Not applicable.
1. Field of the Invention
This invention relates to the field of wireless communication. More particularly, this invention relates to the field of a direct conversion delta-sigma transmitter.
2. Description of Related Art
A continuing challenge in the transmission of wireless signals is the process of modulating a carrier signal, i.e., varying the amplitude and/or phase (frequency) in such a way as to convey information onto that carrier, in a mathematically ideal way without producing noise or distortion. Prior art systems have used either analog circuitry exclusively, or have used digital circuitry to generate a baseband or intermediate frequency (IF) waveform followed by conventional analog circuitry to translate the signal up to the required carrier frequency. In either case, the use of conventional analog circuitry has resulted in the generation of spurious signals due to the inability of analog circuitry to accomplish mathematically ideal operations.
U.S. Pat. No. 6,748,025, assigned to the assignee of the present invention, discloses the use of commutation in combination with delta-sigma conversion in the architecture of an A/D converter based linear receiver.
The present invention uses a delta-sigma modulator in combination with a commutator to realize an ultra-linear transmitter as well. The present invention achieves direction conversion of a signal from a lower frequency such as a baseband frequency to a higher frequency such as an RF frequency using at least one delta-sigma modulator in combination with a commutator. The commutator multiplies the signal by +1 and −1 on alternating half-cycles of a clock such as an RF carrier frequency clock. The resulting waveform contains both the original waveform upconverted to the carrier frequency, plus side lobes that can be filtered out by a bandpass filter (BPF) whose characteristics need not be precisely controlled. A more general transmitter uses a quadrature approach in which two delta-sigma modulators have outputs that are input into respective commutators. The two commutators are clocked 90 degrees apart such that the resulting commutated waveforms are 90 degrees out of phase with each other, i.e., in quadrature, to produce I and Q components. The I and Q components can then be combined. By controlling the generation of the two input waveforms, commutating the two waveforms in quadrature to produce I and Q components and then combining those components, the amplitude and phase and hence frequency of the resulting signal can be arbitrarily controlled. That is, any modulation scheme can be implemented. Because the transmitter circuit does not depend on any precision analog components or filters, the transmitter design is capable of producing virtually any arbitrary waveform over a wide range of frequencies.
When the clock signal is in the radio frequency range, the resulting output is a radio frequency signal which represents the baseband signal upconverted directly to radio frequency without the need for any additional upconversion. Thus, the transmitter eliminates the need for an intermediate frequency conversion step commonly used in radio transmitters.
Additionally, the resulting transmitter can transmit over a wide frequency range merely by changing the clock frequency. If the clock frequency is changed significantly, then the stop band of the BPF will also need to be moved, but moving the stop band of a BPF whose transition bands and other characteristics need not be precisely controlled is a relatively simple task.
Because the invention allows generation of an arbitrary transmitted waveform at an arbitrary carrier frequency merely by programming the device, the present invention achieves a highly flexible and programmable radio frequency transmitter. By reprogramming the device on the fly, the transmitter can be reprogrammed to change the modulation scheme and the carrier frequency essentially instantaneously.
In one aspect therefore, the invention is of a radio frequency transmitter that includes first and second delta-sigma modulators for generating first and second information signals, respectively, clock generation circuitry for generating first and second clocks in the radio frequency range and having a predefined phase difference of 90 degrees, first and second commutators for receiving the information signals and switching the information signals according to the first and second clocks to produce I and Q components of a complex signal, and a summer for combining the first and second commutated signals to produce a combined signal to be transmitted across a wireless network. A controller controls the generation of the first and second information signals such that the resulting transmitted waveform carriers the desired data using the desired modulation scheme.
In another aspect, the invention is of a method of generating a signal for transmission, the method including the steps of generating first and second data words and converting those data words into binary pulses whose pulse widths correspond to the value of the data words respectively, inverting first and second quadrature phase clocks when the first and second binary pulses are present respectively, and not inverting the first and second clocks when the first and second binary pulses are not present respectively, to produce I and Q components, summing the I and Q components, and filtering the summed signal to remove unwanted frequencies, to produce a combined radio frequency transmission signal.
In yet another aspect, the invention is of a transmitter having a delta-sigma modulator for receiving a digital input and for generating a baseband signal in response thereto, a commutator for multiplying the baseband signal alternately by a positive factor and a negative factor thereby commutating the signal to produce a second signal, and a filter for removing from the second signal radio frequency components that are significantly outside the range of one half the rate at which the commutation occurs.
Exemplary embodiments of the invention will be further described below with reference to the drawings, in which like numbers refer to like parts.
By converting the PWM output of the delta-sigma converter 10 to analog via 1-bit D/A converter 20 and then low pass filtering the analog signal via analog low pass filter (LPF) 22, the resulting output represents the magnitude of XIN/XMAX. When LPF 22 suitably suppresses the outband noise, the resulting analog waveform can have an effective resolution that is extremely high, e.g., greater than 22 bits.
Conventional delta-sigma D/A converter technology is capable of clock rates on the order of 10 MHz. Such clock rates are suitable for achieving high resolution output waveforms at audio rates, i.e., significantly lower than 50 kHz. High speed semiconductor technologies such as gallium arsenide (GaAs) and silicon germanium (SiGe) are capable of extending clock rates beyond the 4 GHz realm. However, such clock rates can only enable the generation of waveforms having 5-50 MHz of bandwidth, which is clearly insufficient to render a high speed modulated carrier waveform whose center frequency exceeds 2 GHz.
Commutation has been demonstrated to be an effective way of performing frequency translation without the use of an explicit multiplication operation, which for A/D converters cannot be performed distortion-free and for D/A converters requires the use of large amounts of digital circuitry. For example, the multiplication of two 16-bit words generates an output that is 32 bits wide. Furthermore, if explicit multiplication is used, the resulting parallel output would need to be driven into a GHz rate high resolution flash D/A converter, which although easier to design than an A/D converter, is still nevertheless difficult to implement in current generation semiconductor technologies.
Commutation has been previously shown to be an effective method of downconverting a signal as described in copending U.S. patent application Ser. No. 10/858,611, which is a continuation of Ser. No.09/241,994, now U.S. Pat. No. 6,748,025, which are hereby incorporated by reference.
With respect to what commutation does to the spectrum of an incoming baseband or otherwise low frequency signal, we note that in discrete-time commutation where a signal x(n) is multiplied by the sequence (−1)n where n is an integer:
The above result suggests that an input spectrum centered around ω=Δ will produce an output spectrum with images around ω=π+Δ, 3π+Δ, 5π+Δ, . . . Thus, an image centered around one half of the clock frequency results, with the nearest repetition of that image being centered around one and one-half times the clock frequency. This is effectively a translation of the spectrum by an amount ω=π. Furthermore, aliasing is avoided so long as the bandwidth of the spectrum is limited to Δω<π/2.
The basic concept behind a single-bit delta-sigma modulator is that any static voltage Vout (or time varying voltage at a suitably low frequency) with a value between two limits Vmin<Vout<Vmax can be conveyed by the long-term average of a pulse-width modulated (PWM) waveform that switches between Vmin and Vmax with a duty cycle ρ such that:
where the quantities Nmin and Nmax denote the number of sample periods during which the output of the PWM waveform is held at Vmin and Vmax, respectively. Needless to say, the equation correctly predicts that as the number of samples used to convey an output value (i.e., Nmin+Nmax) grows, the precision with which the value of Vout can be conveyed improves.
This principle also extends to multibit (multilevel) modulators used as part of delta-sigma modulation loops. Theoretically, delta-sigma converters employing multibit modulators can achieve better resolution than their one-bit counterparts because each sample produced by the modulator has greater resolution. However, realizing this improved resolution is difficult in practice because it is far easier to build a one-bit modulator with, say, 16 bits of precision (i.e., the output voltage is precise to within 0.0015%) than to build a multibit modulator with this same precision.
The analysis above has already shown that if a baseband signal conveyed by a delta-sigma modulator is commutated using a sequence (−1)n, then the spectrum of the baseband signal is translated upward in frequency resulting in a spectrum that is centered around one-half of the clock frequency. It follows, then, that if the output waveform of a delta-sigma modulator is upconverted through commutation, then the spectrum of this upconverted waveform will closely approximate that of the desired (ideal) analog waveform at frequencies close to the carrier frequency (half of the clock rate) and would deviate from the desired spectrum at frequencies far removed from the carrier frequency. The simulation in
Given that we have just demonstrated that a PWM waveform generated by a delta-sigma modulator can be driven into a commutator to form a signal whose spectral characteristic in the neighborhood of the carrier frequency closely matches that which would be produced by an ideal mixer, it then follows that two such signals can be added in quadrature to form a generalized complex modulation envelope. That is, we can independently vary the amplitude and phase and hence frequency of a carrier signal. This facilitates the generation of any arbitrary modulated carrier waveform since all modulation schemes involve manipulating the instantaneous amplitude and phase of a carrier.
In a sense, commutation is somewhat similar to biphase shift keying (BPSK) and multibit commutation is similar to quadrature amplitude modulation (QAM). Thus, a commutating delta-sigma converter may be regarded as rendering a sine wave (represented as a point on the complex plane) with a prescribed amplitude and phase by alternating back and forth between its two closest neighbors on a QAM (or QPSK for the single bit case) constellation.
A complete delta-sigma transmitter according to one embodiment of the present invention thus consists of a quadrature clock generator, two delta-sigma modulators (one each for I and Q), a pair of commutators (one for each delta-sigma modulator), a summing amplifier, and a bandpass filter. A block diagram of the delta-sigma transmitter is shown in
The passband of BFP 82 is ideally centered around half the clock frequency, with a passband wide enough to pass the two sidebands each of which have a width equal to the baseband signal. The stop bands need only substantially block the lobes that are produced as a result of the commutation at the clock frequency. As discussed the first lobe to be filtered out will be at a frequency of approximately one and a half times the frequency of the commutation clock signals. The design of band pass filters is well known within the art of filter design. The passband need not be centered exactly at one half the clock frequency. It is only necessary that BPF substantially pass half the clock frequency and the two sidebands produced by the baseband signal, and substantially stop the lobes centered around one and a half times the clock frequency and higher orders.
For radio frequency transmission the clock frequency will also be in the radio frequency range. Thus, the clock frequency may be greater than 1 MHZ, greater than 10 MHz, and greater than 1 GHz. This results in direct conversion of a baseband signal to radio frequency without the need for first converting to an intermediate frequency and then upconverting the intermediate frequency signal to radio frequency.
A controller (not shown) provides the stream of digital inputs to the delta-sigma modulators so that the resulting combined complex waveform represents the information to be transmitted according to the desired modulation scheme. Such a controller may be or may include a microprocessor, a digital signal processor, or other dedicated processor for generating the necessary data stream. Theory and certain details regarding delta-sigma modulators in general can be found in Delta-Sigma Data Converters: Theory, Design, and Simulation by Steven R. Norsworthy, Richard Schreier, and Gabor C. Temes (editors) (IEEE Press 1997). Theory and details regarding modulating carrier signals can be found in Modulation, Detection, and Coding by Tommy Oberg (Wiley 2001) and Electronic Communication Systems: Fundamentals Through Advanced by Wayne Tomasi (Pearson Education 2000) Using these references the practitioner skilled in the relevant art could generate the necessary PWM waveforms for synthesizing the desired complex output signal. In order to offload the processor, the data streams to the delta-sigma modulators could also be provided in part from values stored in sequential access memory (SAM) or randomly accessed access memory such as RAM or ROM accessed via direct memory access (DMA). The design of RF amplifiers and antennas suitable for use in the present application in order to physically transmit the signal produced is also well known.
It will be understood that the multiplication step need not implement multiplication by exactly +1 and −1, as the presence of a fixed amplification or attenuation at any stage in the process does not affect the general operation of the circuit. Thus, as used herein, the terminology of multiplying by +1 or −1 refers generally to multiplying by any positive and negative factor. Similarly, the summing steps and filtering steps, and other steps along the way, can also include amplification or attenuation by some factor without substantially affecting the analysis.
The commutator is a block that multiplies the input by +1 and −1 alternately during opposite half-cycles of the incoming clock signal. More generally, the commutator switches the incoming information signal according to the current state (HI or LO) of the clock signal. For analog signals, the commutator is realized by using a commutating amplifier implemented as a complementary output amplifier that alternates its output terminals during opposite half-cycles of the clock signal. For digital circuits, the concept of a commutator must take on a somewhat different realization. Although many implementation of a commutator are possible and will be apparent to those of ordinary skill in the art, the following addresses two possible implementations.
A simple one-bit commutator is implemented using an exclusive-or (XOR) gate as shown in
For a thermometric converter (i.e., one in which an N-bit word is represented by 2N−1 lines and in which an integer No is represented by No of those lines carrying a logic “1” and 2N−1−No of those lines carrying a logic “0”) a commutator can be implemented by driving an XOR gate with each of the 2N−1 output lines and summing the result.
A variety of methods are available for developing a pair of clock signals that are 90° out of phase with one another as will be apparent to those of skill in the art. Two such methods, both appropriate for implementation on a semiconductor integrated circuit, are elaborated here for illustration.
It will be understood that although a 90 degree phase difference for the quadrature clocks is ideal, a quadrature clock generator which generates clocks whose phase difference deviates only slightly from the ideal 90 degree predefined phase difference will still operate adequately within the transmitter of the present invention under most circumstances, so the phase difference need only be approximately 90 degrees or one quarter cycle.
The present invention allows a flexible, highly programmable transmitter to be implemented that does not require careful tuning or control of components. Using the present invention, a transmitter can be implemented that is capable of transmitting on a carrier frequency that can change by a least a factor of two under software control. The present invention could be implemented in nearly any wireless application, and is well suited for mobile cellular telephones and other spread spectrum applications including frequency hopping spread spectrum.
The present invention could also be used to transmit analog signals. By commutating an analog output such as the output of a one-bit delta-sigma D/A converter, or two delta-sigma D/A converters for the full quadrature case, the present invention allows a flexible analog signal transmitter to be constructed as well.
It will be appreciated that the term “present invention” as used herein should not be construed to mean that only a single invention having a single essential element or group of elements is presented. Similarly, it will also be appreciated that the term “present invention” encompasses a number of separate innovations which can each be considered separate inventions. Although the present invention has thus been described in detail with regard to the preferred embodiments and drawings thereof, it should be apparent to those skilled in the art that various adaptations and modifications of the present invention may be accomplished without departing from the spirit and the scope of the invention. Accordingly, it is to be understood that the detailed description and the accompanying drawings as set forth hereinabove are not intended to limit the breadth of the present invention, which should be inferred only from the following claims and their appropriately construed legal equivalents.