as) United States
`a2) Patent Application Publication co) Pub. No.: US 2006/0083331 Al
`(43) Pub. Date: Apr. 20, 2006
`
`Kaczynski
`
`US 20060083331A1
`
`(54) METHOD AND APPARATUS FOR A
`TRANSCEIVER HAVING A CONSTANT
`POWER OUTPUT
`
`(76)
`
`Inventor: Brian J. Kaczynski, Los Altos, CA
`(US)
`
`Correspondence Address:
`BEVER HOFFMAN & HARMS, LLP
`TRI-VALLEY OFFICE
`1432 CONCANNONBLVD., BLDG. G
`LIVERMORE, CA 94550 (US)
`
`(60) Provisional application No. 60/258,150,filed on Dec.
`22, 2000.
`
`Publication Classification
`
`(51)
`
`Int. CL
`(2006.01)
`HOAL 25/49
`(52) U.S. CMe
`aecsccsstcssssesssssstensssesseoesteneseasesevesee 375/297
`
`(57)
`
`ABSTRACT
`
`The present invention includesa transceiver and a method of
`operating the same that includes in the transmitter a power
`(21) Appl. No.:—11/292,607
`control circuit that operates on an analog differential signal
`containing data packets individually. The power control
`circuit
`initially transmits a series of data symbols with
`known values, periodically strobes the transceiver system
`for correct power levels and incrementally increases the
`power level of the transceiver until
`the optimal gain is
`reached, without exceeding the maximum output power.
`
`(22)
`
`Filed:
`
`Dec. 1, 2005
`
`Related U.S. Application Data
`
`(62) Division of application No. 09/927,425, filed on Aug.
`10, 2001.
`
`
`
`
` Dual
`
`Power
`Detector
`
`Control
`182
`Circuit
`190
`
`APPLE 1029
`
`APPLE 1029
`
`1
`
`

`

`Patent Application Publication Apr. 20,2006 Sheet 1 of 4
`
`US 2006/0083331 A1
`
`Jaxiudy) 061
`
`
`
`JoU0d
`
`]‘sq
`
`YOM
`
`2
`
`
`

`

`Patent Application Publication Apr. 20,2006 Sheet 2 of 4
`
`US 2006/0083331 Al
`
`
`
`Wait for Tx_on
`gainSelect=1
`steadyState=0
`
`PC_RESET
`202
`
`
`
`PC_PAOFF
`204
`
`
`
`PC_INIT_WAIT
`206
`
`timerDone/
`timerReset=1
`
`PC_COMP_WAIT
`208
`
`PC_IDLE
`205
`
`
`
`PC_SUBWAIT
`powerLow/gain=gain-gain_step
`210
`timeSelect=T_SUB
`
`timerReset=0
`
`
`
`If (ovr|gainSelect) gain = Gain_1
`PA_on & ovr
`Else if (steadyState) gain = Gain_F — gain_rev
`
`
`
`
`PA_on & -ovr/timeSelect=T_INIT
`timerReset=0
`
`
`time4uReset=0
`
`
`gainSelect=0
`
`
`
`
`timerDone & -time4uDone/timerReset=1
`
`
`
`
`-powerLow/gain_F=gain
`steadyState=1
`
`
`PC_IDLE
`212
`
`time4uDone|(gain=MAXGAIN)/gainF=gain
`
`-PA_on
`
`PC_PAOFF
`214
`
`Fig. 2
`
`3
`
`

`

`Patent Application Publication Apr. 20,2006 Sheet 3 of 4
`
`US 2006/0083331 Al
`
`116
`
`V3
`
`DN
`Mixer
`nwLL_|
`VV
`
`DP
`
`inp
`
`inn
`
`1g
`
`122
`
`120
`
`124
`
`v2
`
`Vi
`
`\/
`
`on
`
`op
`
`Fig. 3
`
`4
`
`

`

`Patent Application Publication Apr. 20,2006 Sheet 4 of 4
`
`US 2006/0083331 A1
`
`
`
`
`
`
`
`
`
`Dat—rit3(aa
`
`bel
`
`dut
`
`eeeeeeececeteeeceteeeeeeeeeeeOeeeeeeeeeeeeeeesseeHenewaenay
`
`
`
`
`
`
`
`TOOTTOUREDibPePDnOMENImg|cwweeewwwmenemasmeencecewnceeenecesseeneemcccccccemecees|baceecereccscesweetenwnaceeenemanmmeneceeeescsesseed‘ameaeenmcencncccccccercewnenessessensseceneanaanaanees
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`5
`
`
`
`

`

`US 2006/0083331 Al
`
`Apr. 20, 2006
`
`METHOD AND APPARATUS FOR A
`TRANSCEIVER HAVING A CONSTANT POWER
`OUTPUT
`
`RELATED APPLICATIONS
`
`[0001] This application is related to and claimspriority to
`US. Provisional Patent Application No.: 60/258,150 filed on
`Dec. 22, 2000 and is a divisional of U.S. patent application
`Ser. No.: 09/927,425 filed on Aug. 10, 2001.
`FIELD OF THE INVENTION
`
`invention is generally related to a
`[0002] The present
`complementary metal oxide semiconductor (CMOS)trans-
`ceiver. More specifically, the present invention relates to a
`method and apparatus for achieving constant output power
`from a CMOStransceiver.
`
`BACKGROUND OF THE RELATED ART
`
`[0003] A transceiver is a well-knowncircuit, containing a
`transmitter and a receiver, which are thus capable oftrans-
`mitting and receiving communication signals, respectively.
`Conventionally, the transmitter contains a power amplifier
`(PA) that provides the last stage of amplification of the
`signal to be transmitted.
`
`In most conventional designs, the power amplifier
`[0004]
`is implemented as a componentthat is physically separate
`from other parts of the transmitter and/or transceiver. Also,
`power amplifier’s made from gallium arsenide (GaAs) or
`Silicon bipolar junction transistors (SiBJT) are typically
`used because they have an inherently higher breakdown
`voltage than transistors made in a CMOScircuit, whether
`the transistors are n-channel or p-channel transistors. While
`such designs allow for a power amplifier that has the desired
`amplification characteristics, they do so at the expense of
`cost. Not only is a GaAs, SiBJT or other non-CMOSpower
`amplifier costlier than a transistor in a CMOSintegrated
`circuit, but
`the non-CMOS power amplifier cannot be
`formed on the same integrated circuit chip as the compo-
`nents of the transmitter and/or transceiver. Both of these
`
`factors add to the overall cost of the resulting transceiver.
`
`It has been recognized that it would be beneficial to
`[0005]
`have a transceiver in which most of the transmitter and
`receiver circuits are on a single chip, including the power
`amplifier. For example, in the article entitled 4 Single Chip
`CMOSDirect-Conversion Transceiverfor 900 MHz Spread
`Spectrum Digital Cordless Phones by T. Choet al. that was
`presented at the 1999 IEEE International Solid State Circuits
`Conference, there is described a CMOStransceiver chip that
`includes an integrated power amplifier. An improved CMOS
`power amplifier is also described in the application entitled
`CMOS TRANSCEIVER HAVING AN INTEGRATED
`
`POWER AMPLIFIER,bearing application Ser. No. 09/663,
`101, filed on Sep., 15, 2000 and assigned to the same
`assignee as the assignee of the invention described herein,
`which recognizes the advantage of integrating the power
`amplifier.
`
`[0006] Nevertheless, a major disadvantage of CMOS
`power amplifiers is that they exhibit a wide range of power
`levels variation due to their sensitivity to thermal and
`process variations. High efficiency and constant powerlev-
`els in CMOSpower amplifiers is impeded by the technolo-
`gies low breakdown voltage, low current drive, and lossy
`substrate.
`
`transmitter designs
`conventional
`[0007] Furthermore,
`operate so that the output poweris transmitted based upon
`a function of many different variables. In a CDMA envi-
`ronment, for example, the power output of a mobile trans-
`mitter will typically be based uponthe distance between the
`mobile transmitter and the base station currently in used. In
`such an environment, the output power will increase, for
`example, if the mobile transmitter travels closer to the base
`station. In operation, the gain of a variable gain amplifier
`that is part of the transmitter, at either the intermediate
`frequency (IF) or radio frequency (RF) stage, will be
`changed to thereby lowerthe transmit output power.In this
`situation, while the output power may becometoo large for
`a period of time, that is acceptable within the overall system
`requirements.
`
`In other environments, however, it is required, by
`[0008]
`for instance the Federal Communication Commission, that
`the output power must not exceed a pre-specified level at any
`time. In such an environment, the above-described design
`cannot be used. Since in order to take into account instances
`
`in which powerwill exceed the pre-specified maximum,the
`average output power must be much lower than that maxi-
`mum, which degrades system performance to an unaccept-
`able level.
`
`[0009] Accordingly, a transmitter containing a variable
`gain amplifier and a power amplifier integrated with a
`CMOStransceiver chip that overcomes the above disadvan-
`tage would be desirable.
`
`SUMMARY OF THE INVENTION
`
`It is an object of the present invention to provide a
`[0010]
`transmitter on an integrated CMOStransceiver chip that
`provides a substantially constant power output.
`
`It is another object of the invention to provide an
`[0011]
`apparatus and methodthat allows for gradual step increases
`in output power to avoid the output power from exceeding
`a predetermined maximum output power.
`
`It is still another object of the present invention to
`[0012]
`provide a method of operating a variable gain amplifier
`using the sensed output power to determine whether to
`back-off from the current output power.
`
`It is a further object of the present invention to
`[0013]
`provide a variable gain amplifier including a folded cascode
`stage and/or a mirroring array of unit gain cells.
`
`[0014] The above objects of the present invention, among
`others, either alone or in combination are achieved with a
`transceiver and a method of operating the samethat includes
`in the transmitter a power control circuit that operates on an
`analog differential signal containing data packets individu-
`ally. The powercontrol circuit initially transmits a series of
`data symbols within each packet with known deterministic
`values, periodically strobes the transceiver system for cor-
`rect power levels and incrementally increases the power
`level of the transceiver until the optimal gain is reached,
`without exceeding the maximum output power.
`
`circuit
`control
`the power
`specifically,
`[0015] More
`that are
`indicating the output power
`receives
`signals
`obtained from a power detector and comparator combina-
`tion, and based upon the level of the received signals will
`accordingly adjust the variable gain amplifiers. During ini-
`
`6
`
`

`

`US 2006/0083331 Al
`
`Apr. 20, 2006
`
`tial operation, the gain of the variable gain amplifier will be
`set to a predetermined, preferably user-configurable, initial
`gain when transmitting the first symbol in thefirst transmit-
`ted packet. After an appropriate wait time to ensure that the
`variable gain amplifiers stabilize, and a correspondingly
`accurate output poweris achieved, the powercontrol circuit
`strobes the comparator to receive a signal indicating the
`output power while that symbol is being transmitted. If the
`output power, and therefore the gain, is too low, the power
`control circuit will repeatedly increment the gain in order to
`reach but not exceed the predetermined maximum output
`power.
`
`[0016] Once achieved, output power is prevented from
`exceeding the predetermined maximum by decreasing the
`gain by a predetermined amountat the beginning oftrans-
`mission of each subsequent packet,so that the output power
`can be lowered by an amount correspondingto the decreased
`gain on a per-packetbasis. Alternatively, the comparator can
`be strobed during the training sequence of symbols within
`each packetat the results of the comparison used to back-off
`the output powerby at least one or maybe more steps during
`the transmission of the next packet.
`
`[0017] The variable gain amplifier implemented allows for
`the power control circuit to change the gain in small incre-
`mental steps, thereby allowing the power control algorithm
`implemented by the power control circuit to operate.
`
`BRIEF DESCRIPTION OF THE DRAWINGS
`
`[0018] The above, and other objects, features, and advan-
`tages of the present invention are further described in the
`detailed description which follows, with reference to the
`drawings by way of non-limiting exemplary embodiments of
`the present invention, wherein like reference numerals rep-
`resent similar parts of the present
`invention throughout
`several views and wherein:
`
`[0019] FIG. 1 illustrates a block diagram of an embodi-
`ment of the power control circuitry according to the present
`invention.
`
`[0020] FIG. 2 illustrates a state diagram of the power
`control algorithm according to the present invention.
`
`[0021] FIG. 3 shows the inductively-loaded folded-cas-
`code level-shift stage between the upmixer andthe interme-
`diate frequency variable gain amplifier of the power control
`circuit according to the present invention.
`
`[0022] FIG. 4 illustrates the variable gain amplifier of the
`powercontrol circuit according to the present invention.
`
`DETAILED DESCRIPTION OF THE
`PREFERRED EMBODIMENTS
`
`[0025] FIG. 1 illustrates a block diagram of an embodi-
`ment of a power control circuit 100. As shown in FIG.1, IF
`upmixer 110 upconverts signals received by the transceiver
`to an IF frequency as is known, for example a 1 GigaHertz
`IF Frequency and a 5 GigaHertz RF frequency. After the IF
`upmixer 130, the IF variable gain amplifier 130 which,in the
`preferred embodiment contains a 5 bit input control input
`and is configurable from 0 dB to 15.5 dB in steps of 0.5 dB,
`amplifies the IF signal. The amplified IF signal
`is then
`transmitted to an RF upmixer 178, which upconverts the IF
`signal to an RF signal.
`
`[0026] The gain ofthe variable gain amplifier 130 is based
`on several factors such as the die temperature and device
`process corner.
`
`[0027] The output from RF upmixer 178 is then supplied
`to the power amplifier 180, which amplifies the signal to be
`transmitted. To sense the powerof the transmitted signal, a
`powerdetector is employed. In the preferred embodiment, a
`dual matched power detector 182 is used as shown in FIG.
`1. One of the power detectors 182A is used to detect the
`transmitted signal, obtained from the radio frequency (RF)
`signal at the drain of the power amplifier 180, while the other
`powerdetector 182B is used to detect the reference signal.
`The reference signal is generated using a predetermined
`digital value that
`is used to create an analog signal of
`appropriate level using a digital to analog converter 186, as
`shown. Each of the power detectors 182 is essentially a
`source followercircuit biased at very low currents (200 nA)
`with large capacitive loads (4 pF). The outputs of the two
`matched power detectors are compared, and optimal power
`is reached when the power detector outputs match. The use
`of matched power detectors made from the same process
`results in an optimal powerthat is as independent of tem-
`perature and process as is possible.
`
`[0028] The outputs from each of the powerdetectors 182A
`and 182B are supplied to a comparator 188, which is strobed
`at appropriate intervals, described hereinafter, and the dif-
`ference between the transmitted signal and the reference
`signal obtained from the comparator 188 is input to power
`control circuit 190. As described further hereinafter with
`
`respect to FIG. 2, the power control circuit 190 is used to
`achieve and maintain a steady state operation, such that on
`a packet-to~packet basis the gain settings of the variable
`gain amplifier 130 desirably results in a substantially con-
`stant output power. While the variable gain amplifier 130 is
`described hereinafter as a single gain stage containing
`numerous gain cells, it is noted that a number of variable
`gain amplifiers,
`in both IF and RE transmitter portions,
`could instead be used, with the composite gain then being
`determined and used by the power control circuit 190 as
`described herein.
`
`[0023] Reference will now be madein detail to preferred
`embodiments of the invention, examples of which areillus-
`trated in the accompanying drawings. Wherever convenient,
`the same reference numbers will be used throughout the
`drawings to refer to the sameorlike parts.
`
`to those
`[0024] Other embodiments will be apparent
`skilled in the art from consideration of the specification and
`practice of the invention disclosed herein.It is intended that
`the specification and examples be considered as exemplary
`only, with a true scope of the invention being indicated by
`the following claims and equivalents.
`
`[0029] The powercontrol circuit 190 is preferably imple-
`mentedas a finite state machine executing the power control
`algorithm as described herein, which is preferably a hard-
`ware-based logic. Using such a power control circuit 190
`allows operation at a power level which is close to but will
`not exceed the maximum output powerthat can be trans-
`mitted for any given packet by the system andstill be within
`the FCC power requirements.
`
`[0030] Each of the components described above is pref-
`erably made on the sameintegrated circuit chip. Also, while
`the output power detection circuitry is described as being
`
`7
`
`

`

`US 2006/0083331 Al
`
`Apr. 20, 2006
`
`implemented with power detectors 182A and 182B, digital
`to analog converter 186, and the comparator 188, other types
`of circuit elements can be used to detect the output power.
`
`[0031] FIG. 2 illustrates a state diagram of the power
`control algorithm according to the present invention, which
`will be used to describe the operation of the gain control
`process in moredetail.
`
`[0032] After a reset signal is received by the power control
`circuit 190, in step 202, the power control circuit 190 waits
`for TX_onsignal indicating that a transmission is to begin.
`Atthis time a gainSelect flag is set to “1” indicating that a
`default initial gain value will be used and a steadyState flag
`is set to “O”indicating that steady state operation has not yet
`been achieved.
`
`[0033] Once the TX_onsignal is received indicating that
`transmission is desired, step 204 follows, in which a deter-
`mination is made whether a circuit override operation is
`desired, which will typically occur during burn-in testing. In
`this case, the gainSelect flag is maintained at the “1” state,
`thus indicating that a preset burn-in gain value should be
`used. Accordingly, step 205 will follow if the override
`operation takes place, and the burn in gain will be applied to
`the upmixer 110 and the variable gain amplifier 130 once the
`power amplifier 180 is on until the burn-in test operation is
`complete.
`
`Jf, however, in step 204 a normal operation mode
`[0034]
`occurs, then gainSelect flag will be set to “O”, and the gain
`used will set the variable gain amplifier 130 to a normal
`mode operation initial gain value. In this state, the first few
`data symbols, suchas the first eight, that are transmitted will
`preferably have known, deterministic initial values,
`thus
`allowing the power control circuit to achieve a steady state
`condition more accurately. In a preferred embodiment, gain
`control is only performed on someinitial numberofthe first
`few data symbols, such as 5, so that the remaining symbols
`having deterministic values can be used for automatic gain
`control (AGC) in the receiver which is receiving the trans-
`
`mitted signal. Further, in the preferred embodiment, each
`
`
`
`
`symbolis 0.8C)s
`long, such that if gain changes occur during
`
`
`
`the first 5 symbols this provides 40s for obtaining the
`appropriate gain, and each packet is about 1 msin duration.
`
`[0035] Oncethe gain setting for the VGA 130 is obtained
`and the power amplifier 180 becomes turned on, then the
`normal mode operation initial gain value will be used to
`initially operate the variable gain amplifier 130 for the
`remainder of the packet transmission. An initial wait step
`206 then follows, and allows the system to settle at this
`initial gain value. Theinitial wait time can be predetermined,
`but will typically be longer than the wait time that is used
`between gain steps as described hereinafter.
`
`the power control
`time,
`initial wait
`[0036] After that
`circuit 190 enters into its normal mode loop that is used to
`reach an appropriate steady state gain. In step 208, the output
`power is checked by strobing the comparator 188, and if it
`is low is adjusted by increasing the gain by a gain step.
`
`the gain may be
`In the preferred embodiment,
`[0037]
`increased in 0.5 dB increments, although in early steps if the
`gain is lower than the desired gain by some predetermined
`threshold, steps as large as 2.0 dB can beinitially used. No
`matter what gain steps are used, however, an important
`aspect of the present invention is that an individual gain step
`
`will not cause the overall powerof the transmitted signal to
`exceed a predetermined maximum value, which value will
`typically correspond to FCC regulations, as noted above.
`Also during step 208, the next wait interval is set, which
`again can be up to 2 us in 62.5 ns steps in the preferred
`embodiment.
`
`Steps 208 and 210 represent the core of the power
`[0038]
`control algorithm. In step 208, the power control comparator
`is strobed, and it is determined whether or not the output
`poweris too low. If the power is too low, the algorithm
`increases the gain setting by one increment, and proceeds to
`step 210, the post-gain~change wait period. Steps 208 and
`210 subsequently repeat, with the gain increasing, until
`either (1) the time duration allocated for changing the gain
`expires or (2) the optimal output power is reached. When
`either of the two aforementioned conditions is met,
`the
`algorithm enters the hold state 212 and the existing gain
`setting is held for the duration of the packet.
`
`If condition (1) is met, meaning that the optimal
`[0039]
`output power was not reached, then the remainder of the
`packet will be transmitted at the then current gain setting
`until data corresponding to the next packet is ready for
`transmission. At the end of the transmission time of that
`packet, the power amplifier 180 is turned off until the next
`packet is ready to be transmitted. At the time for transmis-
`sion of the next packet the algorithm enters step 204 and the
`gain setting reached in the previous packet will be used as
`the initial gain setting for the new current packet. When the
`next packet is to be transmitted, the gainSelect flag, which
`had been set to “1, “is set to “O” indicating that the default
`initial gain value will not be used, but that the initial gain
`value will be the last gain value from the previous packet.
`This packet-to-packet cycle of increasing the power will
`repeat until the optimal power is reached. When the optimal
`poweris reached, the steadyState flag is set to “1”, enabling
`the power control circuit 190 to reduce the powerif neces-
`sary, as described below.
`
`[0040] Once the optimal gain setting is reached the present
`invention also includes a mechanism for reducing the gain
`setting if operating conditions so require, such as if a
`temperature variation causes an increase in output power. To
`accomplish this, during the PA-OFFstate 204, at the begin-
`ning of each packet following an “optimal-power” packet,
`the gain will be reduced by a user-configurable amount, such
`as 2 dB, and will be allowed to either (1) recover the same
`gain setting through the process of increasing the gain
`described above, or (2) will settle to a lower gain setting if
`operating conditions so require. In either case (1) or (2)
`above, the system should recover the same output power.
`Since the poweris already very close to optimal, this method
`ensures that the output power will be within a user config-
`urable step of the target output power. Alternatively, the
`comparator can be strobed during the training sequence of
`symbols within each packetat the results of the comparison
`used to back-off the output power by at least one or maybe
`more steps during the transmission of the next packet.
`
`[0041] Having described the operation of the power con-
`trol circuit 100, a further discussion will be providedrelating
`to certain of the circuits used in the power control circuit
`100.
`
`[0042] FIG. 3 shows the inductively-loaded folded-cas-
`code level-shift stage between the upmixer 110 and the
`
`8
`
`

`

`US 2006/0083331 Al
`
`Apr. 20, 2006
`
`variable gain amplifier 130 in more detail. The IF upmixer
`110, which will either take the basebandsignal to an IF level
`as described in the preferred embodiment, as well as the RF
`upmixer 178, which will take the IF signal to an RF level,
`can be formed using conventional techniques. The present
`invention provides, however, an inductively tuned level-
`shift stage at the output of the IF mixer 110. The differential
`output signal, shown as DP (positive) and DN (negative),
`that is output from the mixer 112, is transmitted through an
`inductively loaded folded cascode circuit. PMOStransistors
`118 and 120, with each gate thereof biased at a DC bias that
`will result
`in a fixed, predetermined DC drain current
`flowing through the PMOStransistors, complete the level-
`shift circuit at the output of mixer 110. The purpose of the
`level-shift block is to convert the VDD-referenced driver
`output of the upmixer circuit 110 to a ground-referenced
`signal suitable for driving an NMOScurrent mirror, as well
`as to convert the differential outputs of the upmixer circuit
`110 into low-impedance (current-mode) nodes,
`thereby
`making the upmixer circuit less sensitive to the quality
`factor (Q) of the tuned output load. Put another way, the
`level-shift block with the PMOS common gate stage oper-
`ates as a folded-cascode stage with unity current gain,
`redirecting AC current from the upmixer circuit 110 to
`ground.
`
`FIG.4 illustrates the variable gain amplifier 130 of
`[0043]
`the powercontrol circuit 100 in moredetail. Initially, certain
`of the blocks that make up the variable gain amplifier 130
`will be described. The variable gain amplifier includes an
`input current load block 132, a plurality of switch network
`blocks 142-1 to 142-n, and a correspondingplurality of gain
`cells 160-1 to 160-7. The gain cells 160 are each replicated,
`as described hereinafter, to allow the step increments in the
`gain, as mentioned above and described more fully below.
`
`[0044] The input current load block 132 of FIG. 4 will
`first be described in more detail, and contains NMOS
`transistors 134, 136, 138 and 140, with the gates of NMOS
`transistors 134 and 136 being biased bythe first DC voltage
`and which together function as cascode transistors for the
`current mirror transistors 138 and 140. The gate of each of
`transistors 138 and 140, and the drain of each oftransistors
`134 and 135, respectively receive the INP and INN input
`signals, which are output from the input current load block
`as signals GN and GP, as shown.
`
`[0045] The switch network 142 of FIG. 4 will next be
`described in more detail and contains PMOStransistors 144
`and 148, and NMOStransistors 146, and 150. Transistors
`144 and 146 operate as a pair and are used to switch the
`cascodevoltageat the gates of transistors 134 and 136 to the
`outer pair of transistors 162 and 170 of the gain cell block
`160, as described further hereinafter, whereas transistors 148
`and 150 operate as a pair and are used to switch the cascode
`voltage at the gates of transistors 134 and 136 to the inner
`pair of transistors 164 and 168 of the gain cell 160. Each of
`transistors 144 and 146 are switched based upon the POS_B
`input signal, whereas each of transistors 148 and 150 are
`switched based upon the NEG_B inputsignal. In operation,
`either one of POS_B or NEG_B maybeonat the sametime,
`but both will not be on at the sametime.It is also noted that
`PMOStransistors 144 and 148 have their bulk nodetied to
`
`their source nodes providing lower on-resistance, which
`improvestheir switch characteristics, and that the size of the
`transistors 144, 146, 148 and 150is fixed, and notrelated to
`
`the size of any other devices, unlike the transistors in the
`input current load block 132 and gain cell 160, which are
`chosen to mirror each other, as described further herein.
`
`[0046] Each gain cell 160, such as the gain cell 160-1 of
`FIG. 4, is essentially an NMOScurrent mirror, formed of
`transistors 162-172. Before further describing a gain cell
`160, it is noted thatthe current outputs from the gain cell 160
`mirror the current inputs INP and INN presented to the input
`current load block 132. The sizing of the transistors 162,
`164, 168 and 170 thus mirror the size of the transistors
`134-140 from the input current load block 132.
`
`[0047] With transistor 166 having its gate controlled by
`the GPsignal, and transistor 172 having its gate controlled
`by the GN signal, and possibly either transistors 162 and
`170, or 164 and 168 tured on, depending uponthestate of
`the POS_B and NEG_Bsignals, each gain cell is provided
`with two gain settings: a positive polarity setting and a
`negative polarity selling. In the positive gain setting, current
`from transistors 166 and 172 flows through transistors 162
`and 170, respectively,
`in a conventional current mirror
`configuration. In the negative gain setting, the drain outputs
`of the current mirrors are reversed, and current from tran-
`sistors 166 and 172 flows through transistors 164 and 168,
`respectively, resulting in a current mirror cell with the same
`AC gain, but opposite polarity.
`
`In operation, as noted, multiple ones of the gain
`[0048]
`cells 160 in the variable gain amplifier 130 will be connected
`in parallel such that two signals, GN and GP, drive the
`common GN and GPinputofall of the gain cells, and two
`outputs, OUTN and OUTP, will be driven by the common
`OUTNand OUTPoutputs of all of the gain cells. This type
`of parallel connection of multiple gain cells allows for small
`incremental gain steps. In operation, there should always be
`more“positively-connected”gain cells than “negativelycon-
`nected” gain cells, resulting in an overall positive configu-
`ration.
`
`[0049] Within the variable gain amplifier 130, a single
`gain cell 160', which is constructed the sameas the gain cell
`160 previously described is included. In operation, the gain
`cell 160' can be both “positively-connected” and “nega-
`tively-connected,” thereby allowing it to be placed in a
`neutral gain configuration and allowing for fine adjustments
`to be made by simply turning this gain cell on or off, and
`effectively allowing the gain incrementto be half of what it
`would be without this gain cell 160'. Thus, for example, if
`gains are stepped throughat 0.5 dB, 1.0 dB, 1.5 dB, 2.0 dB,
`2.5 dB, and 3.0 dB, the fine adjust cell will change state
`severaltimes. It should also be noted that each of the various
`POS_B and NEG_Bsignals is controlled by the power
`control circuit, which, as described, operates digitally.
`Accordingly, it will be appreciated that the relative size of
`each gain step can be precisely controlled, since each gain
`step may be a combination of both positively connected gain
`cells and negatively connected gain cells.
`
`[0050] While the present invention has been described
`herein with reference to particular embodiments thereof, a
`latitude of modification, various changes and substitutions
`are intended in the foregoing disclosure, and it will be
`appreciated that in some instances some features of the
`invention will be employed without a corresponding use of
`other features without departing from the spirit and scope of
`the invention.
`
`9
`
`

`

`US 2006/0083331 Al
`
`Apr. 20, 2006
`
`1. An upconversion and amplification apparatus for use in
`a transmitter that provides an amplified differential signal
`having substantially constant output power and containing,
`over time, data from a plurality of data packets, each data
`packet including a plurality of training symbols towards a
`beginning of the packet, the apparatus comprising:
`
`the
`a variable gain amplifier having a variable gain,
`variable gain amplifier inputting a differential signal
`and outputting an amplified differential signal;
`
`a poweramplifier electrically coupled to the variable gain
`amplifier, the power amplifier inputting the differential
`signal or a signal deriving from the amplified differen-
`tial signal and outputting a power amplified differential
`signal having an output power;
`
`a powerdetection circuit including a power detector that
`is coupled to the power amplification stage and pro-
`vides a detected power signal
`indicating the output
`powersuch that the detected powersignal is created at
`an interval that is at least once per symbol during some
`of the plurality of training symbols; and
`
`a powercontrol circuit electrically coupled to the power
`detection circuit, and the variable gain amplifier, the
`powercontrol circuit:
`
`increasing the variable gain of the variable gain amplifier
`in a step increment if the detected power signal, as
`detected during each of the some of the plurality of
`training symbols during each packet transmission, is
`lower than a predetermined power; and
`
`decreasing the variable gain of the variable gain amplifier
`in a step decrement if conditions require so that the
`detected powersignal will not exceed an output power
`threshold.
`
`2. An apparatus according to claim 1, wherein the power
`control circuit decreases the variable gain after the prede-
`termined power was reached once per packet, at the begin-
`ning of the packetprior to the transmissionofa first training
`symbol.
`3. An apparatus according to claim 2 wherein the power
`control circuit maintains the variable gain amplifier at a last
`gain established at the last of the training symbols used to
`increase the variable gain for the remainderofa first packet.
`4. The apparatus according to claim 1, wherein the power
`control circuit repeats the increasingthefirst variable gain in
`the step increment a maximum of a predetermined number
`of times for the input differential signal representing each
`packet of data.
`5. The apparatus according to claim 1 wherein the vari-
`able gain amplifier,
`the power amplifier and the power
`control circuit are formed on an integrated circuit chip.
`6. The apparatus according to claim 1, wherein:
`
`the power detection circuit includes:
`
`two matched power detectors, the power detector that
`produces a current powerdetection signal and