Circuits, systems and methods for digital correction of phase and magnitude errors in image reject mixers

[0002] 1. Field of the Invention

[0003] The present invention relates in general to wireless communications and in particular to circuits, systems and methods for correction of phases and magnitude errors im image reject mixers.

[0004] 2. Description of the Related Art

[0005] In conventional wireless communications systems, data is transmitted by modulating a carrier signal at a frequency f_cusing any one of a number of well known modulation techniques, for example phase shift keying or frequency shift keying. This signal typically represents a single channel in a multiple channel frequency band allocated for the particular application or purpose. After receipt by the receiver antenna, the modulated signal is typically passed through a low noise amplifier to improve the overall receiver noise figure and then through a front end filter. The signal next is either upconverted or downconverted using one or more mixing stages to a desired frequency which can more easily be processed. In a digital receiver, this processing includes analog to digital conversion following any analog processing, such as analog filtering.

[0006] The mixing taking place during up- or down-conversion mixes the received signal with a local oscillator signal at a desired local oscillator frequency f_lo. The result is multiple intermediate frequency (IF) signals, for example one at f_c+f_loand another at f_c−f_lo. One of these signals is the desired signal and the other is its image. The image, along with f_loand various subharmonics resulting from the mixing are filtered out, for example using preselection filtering. While this type of system if suitable for narrow band applications, it is not practical in wide band applications where the image could fall in the band of interest and appear as a legitimate signal.

[0007] In some wideband applications, image rejection mixers are commonly used. In an image rejection mixer, the incoming signal is mixed with the local oscillator signal and in parallel with the local oscillator signal shifted by ninety (90) degrees. The resulting in-phase (I) and quadrature (Q) signals together represent a complex signal (with I the real part and Q the complex part) which is for further processing (in a digital receiver these signals are normally sent on a pair of analog to digital converters (ADCs)).

[0008] Image rejection mixers have some serious limitations. Among other things, to achieve acceptable levels of image rejection, the pair of mixers must be closely matched in output amplitude and phase. Even a very small mismatch of 1% or less in amplitude or phase across the entire band can cause the image amplitude to rise significantly within the complex signal. This degree of matching is impractical to implement in most systems.

[0009] Some current image rejection schemes down-convert each channel in the given frequency band to baseband on an individual basis. Magnitude and phase errors between the I and Q components can then be corrected at lower frequencies where filter matching is less critical. This however requires one or more variable frequency local oscillators and corresponding loop filters, at either the RF or IF conversion stages, for mixing each channel of interest to baseband. In other words, in these systems, the entire band of interest (containing many potential channels) cannot be converted with a fixed local oscillator while maintaining satisfactory image rejection.

[0010] Therefore, a need has arisen for circuits, systems and methods for maintaining high image rejection in broadband wireless applications.

[0011] Image rejection circuitry is disclosed including a first digital path having a finite impulse response filter having a variable gain stage operable to apply a gain to a first digital signal in response to a coefficient selected from a coefficient table and an adder for summing the first signal to the output of the variable gain stage to produce a corrected first digital signal. A second digital path is disclosed for processing a second digital signal along with error calculation circuitry for determining an error between the first and second signals and selecting the coefficient from the coefficient table.

[0012] The present inventive concepts allow for maintaining high image rejection in broadband wireless systems. Among other things, these concepts allow an entire multiple channel band to be downconverted in response to one or more fixed frequency local oscillators. Moreover, the RF to IF conversion stage can be eliminated if desired and only a single image rejection mixer used at the receiver front-end.

[0013] For a more complete understanding of the present invention, and the advantages thereof, reference is now made to the following descriptions taken in conjunction with the accompanying drawings, in which:

[0014] FIG. 1 is a diagram depicting the major components of a typical radio frequency or microwave (collectively “RF”) receiving; and

[0015] FIG. 2 is a diagram of an exemplary wireless receiving system embodying the present inventive concepts.

[0016] FIG. 1 is a diagram depicting the major components of a typical radio frequency or microwave (collectively “RF”) receiving system 100. Systems similar to this are used in such applications as cellular telephone systems and other wireless communications systems. Modulated signals carrying information in the RF or microwave bands are received through an appropriately tuned antenna 101. A preselector filter 102 is used to filter out the band of interest from extraneous noise and signals, as well as, among other things, reduce the response of the receiver to the image frequency, the intermediate frequency and spurious noise and signals. The output of filter 102 is passed through a low noise amplifier to improve the overall noise figure of system 100.

[0017] RF mixer 104, in a down-converter receiver, mixes the RF signal at frequency f_RFwith a first local oscillator (LO) signal at a frequency f_LO1. The result is a signal at a desired intermediate frequency (IF) along with its image, and a corresponding number of subharmonics. Filter 105 alternates the image, subharmonics and noise and passes the IF signal for further amplification at stage 106.

[0018] In a downconverter system, the IF signal is next brought down to the baseband frequency using a image rejection mixer 107. Specifically, the IF signal is split between two mixers 108aand 108b,with mixer 108amixing the IF signal with a second LO signal at a frequency of f_LO2to generate an in-phase signal I and mixer 108bmixing the IF signal with the second LO signal shifted by 90 degrees in phase to generate an out-of-phase signal Q. These two signals combined compose a complex signal having an image attenuated to a given level by cancellation. The I and Q signals are then sent in parallel through a matched pair of low pass filters 109aand 109band analog to digital converters (ADCs) 110 at the front end of the digital processing section of the system.

[0019] The system shown in FIG. 1 is suitable for narrow band systems. However, in high bandwidth systems problems arise due to insufficient image rejection. Mismatching between low pass filters 109aand 109b,mixers 108aand 108band ADCs 110aand 110b,can cause phase and magnitude errors between the I and Q paths. Consequently, the image cancellation is reduced and an image of an unacceptably high level arises in the complex I/Q composite signal. Assume for example that the RF signal is any one of a number of channels in an 80 Mhz band on a 2.4 Ghz carrier. To achieve image rejection of 50 db or greater for any channel across the 80 Mhz band, and if low pass filters 109 are selected to be second order filters, the poles of the filters must be matched to approximately 0.1%. Filter matching to this degree is too difficult to achieve for practical applications. Moreover, in order to tune into each channel separately for RF to baseband down conversion, the source of local oscillator signals LO₁and/or LO₂must be frequency variable. For example, frequency synthesizer 111 and 112 or similar frequency variable local oscillator sources are necessary.

[0020] FIG. 2 is a diagram of an exemplary RF receiving system 200 embodying the present inventive concepts. System 200 can advantageously be used in a wide variety of wireless systems, including cellular communications, wireless local area networks (LANs), consumer electronics, etc. In this example the RF to IF conversion stage is not used, with the amplified and filtered RF signal being passed directly to the image reject mixer 107. Because of the inventive concepts, the image rejection at the digital processing stage can be sufficiently high even for broadband applications, such that even direct RF to baseband down-conversion is possible. If desired, or if the application requires it, RF to IF down-conversion can still be used in systems embodying the present invention. Moreover, the improvement in image rejection across the entire band allows that entire band to be down converted at once using fixed frequency local oscillator 201.

[0021] Again, the outputs from the image rejection mixer are passed through low pass filters 109aand 109band then converted to digital form by ADCs 220. The I and Q signals, output from ADCs 220 represent the real and complex components of a complex wave function.

[0022] The I output from ADC 220ais passed through a finite impulse response (FIR) filter 202 which includes a single delay element 203, a variable gain stage 204, and an adder 205. In this embodiment, delay element delays the digitized signal by one period of the sampling clock used in the analog to digital conversion process by ADCs 220. The variable gain stage 204, which could be a shifter, adder or submultiplier, corrects the phase and magnitude errors (caused by magnitude and/or phase differences between the I and Q paths) in the complex signal by moving the zero of FIR filter 202 as close as possible to the origin of the unit circle in the complex plane defined by complex signal. The gain in variable gain stage 204 set in response to FIR filter coefficient selected from a coefficient lookup table 206, discussed further below. Specifically, the coefficient is used to add or subtract a value from the I component approximately equal to the magnitude difference between the I and Q paths measured at a given frequency during a calibration sequence. The direct I component from gain stage 204 is added by adder 205 to the delayed I component directly from ADC 110ato arrive at the corrected component I_corr.

[0023] The uncorrected Q component out ADC 220bis compared in magnitude with the unfiltered I component out of ADC 220aby error calculation circuitry 207 to determine the magnitude error between the uncorrected I and Q paths discussed above, for example through a digital subtraction operation. The calculated magnitude error is then used to lookup a corresponding coefficient from coefficient table 206 for use by variable gain stage 204. Concurrently, a coefficient is selected from the lookup table to correct DC gain in the uncorrected Q path. In particular, the uncorrected output from ADC 220bis multiplied by a multiplier 208 by the retrieved coefficient to set the DC gain at a desired value and produce a corrected Q signal Qcorr.

[0024] Although the present invention and its advantages have been described in detail, it should be understood that various changes, substitutions and alterations can be made herein without departing from the spirit and scope of the invention as defined by the appended claims.