Parallel-process digital modulator structures and methods

Information

  • Patent Grant
  • 6621366
  • Patent Number
    6,621,366
  • Date Filed
    Friday, December 7, 2001
    24 years ago
  • Date Issued
    Tuesday, September 16, 2003
    23 years ago
Abstract
In order to generate a carrier frequency Fc, a digital modulator must operate (to satisfy the Nyquist criteria) at a system sample rate Rs which is at least twice the carrier frequency Fc. However, digital modulator structures are provided herein that facilitate the use of M quadrature modulators which modulate, at a reduced modulation rate Rs/M, respective ones of M polyphase cosine elements and M polyphase sine elements with respective ones of interpolated I elements and interpolated Q elements to thereby form M polyphase modulated elements. The modulated elements are then sequentially selected in a multiplexer to form a modulated digital signal. The reduced modulation rate simplifies modulator design and lowers fabrication costs.
Description




BACKGROUND OF THE INVENTION




1. Field of the Invention




The present invention relates generally to digital modulator systems.




2. Description of the Related Art




Digital modulators find use in a variety of modern communication systems. Cable modems, for example, employ digital modulators to provide high-speed Internet service over bidirectional transmission systems (e.g., coaxial cables and optical fibers). Downstream data has typically been processed with quadrature amplitude modulation (QAM). Although quadrature phase-shift keying (QPSK) is a typical upstream modulation process, other processes (e.g., QAM) are often used to enhance the efficiency (bits/Hz) of the upstream data path. Cable modems have typically operated with an upstream data rate (e.g., 10 Mb/s) in an upstream communications band (e.g., 5-42 MHz) and with a higher downstream data rate (e.g., 36 Mb/s) in a different downstream communications band (e.g., 42-850 MHz).




Realizing digital modulators with extremely high carrier frequencies F


c


is challenging because the Nyquist criteria requires that the system sampling rate R


s


be at least twice the carrier frequency F


c


. For a carrier frequency in the region of 500 MHz, a modem's quadrature modulator must therefore operate at a system rate R


s


of 1 GHz. Design of a 1 GHz modulator generally requires extensive pipelining of the modulator's adders and multipliers with consequent increase of modulator costs.




BRIEF SUMMARY OF THE INVENTION




The present invention is directed to digital modulator structures and methods that can operate at extended system rates R


s


and thus realize high carrier frequencies F


c


.




These goals are realized with modulator structures and methods that facilitate the use of M quadrature modulators which modulate, at a reduced modulation rate R


s


/M, respective ones of M polyphase cosine elements and M polyphase sine elements with respective ones of interpolated I elements and interpolated Q elements to thereby form M polyphase modulated elements.




Modulator embodiments of the invention are thus significantly easier to realize for a high carrier frequencies F


c


and their structure can be substantially simplified.




The novel features of the invention are set forth with particularity in the appended claims. The invention will be best understood from the following description when read in conjunction with the accompanying drawings.











BRIEF DESCRIPTION OF THE DRAWINGS





FIG. 1

is a block diagram of a digital modulator embodiment of the present invention;





FIG. 2

is a flow chart that recites exemplary process steps in the digital modulator of

FIG. 1

;





FIGS. 3A and 3B

are diagrams of modulation constellations in embodiments of an encoder of the digital modulator of

FIG. 1

; and





FIG. 4

is a block diagram of a digital modulator that illustrates advantages of modulator embodiments of the present invention.











DETAILED DESCRIPTION OF THE INVENTION





FIG. 1

illustrates a digital modulator embodiment


20


of the current invention. Although a digital modulator that generates a carrier frequency F


c


must operate at a system sample rate R


s


which is at least twice the carrier frequency F


c


(to satisfy the Nyquist criteria), the structure of the modulator


20


facilitates the use of quadrature modulators which operate at a reduced rate R


s


/M. Accordingly, digital modulator embodiments of the invention are significantly easier to realize for a selected carrier frequency F


c


and their structure is significantly simplified.




In particular, the digital modulator embodiment


20


of

FIG. 1

includes an encoder


22


, an I polyphase filter


24


and a Q polyphase filter


25


, M quadrature modulators


28


, a multiplexer


30


, a digital-to-analog converter (DAC)


32


and a digital synthesizer


34


. The encoder


22


is coupled between the I and Q polyphase filters


24


and


25


and an input data port


36


and receives a data stream D


strm


from the port


36


.




Each of the M quadrature modulators


28


has a summer


38


and an I multiplier


40


that is coupled between the summer and a respective output of the I polyphase filter


24


. Each of the quadrature modulators also has a Q multiplier


42


that is coupled between its summer and a respective output of the Q polyphase filter


25


. The DAC


32


delivers a modulated analog signal S


mod






anlg




to an output port


44


and the multiplexer


30


is coupled between the DAC and the M quadrature modulators


28


.




The multiplexer


30


responds to an address generator


46


at a system rate R


s


which is supplied by a rate (clock) generator


48


. A divide-by-M divider


49


responds to the rate generator and supplies a reduced rate R


s


/M to the M quadrature modulators


28


and the digital synthesizer


34


.




An operational description of the digital modulator


20


is enhanced by preceding it with the following investigation of the digital synthesizer


34


. As shown in

FIG. 1

, this synthesizer includes a phase (φ) accumulator


50


which is coupled to a phase offset generator


54


and also includes M phase-to-amplitude (φ/A) generators


52


. One of the φ/A generators


52


is driven by the phase accumulator


50


and M−1 of the φ/A generators


52


are driven by the phase offset generator


54


.




The phase accumulator


50


is an N-bit accumulator that generates a primary periodic stream


51


of digital words at the reduced rate R


s


/M that is supplied by the divider


49


. The phase accumulator


50


counts modulo Mφ


s


wherein φ


s


is a programmable phase step and the accumulator has a capacity of C=2


N


so that the primary periodic stream of words repeats at an accumulator frequency of f


acmltr


=(φ


s


R


s


)/C.




The phase offset generator


54


offsets each word of the primary stream by φ


s


, 2φ


s


- - - (M−1)φ


s


to thereby provide M−1 secondary streams


53


in which each word in the primary stream has corresponding words in the secondary streams that are progressively offset by φ


s


, 2φ


s


- - - (M−1)φ


s


.




The primary and secondary word streams are applied to the φ/A converters


54


which convert them to words that represent respective amplitudes in a quantized cosine signal and a quantized sine signal. Therefore, at the reduced rate of R


s


/M, the digital synthesizer


34


provides M polyphase cosine elements of a quantized cosine signal and M polyphase sine elements of a quantized sine signal to the quadrature modulators


28


.




Attention is now directed to the flow chart


60


of

FIG. 2

which illustrates operational processes of the digital modulator


20


of FIG.


1


. In a first process step


61


, an encoder (


22


in

FIG. 1

) encodes data elements of the input data stream D


strm


into respective I and Q baseband symbols (indicated as I and Q at the outputs of the encoder


22


in

FIG. 1

) which map the data elements to corresponding states in a selected modulation constellation.




The selected modulation constellation can be any of various quadrature modulations, e.g., quadrature phase shift keying (QPSK), offset QPSK, π/4 QPSK, M-ary PSK and quadrature amplitude modulation (QAM). For example,

FIGS. 3A and 3B

respectively show a π/4 QPSK modulation constellation and a 16-state QAM constellation. The constellation


70


of

FIG. 3A

has four states


71


that lie on a common circle


72


and are rotated 45° from the I/Q axes. The modulation of

FIG. 3A

can represent 2 bits of data (2


2


=4) and, because the states


71


are equidistant from the constellation center, it generates a constant-amplitude modulated signal.




In contrast, the constellation


76


of

FIG. 3B

has 16 states


77


that are spaced apart by equal distances along the I and Q axes. The modulation of

FIG. 3B

can represent 4 bits of data (2


4


=16) and, because the states


77


have different distances from the constellation center, it generates a modulated signal whose amplitude is not constant. Many other quadrature modulation constellations are well known and may be practiced with the structure and method embodiments of the invention.




Process step


62


of

FIG. 2

describes operation of an M-branch I polyphase filter (


24


in

FIG. 1

) which interpolates the I baseband symbols into M interpolated I elements (which are indicated as


80


in

FIG. 1

) and an M-stage Q polyphase filter (


25


in

FIG. 1

) which interpolates the Q baseband symbols into M interpolated Q elements (which are indicated as


81


in FIG.


1


). If M=8, for example, the filters


24


and


25


would be finite impulse response (FIR) filters whose sequence of filter coefficients is partitioned into 8 subsequences (one on each filter branch) which are generally referred to as the polyphase components of the original sequence.




Polyphase filters are rate-conversion filters that interpolate an input stream of elements. When the polyphase components are multiplexed at the filter output, the input sample rate is increased by the number of polyphase components. In the M-branch I and Q polyphase filters of

FIG. 1

, however, the interpolated elements are not combined at the filter outputs but, rather, are provided to respective quadrature modulators


28


. Thus, the interpolated elements


80


and


81


in

FIG. 1

are delivered in parallel to the quadrature modulators.




Process step


63


summarizes the previously-described operation of the digital synthesizer


34


of FIG.


1


. In particular, this process step refers to a digital synthesizer that repeatably provides M polyphase cosine elements (shown as


82


in

FIG. 1

) of a quantized cosine signal and M polyphase sine elements (shown as


83


in

FIG. 1

) of a quantized sine signal.




In process step


64


, the M quadrature modulators (


28


in

FIG. 1

) modulate, at the reduced rate R


s


/M, respective ones of the M polyphase cosine elements and the M polyphase sine elements with respective ones of the interpolated I elements and the interpolated Q elements to thereby form M polyphase modulated elements that are indicated as


84


in FIG.


1


.




In process step


65


, the multiplexer (


30


in

FIG. 1

) multiplexes (i.e., sequentially selects) the polyphase modulated elements (


84


in

FIG. 1

) into a modulated digital signal S


mod






dig




that is indicated as


86


in FIG.


1


. Finally, the digital-to-analog converter


32


of

FIG. 1

converts, as recited in process step


66


, the modulated digital signal S


mod






dig




into a modulated analog signal S


mod






anlg




at the digital modulator's output port (


44


in FIG.


1


).




In another digital modulator embodiment, a digital inverse sinc filter


90


is inserted between the mulitplexer


30


and the DAC


32


as indicated by insertion arrow


91


in FIG.


1


. By imposing an x/sin x response on the modulated digital signal


86


, this filter compensates the inherent sin x/x (sinc) rolloff which is necessarily generated by the sampling process of the DAC


32


.




In another digital modulator embodiment, the sequential selection function of the multiplexer


30


can be replaced with a multi-input DAC which would replace the multiplexer


30


and the single-input DAC


32


of FIG.


1


. The multi-input DAC is preferably configured to simultaneously process its multiple input signals and form the modulated analog signal S


mod






anlg




by sequentially ordering the processed signals.




Another digital modulator embodiment is formed by replacing the multiplexer


30


(and the inserted inverse sinc filter


32


) with an interleaved inverse sinc filter


92


as indicated by replacement arrow


93


in FIG.


1


. The filter


92


is formed with M components that each process the M polyphase modulated elements


84


in a respective order that realizes a respective convolution of the modulated elements and the filter's impulse response coefficients. The interleaved filter


92


includes a multiplexer that sequentially selects the convolution products.





FIG. 4

illustrates a digital modulator


100


which processes the I and Q baseband symbols of

FIG. 1

with a single quadrature modulator


102


that is followed by an inverse sinc filter


103


and a DAC


104


to provide a modulated analog signal S


mod






anlg




at an output port


106


. A digital synthesizer


108


provides quantized cosine and sine signals


110


and


111


to the multipliers of the quadrature modulator


102


. The I and Q baseband symbols are initially processed through three rate-conversion filters


114


. Preferably, these filters are polyphase filters (typically referred to as halfband filters) that each have two polyphase components and realize an interpolation increase of two in the process sample rate.




A rate (clock) generator


116


is directly coupled to the DAC


104


, the inverse sinc filter


103


, the digital synthesizer


108


and the final pair of halfband filters. A divide-by-two divider


118


couples the rate generator


116


to the middle pair of halfband filters and a final divide-by-two divider


119


couples the rate generator


116


to the initial pair of halfband filters.




A frequency response diagram


120


indicates (with indication arrow


130


) a selected carrier frequency F


c


that is generated by the quadrature output signals


110


and


111


of the digital synthesizer. Given that carrier frequency selection, another frequency response diagram


122


indicates (with indication arrow


131


) the selected carrier frequency F


c


with baseband signals S


bb


that are mirrored about the carrier by the modulation process of the quadrature modulator


102


.




The Nyquist criteria requires that a signal having an upper band limit must be sampled at a rate equal to or greater than that limit. Accordingly, the system (sample) rate R


s


of the rate generator


116


is selected as shown in the diagram


122


so that R


s


/2 is above the carrier and its mirrored baseband signals. The carrier and its baseband signals will also be mirrored about the system rate R


s


and the lower of these mirrored signals is indicated in broken lines spaced just to the right of R


s


/2. The frequency range below R


s


/2 is typically referred to as the first Nyquist zone and the Nyquist criteria requires that the carrier and its mirrored baseband signals be completely within the first Nyquist zone (as in FIG.


4


).




Another frequency response diagram


124


indicates (with indication arrow


132


) the relationship (at the output of the rate-conversion filters


114


) between the baseband signals S


bb


, the selected system rate R


s


and the consequent R


s


/2 location. Note that the baseband signals S


bb


are mirrored about the system rate R


s


as shown in broken lines. Because the three halfband filters


114


increase the sample rate by a factor of 8, the sample rate at the input of the filters is R


s


/8. Accordingly, another frequency response diagram


126


indicates (with indication arrow


133


) the relationship (at the input of the rate-conversion filters


114


) between the baseband signals S


bb


, the reduced sample rate R


s


/8 and one half of this rate which is R


s


/16. As indicated in broken lines, the baseband signals are mirrored about the reduced sample rate R


s


/8.





FIG. 4

shows that a selected carrier frequency F


c


requires that the quadrature modulator


102


operate at a system rate R


s


that is at least twice the carrier frequency F


c


. If, for example, the selected carrier frequency F


c


is 500 MHz, the quadrature modulator must process signals at a rate not less than 1 GHz. As previously mentioned, it is quite difficult to realize a quadrature modulator that can operate at 1 GHz and if such a modulator is realized, it typically requires extensive pipeplined structures. Because the M quadrature modulators


28


of

FIG. 1

operate at a rate of R


s


/M, their realization is significantly easier and their structure is significantly simplified. Accordingly, the operation range of digital modulator embodiments of the invention is substantially enhanced.




It is noted that the data stream D


strm


at the input port


36


of

FIG. 1

may also be processed through rate-conversion filters to thereby lower the input sample rate. This is accomplished by inserting (shown by insertion arrow


140


) rate-conversion filters


142


(e.g., polyphase, halfband and cascaded integrator-comb) between the encoder


22


and the I and Q polyphase filters


24


and


25


.




The embodiments of the invention described herein are exemplary and numerous modifications, variations and rearrangements can be readily envisioned to achieve an equivalent result, all of which are intended to be embraced within the scope of the appended claims.



Claims
  • 1. A digital modulator that generates, at a system rate Rs and in accordance with a modulation constellation, a modulated analog signal Smodanlg from an input data stream Dstrm, the system comprising:an encoder that encodes data elements of said input data stream Dstrm into respective I and Q baseband symbols which map said data elements to corresponding states in said modulation constellation; an M-stage I polyphase filter that interpolates said I baseband symbols into M interpolated I elements; an M-stage Q polyphase filter that interpolates said Q baseband symbols into M interpolated Q elements; a digital synthesizer that repeatably provides M polyphase cosine elements of a quantized cosine signal and M polyphase sine elements of a quantized sine signal; M quadrature modulators that modulate, at a modulation rate Rs/M, respective ones of said M polyphase cosine elements and said M polyphase sine elements with respective ones of said interpolated I elements and said interpolated Q elements to thereby form M polyphase modulated elements; and a digital-to-analog converter that converts said M polyphase modulated elements into said modulated analog signal Smodanlg.
  • 2. The modulator of claim 1, further including at least one first rate-conversion filter that couples said encoder to said I polyphase filter and at least one second rate-conversion filter that couples said encoder to said Q polyphase filter to thereby reduce the rate of said data elements.
  • 3. The modulator of claim 2, wherein said first and second rate-conversion filters are polyphase filters.
  • 4. The modulator of claim 1, wherein said encoder is configured as a quadrature phase shift keying encoder.
  • 5. The modulator of claim 1, wherein said encoder is configured as a quadrature amplitude modulator encoder.
  • 6. The modulator of claim 1, wherein said synthesizer includes:a phase accumulator that generates a primary substream of digital words; M−1phase offset generators that offset said primary substream to thereby generate M−1 secondary substreams of digital words; and M phase-to-amplitude converters that convert respective digital words of said primary and secondary substreams to respective ones of said polyphase cosine elements and said polyphase sine elements.
  • 7. The modulator of claim 1, wherein each of said quadrature modulators includes:an I multiplier that multiplies interpolated I elements and cosine elements; a Q multiplier that multiplies interpolated Q elements and sine elements; and a summer that sums products of said I and Q multipliers.
  • 8. The modulator of claim 1, wherein said digital-to-analog converter is configured to process M polyphase modulated elements.
  • 9. The modulator of claim 1, further including a multiplexer that couples said M quadrature modulators to said digital-to-analog converter and that sequentially selects said polyphase modulated elements to form an output stream of modulated elements and wherein said digital-to-analog converter is configured to process said output stream into said modulated analog signal Smodanlg.
  • 10. The modulator of claim 1, further including an interleaved inverse sinc filter that couples said M quadrature modulators to said digital-to-analog converter and that multiplexes said polyphase modulated elements into an output stream of modulated elements and wherein said digital-to-analog converter is configured to process said output stream into said modulated analog signal Smodanlg.
  • 11. The modulator of claim 1, wherein M is 2.
  • 12. A digital modulator that generates a modulated analog signal Smodanlg at a system rate Rs from I and Q baseband symbols which map data elements of an input data stream to corresponding states in a modulation constellation, the system comprising:an M-stage I polyphase filter that interpolates said I baseband symbols into M interpolated I elements; an M-stage Q polyphase filter that interpolates said Q baseband symbols into M interpolated Q elements; a digital synthesizer that repeatably provides M polyphase cosine elements of a quantized cosine signal and M polyphase sine elements of a quantized sine signal; M quadrature modulators that modulate, at a modulation rate Rs/M, respective ones of said M polyphase cosine elements and said M polyphase sine elements with respective ones of said interpolated I elements and said interpolated Q elements to thereby form M polyphase modulated elements; and a digital-to-analog converter that converts said M polyphase modulated elements into said modulated analog signal Smodanlg.
  • 13. The modulator of claim 12, further including at least one first rate-conversion filter that precedes said I polyphase filter and at least one second rate-conversion filter that precedes said Q polyphase filter to thereby reduce the rate of said data elements.
  • 14. The modulator of claim 12, wherein said first and second rate-conversion filters are polyphase filters.
  • 15. The modulator of claim 12, wherein said synthesizer includes:a phase accumulator that generates a primary substream of digital words; M−1phase offset generators that offset said primary substream to thereby generate M−1 secondary substreams of digital words; and M phase-to-amplitude converters that convert respective digital words of said primary and secondary substreams to respective ones of said polyphase cosine elements and said polyphase sine elements.
  • 16. The modulator of claim 12, wherein each of said quadrature modulators includes:an I multiplier that multiplies interpolated I elements and cosine elements; a Q multiplier that multiplies interpolated Q elements and sine elements; and a summer that sums products of said I and Q multipliers.
  • 17. The modulator of claim 12, wherein said digital-to-analog converter is configured to process M polyphase modulated elements.
  • 18. The modulator of claim 12, further including a multiplexer that couples said M quadrature modulators to said digital-to-analog converter and that sequentially selects said polyphase modulated elements to form an output stream of modulated elements and wherein said digital-to-analog converter is configured to process said output stream into said modulated analog signal Smodanlg.
  • 19. The modulator of claim 12, further including an interleaved inverse sinc filter that couples said M quadrature modulators to said digital-to-analog converter and that multiplexes said polyphase modulated elements into an output stream of modulated elements and wherein said digital-to-analog converter is configured to process said output stream into said modulated analog signal Smodanlg.
  • 20. The modulator of claim 12, wherein M is 2.
  • 21. A digital modulator that generates M polyphase modulated elements at a system rate Rs from I and Q baseband symbols which map data elements of an input data stream to corresponding states in a modulation constellation, the system comprising:an M-stage I polyphase filter that interpolates said I baseband symbols into M interpolated I elements; an M-stage Q polyphase filter that interpolates said Q baseband symbols into M interpolated Q elements; a digital synthesizer that repeatably provides M polyphase cosine elements of a quantized cosine signal and M polyphase sine elements of a quantized sine signal; and M quadrature modulators that modulate, at a modulation rate Rs/M, respective ones of said M polyphase cosine elements and said M polyphase sine elements with respective ones of said interpolated I elements and said interpolated Q elements to thereby form M polyphase modulated elements.
  • 22. The modulator of claim 21, further including at least one first rate-conversion filter that precedes said I polyphase filter and at least one second rate-conversion filter that precedes said Q polyphase filter to thereby reduce the rate of said data elements.
  • 23. The modulator of claim 21, wherein said first and second rate-conversion filters are polyphase filters.
  • 24. The modulator of claim 21, wherein said synthesizer includes:a phase accumulator that generates a primary substream of digital words; M−1phase offset generators that offset said primary substream to thereby generate M−1 secondary substreams of digital words; and M phase-to-amplitude converters that convert respective digital words of said primary and secondary substreams to respective ones of said polyphase cosine elements and said polyphase sine elements.
  • 25. The modulator of claim 21, wherein each of said quadrature modulators includes:an I multiplier that multiplies interpolated I elements and cosine elements; a Q multiplier that multiplies interpolated Q elements and sine elements; and a summer that sums products of said I and Q multipliers.
  • 26. The modulator of claim 21, further including a multiplexer that sequentially selects said polyphase modulated elements to form a modulated digital signal Smoddig.
  • 27. The modulator of claim 21, wherein M is 2.
  • 28. A method of generating a modulated digital signal Smoddig at a system rate Rs from I and Q baseband symbols which map data elements of an input data stream to corresponding states in a modulation constellation, the method comprising the steps of:interpolating said I and Q baseband symbols into M interpolated I elements and M interpolated Q elements; synthesizing M polyphase cosine elements of a quantized cosine signal and M polyphase sine elements of a quantized sine signal; at a modulation rate Rs/M, modulating respective ones of said M polyphase cosine elements and said M polyphase sine elements with respective ones of said interpolated I elements and said interpolated Q elements to thereby form M polyphase modulated elements; and multiplexing said M polyphase modulated elements into said modulated digital signal Smoddig.
  • 29. The method of claim 28, wherein said interpolating step includes the step of processing said I and Q baseband symbols with polyphase filters.
  • 30. The modulator of claim 28, wherein M is 2.
US Referenced Citations (4)
Number Name Date Kind
5638401 Jones Jun 1997 A
5781076 Iwamatsu et al. Jul 1998 A
6240142 Kaufman et al. May 2001 B1
6384677 Yamamoto May 2002 B2