(19) United States
`(12) Patent Application Publication (10) Pub. No.: US 2007/0139122 A1
`Nagarkatti et al.
`(43) Pub. Date:
`Jun. 21, 2007
`
`US 20070139122A1
`
`(54) RADIO FREQUENCY POWER DELIVERY
`SYSTEM
`
`(75) Inventors: Siddharth P. Nagarkatti, Acton, MA
`(US); Michael Kishinevsky, North
`Andover, MA (US); Ali Shaji, Canton,
`MA (US); Timothy E. Kalvaitis,
`Nashua, NH (US); William S.
`McKinney JR., Cambridge, MA (US);
`Daniel Goodman, Lexington, MA
`(US); William M. Holber, Winchester,
`MA (US); John A. Smith, North
`Andover, MA (US)
`Correspondence Address:
`PROSKAUER ROSE LLP
`ONE INTERNATIONAL PLACE 14TH FL
`BOSTON, MA 02110 (US)
`(73) Assignee: MKS INSTRUMENTS, INC., WILM
`INGTON, MA (US)
`(21) Appl. No.:
`11/554,979
`
`(22) Filed:
`
`Oct. 31, 2006
`Related U.S. Application Data
`(60) Provisional application No. 60/731,797, filed on Oct.
`31, 2005.
`
`Publication Classification
`
`(51) Int. Cl.
`HO3F 3/19
`
`(2006.01)
`
`(52) U.S. Cl. .............................................................. 330/302
`
`(57)
`
`ABSTRACT
`
`A system and method are provided for delivering power to
`a dynamic load. The system includes a power Supply pro
`viding DC power having a substantially constant power
`open loop response, a power amplifier for converting the DC
`power to RF power, a sensor for measuring Voltage, current
`and phase angle between voltage and current vectors asso
`ciated with the RF power, an electrically controllable imped
`ance matching system to modify the impedance of the power
`amplifier to at least a Substantially matched impedance of a
`dynamic load, and a controller for controlling the electrically
`controllable impedance matching system. The system fur
`ther includes a sensor calibration measuring module for
`determining power delivered by the power amplifier, an
`electronic matching system calibration module for determin
`ing power delivered to a dynamic load, and a power dissi
`pation module for calculating power dissipated in the elec
`trically controllable impedance matching system.
`
`AC
`Line in
`
`210
`
`
`
`222
`
`220
`
`
`
`240
`
`250
`
`260
`
`-
`
`Fast DC
`BuS
`
`Fast Bus
`Controller
`
`PA Gate RF Power
`Drive
`Amplifier
`
`RF Voltage
`and Current
`
`Y.
`
`Electronic
`Match
`
`E-Match
`Controller
`
`Power Setpoint
`Control Parameters
`Safety Parameters
`
`Frequency
`13.56 5% MHZ
`
`High-speed
`Digital Link
`
`
`
`Input AC Line Voltage
`Output DC Voltage
`Status
`
`DSP
`Compensator
`Board
`
`RS485
`to USB
`User interface
`
`Digital Bit
`Switching
`Command (0-15)
`to change shunt
`Capacitance in
`E-Match
`
`DSP Seria
`
`
`
`
`
`230
`
`RENO EXHIBIT 2026
`Advanced Energy v. Reno, IPR2021-01397
`
`

`

`Patent Application Publication Jun. 21, 2007 Sheet 1 of 14
`
`US 2007/O139122 A1
`
`
`
`|--------------------------------------]09||
`
`

`

`Patent Application Publication Jun. 21, 2007 Sheet 2 of 14
`
`US 2007/O139122 A1
`
`
`
`
`
`peoT euse|-
`
`o», ------------|-------------------?u?T
`
`,07||uo|ss|uusueu L
`-]09||
`
`

`

`Patent Application Publication Jun. 21, 2007 Sheet 3 of 14
`
`US 2007/O139122 A1
`
`
`
`

`

`Patent Application Publication Jun. 21, 2007 Sheet 4 of 14
`
`US 2007/O139122 A1
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`dSC]
`
`pueO9
`
`JO?eSueduuOO
`
`ZHWN %9 RF 99’9||
`
`Kouanbau
`
`

`

`Patent Application Publication Jun. 21, 2007 Sheet 5 of 14
`
`US 2007/O139122 A1
`
`
`
`009
`
`

`

`Patent Application Publication Jun. 21, 2007 Sheet 6 of 14
`
`US 2007/0139122 A1
`
`
`
`s
`
`N
`
`s
`
`

`

`Patent Application Publication Jun. 21, 2007 Sheet 7 of 14
`
`US 2007/O139122 A1
`
`
`
`g
`
`s
`
`N
`
`C
`h
`(s
`O
`
`.
`2.
`
`CC
`
`

`

`Patent Application Publication Jun. 21, 2007 Sheet 8 of 14
`
`US 2007/O139122 A1
`
`
`
`peo" euuselº
`
`
`
`
`
`
`
`
`
`

`

`Patent Application Publication Jun. 21, 2007 Sheet 9 of 14
`
`US 2007/O139122 A1
`
`092/
`
`
`
`
`
`
`
`
`
`

`

`Patent Application Publication Jun. 21, 2007 Sheet 10 of 14
`
`US 2007/O139122 A1
`
`Calibrate POWer Meter
`to 50 C2 Calorimeter
`POWer Reference
`310
`
`
`
`
`
`
`
`
`
`Calibrate Load
`Simulator Calorimeter
`to DC Power Reference
`320
`
`
`
`
`
`
`
`
`
`
`
`
`
`POWer Delivered
`to 50 O Load
`
`Power Dissipated
`Inside Load Simulator
`
`Calibrate
`V Probe
`into 50 O Load
`
`
`
`POWer Delivered
`by the PA
`
`Calibrate Output of
`On-Chamber DUT
`into Load Simulator
`340
`POWer Delivered
`to Z = R + j X.
`
`
`
`
`
`
`
`
`
`Power Dissipated
`in the E-Match
`
`

`

`Patent Application Publication Jun. 21, 2007 Sheet 11 of 14
`
`US 2007/O139122 A1
`
`
`
`N
`N
`cy)
`
`O
`
`v Q
`
`.C.
`
`O
`w
`N
`
`O
`cN
`cN
`
`c
`w
`CN
`
`
`
`D
`o
`NC
`S 5
`(s
`
`CC
`O)
`
`d
`Ol
`2
`
`s O 3
`
`'S
`g
`o
`
`

`

`Patent Application Publication Jun. 21, 2007 Sheet 12 of 14
`
`US 2007/0139122 A1
`
`
`
`us
`2000)
`DC in POrt 8OOW
`
`

`

`Patent Application Publication Jun. 21, 2007 Sheet 13 of 14
`
`US 2007/0139122 A1
`
`occ/
`
`
`
`

`

`Patent Application Publication Jun. 21, 2007 Sheet 14 of 14
`
`US 2007/O139122 A1
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`uOSJe3e3
`
`?O
`
`p1808 dSO
`
`OZZ
`
`0 || Z.
`
`

`

`US 2007/O 1391 22 A1
`
`Jun. 21, 2007
`
`RADIO FREQUENCY POWER DELIVERY
`SYSTEM
`
`RELATED APPLICATION
`0001) This application claims the benefit of U.S. Provi
`sional Application No. 60/731,797, filed on Oct. 31, 2005,
`the entire teachings of which are incorporated herein by
`reference.
`
`BACKGROUND
`0002 Various approaches exist for providing RF power
`to dynamic loads. RF generators provide power to dynamic
`loads typically at frequencies between about 400 kHz and
`about 200 MHz. Frequencies used in some scientific, indus
`trial and medical applications are approximately 2 MHZ,
`13.56 MHZ and 27 MHZ.
`0003) As shown in FIG. 1A, one system 100 for provid
`ing RF power to dynamic loads (i.e., a plasma load 140)
`involves a fixed frequency RF generator 110 and a two-axis
`tunable matching network 120 connected by a 50 C2 trans
`mission line 130. The tunable matching network 120
`includes a series motorized vacuum variable capacitor 122
`and inductor 124 and a shunt motorized vacuum variable
`capacitor 126. The algorithm used to determine the series
`and shunt capacitance is based on impedance measurements
`typically made using a magnitude and phase detector 150.
`Independent power control is based on power measurements
`at the RF generator 110. The power control loop 160 and
`impedance control loop 162 are independent.
`0004 As shown in FIG. 1B, another system 100' for
`providing RF power to dynamic loads involves a fixed
`element matching network 120' fed by an RF generator 110
`and connected by a 50 C2 transmission line 130. The fixed
`element matching network 120' includes a series capacitor
`122 and inductor 124 and a shunt capacitor 126. The
`frequency of the RF generator 110 can be tuned to a certain
`range (e.g., 13.56 MHz +5%). The RF generator 110 fre
`quency command is based on the value of Voltage standing
`wave ratio (VSWR). The independent power loop and
`VSWR (impedance) control loop 160" are based on mea
`surements at the output of the RF generator 110.
`0005. As shown in FIG. 1C, another system 100" for
`providing RF power to dynamic loads involves an integrated
`RF generator-impedance matching network 120". The RF
`generator-impedance matching network 120" includes a
`series capacitor 122 and inductor 124 and a plurality of shunt
`capacitor 126a. . . 126 n. The shunt capacitor 126a. . . 126n
`are coupled to a switching circuit 127a ... 127n that couples
`and decouples the capacitors 126 to ground. The power
`control and frequency control 160" of the system 100" are
`not conducted simultaneously.
`SUMMARY
`0006 These prior art techniques and methods have dis
`advantages. Higher cost is typically associated with prior art
`techniques and methods due to the need for at least two
`separate modules: 1) the RF generator/amplifier and 2) the
`impedance matching network, which are to be connected via
`a transmission line. Furthermore, each module requires a RF
`Voltage/current sensor or a magnitude/phase detector.
`0007 Plasma impedance is a function of the power
`delivered to the plasma. Furthermore, the power delivered
`
`by the RF generator is a function of the impedance “seen
`by the generator. As a result, a clear circular interdependence
`exists between delivered power and load impedance yielding
`a multi-input-multi-output (MIMO) system with cross-cou
`pling. In prior art systems, the RF generator control loop and
`the impedance matching control loop are independent and
`thus cannot compensate for the cross-coupling between
`power control and impedance matching control loops. This
`leads to poor closed-loop performance.
`0008. The dynamic response of any controlled system is
`only as fast as the slowest functional module (sensor,
`actuator, or control system parameters). In prior art systems,
`the slowest functional module is typically the DC power
`supply. Specifically, the DC power supplied to the input of
`the RF power amplifier usually includes a large electrolytic
`capacitor that is used to filter higher frequencies. The
`downside of using Such a filter network is that the dynamic
`response (e.g., response to a step change in power com
`mand) is slow regardless of the control update rate. The
`system is therefore unable to sufficiently compensate for
`plasma instabilities.
`0009. In systems that use a vacuum capacitor driven by
`motors, the response time is on the order of hundreds of
`milliseconds. Owing to the fact that plasma transients (Sud
`den and rapid change of impedance) of interest occur within
`hundreds of microseconds, the vacuum capacitor cannot be
`used to match load changes attributed to plasma transients.
`0010 Control algorithms for matching networks used in
`the prior art have relied upon the real and imaginary com
`ponents of the measured impedance. Impedance measure
`ment-based matching control Suffers from an inherent dis
`advantage. For example, a change in shunt capacitance to
`correct or modify the real component of the impedance
`results in an undesirable change in the imaginary component
`of the impedance. Similarly, a change in the series capaci
`tance or frequency to correct or modify the imaginary
`component of the impedance results in an undesirable
`change in the real component of the impedance. The matrix
`that relates the controlled variable vector (formulated by the
`real and imaginary components of the impedance) and the
`controlling variable vector (formulated by the shunt and
`series capacitance or the shunt capacitance and frequency) is
`non-diagonal. Impedance measurement-based control algo
`rithms are therefore not effective. Control algorithms based
`on the impedance formulated by using magnitude and phase
`measurements of the impedance are similarly ineffective.
`0011 Calibration methods for prior art systems calibrate
`the RF impedance analyzer or VI probe at the input of the
`electronic matching network. These calibration methods
`assume the power loss in the electronic matching network is
`fixed for all states of the electronic matching network and
`operating frequencies. However, the losses of the electronic
`matching network contribute significantly to the overall
`system operation.
`0012. Accordingly, a need therefore exists for improved
`methods and systems for controlling power Supplied to a
`dynamic plasma load and the losses associated therewith.
`0013 There is provided a system for delivering power to
`a dynamic load. The system includes a power Supply pro
`viding DC power having a substantially constant power
`open loop response, a power amplifier for converting the DC
`
`

`

`US 2007/O 1391 22 A1
`
`Jun. 21, 2007
`
`power to RF power, a sensor for measuring Voltage, current
`and phase angle between voltage and current vectors asso
`ciated with the RF power, an electrically controllable imped
`ance matching system to modify the impedance of the power
`amplifier to at lease Substantially match an impedance of a
`dynamic load, and a controller for controlling the electrically
`controllable impedance matching system. The system fur
`ther includes a sensor calibration measuring module for
`determining power delivered by the power amplifier, an
`electronic matching system calibration module for determin
`ing power delivered to a dynamic load, and a power dissi
`pation module for calculating power dissipated in the elec
`trically controllable impedance matching system.
`0014. In one embodiment, the electrically controllable
`impedance matching system can include an inductor, a
`capacitor in series with the inductor, and a plurality of
`switched capacitors in parallel with the dynamic load. The
`inductor can be a multiple tap-type inductor or a variable
`type inductor. Each of the plurality of switched capacitors
`can be in series with a Switch and an additional capacitor. In
`another embodiment, the electrically controllable imped
`ance matching system can include a capacitor, and a plural
`ity of Switched capacitors in parallel with the dynamic load,
`wherein each of the plurality of capacitors is in series with
`a Switch and an additional capacitor. In yet another embodi
`ment, the electrically controllable impedance matching sys
`tem can control the frequency of the impedance matching
`between the power amplifier and the dynamic load.
`0015. In one embodiment, the controller can control the
`electrically controllable impedance matching system for
`simultaneous control of conductance and Susceptance asso
`ciated with the impedance between the power amplifier and
`the dynamic load. In another embodiment, the controller can
`simultaneously control RF power frequency, RF power
`magnitude and the impedance between the power amplifier
`and the dynamic load. In yet another embodiment, the
`controller can control the electrically controllable imped
`ance matching system for regulating conductance and Sus
`ceptance to setpoints that stabilize an unstable dynamic load.
`0016. The power dissipated in the electrically control
`lable impedance matching system is the difference between
`the power delivered by the power amplifier and the power
`delivered to the dynamic load. The power delivered to the
`dynamic load is a sum of the power delivered to a resistive
`load and the power dissipated inside the load simulator.
`0017. The sensor calibration measuring module cali
`brates the sensor into a resistive load, wherein the resistive
`load is 50 C2. The electronic matching module calibrates an
`output of the electrically controllable impedance matching
`system into a load simulator. The load simulator can be an
`inverse electrically controllable impedance matching sys
`tem. The electronic matching system calibration module can
`include a power meter calibration module for determining
`power delivered to a resistive load; and a load simulator
`calibration module for determining power dissipated inside
`the load simulator. The resistive load can be 50 C2. The radio
`frequency power delivery system provides at least the fol
`lowing advantages over prior art systems. The system can
`enhance power setpoint regulation, impedance matching,
`and load disturbance mitigation using high-speed (e.g., in
`excess of 50 kHz in one embodiment) digital multi-input
`multi-output (MIMO) control. The system can operate in the
`
`presence of transient changes in plasma load properties and
`under conditions involving fast plasma stabilization. The
`system can provide a RF power delivery system that is
`robust to transients during startup of the system. The system
`can provide a high power step-up ratio, wherein the high
`power step-up ratio is 100 (e.g., 15 W to 1500 W). The
`system can measure power delivered to the load connected
`to the output of the integrated generator system. The system
`can allow for regulation of power that is independent of the
`power loss variation associated with the state/value of
`various controlled variables. The system can eliminate the
`need for recipe-based calibration for plasma loads.
`
`BRIEF DESCRIPTIONS OF THE DRAWINGS
`0018. The foregoing and other objects, features and
`advantages of the invention will be apparent from the
`following more particular description of preferred embodi
`ments of the invention, as illustrated in the accompanying
`drawings. The drawings are not necessarily to Scale, empha
`sis instead being placed upon illustrating the principles of
`the invention.
`0.019
`FIG. 1A is a diagram of an RF power delivery
`system having a two-axis tunable matching network accord
`ing to the prior art;
`0020 FIG. 1B is a diagram of an RF power delivery
`system having a fixed matching network according to the
`prior art;
`0021
`FIG. 1C is a diagram of an RF power delivery
`system having an integrated RF generator-impedance
`matching network according to the prior art;
`0022 FIG. 2 is a module-based diagram of the On
`Chamber RF power delivery system;
`0023 FIG. 3 is a plasma stability graph;
`0024 FIG. 4 is one embodiment of a fast DC bus of FIG.
`2:
`FIG. 5 is one embodiment of an RF impedance
`0.025
`analyzer or VI Probe of FIG. 2
`0026 FIG. 6 is one embodiment of an electronic match
`ing network of FIG. 2;
`0027 FIG. 7 is one embodiment of a module-based
`diagram of a DSP compensator board of FIG. 2;
`0028 FIG. 8 is a block diagram for calibrating the
`On-Chamber RF power delivery system;
`0029 FIG. 9A is one embodiment for calibrating a power
`meter to a 50 C2 calorimeter power reference;
`0030 FIG.9B is one embodiment for calibrating a load
`simulator to a DC power reference:
`0031 FIG. 9C is one embodiment for calibrating an RF
`impedance analyzer into a 50 C2 load; and
`0032 FIG. 9D is one embodiment for calibrating power
`delivered into the load simulator.
`
`DETAILED DESCRIPTION
`0033 Generally, an integrated radio frequency (RF)
`power delivery system is provided for dynamic load appli
`cations (e.g., inductive and/or capacitive plasma load). FIG.
`
`

`

`US 2007/O 1391 22 A1
`
`Jun. 21, 2007
`
`2 is an illustration of the integrated radio frequency (RF)
`power delivery system 200. Representative functional mod
`ules of the integrated system 200 include a fast DC bus 210,
`an RF power amplifier (“PA) 220, a digital signal processor
`(“DSP) compensator board 230, an RF impedance analyzer
`or VI probe 240, and an electronic matching network 250.
`The system 200 is coupled to a plasma load 260. It should
`be understood by one skilled in the art that the integrated
`system 200 can be implemented for a wide range of resistive
`and reactive loads.
`0034 Generally, the fast DC bus 210 delivers DC power
`to the power amplifier 220. The power amplifier 220 con
`verts the DC power from the fast DC bus 210 to an RF power
`at a frequency. The electronic matching system 250 switches
`shunt capacitors (not shown) to match the impedance
`between the power amplifier 220 and the plasma load 260 to
`facilitate stable and maximum power transfer from the
`power amplifier 220 to the plasma load 260. The DSP
`compensator board 230 controls the operation of the system
`200 based on measurements received from the fast bus
`controller 212 and RF impedance analyzer 240. The RF
`impedance analyzer 240 measures the RMS voltage, RMS
`current, and phase angle between the RF voltage and current
`vectors. Based on these measurements, relevant RF param
`eters are computed by the DSP compensator board 230.
`These parameters include, but are not limited to impedance
`vector Z, admittance vector y, delivered power P, and
`voltage-standing wave ratio (“VSWR). Typical operations
`of the DSP compensator board include power setpoints
`through the fast bus controller 212, RF power frequency
`setpoints through the power amplifier driver 222, and
`Switching frequency through the electronic match controller
`252.
`0035) In one aspect, the system 200 achieves simulta
`neous power and impedance regulation. Independent Sus
`ceptance regulation allows for the implementation of a
`frequency control algorithm based only on the deviation of
`the conductance from the conductance setpoint. As a result,
`both control loops can be operated simultaneously and at
`high-speed resulting in improved robustness. Further, well
`known instabilities for electronegative plasmas at low-pres
`sure (e.g., SF6 at 5 mT at 300 Was illustrated in FIG. 3) can
`be stabilized by setting arbitrary conductance and Suscep
`tance setpoints in conjunction with operation of the Fast DC
`bus 210.
`0.036
`FIG. 4 is a diagram of a partial resonant inverter
`power supply type fast DC bus 210. The fast DC bus 210
`provides process stability due to its associated constant
`power open loop response. The fast DC bus 210 improves
`FET utilization over the entire load space which results in
`more power being delivered to the load with the same PA
`220 (FIG. 2). The fast DC bus 210 has a fast response rate
`allowing it to deliver increased power to the plasma So it
`does not extinguish while also allowing the flexibility to
`reduce the bus voltage to ensure the FETs on the PA 220
`operate in a safe mode. Other types of topologies can for the
`fast DC bus 210 can be used. See for example, co-pending
`continuation-in-part application its parent U.S. application
`Ser. No. 10/947,397 filed Sep. 22, 2004, the entire teaching
`of each application are herein incorporated by reference.
`0037. In one embodiment, the fast DC bus can be a partial
`resonant inverter 210 that includes a pair of switches (MOS
`
`FETs) 302a, 302b, an inductor (L) 306, a capacitor (C) 308,
`and four diodes 310a, 310b, 310c, and 310d. In operation,
`the partial resonant inverter 300 converts the input voltage
`into a square wave or other known type DC wave form. The
`square wave is passed through the inductor 306 and capaci
`tor 308, the combination of which form an LC filter,
`clamped by the diodes 310c. 310d, coupled and rectified by
`a transformer rectifier 304 and filtered to obtain a desired DC
`voltage (power setpoint). The DC power setpoint is provided
`from the DSP compensator board 230 (FIG. 2). The desired
`impedance setpoint can be specified in terms of its vector
`inverse (referred to as admittance) and which constitutes
`simultaneous regulation of conductance to an arbitrary con
`ductance setpoint and regulation of Susceptance to an arbi
`trary Susceptance setpoint. The output of the partial resonant
`inverter 300 (DC-DC converter) is connected to DC input of
`the RF power generator/amplifier 220.
`0038. In operation, the capacitor 308 is periodically
`charged to an input rail voltage (+Vin) and discharged while
`the capacitor current is passed via the plasma load 260 (FIG.
`2). Every charge or discharge cycle, the energy deposited in
`the resistive load is equal to CV/2, independent of load
`resistance. Thus, the power is equal to FswxCV/2, where
`F is the Switching frequency and V is the input voltage. The
`inductor 306 ensures that the capacitor 308 is fully charged
`and discharged in finite time. One advantage of the partial
`resonant inverter 300 design is the ability to control the
`output voltage by varying either V or/and Fsw.
`0.039
`FIG. 5 is a diagram of one embodiment of an RF
`impedance analyzer or VI Probe 240. The VI Probe 240
`includes a DC power Supply 242, an analysis board assem
`bly 244, and a probe head assembly 246. The analysis board
`assembly 244 receives low-level RF signals from the probe
`head assembly 246. The probe head assembly 246 provides
`two Voltage outputs: 1) a Voltage representation of the time
`varying electric field present in the probe head assembly 246
`(voltage signal); and 2) a Voltage representation of the time
`varying magnetic field present in the probe head assembly
`246 (current signal). The analysis board assembly 244
`receives and processes the two voltage outputs of the probe
`head assembly 246 and outputs the RF parameters to the
`DSP compensator board 230 (FIG. 2). MKS Instruments,
`Inc. VI-Probe-4 100 and VI-Probe-350 are exemplary ana
`lyZers that can be used for this purpose.
`0040 FIG. 6 is a diagram of one embodiment of an
`electronic matching network 250. In one embodiment, the
`electronic matching 250 includes an inductance 254 in series
`with the load 260 (e.g., a compact inductor with multiple tap
`points), a fixed or variable series-padding capacitor 252, and
`field effect transistors (“FETs') 256a . . . 256n that switch
`one or more upper capacitors C(i) 258a . . . 258n to a
`corresponding lower capacitor C(i) 258a' . . . 258n', which
`is terminated to ground. In some embodiments, the elec
`tronic matching 250 network does not include the induc
`tance 254 in series with the load 260. Other types of
`electronic matching networks can be used. See for example,
`U.S. Pat. No. 6,887,339, the entire teaching of which is
`herein incorporated by reference.
`0041
`FIG. 7 shows a module-based diagram of a DSP
`compensator board 230. The DSP compensator board 230
`incorporates both a digital signal processor (“DSP) and a
`field programmable gate array (“FPGA'), and together con
`
`

`

`US 2007/O 1391 22 A1
`
`Jun. 21, 2007
`
`trols the entire integrated system 200. The DSP compensator
`board includes an admittance compensation module 232, a
`frequency control module 234, an electronic match control
`module 236, an RF power computation module 237, and an
`RF power control module 238. Generally, the DSP compen
`sator board receives the output from the VP probe 240. The
`admittance computation module 232 uses the VI probe
`outputs to calculate the admittance of the system 200. The
`frequency control module 234 uses the admittance to vary
`the frequency of the power amplifier 220. The electronic
`match control module 236 uses the admittance to switch the
`FETs 256 of the electronic matching network 250 on or off.
`The RF power computation module 237 uses the VI probe
`outputs to calculate the RF power of the system 200. The RF
`power control module 234 uses the RF power computation
`to regulate the power supplied from the fast DC bus power
`210. A more detailed description of the operation of the
`system 200 is set forth below.
`0042. One embodiment of the power regulation objective
`and algorithm is set forth below: The objective is to regulate
`the delivered power Pie to a user-defined setpoint P. To
`ensure Smooth transitions, trajectory generators are used. In
`one embodiment, a first-order trajectory is generated as
`follows:
`
`P = (Pin-P
`, = , (t) - P.)
`
`EQN. 1
`
`where t, is the trajectory time constant and P, is the desired
`power trajectory. The delivered-power control algorithm, in
`terms of the change in power commanded to the Fast Bus,
`is given by the following relationship:
`
`where k and k are the proportional and integral gains,
`respectively.
`0.043 Admittance regulation objective: A normalized
`admittance vector is defined as follows: y=g+b where g is
`the normalized conductance and b is the normalized Sus
`ceptance. The impedance matching control objective is
`formulated as follows: g->g, and b->b, where g and b,
`are arbitrary setpoints selected to improve plasma stability.
`The above objective is reinterpreted in terms of impedance
`by noting that impedance is defined as the reciprocal of
`admittance, according to the following relationship:
`
`--
`
`EQN. 3
`
`where Z is the normalized impedance, r and X are the
`resistance and reactance, respectively, Zo=Ro-0 denotes a
`nominal RF amplifier characteristic impedance. It follows
`that wheng->1 and b->0, we obtain R->R and X->0.
`0044 Admittance regulation algorithm: The frequency
`control loop is designed by using conductance measure
`ments, for example, as a PI control algorithm as follows:
`
`where ker and kit are scalar proportional and integral control
`gains. The shunt capacitance control loop is designed by
`using conductance measurements, for example, as a PI
`control algorithm as follows:
`
`where des and k are scalar proportional and integral control
`gains.
`0045. In operation, referring now to FIGS. 2, 3 and 6.
`after the user provides a non-zero setpoint, the trajectory
`generator and the power and admittance control algorithms
`are simultaneously activated and executed. The VI probe
`240 provides analog signals proportional to the RF voltage
`and RF current, which are synchronously sampled by the
`analog-to-digital converters, sent to a mixer and CIC filter
`(not shown) and ultimately sent through a calibration matrix
`to yield RF voltage and RF current measurements given by
`the following relationships:
`
`EQN. 6
`l=Itil,
`where V, Idenote vector representations of the instantaneous
`RF Voltage and current, respectively, and Subscripts r and i
`are used to denote the Scalar values of the real and imaginary
`components.
`0046) The average delivered power is computed as fol
`lows:
`
`. . .
`.
`Piet = Re(VI } = VI, + VI;
`
`EQN. 7
`
`where Re} denotes the real component of the vector, and
`SuperScript * is used to denote the complex conjugate of the
`Vector.
`0047 The admittance vector Yis then computed as fol
`lows:
`
`(I, V, +IV)
`I
`W V2 + V?
`
`(I; V - IV) - E G + i B
`V2+ V2
`
`EQN. 8
`
`where the conductance G and the Susceptance B are real and
`imaginary components of the admittance Y.
`0048. The normalized conductance g and nonnalized
`Susceptance b are computed as follows:
`
`(IV, +IV)
`g = Zo G = Zo
`-
`V2+ V?
`
`and
`
`(I; V - IV)
`b = ZoB = Zn
`-
`O V2 + V?
`O
`
`EQN. 9
`
`where Zo denotes the characteristic impedance of the RF
`amplifier. The measurements of P. g., b are respectively
`sent to the control algorithms for P
`find Cend respec
`tively.
`
`cmd
`
`tc.
`
`

`

`US 2007/O 1391 22 A1
`
`Jun. 21, 2007
`
`0049. The electronic match controller 252 switches the
`FETs 256 (FIG. 6) thereby switching the shunt capacitors
`258 to match the impedance between the power amplifier 20
`and the dynamic load 260. The absence of moving mechani
`cal parts leads to higher reliability. In one embodiment, the
`step response of the system 200 is faster than about 1 ms
`because the speed of the response is governed by the
`electronics and not by the mechanical response.
`0050. A change in frequency results in a change in both
`the conductance and the Susceptance. However, for an
`integrated System without transmission line cables, a change
`in shunt capacitance results only in a change in the Suscep
`tance and does not affect the conductance value. Thus, the
`matrix that relates the controlled variable vector (formulated
`by the real and imaginary components of the admittance)
`and the controlling variable vector (formulated by the shunt
`and series capacitance or the shunt and frequency) is trian
`gular. As a result, independent Susceptance regulation is
`achieved by varying the shunt capacitance.
`0051
`Independent susceptance regulation allows for the
`implementation of a frequency control algorithm based only
`on the deviation of the conductance from the conductance
`setpoint. As a result, both the conductance-based frequency
`control loop and the Susceptance-based shunt capacitance
`control loop can be operated simultaneously and at high
`speed, resulting in improved robustness.
`0.052
`FIG. 8 is a block diagram 300 of a method for
`determining the power dissipated (loss) in the electronic
`matching network 250 (FIG. 2) to improve the efficiency of
`the system 200. Step one (310), a power meter 314 (FIG.
`9A) is calibrated into a 50 C2 calorimeter power reference to
`determine the power delivered to the 50 C2 load. Step two
`(320), a load simulator calorimeter 332 (FIG. 9B) is cali
`brated to a DC power reference to determine the power
`dissipated inside a load simulator 342 (FIG. 9D). Step three
`(330), the VI probe 240 (FIG. 2) is calibrated into a 50 C2
`load to determine the power delivered by the power ampli
`fier 220 (FIG. 2). Step four (340), the output of the system
`200 is calibrated into the load simulator 342 to determine the
`power delivered to Z =R+X. Step 5 (350), the power
`dissipated in the electronic matching system is calculated by
`difference between the he power delivered by the power
`amplifier 220 and the power delivered to. Z=R+X.
`0053 FIG.9A is detailed implementation diagram of step
`310 for calibrating the power meter 314. A calorimeter 322
`is coupled to the output of the VI Probe 240, RF power is
`applied from the power amplifier 220, and the power meter
`314 is calibrated. Calorimetry is the measurement of thermal
`losses. It is implemented by thermally insulating the 50 C2
`load in the calorimeter (322) to prevent ambient thermal
`losses and measuring the flow rate and the temperature rise
`of the cooling water. The power meter is calibrated to the
`power dissipation in the load computed by
`
`(in
`(in
`Q =
`C(Tout -T), where t
`
`denotes the mass flow rate, C denotes the specific heat of
`water, and T. T. denote the inlet and outlet temperatures,
`respectively. A comp