially from a
variety of
manufacturers.
A gyrotron, together with a typical configuration of its power supplies, is
illustrated in
highly schematic form in Fig. 2. A typical gyrotron comprises a filament F
heating a
cathode K for emitting electrons. The electrons are accelerated in the
acceleration chamber
past an anode A and past a body electrode B to reach the resonance cavity. The
electrons
finally hit a collector C, which is typically kept at a lower potential than
the body electrode
to decelerate the electrons ("depressed collector"). Several power supplies
are employed to
operate the gyrotron. A filament power supply FPS powers the __filament F. A
_cathode .
power supply MPS provides a negative voltage between the collector C and the
cathode K,
this voltage being in the range of several tens of kilovolts. A body power
supply BPS
provides a positive voltage, which is also typically in the range of a few
tens of kilovolts,
between the collector C and the body electrode B. An anode power supply APS
allows to "
CA 02792602 2012-10-19
3
selectively switch a voltage between the cathode K and the anode A to modulate
the beam
current. Other configurations of electrodes and power supplies have been
suggested, which
however need not be discussed here.
An extremely important parameter for the operation of the gyrotron is the so-
called beam
voltage between the cathode K and the body electrode B. The electrons are
accelerated
between the cathode and the body electrode by this voltage, i.e., the kinetic
energy of the
electrons and therefore their relativistic mass increase is determined by this
voltage.
Through the relativistic mass increase, this voltage influences the cyclotron
frequency of
the electrons. Since the cavity of a gyrotron typically has a high quality
factor Q, already
small variations in the cyclotron frequency can dramatically influence the
output power of
1
the gyrotron. The beam voltage should therefore be as accurate as possible and
should have
as little ripple as possible.
In the configuration of Fig. 2, the quality of the beam voltage is determined
both by the
quality of the cathode power supply MPS and of the body power-supply BPS.
Whereas the
current load on the body power supply BPS is relatively low (typically in the
range of a
few tens of milliamperes), the current load on the cathode power supply MPS is
typically
in the range of several tens of amperes. In addition, the cathode power supply
MPS must
be capable of rapidly responding to large, rapid current changes while
maintaining the
beam voltage at a predetermined value as accurately as possible. This presents
a
considerable challenge to the successful design of a cathode power supply.
In particular, if PWM is employed for regulating the MPS output voltage, the
PWM output
filter must be designed to ensure both a rapid response to a change in current
load and a
low voltage ripple due to PWM. These two requirements are difficult if not
impossible to
satisfy simultaneously. Furthermore, parasitic capacitances of the transformer
may
propagate the PWM frequency and its harmonics back into the mains grid, which
may lead
to EMC problems. _ _ ...
If, on the other hand, only coarse-step modulation is used without employing
pwm,
considerable ripple from the mains may result, and voltage regulation can only
occur
stepwise.
CA 02792602 2012-10-19
4
In J. Alex et al., "A new klystron modulator for XFEL based on PSM
technology",
Proceedings of PAC07, Albuquerque, New Mexico, USA (2007), a particular type
of PSM
has been suggested to drive a klystron tube. A plurality of DC power modules
are
connected in series. Each module comprises a rectifier with a smoothing
capacitance,
followed by a boost converter (step-up converter), acting as a constant power
converter to
charge a large output capacitance (20 mF). The voltage across the output
capacitance can
be selectively connected to the output of the module by an output switching
circuit. The
power supply acts to provide short high-voltage pulses to a load
(specifically, a klystron
connected to the power supply via a pulse transformer). During each pulse,
power is drawn
from the output capacitances of the modules. Consequently, the voltage at the
output of
each module drops significantly during each pulse. Subsequently the output
capacitance of
each module is recharged by the boost converter.- The boost converter is
controlled in a
mariner to ensure that constant power is drawn from the mains supply to keep
power
variations ("flicker") on the mains grid due to the pulses as low as possible.
A further development of this type of power supply is disclosed in J. Alex et
al., "A new
prototype modulator for the European XFEL Project in pulse step modulator
technology",
Proceedings of PAC09, Vancouver, BC, Canada, May 4-8, 2009. Again, each module
employs a boost converter to ensure constant power consumption on the mains
supply.
Pulse-width modulation is employed to compensate for the voltage droop on the
storage
capacitance during the pulse in order to improve the flatness of the pulse.
A high-voltage power supply of the same general type, wherein each module
comprises a
boost converter acting as a constant-power converter, is also disclosed in EP
2 099 127 Al.
These power supplies cannot overcome the deficiencies noted above in
connection with the
more traditional kinds of PSM power supplies if both a rapid response to
current changes
and low voltage ripple are needed, since stabilization of the total output
v_oltage still either
requires PWM or, in the alternative, can be only carried out stepwise.
CA 02792602 2012-10-19
SUMMARY OF THE INVENTION
It is therefore an object of the present invention to provide a high-voltage
power supply
that is capable of providing a stable total output voltage having low voltage
ripple while
5 being able to handle large, rapid current changes.
This object is achieved by a high-voltage power supply as specified in claim
1. The
invention further provides a method of operation of such a power supply, as
laid down in
claim 11, a method of use as laid down in claim 13, and a gyrotron employing
such a
power supply as laid down in claim 14. Further embodiments of the invention
are laid
down in the dependent claims.
, According to the present invention, a high-voltage power supply is
provided, comprising:
a plurality of DC power modules having their outputs connected in a series
configuration,
each power module comprising a DC voltage source, a DC-DC converter receiving
an input voltage from said DC voltage source and providing a DC link voltage,
and an
output switching circuit for selectively connecting said DC link voltage to
the output of
said power module,
characterized in that the power supply comprises, for each DC power module, a
feedback control circuit operable to provide driving signals to the DC-DC
converter of said
power module in a manner that regulates said DC link voltage to a
predetermined reference
voltage.
Therefore, in contrast to the prior art, the DC-DC converter is not operated
to ensure that
constant power is drawn from the DC voltage source of the module, but it is
operated as a
voltage regulator, ensuring that the DC link voltage takes a predetermined
reference value.
The reference voltages of all involved modules are preferably identical, such
that the total
output voltage of the power supply is the reference voltage, multiplied by_the
number of
modules involved.
Each module will therefore normally comprise a DC link voltage sensor for
measuring the
DC link voltage, and the feedback control circuit will normally receive the
measured DC
CA 02792602 2012-10-19
6
link voltage from the sensor, optionally process the measured voltage (e.g.,
by subjecting it
to a low-pass filter), and compare it to the (optionally also pre-processed)
reference
voltage. A difference signal will be fed to a suitable controller, e.g., a PI
or PD controller.
The controller output will be fed to an actuator, which calculates the driving
signals for the
DC-DC converter.
Advantageously, when deriving the driving signals, the actuator may directly
take into
account the major disturbance variables of the control loop, in particular,
the actual output
current of the power modules and/or the actual input voltage of the power
modules. In this
manner, a very rapid response to changes in these disturbance variables may be
achieved.
In particular, the power supply may comprise at least one current sensor for
measuring an
output current of said DC power modules, and the feedback control circuit may
then be
adapted to derive driving signals for the DC-DC converter of each power module
taking
into account the measured output current. A single current sensor for the
complete power
supply may be sufficient; however, it is preferred that each module comprises
its own
current sensor. Furthermore, each power module may comprise an input voltage
sensor for
determining an input voltage of its DC-DC converter, and the voltage control
circuit may
then be adapted to derive driving signals for the DC-DC converter of each
power module
taking into account the measured input voltage. The DC-DC converter will
normally be a
switched converter operable at a variable duty cycle. The feedback control
circuit may then
be adapted to calculate the duty cycle taking into account the measured
disturbance
variables, i.e., the measured output current and/or the measured input
voltage, and to
operate the DC-DC converter at that duty cycle. A possible formula for the
duty cycle in
the case of a boost converter will be given in section "Detailed Description
of Preferred
Embodiments" below; however, other relationships taking into account these
disturbance
variables are also conceivable.
Preferably, the DC-DC converter of each power module is a boost converter.
However, it
is conceivable to employ any other type of switched DC-DC converter ,suck,
e.g, a
buck converter, a buck-boost converter, or a SEPIC. For practical reasons, it
is preferred
that the converter is capable of delivering an output voltage that is higher
than the input
voltage, which would render buck converters less preferred. Since it is
furthermore
preferred to minimize any losses, a boost converter with its simple topology
is the
CA 02792602 2012-10-19
7
preferred converter type.
A boost converter comprises at least one output capacitance and at least one
converter
switching element cooperating with at least one choke inductance and at least
one diode (or
any other unidirectional switching element) to charge the output capacitance
to the DC link
voltage. The feedback control circuit is then operable to control the DC link
voltage by
switching the converter switching element at a variable duty cycle.
An interleaved topology of the boost converter may be chosen to reduce AC
currents in the
input and output capacitances, and to further reduce voltage ripple. The boost
converter of
each power module may therefore comprise at least two interleaved boost
converter
circuits adapted to charge a common output capacitance. The control circuit is
preferably
operable to operate the interleaved boost converter circuits in a synchronous
but phase-
shifted manner.
In a concrete setup, such an interleaved boost converter topology may comprise
at least
one common output capacitance, at least one first converter switching element
(possibly
two or more such elements connected in series) cooperating with at least one
first choke
inductance (possibly two or more such inductances in a series configuration
with the first
switch elements) and diode (possibly two or more diodes in a series
configuration with the
first choke inductances) to charge said output capacitance to the DC link
voltage, and at
least one second converter switching element (possibly two or more such
elements
connected in series) cooperating with at least one second choke inductance
(possibly two
or more such inductances in a series configuration with the second switch
elements) and
diode (possibly two or more diodes in a series configuration with the second
choke
inductances) to charge the same output capacitance. The control circuit is
then operable to
control the DC link voltage by driving the first and second converter
switching elements at
a variable duty cycle, preferably in a synchronous but phase-shifted manner.
= _
_ _
In order to ensure a rapid response to current changes and to reduce switching
losses, the
control circuit preferably operates the boost converter of each power module
in
discontinuous mode during voltage regulation. This can be ensured by
appropriately
choosing the choke inductance to be lower than the so-called critical choke
inductance in
CA 02792602 2012-10-19
8
the intended operating regime of the module. An example for the calculation of
the critical
choke inductance is provided below in section "Detailed Description of
Preferred
Embodiments".
The control circuits of the individual modules may be commonly controlled by a
main
control system for the modules. In particular, the overall step-response
behavior of the
power supply to disturbances such as current changes may be improved by
driving the DC-
DC converters of different DC power modules in a synchronous but phase-shifted
manner.
In this manner, the DC-DC converters of some modules in the appropriate phase
within the
converter cycle time can react to disturbances earlier than those lagging
behind in phase.
Preferably, the total output voltage of the power supply during each time
period in which
, the power supply is supposed to deliver a certain predetermined, non-zero
voltage value
("voltage pulse") is regulated by only controlling the DC-DC converters of the
individual
power modules, without applying coarse-step modulation, and without applying
pulse-
width modulation. In other words, the modules are preferably operated in a
manner in
which the output switching circuits of all power modules remain in the same
(active or
passive) state over prolonged times, in particular, as long as the desired
total output voltage
remains unchanged, unless a fault condition occurs. Such time periods will
normally cover
at least 10, often at least 50 switching cycles of the DC-DC converter. They
can last, e.g.,
from less than 1 ms to several seconds or even longer.
In order to enable a rapid discharge of the output capacitance in cases such
as a missing
load, each DC power module may comprise a discharge resistor and a discharge
switch for
selectively discharging the output capacitance.
In preferred embodiments, the power source may comprise at least one multi-
secondary
transformer having a plurality of sets of secondary windings, and each DC
power module
may then comprise a rectifier circuit connected to one set of secondary
windings- The_DC
voltage source of each module may thus be considered to comprise this set of
secondary
windings and the corresponding rectifier circuit. Alternatively, a separate
transformer may
be used for each module. Depending on the field of application, it is even
conceivable to
use other kinds of DC voltage sources, even batteries.
CA 02792602 2012-10-19
9
To improve power ratio, the high-voltage power supply may comprise a first and
a second
multi-secondary transformer, the first transformer and the second transformer
being
configured to provide secondary voltages that are phase-shifted between the
transformers.
In the case of a three-phase mains grid, this may be accomplished, e.g., by
using a star-star
configuration for the first transformer and a delta-star configuration for the
second
transformer, resulting in twelve-pulse rectification. Other suitable schemes
for achieving a
phase shift are well known in the art.
The present invention also provides a method of operating a high-voltage power
supply as
described above. The method comprises, for each power module:
setting the reference voltage;
measuring and processing the actual DC link voltage;
comparing the DC link voltage and the reference voltage to derive a difference
signal;
from the difference signal, deriving an actuating signal; -
measuring an actual output current and/or the actual input voltage;
from the actuating signal, deriving driving signals for the DC-DC converter,
taking
into account the measured output current and/or input voltage; and
driving each DC-DC converter by said driving signals to actively control
said output voltage.
In particular, as outlined above, the driving signals may implement a duty
cycle
determined, inter alia, by the actuating signal, the output current and the
input voltage.
The high-voltage power supply described above may advantageously be used as a
power
supply for a gyrotron, in particular, as its cathode power supply.
Accordingly, the present
invention also relates to a gyrotron. Generally, a gyrotron has at least a
cathode, a body
electrodeõ and a collector electrode, and advantageously_ the power supply_
is_connected
between the cathode and the collector electrode.
While a particular application for a gyrotron has been described, there are
also other fields
of applications where a highly stable high-voltage power supply capable of
handling rapid
CA 02792602 2012-10-19
current changes is needed, and the power supply of the present invention may
be employed
in such other fields as well.
BRIEF DESCRIPTION OF THE DRAWINGS
5
Preferred embodiments of the invention are described in the following with
reference to
the drawings, which are for the purpose of illustrating the present preferred
embodiments
of the invention and not for the purpose of limiting the same. In the
drawings,
10 Fig. 1 shows a schematic block diagram of a PSM power supply
according to the
prior art;
Fig. 2 shows a highly schematic sketch of a typical gyrotron together
with its
associated power supplies;
Fig. 3 shows, in a highly schematic fashion and not to scale, a
typical sequence of
voltages, currents and rf power in a gyrotron;
Fig. 4 ¨ shows a schematic block diagram of a power supply according to
the
present invention;
Fig. 5 shows a simplified circuit diagram of a single module of the
power supply
of Fig. 4; and
Fig. 6 shows a diagram illustrating the control loop for controlling the
output
voltage of a module as shown in Fig. 5.
DESCRIPTION OF PREFERRED EMBODIMENTS
Figure 3 illustrates, in a highly schematic fashion and not to scale, a
typical sequence of the
various voltages and currents involved in the operation of a gyrotron. At the
beginning of
the sequence, the cathode power supply MPS and the body power supply BPS are
switched
to their nominal output voltages to provide a predetermined beam voltage VBK.
Both
voltages are actively controlled to ensure stability of these voltag_e_s. The
anode-cathode
voltage VAK provided by the anode power supply APS is initially kept at a
value which
avoids any significant beam current between the cathode K and the collector C.
Only when
the gyrotron is to generate electromagnetic radiation, the anode-cathode
voltage VAK is
switched to a positive value, leading to a rapid rise of the beam current IK
and to the
CA 02792602 2012-10-19
11
emission of electromagnetic radiation with power P.
From this diagram it is apparent that the cathode power supply MPS must be
capable of
rapidly reacting to large variations in beam current IK while providing a
stable output
voltage with low voltage ripple.
Figure 4 shows a simplified block diagram of a high-voltage power supply
according to the
present invention, which is adapted to satisfy these requirements. The general
setup is
similar to the setup of Fig. 1. However, instead of one single multi-secondary
transformer,
in the present example two multi-secondary transformers 44, 45 in different
configurations
(YY vs. DY) are used to improve overall power factor. Both transformers 44,45
are fed by
a three-phase mains voltage Vm. The three primary windings of the first
transformer 44 are
- connected in a star ("Y") configuration, while the three primary windings
of the second
transformer 45 are connected in a delta ("D") configuration. Each transformer
comprises a
plurality of sets of three secondaries (typically 20-30 such sets), each set
being connected
in a star configuration. Overall, the first transformer 44 thus has a star-
star ("YY")
configuration, while the second transformer has a delta-star ("DY")
configuration. By this
overall configuration of the two transformers 14, 15, the secondary voltages
of the two
transformers are appropriately phase-shifted with respect to each other to
achieve twelve-
pulse rectification, which leads to a high power factor of typically more than
0.95.
The power supply comprises a plurality of identical DC power modules 40
connected to
the secondaries of the transformers 44, 45. Each DC power module comprises a
rectifier
circuit 41, a boost converter 42, and an output switching circuit 43. The
outputs of the
modules 40 are connected in a series configuration.
Fig. 5 shows a simplified circuit diagram for a single module 40, which will
now be
described in more detail.
_
Input terminals X10, X11, X12 are connected to a set of three secondaries of
one of the
transformers 44, 45. The rectifier circuit 41 in the form of a full-bridge
rectifier, consisting
of six diodes D1, rectifies the three-phase secondary voltage.
CA 02792602 2012-10-19
12
The boost converter 42 comprises two interleaved boost converter circuits
having a
common input capacitance and charging a common output capacitance, each of the
interleaved converter circuits being in turn designed with two switching
elements in series
to reduce the required voltage rating of each switching element. In more
detail, the
common input capacitance of the boost converter is formed by a network of
capacitors
C10, C11, connected in series to enable the use of capacitors having a reduced
voltage
rating, and symmetizing resistors R10, R11 in parallel to the two capacitors.
The voltage
across this input capacitance is called the input voltage \Tin. The common
output
capacitance of the boost converter is formed by two series-connected
capacitors C20, C21.
Each of the upper and lower terminals of the output capacitance is connected
to the input
capacitance via two groups of elements consisting of a series-connected choke
inductance
and diode each, the two groups being connected in parallel. The diodes are
forward-biased
- to allow energy to flow from the input capacitance through the choke
inductances into the
output capacitance, but not in the reverse direction from the output
capacitance back into
the input capacitance. Two pairs of series-connected actively controlled
converter switches
V10, V11 and V20, V21, respectively, here in the form of IGBTs, connect the
connection
node between an upper choke inductance and upper diode with the connection
node
between a lower choke inductance and lower diode. By closing these switches,
current can
flow through these choke inductances while the output capacitance is bypassed.
As a
result, the upper and lower choke inductance together are subjected to the
input voltage,
and inductive energy builds up in the choke inductances due to this voltage.
When the
switches are opened again, the stored energy in the choke inductances is
transferred to the
output capacitance through the diodes. The resulting output voltage (DC link
voltage) Vout
across the output capacitance can be much larger than the input voltage. This
principle of
operation of a boost converter is as such well known in the art.
A discharge resistor R40 and a switch K20 enable a rapid discharge of the
output
capacitance if needed. Output switches V30, V31 in the form of IGBTs
Selectively provide
the DC link voltage Nr_out to the output terminals X20, X21 of the module. Two
reverse-
biased, series-connected freewheeling diodes D30, D31 enable a unidirectional
current to
flow between the output terminals even when the output switches V30, V31 are
open. The
common node between the two diodes D30, D31 is connected to the common node
between the converter switches V10, V11 and V20, V21, respectively, of each
pair of these
CA 02792602 2012-10-19
13
switches as well as to the common node between the capacitors C20, C21 to
provide
improved symmetrization. Additional symmetrization resistors (not shown in
Fig. 5) may
optionally be provided in parallel to the capacitors C20, C21.
A module controller 51 is fed from input terminals X10, X11 via a small
transformer T10.
Control signals and diagnostic signals are exchanged between the module
controller 51 and
an external main control system 53 via a fiber-optic link 52 ensuring galvanic
isolation.
The module controller 51 controls the booster switches VIO, V11, V20, V21, the
discharge
switch K20, and the output switches V30, V31 via leads that have been omitted
in Fig. 5
for the sake of clarity. In the present example, all switches are implemented
as
semiconductor switches, in particular, as IGBTs, which will usually be
equipped with
additional reverse-biased freewheeling diodes in parallel to the collector-
emitter path (not
shown). However, other types of actively controlled semiconductor switches may
be
employed, depending on the actual load requirements, such as power MOSFETs
etc.
The power supply is operated as follows: Depending on the desired total output
voltage
Vtot and on the status of the modules, the main control system selects whether
all or only a
part of the modules shall be involved in providing the desired total output
voltage Vtot. The
main control system accordingly provides control signals to the module
controllers 51 of
the individual modules via the fiber-optic link 52. Each module controller
drives the
converter switches V10-V21 of its associated module to charge the output
capacitance
C20, C21 to a reference DC link voltage Vref determined by the main control
system. This
reference voltage is set to the same value in all modules. It corresponds to
the desired total
output voltage Vtot, divided by the number of involved modules. The output
switches of the
involved modules are then closed to provide the DC link voltages of the
modules to their
outputs, so as to provide the sum of the DC link voltages at the output of the
power supply.
During normal operation, and in particular during individual current pulses
delivered by
the power supply, the output switches remain closed and are not operated, in
contrast to
prior-art devices, where PWM is implemented at the output switches to provide
voltage
regulation. Instead, regulation of the total output voltage is carried out by
regulating the
DC link voltages Vow supplied by the boost converters in a feedback control
circuit
implemented in module controller 51.
CA 02792602 2012-10-19
14
The feedback control circuit 60 is illustrated schematically in Fig. 6. The
actual DC link
voltage Vow (in terms of control theory, the controlled variable of the
control loop) is
determined by a suitable voltage sensor 65. The measured DC link voltage is
subjected to a
low-pass filter 64 to filter out the switching frequency of the boost
converter and its
harmonics, and the filtered DC link voltage is compared to the reference
voltage Vref in a
comparator 61. The difference Vdiff of these two voltages is fed to a PI
controller 62, whose
controller output signal Vad (in terms of control theory, the actuating
variable of the
control loop) is fed to a calculating unit 63 (the actuator of the control
loop). An input
voltage sensor 64 measures the input voltage Via, and an output current sensor
measures
the output current I. These signals (corresponding to the main disturbance
variables of
the control loop) are also fed to the calculating unit 63. From the controller
signal Vetrl,
from the input voltage Vii, and from the output current loa, the calculating
unit 63
- calculates a duty cycle (i.e., the ratio of the time during which the
switches of the converter
are closed and the cycle time of the converter) and drives the converter
switches V10-V21
according to this duty cycle.
The controlled variable Vow is a quadratic function of the duty cycle. In
order to achieve a
linear control path, the duty cycle should therefore be a square root function
of the
actuating variable V. In this manner, a linear dependence between controller
signal Vali
(actuating variable) and DC link voltage Voot (controlled variable) results.
As will be
detailed further below, the boost converter is preferably operated in
discontinuous mode
(i.e., the current in the booster choke inductances substantially decreases to
zero before the
next booster cycle starts). The duty cycle in discontinuous mode may be
calculated
according to the following formula:
D = 112 = / - = (Vm ¨
(Equation 1).
Tbooster =Vin2
Here, T- is the booster cycle time (the inverse of the-boost-converter
operating
frequency, which is kept constant during operation), and ',booster is the
total choke
inductance of the boost converter (in the specific arrangement illustrated
here, Lboost, =
L10 = L11 =- L20 = L21). Both pairs of converter switches in the interleaved
converter
circuits are operated at the same duty cycle, but phase-shifted by 180
relative to teach
CA 02792602 2012-10-19
other.
An important property of the calculating unit is that the most important
disturbance
variables, Lut and Vin, directly act on the actuator of the control loop. For
example, in the
5 case of a sudden rise of the output load, the output current will also
rise rapidly, while the
DC link voltage will drop slowly due to the presence of the large output
capacitance. Since
the increased output current directly acts on the actuator, the duty cycle
will be increased
almost instantaneously, and the voltage drop will stop within a single booster
cycle time
The PI controller can now correct the (relatively small) voltage drop that has
10 already occurred.
In order to ensure a rapid response to load changes, the boost converter
should be operated
- in discontinuous mode during normal operation. This measure also improves
stability of
the regulation of the DC link voltage, and minimizes switching losses on the
converter
15 switches, since these are always switched on at zero current. This poses
certain restrictions
on the booster choke inductances. In particular, the total choke inductance
should not
exceed a certain critical value, which is well known in the art and depends on
the desired
operating point as follows:
20Viõ2 = (V ¨ V. ) = T
_ out in booster
(Equation 2).
On the other hand, the choke inductance should not be too small in order to
keep the input
current to low figures. This calls for a choice of choke inductance below but
close to the
critical choke inductance Lcrit.
The PI controller 62 can be disabled selectively in certain situations where
feedback
control would be inappropriate. One such situation is the case of a missing
output load (no-
load case), if at the same time the DC link voltage Vout is higher than the
reference voltage
Vref. While. the controller is disabled, the output capacitance may be
discharged down to
the reference voltage by closing the discharge switch K20. Once the reference
voltage Võf
is reached again, the controller is enabled again.
In order to enable controlled ramp-up of the output voltage in the no-load
case, e.g., after a
CA 02792602 2012-10-19
16
positive change of the reference voltage Vref, particular measures are
required, since
Equation (1) implies that the duty cycle will be zero as long as the output
current is zero. In
order to overcome this problem, it is possible to set the current in Equation
(1) to some
predetermined minimum value (e.g., 1-2% of the maximum output current) if the
actual
Particular measures are required for power-up. A possible power-up sequence
may be
implemented as follows: The input capacitance Cl 0, Cl I and the output
capacitance C20,
C21 are initially charged to the nominal voltage of Vin via step-start
switches and a
charging resistor (not shown in Fig. 5), as they are well known in the art.
The boost
converter switches remain disabled until all capacitances are charged to
approximately the
nominal voltage of \Tin. Only then the boost converter starts to operate. The
boost converter
is initially operated at constant duty cycle, until the reference voltage is
reached across the
output capacitance. Only then closed-loop control starts.
Operating conditions of the modules are continuously supervised- and any
module is
switched off and possibly replaced by another (so far idle) module if a fault
condition is
detected. In particular, a fault condition is assumed if the booster input
voltage or the
booster output voltage is outside a predetermined range, or if the output
current exceeds a
predetermined maximum value. In addition, temperature, desaturation etc. may
also be
supervised.
The output of each module may be provided with a small output snubber (not
shown in
Fig. 5) in order to limit the current surge in the module in the case of a
short circuit. This
Actual values of capacitances, inductances, resistors etc. will largely depend
on the
concrete application and on the desired operating point. _
The above description is only for illustrative purposes, and a number of
modifications can
be made without departing from the scope of the present invention. In
particular, the boost
converter design can be different from the design as described above. In the
simplest case,
CA 02792602 2012-10-19
17
a single converter switch may be used in conjunction with a single choke
inductance, a
single diode and a single output capacitor, as it is well known in the art and
illustrated
schematically in the box symbolizing module 40 in Fig. 6. Different controller
types than
PI controllers may be employed, such as PID controllers. All diodes (acting as
passive
switches) may be replaced by active switching elements such as transistors if
desired. The
rectifier circuit may be designed differently, e.g., as an actively controlled
thyristor
rectifier circuit. Instead of .single-quadrant output switching circuits, as
in the above-
described embodiment, which allow only for unipolar voltage and unidirectional
current,
also two-quadrant output switching circuits allowing for bipolar voltages at
unidirectional
current or for unipolar voltage at bidirectional currents or even four-
quadrant output
switching circuits allowing for arbitrary sign of both output voltage and
output current may
be employed. Two-quadrant switching may be useful, e.g., for inverse voltage
operation to
-reduce currents after a short circuit has occurred, or for driving capacitive
loads such as a
control electrode of a vacuum tube. Suitable output switching circuits for two-
quadrant or
four-quadrant operation are disclosed, e.g., in EP 2 099 127 Al, in particular
in its Figures
5-7, and the disclosure of that document is incorporated herein by reference
in its _entirety
for teaching suitable output switching circuits for two- and four-quadrant
operation.
Suitable output switching circuits and modes of operation of two- and four-
quadrant output
switching circuits are also disclosed in WO 95/10881 Al and EP 1 553 686 Al.
In other embodiments, depending on the intended field of use, the boost
converter may be
replaced by any other form of switched DC-DC converter. This might be a buck
converter,
a buck-boost converter, a SEPIC etc. Such switched-mode DC-DC converters are
well
known in the art. The operating principles as outlined above remain the same
with such
DC-DC converters. In particular, also with other types of DC-DC converters it
is possible
to regulate the DC link voltage of each module by controlling the DC-DC
converter,
instead of employing PWM and/or CSM schemes to regulate the total output
voltage of the
complete power supply.
The proposed power supply may not only be employed as the main power supply of
a
gyrotron, but may be used in any application which require a stabilized high
voltage which
is stable even under rapid load changes. Examples include the cathode or anode
power
supply of any other type of vacuum tube having a control electrode which may
rapidly
CA 02792602 2012-10-19
18
change the current in the tube.
LIST OF REFERENCE SIGNS
power module C10, C11 input capacitors .
11 rectifier circuit R10, R11 divider resistors
12 smoothing capacitance Ll 0, Lll, L20, L21 choke inductances
13 output switching circuit V10, V11, V20, V21 converter switches
14 transformer D10, Dll, D20, D21 converter diodes
2 gyrotron C20, C21 output capacitors
FPS filament power supply R20 dissipating resistor
MPS cathode power supply K40 dissipating switch
BPS body power supply V30, V31 output switches
APS anode power supply D30, D31 freewheeling diodes
F filament 51 module controller
K cathode 52 fiber optic link
A anode Vin input voltage
B body electrode Void DC link voltage
C collector Vtot total output voltage
VK cathode voltage Vref reference voltage
VBK beam voltage Vdiff voltage difference
YAK anode-cathode voltage Vi controller output signal
Ix beam current lout output current
Pif radiated power 60 control circuit
40 power module 61 comparator
41 rectifier circuit 62 PI controller
42 boost converter 63 calculating unit
43 output switching circuit 64 input voltage sensor
X10, X11, X12 input terminals 65 output voltage sensor
X20, X21 output terminals 66 output current sensor
D1 rectifier diode
