TITLE: RF TRANSMITTER AND METHOD OF OPERATION

FIELD OF THE INVENTION

The present invention relates to an RF transmitter and a method of operation using the transmitter. In particular, the invention relates to a multi-carrier transmitter suitable for use in wideband communications.

BACKGROUND OF THE INVENTION

Many modern wideband communications systems use multi-carrier modulation schemes. In these schemes a high bit rate data stream is split into a number of lower bit rate streams, each lower bit rate stream being sent using an independent carrier referred to herein as a 'sub-carrier'. If for example n symbols per second have to be transmitted and N sub-carriers are used each sub-carrier carries n/N symbols per second, i.e. this is the sub-carrier symbol rate.

For example, the IEEE 802.11 (a) standard for WLAN (Wireless Local Area Network) specifies use of OFDM (Orthogonal Frequency Division Multiplexing), which is a form of multi-carrier modulation that provides good bandwidth efficiency whilst maintaining a relatively low symbol rate. This helps to combat the destructive effects of multi-path fading which occurs when a signal is sent over the air between a transmitter and a receiver.

As the number of sub-carriers in a multi-carrier modulation scheme increases, a power peak-to-average ratio (PAR) of the transmitted signal also increases. This increases the required power handling capacity of the RFPA (radio frequency power amplifier) employed in the transmitter. Multi-carrier modulation schemes such as OFDM implement linear modulation, such as such as binary phase-shift keying (BPSK), quadrature PSK (QPSK), 16-level quadrature amplitude modulation (16- QAM), or 64-level QAM (64-QAM), and therefore require linear power amplification. Conventional class "A" or "B" RFPAs which are conventionally used to provide linear power amplification need to operate at a point which is sufficiently below saturation in order to prevent clipping of the modulation peaks. For multi-

carrier modulation this results in several disadvantages. For example, (1) a low allowable transmitted average power, owing to a high PAR, limits the transmission range and (2) the need to use high power amplifying devices, owing to the high peak power required, leads to large physical size and high cost of the RFPA. In any case, RFPA efficiency is reduced. The reduced efficiency in particular presents a serious problem in RF transmitters used in battery-powered portable terminals owing to the high power consumption which takes place in such transmitters, particularly in the RFPA, which leads to a short battery life. Heat generation by the RFPA and dissipation of such heat is also a considerable issue for such transmitters in portable terminals.

In the future, the problems which have been described above will become more severe since, according to proposed developments of OFDM modulation standards, it is planned to increase the number of sub-carriers used in RF signals. For example, the proposed WiMAX (802.16e) standard specifies use of 1500 sub-carriers. Several methods of improving the efficiency of an RFPA (radio frequency power amplifier) in a multi-carrier, e.g. OFDM, transmitter are currently used or have been suggested. One method is to use an enhanced-efficiency RFPA such as an envelope-following amplifier, or a Doherty amplifier, etc. Envelope following, also called supply modulation, provides an efficiency improvement technique for linear RFPAs, but is limited in terms of bandwidth since it requires an agile switching power supply whose response bandwidth is significantly greater than the modulation bandwidth. The Doherty amplifier provides limited efficiency improvement and is not a practical solution for use with high PAR signals.

Another method of improving RFPA efficiency is to employ PAR reduction techniques. Several methods for reducing the PAR have been suggested and are currently under research. However, these methods do not reduce the PAR by a suitable margin, so do not provide a satisfactory solution to the problems described above relating to limited RFPA efficiency in a multi-carrier modulation transmitter.

SUMMARY OF THE INVENTION

According to the present invention in a first aspect there is provided an RF transmitter as defined in claim 1 of the accompanying claims. According to the present invention in a second aspect there is provided a method as defined in claim 20 of the accompanying claims.

Further features of the invention are defined in the accompanying dependent claims and are disclosed in the embodiments of the invention to be described.

Embodiments of the present invention will now be described by way of example with reference to the accompanying drawings, in which:

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a block schematic diagram of a known OFDM transmitter. FIG. 2 is a block schematic diagram of an OFDM transmitter embodying the present invention.

FIG. 3 is a block schematic diagram of a sub-signal generator suitable for use in the OFDM transmitter of FIG. 2.

FIG. 4 is a waveform diagram in the frequency domain (i.e. several waveforms indicating amplitude plotted against frequency) showing how a single multi-carrier OFDM signal as used in the prior art is split in accordance with an embodiment of the invention into a plurality of sub-signals.

FIG. 5 is a block schematic diagram of part of an alternative sub-signal generator suitable for use in the OFDM transmitter of FIG. 2.

DESCRIPTION OF EMBODIMENTS OF THE INVENTION

FIG. 1 is a block schematic diagram illustrating a common implementation of a known OFDM transmitter 100 using N sub-carriers equally spaced in frequency and each having an amplitude and phase representing a unit of one or more bits of input data to be transmitted. The transmitter 100 includes a signal generator 101 connected in turn to an RFPA (radio frequency power amplifier) 113 and a transmitter front end

115. The signal generator 101 includes a symbol mapper 103 connected in turn to an IDFT 105 (Inverse Discrete Fourier Transformer or operator which produces an Inverse Discrete Fourier Transform, IDFT), a DAC (digital to analog converter) 107, an I/Q modulator 109 and a synthesizer 111. Generation and transmission of a single complex OFDM symbol by the

OFDM transmitter 100 is carried out as follows. Input digital data to be transmitted by RF communication enters the symbol mapper 103, where it is mapped into OFDM symbols each comprising N sub-carriers having different amplitude and phase to represent the input data. The resulting amplitude and phase of each of the sub-carriers k may be represented by a complex coefficient, X[k]. The complex coefficients X[k] are fed into the EDFT 105 which is an Nx N IDFT (i.e., a processor which transforms N input points into N output points) and produces digital samples x[n] of the complex OFDM signal in the time domain. The mathematical representation of these samples is (where j is the square root of -1):

₁ N-I _nπ ±ι 4.n] = η~∑ X[k]e ^N

The digital samples x[n] are fed into the DAC (digital to analog converter) 107 in which they are converted into two continuous-time or analog voltage waveforms xι{t) and λ _Q (0 which represent in-phase (I) and quadrature phase (Q) components of the complex OFDM signal. The voltage waveforms xiit) and xφ) are applied to the FQ modulator 109 together with an RF carrier signal, c(t) produced by the synthesiser 111 in a known manner. The I/Q modulator 109 modulates the carrier signal c(f) with the in-phase (I) and quadrature phase (Q) components of the complex OFDM signal represented by the component waveforms xtf) and X _Q (O- The output of the I/Q modulator 109 is thus an OFDM-modulated RF (radio frequency) signal, y{i).

It should be noted that in some cases, the I/Q modulator 109 may optionally produce an OFDM-modulated IF (intermediate frequency) signal. This IF signal, which may have a frequency in a wide range of intermediate frequencies, may then be filtered and up-converted using a filter and an up-converter (not shown) in a known manner to produce the required modulated RF signal y(ή.

Finally, the OFDM-modulated RF signal y(t) is amplified by the RFPA 113 and the amplified signal produced by the RFPA 113 is transmitted as an output RF signal by a transmitter front end 115 comprising an antenna or RF radiator and other associated known front end components. The output RF signal is transmitted over the air to a receiver (not shown) designed to receive such a signal.

In accordance with an embodiment of the invention, a novel RF transmitter employing multi-carrier modulation is provided. This novel transmitter is illustrated as follows in terms of OFDM modulation. However, the invention is not limited to use with OFDM modulation. The same principle may be applied to any type of RF multi-carrier modulation.

FIG. 2 is a block schematic diagram of a transmitter 200 embodying the present invention. Components which are the same as components in FIG. 1 have the same reference numerals as such components in FIG. 1. In the transmitter 200 the OFDM signal generator 101 of FIG. 1 is replaced by an OFDM sub-signal generator 201 (to be described in detail with reference to FIG. 3). The sub-signal generator 201 produces a set of M output sub-signals y _o (t), yi(t)...y(M-i)in parallel, each of which is applied to a dedicated one of a plurality of RFPAs 203 by modulating the input data with a carrier signal c(t) from the synthesizer 111. The RFPAs 203 produce amplified versions of the sub-signals y _o (t), yi(t)...y _(M -i)respectrveVy applied to them, and the amplified signals are applied to a power combiner 205. The power combiner 205 is a passive, preferably non-resistive, RF power combiner such as a Wilkinson combiner, which is a well known power combiner. This may be implemented for example by use of transmission lines such as, but not limited to, microstrip lines. The power combiner 205 combines the amplified versions of the sub-signals yo(t), yi(t)...V( _M -i)to produce a single output RF signal which, as in the transmitter 100 of FIG. 1, is applied to the transmitter front end 115 for over-the-air transmission as an RF signal.

The novel RF transmitter 200 operates by carrying out the following novel steps: ( 1 ) An OFDM-modulated signal formed by symbol mapping of the input data for a given symbol in the symbol mapper 103 (FIG. 1) is split in the sub-signal generator 201 into M time-domain sub-signals (namely, yo(t), yi(t)...V(M-i)) _> such that each of these sub-signals represents only a portion of the total number of OFDM sub-carriers included in the OFDM symbol. As illustrated later, this results in each sub-signal

beneficially possessing a lower PAR than the PAR of the corresponding total OFDM- modulated signal which is split into sub-signals.

( 2 ) Each of the M sub-signals yo(t), yi(t)...y(M-i) is amplified using a separate dedicated one of the RFPAs 203, each RFPA 203 thereby operating on a sub-signal of reduced PAR.

( 3 ) The amplified sub-signals are then combined in the RF power combiner 205 to obtain a single total OFDM-modulated signal for over-air-the transmission by the transmitter front end 115.

Beneficially, reducing the PAR by forming sub-signals in this way allows the power-added efficiency of each power amplifier (namely each RFPA 203) to be increased in inverse proportion to the decrease in PAR obtained by splitting the OFDM signal into sub-signals. Furthermore, distributing the power over several lower-power devices allows the use of transmitter components capable of processing lower peak powers, with a consequential advantageous reduction of size and cost. In addition, the problem of generation of RFPA heat and dissipation of such heat is reduced. These benefits are illustrated later.

FIG. 3 is a block schematic diagram of an arrangement 301 suitable for use as the sub-signal generator 201 in the transmitter 200 of FIG. 2. In the arrangement 301, as in the signal generator 101 in the known transmitter 100 of FIG. 1, input data is mapped into OFDM symbols by the symbol mapper 103. An output signal from the symbol mapper 103 of the arrangement 301 is applied in parallel to each of a set of M EDFTs 305, where M is a number of sub-signals to be produced. Each of the IDFTs 305 operates, in a manner to be described in more detail later, to produce its own complex output signal x _m [n], where m is an index identifying each of the M output signals x[n] from the IDFTSs 305. Each output signal x _m [n] is applied as an input digital signal to a corresponding one of a set of M dual-channel DACs 307. Each of the DACs 307 produces two voltage waveforms Xi _m (t) and X _Qm (t) in the same manner as the single DAC 107 of the signal generator 101. The two voltage waveforms Xi _m (t) and XQ _m (t) produced by each DAC 307 are respectively in-phase (I) and quadrature phase (Q) analog components of the input digital signal x _m [n]. The pair of waveforms x\{t) and xφ) produced by each of the DACs 307 is applied to a corresponding one of a set of M I/Q modulators 309n together with the RF carrier signal c(t) produced by

the synthesiser 111 as in the signal generator 101. Each of the FQ modulators 309 modulates the carrier signal c(t) with the modulation signal represented by the particular waveforms xι(t) and X _Q (?) it receives as input signals. The output of each of the I/Q modulators 309 is thus a separate sub-signal in the time domain representing part of the total OFDM-modulated RF (radio frequency) signal required to be transmitted. These sub-signals are the output signals y _o (t), y ₁ (t)...y(M-i) (t) (referred to earlier with reference to FIG. 2) provided as output signals by the OFDM sub-signal generator 201. As described earlier with reference to FIG. 2, each of these output signals is then amplified and the amplified signals are combined and transmitted over the air.

Operation of the sub-signal generator 301 is further described as follows. Let us assume that the OFDM signal for a symbol produced by the symbol mapper 103 contains N sub-carriers. Assume also that the OFDM signal is split into M sub-signals at the IDFTs 305, such that each sub-signal contains S sub-carriers equally spaced in frequency. Assume also that M - 2 ^R . These definitions are summarized as follows: N = total number of sub-carriers; N = 2 ^T , where T is a positive integer. M = number of sub-signals; M- 2 ^R where R is a positive integer; S = number of sub-carriers in each sub-signal; S = MM and S = 2 ^U , where U is a positive integer. We define x _m as the m-th sub-signal, where m is an index ranging from 0 in integer steps to M-I. These index designations are shown in FIG. 3 in relation the outputs x[n] from the IDFTs 305 (and also to the IDFTs 305). We also define a process called IDFT _OT (Sχw), which operates in each of the IDFTs 305. In this process, a vector of N complex frequency-domain samples is received and S equally spaced samples are selected using the offset index in to produce M complex time-domain samples by discrete Fourier transformation. This process is described by the following equation:

x _m [n] = IDVr _nKSxM) {X[k]} = -∑ X[s - M + m]e

This process is equivalent to performing a conventional discrete Fourier transformation IDFT( _NXN ), as in the IDFT 105 of FIG. 1, with some of the frequency- domain samples X[Jc] set to zero. These zero-value samples in a given transformation

in one of the IDFTs 305 of FIG. 3 represent the sub-carriers that are not included in a particular sub-signal being processed by the E)FT 305. Another way of considering this process is to say that each IDFT 305 only processes the sub-carriers of the original signal that it is designated to process and it skips the other sub-carriers. In practice, each of the IDFTs 305 is operated by a software program in a signal processor, so the part of the original signal processed by each IDFT 305 is defined by the software program of that IDFT 305. The result is equivalent to a partial OFDM spectrum that contains S equally spaced sub-carriers. This is illustrated in Figure 4. FIG. 4 is a waveform diagram showing several waveforms in the frequency domain (i.e. several waveforms of amplitude plotted against frequency) showing how a single OFDM signal as produced by the symbol mapper 103 as used in the prior art is split in accordance with the embodiment of the invention described with reference to FIGS. 2 and 3 into a plurality of sub-signals. FIG. 4 illustrates the case in which the number of sub-carriers N in the OFDM signal is 16, the number M of sub-signals produced is 4 and the number >S of sub-carriers in each sub-signal is 4. A total OFDM signal having sixteen sub-carriers is represented by waveform (1). Each of the individual sub-carriers included in the total OFDM signal has a form shown as waveform (2). The total OFDM signal is split into four separate sub-signals represented by waveforms (3) to (6) respectively, the waveforms (3) to (6) representing equal spacings of the index m from m = 0 to m = 3. It can be seen that the four waveforms (3) to (6) in FIG. 4 when combined add to give the waveform (1). Of course, by processing in the IDFTs 305, the sub-signals represented by the waveforms (3) to (6) are also transformed into time domain signals. FIG. 4 illustrates the respective frequency spectra of these signals. It is to be noted that if the sixteen sub-carriers in waveform (1) are designated as sub-carriers 1, 2, 3 ... 16 by their position in the set of sub-carriers then the sub- carriers which appear in waveform (3) are sub-carriers 1, 5, 9 and 13; the sub-carriers which appear in waveform (4) are sub-carriers 2, 6, 10 and 14; the sub-carriers which appear in waveform (5) are sub-carriers 3, 7, 11 and 15 and the sub-carriers which appear in waveform (4) are sub-carriers 4, 8, 12 and 16. In other words the sub- carriers in each sub-signal are equally spaced by a frequency difference 4.s where s is the frequency spacing between adjacent sub-carriers in the waveform (1). By suitable

programming of the IDFTs 305, each of the M sub-signals could receive from the total set of N sub-carriers in waveform (1) a sub-set of S sub-carriers in a different order from that illustrated in FIG. 4.

The individual voltage waveforms x _\ (t) and xq(t) produced by the individual DACs 307 in the arrangement 301 are designated as waveforms x _m ι(t) and x, _n Q(t) in FIG. 3, where each index m is in the range 0 to M-I. Each of these waveforms is processed in an individual one of the VQ modulators 309 in a manner similar to the processing of the signals x _\ {f) and xφ) by the single 1/Q modulator 109 of FIG. 1. The resulting modulated output signals produced by the I/Q modulators 309 are designated in FIG. 3 as output signals y _m (t), where m is the same index as defined above.

The improvement in efficiency obtained in the embodiment of the invention described with reference to FIG.s 2 and 3 may be further demonstrated as follows. The power combiner 205 shown in FIG. 2 will be assumed to have an insertion loss L from its multiple input ports to a common output port (L<1). Firstly, we will examine the case of a single, class B amplifier for use as the

RFPA 113. Then we will examine the case of a set of multiple class B amplifiers as the RFPAs 203. In each case we will assume that the following parameters are given (in linear units) for each amplifier:

P _ouτ = Average output power; PAR = Peak-to-average power ratio of the modulation for a single amplifier, i.e. the RFPA 113;

PAR ' = Peak-to-average power ratio of the modulation for each of the multiple amplifiers, RFPAs 203;

EFF _SAT = Power amplifier efficiency at saturation. We will assume also that the saturation point of each amplifier considered must be designed to pass the modulation peaks of a signal applied to it without clipping; thus, in the case of the single amplifier RFPA 113: p _ p . OA T?

¹ SAT ~~ ^L OUT ^r "^^

We will assume also that the efficiency of each power amplifier considered is proportional to the square root of the output power of the amplifier. This is a reasonable assumption for class B amplifiers. This efficiency is given by:

EFF

We will assume again that the OFDM signal is split into M sub-signals, such that M = 2 . Owing to the loss L in the power combiner 205 as described above, the average power P' _O U _T required at the output of each RFPA 203 is:

As noted above, the sub-signal peak-to-average ratio at each RFPA 203 is PAR'. In view of the reduced number of sub-carriers in each sub-signals we can assume that: PAR' < PAR The required saturation power, P' _SAT , of each RFPA 203 is:

The efficiency EFF' of each RFPA 203 is:

EFF' = EFF _SAT

The overall efficiency of the combination of the RFPAs 203 and power combiner 205, EFF _TOT , is the efficiency of each RFPA 203, degraded by the insertion loss L of the combiner 205:

An efficiency improvement factor, k, obtained using multiple sub-signals and multiple RFPAs 203 for each sub-signal (in the embodiment of the invention) compared with use of a single signal and single RFPA 113 (in the prior art) is given by:

EFF _τnτ L- EFF, _δT -/PAR , I PAR k = - ^■ TOT _ SAT

EFF y/PAR' EFF _SAT \ PAR'

A specific illustration of this efficiency improvement is as follows. Assume that the OFDM signal is a 64-carrier QPSK-modulated OFDM signal having the following typical values: P ₀ OT= I W;

PAR = 10 (10 dB);

EFFsAT = 50%;

For a single amplifier, e.g. RFPA 113:

^P sAτ = ^p ouτ ^■ PAR = 1*10 = 10 W

Thus, the saturated power of the single amplifier RFPA 113 needs to be 1OW in order to transmit an average power of only IW. The efficiency for the single amplifier RFPA 113 in this specific example is:

This means that the DC power needed at the input of the single amplifier RFPA 113 is 6.33 W and the dissipated power is 5.33 W.

Assume instead for this specific illustration that the same OFDM signal is split into sub-signals in accordance with the embodiment of the invention described above. Based on simulations and measurements of a QPSK-modulated OFDM signal, we have found that the peak-to-average ratio (PAR') for different numbers of sub-carriers is as summarized in Table 1 below:

TABLE 1

We will assume a loss in the power combiner 205 of 0.5 dB (equivalent to L - 0.891). The average output power and saturated output power of each RFPA 203 in this case is: τ>> _ "OUT . J_ _ 2. 1

= 0.14 W ^ouτ M ^' L 8 0.891

P _S ' _AT = P _O ' _UT ^■ PAR' = 0.14 ^• 5.62 = 0.79 W

Note that there is a lower saturation power requirement for each of the multiple amplifiers, i.e. each RFPA 203, compared with that for the single amplifier, RFPA 113 (0.79 W compared with 10 W). The efficiency improvement factor for this illustration is:

Therefore, the total efficiency using a 64-sub-carrier OFDM signal split into eight sub-signals is estimated to be: EFF _T0T = EFF - k = 0.158 -1.188 = 0.188 = 18.8 %

If this OFDM signal were to be split instead into sixteen sub-signals, each containing four sub-carriers, the efficiency estimated in the same manner would increase to 22.8%, which is a considerable improvement from the original value of 15.8% estimated when using a single power amplifier, RFPA 113.

A significant efficiency improvement has been illustrated in these calculations for the case of QPSK-modulated OFDM. Further improvements may be obtained by implementation optimization. The efficiency improvement principle which has been illustrated remains valid for other multi-carrier modulation schemes as well, since reducing the number of sub-carriers in each sub-signal will still reduce the PAR for each sub-signal.

In the embodiment of the invention described earlier with reference to FIGS. 2 and 3, the sub-carriers are distributed symmetrically between sub-signals and the sub- signals are distributed symmetrically among the RFPAs 203. This is for architectural convenience. It should be noted that other distributions may be used. However, the number of sub-carriers per RFPA should preferably be the same for all of the RFPAs 203. FIG. 5 shows components included in an alternative sub-signal generator 400 which may be employed in place of the sub-signal generator arrangement 300 of FIG. 3. In the case of the sub-signal generator 400, M voltage waveform pairs x _oi (t) and x- ₀ Q(t); xπ(t) and X _I QOO; ... x _mI (t) and x _mQ (t);... and x _M -π(t) and X _M - _I QOO are produced in the same manner as described with reference to FIG. 3. However, each of these voltage waveform pairs is separately applied in the sub-signal generator 400 to a

corresponding one of a set of M I/Q modulators 401 each of which also receives as an input an IF (intermediate frequency) signal i(t) generated by a synthesizer 403. As in the prior art, the frequency of the IF signal i(t) may be selected from a wide range of possible EF frequencies. The I/Q modulators 401 form a set of M output signals in which the IF signal i(t) is modulated respectively by each of the modulation signals represented by the set of M respective waveform pairs x _m i(t) and x _m o_(t). Each of the output signals produced by the I/Q modulators 401 is applied in turn to (i) one of a set of M filters 405 each of which passes a band which includes the modulated IF signals and then to (ii) one of a set of M mixers 407. Each of the mixers 407 receives as an input a signal c(t) which is an RF carrier signal produced by a synthesiser 409 at a frequency equivalent to that of the RF carrier signal c(t) used in the arrangement 301. Each of the mixers 407 up-converts the modulated IF signal it receives from a corresponding one of the I/Q modulators 401 to a form in which the RF carrier signal c(t) is modulated. The mixers 407 thereby produce a set of M output signals which are the same as the signals y _o (t), y^t)...y _m (t)... y _M -i(t) referred to earlier. These output signals form the outputs of the sub-signal generator 400 and are then each separately amplified by the RFPAs 203, combined by the RF power combiner 205 and sent as a single RF signal over the air by the transmitter front end 115 as described earlier with reference to FIG. 2. Thus, in summary, by splitting a multi-carrier signal into several time-domain sub-signals, each generated by only a portion of the sub-carriers, and then amplifying each sub-signal with a separate RF power amplifier and combining the amplified signals so produced, the efficiency can be improved significantly. This is because each sub-signal has lower PAR than the PAR of the overall multi-carrier signal. This new approach is very suitable for implementing in the form of an integrated circuit in a known manner. The parallel paths required to produce and process the sub-signals will be inherently balanced if they are implemented on the same substrate of such an integrated circuit. Furthermore, the improved efficiency and lower saturation power requirement for each of the RF power amplifiers will benefit implementation on a single chip. This integrated circuit approach will also facilitate product size reduction.

The transmitter embodying the invention is suitable for use as a multi-carrier transmitter in any known application for such transmitters, especially when using OFDM. For example, the transmitter may be used in a mobile communication system, especially a system which is a local area network in accordance with the EEEE 802.11 (a) standard for WLAN (wireless local area networks).

The RF signal which is transmitted by the transmitter as described in the above embodiments of the invention will be the same as a signal transmitted by the prior art OFDM transmitter shown in FIG. 1, so the signal may be received in an OFDM receiver as known in the art without modification.