The present invention relates generally to the field of radios that include both a transmitter and a receiver. More specifically, the present invention relates to such radios in which cancellation techniques reduce the corruption of a receive signal caused by a transmit signal. And, the present invention relates to such radios in which compensation is provided for signal-generated component heating.
Radios that include both a transceiver and a receiver (i.e., transceivers) often have an ability to transmit and receive simultaneously. This ability requires that transmission and reception occur in different frequency bands. Such a transceiver should be configured so that its transmitter's outgoing transmit signal does not interfere with the incoming receive signal intended for its receiver. Such interference becomes less of a problem when the transmit frequency band is spaced farther apart from the receive frequency band in the electromagnetic spectrum, when the radio's transmit power is lower or the remote device's transmit power is higher, and/or when a higher quality duplexer is used. But in some radio applications, the available portions of electromagnetic spectrum simply do not permit as much spacing between transmit and receive bands as might be desired. And, in some radio applications, higher local transmit power is required to achieve link margins over the desired radio coverage area, or remote devices simply cannot transmit at higher power due to regulations and/or a need to preserve battery reserves. And, high quality duplexers may simply be impractical in some radio applications because high quality duplexers tend to be very expensive and/or very large.
In situations were transmit and receive bands are near each other, where transmit power is high or the remote device's transmit power is low, and/or where lower quality duplexers are called for, cancellation techniques have been applied to further reduce corruption of the receive signal caused by the transmit signal. Generally, cancellation techniques call for extracting a small portion of the transmit signal, processing this extracted portion, and then subtracting it from the receive signal. The processing alters the extracted transmit signal's timing, amplitude, and phase to match the transmit-signal interference that has corrupted the receive signal. Often, the processing takes place using an adaptive equalizer whose filter characteristics are continuously adjusted to improve the match.
But conventional canceling techniques have failed to adequately reduce corruption of a receive signal by a transmit signal. As a result, in order to use conventional cancellation techniques and sufficiently reduce corruption of a receive signal by a transmit signal, undesirably expensive or large duplexers are still required, transmit and receive bands are separated from one another by spectral distances that result in painful wastes of spectrum, and remote devices are forced to transmit at undesirably high power levels to overcome the remaining interference.
It is believed the poor performance demonstrated by conventional cancellation techniques is due, at least in part, to the large temporal scale over which precision in processing the transmit signal should be maintained. Better cancellation requires a better match between a reference transmit signal and the interfering portion of the receive signal. In order to get a more precise match, a feedback loop that drives the adaptive equalizer processing the transmit signal should have a narrower loop bandwidth. In other words, the feedback loop should interpret almost all mismatches as noise, and ignore them unless they persist in a consistent way for a very long duration.
On the other hand, the amplitude and/or phase characteristics of the corrupted receive signal and reference transmit signal change as a result of temperature and humidity changes, component aging, and the like. In order to maintain a precise match, corresponding characteristics of the processed transmit signal should track such changes. Many of these agents of change occur slowly and may be tracked even by a narrow loop bandwidth in the feedback loop that drives the adaptive equalizer processing the transmit signal.
But a portion of the temperature changes may be due to signal-generated component heating. In other words, an analog component's temperature rises and falls in response to the power level of the signal it processes. This type of temperature change can rapidly produce a mismatching effect. In order to better track such temperature changes, the feedback loop that drives the adaptive equalizer processing the transmit signal should have a wider loop bandwidth. Thus, conventional canceling techniques strike a compromise between these two competing design considerations that simultaneously call for both a narrower and a wider loop bandwidth. As a result of the compromise, the reference transmit signal is not processed adequately to achieve both a precise match and to track changes.
A more complete understanding of the present invention may be derived by referring to the detailed description and claims when considered in connection with the Figures, wherein like reference numbers refer to similar items throughout the Figures, and:
A modulator 16 receives raw data 18 and digitally modulates raw data 18 into a modulated data stream 20 that is provided to an input of RF transmitter 12. In a preferred embodiment, modulator 16 is configured so that modulated data stream 20 conveys raw data 18 in a coded form and is arranged as a complex data stream having quadrature phase components. Complex notation is omitted from the figures in order to simplify the presentation of this material. The type of modulation used by modulator 16 to produce modulated data stream 20 is not a critical parameter of the present invention. Examples of modulations that may be implemented in modulator 16 include any type of quadrature amplitude modulation (QAM), code-division-multiple-access (CDMA), orthogonal frequency division modulation (OFDM), multiple-input, multiple-output (MIMO), and the like. Modulated data stream 20 may be viewed as a wideband data stream and may have resulted from combining a plurality of independently modulated, complex data streams together into a single modulated data stream 20. The plurality of data streams may correspond to different channels, and the different channels may be configured for a frequency division duplex (FDD) or a multichannel time division duplex (TDD) communication system. In addition, other processing may have been applied to modulated data stream 20 in modulator 16. Such other processing may include pulse shaping filters that are configured to minimize inter-symbol interference (ISI) and peak or crest reduction circuits that reduce the peak-to-average power ratio of modulated data stream 20.
RF transmitter 12 generates an RF transmit signal 22 from modulated data stream 20. In a preferred embodiment, RF transmitter 12 includes predistortion circuits, digital-to-analog conversion circuits, upconversion circuits, and a bandpass filter, each of which is omitted from
In the embodiment depicted in
An input of transmit filter 28 serves as a transmitter port 34 for duplexer 32 and couples to an output of transmitter 12 through directional coupler 26. Within duplexer 32, an output of transmit filter 28 couples to an input of receive filter 30 and serves as an antenna port 36 for duplexer 32. Thus, transmit signal 22 passes through directional coupler 26 and duplexer 32 at transmit filter 28, to an antenna 38, where it is broadcast from transceiver 10 with the intention of being received by some remotely located transceiver or receiver. To the maximum extent practical, transmit filter 28 passes signal energy in the frequency band where the transmit channels are located and attenuates signal energy outside this transmit band. In one an alternate embodiment, directional coupler 26 may be located between duplexer 32 and antenna 38, rather than as shown in
RF receive signal 40 is received at antenna 38. In the preferred embodiments, receive signal 40 may be received at antenna 38 simultaneously with the broadcasting of transmit signal 22 from antenna 38. The signal level of receive signal 40 is likely to be far, far lower than that of transmit signal 22. Receive signal 40 passes to antenna port 36 of duplexer 32 and an input of receive filter 30.
Due to the multiport operation of duplexer 32, the signal level of transmit signal 22 propagating toward receive filter 30 may be considerably reduced from the signal level of transmit signal 22 propagating toward antenna 38. And, due to the spacing of the transmit band apart from a receive band, and to the filtering operation of transmit filter 28, the portion of the signal energy from transmit signal 22 that resides in the receive band with receive signal 40 may be far reduced from the transmit band energy propagating toward antenna 38. But, due to the far, far greater signal level of transmit signal 22 compared to receive signal 40, transmit signal 22 may still potentially interfere with receive signal 40, and the potential for interference is greater the spectrally closer the transmit band is to the receive band.
Receive filter 30 thus receives a composite of transmit signal 22 and receive signal 40. This composite signal is referred to herein as a transmit-corrupted receive signal (RX+) 42. For the purposes of this description, transmit-corrupted receive signal 42 and all variants thereof produced in transceiver 10 subsequent to the reception of receive signal 40 at antenna 38 are referred to as transmit-corrupted receive signal 42 to distinguish them from transmit signal 22.
An output of receive filter 30 couples to a signal input of RF receiver 14 and serves as a receiver port 44 for duplexer 32. To the maximum extent practical, receive filter 30 blocks RF transmit signal 22 but passes RF receive signal 40. In particular, to the maximum extent practical, receive filter 30 passes signal energy in the frequency band where the receive channels are located and attenuates signal energy outside the receive band, and particularly in the transmit band. But to some extent, transmit signal 22 leaks into transmit-corrupted receive signal 42 at the output of receive filter 30.
Desirably, the amount of leakage is as low as practical. In addition to spectrally spacing transmit and receive bands as far apart as possible, leakage may be further reduced through the use of a high quality duplexer. But the benefits of the present invention are best appreciated in situations where it may be impractical to use a high quality duplexer, whether due to the high cost traditionally associated with high quality duplexers or to space limitations. Accordingly, at the output of receive filter 30, transmit-corrupted receive signal 42 may be expected to include a troublesome amount of transmit signal 22 combined with receive signal 40.
A portion, and preferably a very small portion, of transmit signal 22 is extracted at a coupled port of directional coupler 26 and routed to a control input of receiver 14. Receiver 14 includes a low noise amplifier (LNA) 46 along with other conventional circuits (not shown) which serve to downconvert and digitize transmit-corrupted receive signal 42. In addition, receiver 14 includes circuits to process transmit signal 22 and then combine the processed transmit signal with transmit-corrupted receive signal 42 to form a transmit-cancelled receive signal (RX−) 48. Desirably, the combining operation reduces transmit signal 22 corruption in receive signal 40 by canceling a significant amount of transmit signal 22 that was included with receive signal 40 in transmit-corrupted receive signal 42. The processing and combining operations for a preferred embodiment of the present invention are discussed below in connection with
Transmit-cancelled receive signal 48 then passes to a detector 50. Detector 50 receives and demodulates transmit-cancelled receive signal 48 to produce raw data 18′. Desirably, the interference caused by transmit signal 22 with receive signal 40 has been ameliorated through the cancellation operation carried out in receiver 14 so that detector 50 provides raw data 18′ that is substantially equivalent to raw data 18. Detector 50 desirably performs a complementary operation to modulator 16.
Unlike signals processed by digital components, signals processed by analog components, whether active or passive, are influenced by temperature changes normally experienced within the components' operational temperature ranges. Such influences may include phase changes, gain changes, group delay changes, bandwidth alterations, DC offset changes, and the like. For higher quality analog components the influence of temperature tends to be less than for lower quality analog components. But temperature influences nevertheless occur. And, the use of higher quality analog components is simply impractical for many transceiver applications due to the higher costs typically associated with higher quality components. Accordingly, transmit signal 22 and transmit-corrupted receive signal 42 are likely to be influenced by temperature changes.
Transmit signal 22 thus propagates toward and into receiver 14 through at least two separate signal paths 52 and 54. Through much of receive signal path 54, transmit signal 22 propagates as one component of transmit-corrupted receive signal 42. Under steady state temperature conditions, processing which takes place in receiver 14 desirably causes transmit signal 22 to precisely match the interfering portion of transmit-corrupted receive signal 42 in amplitude, phase, delay, and over the entire relevant bandwidth. But steady state temperature conditions are not normal operating conditions. When an analog component unique to either first or second set 56 or 58 of analog components experiences a temperature change, particularly a temperature change not experienced by the other set of analog components, the precision with which transmit signal 22 matches the interfering portion of transmit-corrupted receive signal 42 deteriorates. As a result, unless receiver 14 is able to track the effects of such temperature changes, the ability of receiver 14 to cancel interference caused by transmit signal 22 likewise deteriorates.
Receive filter 30, located in receive signal path 54, provides one example of an analog component which is likely to experience temperature changes not experienced by analog components in transmit signal path 52. Receive filter 30 may be configured to block a significant portion of the energy from transmit signal 22. As the power level of transmit signal 22 changes, signal-generated component heating results. In other words, an increased power level of transmit signal 22 results in increased power dissipation in receive filter 30 and an increase in temperature for receive filter 30. Then, when the power level of transmit signal 22 decreases, the temperature likewise decreases. When temperature changes at receive filter 30, center and corner frequencies change, causing phase and amplitude shifts. But receive filter 30 is only one example of an analog component which suffers from temperature influences. The temperature influences experienced by each component in each of signal paths 52 and 54 are experienced independently and in different amounts for each component. Consequently, the overall temperature influence results from the cumulative temperature influences of each component, and is in a constant state of flux as power levels and the ambient temperature change.
The ambient temperature may be assumed to change at a sufficiently slow rate so that a feedback loop which processes transmit signal 22 to achieve a precise match with the interfering portion of transmit-corrupted receive signal 42 can track ambient temperature changes. But signal-generated component heating occurs much more quickly. A feedback loop which processes transmit signal 22 may not be able to both achieve a sufficiently precise match with the interfering portion of transmit-corrupted receive signal 42 and track the effects of signal-generated component heating. The embodiment of receiver 14 discussed below in connection with
Downconvert and digitize section 64 downconverts transmit-corrupted receive signal 42 to baseband and digitizes the signal. Section 64 may, but is not required to, process transmit-corrupted receive signal 42 through an intermediate frequency in forming the baseband signal. But it is desirable that the downconversion be coherent with the upconversion performed in transmitter 12 (
Transmit signal 22 is directed along transmit signal path 52 to a bandpass filter 70. Preferably, bandpass filter 70 exhibits a transfer function approximately equal to the transfer function of receive filter 30 within duplexer 32. In other words, bandpass filter 70 passes the receive frequency band and attenuates the transmit frequency band. Transmit-corrupted receive signal 42 passes through receive filter 30 prior to arriving at combiner 68. With bandpass filter 70 having an approximately equal transfer function, transmit signal 22 after filtering in bandpass filter 70 should include components spectrally close to the interfering portion of transmit-corrupted receive signal 42. But bandpass filter 70 need not exhibit an exactly equal transfer function to that of receive filter 30. For example, bandpass filter 70 may be formed from less expensive filter components than receive filter 30. Any inequality in transfer function will be compensated for in an equalizer (discussed below) which is adaptively adjusted in a feedback loop to maximize the achievable cancellation.
An output of bandpass filter 70 couples to an input of a downconversion and digitizing section 72. Downconvert and digitize section 72 downconverts transmit signal 22 to baseband and digitizes the signal. Section 72 may, but is not required to, process transmit signal 22 through an intermediate frequency in forming the baseband signal. But it is desirable that the downconversion be coherent with the upconversion performed in transmitter 12 (
Equalizer 74 is desirably a digital filter. The precise type of filter may vary from application to application, but finite impulse response (FIR) or transversal forms of filters will be adequate for many applications. The filtering characteristics implemented by equalizer 74 are determined by a set of taps 77 provided to equalizer 74 by a tap update section 78.
Equalizer 74 will impose delay on equalized transmit signal 76. Consequently, variable delay section 66 is adjusted to achieve temporal alignment between equalized transmit signal 76 and the interfering portion of transmit-corrupted receive signal 22 at combiner 68. In other words, variable delay element 66 is adjusted so that transmit signal 22 arrives at the positive input of combiner 68 after propagating through receive signal path 54 as a part of transmit-corrupted receive signal 42 at the same instant it arrives at the negative input of combiner 68 after propagating through transmit signal path 52. Desirably, variable delay section 66 has the ability to delay samples by both an integral number of clock cycles and a fractional portion of a clock cycle to achieve high precision in temporally aligning equalized transmit signal 76 with transmit-corrupted receive signal 42.
Tap update section 78 is configured to adaptively generate taps 77 provided to equalizer 74. Taps 77 define the filter characteristics implemented by equalizer 74. Tap update section 78 includes a heat adjustment section 80 and a coefficient update section 82. Heat adjustment section 80 receives and is responsive to a heat signal (ΔH) 84. Heat signal 84 is generated in response to temperatures experience by components which process transmit signal 22 or an estimate of those temperatures.
Coefficient update section 82 generates coefficients 86 that are substantially unresponsive to heat signal 84. In one embodiment, the coefficient update section 82 implements a least-means square (LMS) coefficient adaptation algorithm. Thus, the output of combiner 68 couples to a first input of coefficient update section 82 to provide transmit-cancelled receive signal 48. And, the digitized, baseband form of transmit signal 22 from downconvert and digitize section 72 is routed to a second input of coefficient update section 82 to provide transmit-cancelled receive signal 48. Furthermore, a loop bandwidth constant 88, which is given the label “μ” in
But at least some of the coefficients 86 generated by coefficient update section 82 are not provided directly to equalizer 74. Rather, heat adjustment section 80 forms the taps 77 provided to equalizer 74 by adjusting the coefficients 86 generated in coefficient update section 82 to account for signal-generated component heating.
In one embodiment, a heating estimator 90 generates heat signal 84 in response to temperature measurements taken at one of the analog components from either of the first or second sets 56 or 58 of analog components discussed above. A dotted line in
But in another, and presently more preferred, embodiment, heating estimator 90 generates heat signal 84 in response to some form of transmit signal 22. In the embodiment shown in
In another embodiment (not shown), a version of modulated data stream 20, or a processed version of modulated data stream 20, may be routed to magnitude circuit 98, and section 96 may be omitted. In yet another embodiment (not shown), the form of transmit signal 22 which drives section 96 may be extracted between duplexer 32 and antenna 38 (
An output from magnitude circuit 98 is routed to high pass filter 92. In this embodiment, temperature measurement device 60 (
An LMS cell 108 is provided for each one-sample delay stage in second delay line 102. Each LMS cell 108 is configured substantially the same as the others. Each LMS cell 108 includes a multiplier 110 at which the feedback loop is closed. Multiplier 110 receives a delayed form of transmit signal 22 from second delay line 102, with the amount of delay corresponding to the position of the LMS cell 108 relative to second delay line 102. Multiplier 110 in each LMS cell 108 also receives the form of transmit-cancelled receive signal 48 that has been delayed to about the middle of second delay line 102. Multiplier 110 generates a correlation stream 112 that signals an amount of correlation between transmit signal 22 and transmit-cancelled receive signal 48, at varying stages of relative delay.
In each LMS cell 108, the corresponding correlation stream 112 is routed to a first input of a multiplier 114, and a second input of the multiplier 114 receives loop bandwidth constant 88. The smaller of a value provided for loop bandwidth constant 88, the narrower the loop bandwidth. An output of multiplier 114 in each LMS cell 108 couples to an input of an integrator 116, and an output of each integrator 116 provides a coefficient 86.
In the preferred embodiment, a sufficiently small value for loop bandwidth constant 88 is provided so that the loop bandwidth of each LMS cell 108 is too narrow to track the influences of signal-generated heating in components which process transmit signal 22, even should heat adjustment section 80 be disabled and make no adjustments to coefficients 86. But such a narrow loop bandwidth is desirable because it allows the feedback loop closed in coefficient section 82 to identify very small amounts of correlation between transmit signal 22 and transmit-cancelled receive signal 48. Consequently, coefficient update section 82 is able to form coefficients 86 that allow equalizer 74 (
But, due at least in part to this narrow loop bandwidth, coefficients 86 are largely unresponsive to the effects of signal-generated component heating and to heat signal 84 (
For each second heat adjustment cell 118″, the coefficient 86 is applied at an input to a high pass filter 120 and to a first positive input of a combiner 122. The output of high pass filter 120 provides a data stream which signals changes in the coefficient 86 with respect to a long term nominal value. The output of high pass filter 120 and heat signal 84 respectively drive integrate and dump sections 124 and 126. Integrate and dump sections 124 and 126 retard the rate of the data flowing through heat adjustment cell 118″. In the preferred embodiment, heat signal 84 and coefficient 86 may each provide updated samples at a rate that is two or more orders of magnitude faster than can be perceived in signal-generated component heating. Integrate and dump sections 124 and 126 slow this data to a rate compatible with signal-generated component heating for a power savings.
Integrate and dump section 126 generates an nominal heat signal [H′(n)], and integrate and dump section 124 generates a nominal coefficient signal [C′(n)]. After scaling by a heat coefficient [α(n)] (discussed below) in a multiplier 128, a predicted nominal coefficient [α(n)H′(n)] is then compared with the actual nominal coefficient signal [C′(n)] in a subtraction circuit 130. The predicted nominal coefficient [α(n)H′(n)] is based on heating and on previous values for the coefficient. Subtraction circuit 130 closes a feedback loop and produces an error signal [ε(n)] that corresponds to the difference between the nominal coefficient [C′(n)] and the predicted nominal coefficient based on heating.
The error signal [ε(n)] is then multiplied by the nominal heat signal [H′(n)] in a multiplier 132. The product output by multiplier 132 signals correlation between the error signal [ε(n)] and the nominal heat signal [H′(n)]. The output from multiplier 132 is scaled by a suitable loop constant “λ” in a multiplier 134, and the result integrated in an integrator formed from a combining circuit 136 and a delay element 138. An output of multiplier 134 couples to a first positive input of combining circuit 136, an output of combining element 136 couples to an input of delay element 138, and an output of delay element 138 couples to a second positive input of combining circuit 136. It is the output of this integrator at delay element 138 that provides a current heat coefficient [α(n)] and couples to an input of multiplier 128. The heat coefficient [α(n)] is an estimate of the sensitivity of the coefficient 86 to heat, as expressed by heat signal 84. Polarities are arranged, particularly at subtraction circuit 130, so that the heat coefficient [α(n)] increases or decreases as necessary to reduce the correlation instantaneously signaled at the output of multiplier 132 and accumulated through the operation of the integrator.
The current heat coefficient [α(n)], which remains valid for the integration period of integrate and dump sections 124 and 126, is used to scale the heat signal 84 samples in a multiplier 140, and each scaled current heat signal sample output from multiplier 140 forms an offset 142. An output from multiplier 140 couples to a second positive input of combining circuit 122. Offset 142 is then added to the current coefficient 86 to adjust coefficient 86 and form a current tap 77, which is provided to equalizer 74 (
Thus, the embodiment of heat adjustment cell 118″ shown in
Referring back to
A variety of alternate embodiments may be provided in connection with the implementation of equalizer 74.
In this embodiment, the digital, baseband version of transmit signal 22 from downconvert and digitize section 72 (
Accordingly, the
In summary, at least one embodiment of transceiver 10 provides an improved transmit-canceling transceiver that is responsive to a heat signal and an improved method for operating a transmit-canceling transmitter. In accordance with at least one embodiment of transceiver 10, greater precision is achieved in canceling transmit signal corruption in a received signal. In accordance with at least one embodiment of transceiver 10, greater cancellation precision is maintained both for steady state conditions and for dynamic conditions where analog components experience signal-generated component heating. In accordance with at least one embodiment of transceiver 10, separate feedback loops are provided to achieve a highly precise long-term average match between a transmit signal and a transmit-corrupted receive signal and to track deviations from that long-term average due to signal-generated component heating.
Although the preferred embodiments of the invention have been illustrated and described in detail, it will be readily apparent to those skilled in the art that various modifications may be made therein without departing from the spirit of the invention or from the scope of the appended claims. For example, those skilled in the art may readily adapt sections and components discussed herein to process complex signals, as may be called for in any specific application. Moreover, those skilled in the art may readily combine the teaching presented herein with other transmit signal cancellation techniques to achieve as complete a cancellation as may be required by a given application. Such modifications and adaptations which are obvious to those skilled in the art are to be included within the scope of the present invention.