The present invention relates generally to radio frequency signal receivers, and more particularly to mixers and related circuits for use in radio frequency receivers.
Radio frequency (RF) receivers are used in a wide variety of applications such as television, cellular telephones, pagers, global positioning system (GPS) receivers, cable modems, cordless phones, radios and other devices that receive RF signals. RF receivers all require frequency translation or mixing. For example, a television receiver may translate one channel in the band of 48 MHz to 870 MHz to an intermediate frequency of 44 MHz. And within the United States, FM radios will typically translate FM audio signals, which are broadcast in 200 KHz channels in the frequency band from 88.1 MHz to 107.9 MHz, to an intermediate frequency of 10.7 MHz.
The majority of today's RF receivers perform the necessary frequency translation or mixing using an oscillator and analog multiplier.
Mathematically, the RF input signal can be expressed as:
RF=AIN(cos ωINt) [1]
where AIN is the amplitude of the RF input signal, and ωIN is the radian frequency. Similarly the local oscillator signal can be expressed as:
LO=ALO(cos ωLOt) [2]
where ALO is the amplitude of the local oscillator signal, and ωLO is the radian frequency. The product formed at the output of the mixer can be expressed as:
One of these components forms the channel spectrum translated to the desired frequency and the other component can be filtered out. Oscillator 62 can be implemented, for example, by a tuned inductor-capacitor (LC) oscillator, a charge relaxation oscillator, or a ring oscillator.
An alternative is shown in
It would be desirable to have circuitry for use in an RF receiver that is suitable for higher frequency applications. Such circuitry and related methods are provided by the present invention, whose features and characteristics will become apparent from the subsequent detailed description of the invention and the appended claims, taken in conjunction with the accompanying drawings and this background of the invention.
The present invention is a unique tuner architecture that mixes an analog RF input signal and a digital local oscillator signal to generate a output signal at a desired IF frequency, including low-IF and zero-IF solutions. The unique tuner of the present invention has a number of advantages over previous implementations, such as improved performance for low-IF and zero-IF architectures and a significant reduction in interference between adjacent paths in a multiple tuner solution.
In part, the present invention is a tuner including a direct digital frequency synthesizer having an output terminal for providing a digital local oscillator signal having a frequency chosen to mix a channel to a desired frequency and including a mixer having a first input terminal for receiving a radio frequency signal, a second input terminal coupled to the output terminal of the direct digital frequency synthesizer, and an output terminal for providing an output signal at a desired frequency. In addition, the direct digital frequency synthesizer and the mixer can be combined in a single integrated circuit. And the mixer can include a transconductance amplifier having an input terminal for receiving the radio frequency signal and an output terminal for providing at least one current signal and a mixing digital-to-analog converter having a first input terminal coupled to the output terminal of the transconductance amplifier, a second input terminal coupled to the output terminal of the direct digital frequency synthesizer, and an output terminal for providing the output signal at the desired frequency. Still further, the transconductance amplifier can be made up of a plurality of current cells, and these current cells can be characterized as being binarily weighted. The tuner may also include at least one additional receive path on the single integrated circuit. And the one or more integrated tuners can received a variety of RF signal bands, including terrestrial television broadcasts and FM radio broadcasts. As described below, other features and variations can be implemented, if desired, and related methods can be utilized, as well.
In another embodiment, the present invention is a method for tuning a signal comprising the steps of generating a digital local oscillator signal having a frequency chosen to mix a channel to a desired frequency, receiving a radio frequency signal and mixing the radio frequency signal with the digital local oscillator signal to provide an output signal at the desired frequency. More particularly, the generating and mixing steps can be performed within a single integrated circuit. And the mixing step can include converting the radio frequency signal to at least one current signal and mixing the at least one current signal with the output from the direct digital frequency synthesizer. Still further, the converting step can include generating a plurality of current signals using a plurality of transconductor cells. Still further, additional tuning paths can be provide, and a variety of RF signal bands can be tuned, including terrestrial television broadcasts and FM radio broadcasts. As described below, other features and variations can be implemented, if desired, and related systems can be utilized, as well.
It is noted that the appended drawings illustrate only exemplary embodiments of the invention and are, therefore, not to be considered limiting of its scope, for the invention may admit to other equally effective embodiments.
The present invention provides a unique tuner architecture that mixes an analog RF input signal and a digital local oscillator signal to generate a output signal at a desired IF frequency, including low-IF and zero-IF solutions. The unique tuner of the present invention has a number of advantages over previous implementations, which will be described in more detail below.
In the discussion below,
In basic operation, transconductance amplifier 110 converts the RF signal from a voltage signal into a current signal. DAC 120 is a current multiplying DAC that receives the current signal at the output of transconductance amplifier 110, mixes it bit-by-bit with the N-bit mixing signal from DDFS 130, and sums the output current components to form IOUT. As a result of the mixing operation IOUT moves the spectral content of RF to sum and difference frequencies, namely fIN+fLO and fIN−fLO. DDFS 130 provides signal LO 306 at a frequency chosen to mix the DESIRED CHANNEL to another frequency of interest, such as baseband or another suitable intermediate frequency (IF). Current multiplying DAC 120 includes multiple mixing cells each weighted according to the order of the cell.
Converter cell 160 includes transistors 162, 164, 166, and 168. Transistor 162 has a drain for providing signal IOUT+, a gate for receiving a true data signal labeled “D”, and a source connected to the drain of transistor 152. Transistor 164 has a drain connected to the drain of transistor 162, a gate for receiving a complementary data signal labeled “DB”, and a source connected to the drain of transistor 156. Transistor 166 has a drain for providing signal IOUT−, a gate for receiving signal DB, and a source connected to the drain of transistor 152. Transistor 168 has a drain connected to the drain of transistor 166, a gate for receiving signal D, and a source connected to the drain of transistor 156.
Interface circuit 170 is shared between all current cells in DAC 120 and has an input terminal for receiving the N-bit LO signal from DDFS 130, and an output terminal for providing a multiple bit output signal labeled “D, DB”. D and DB are true and complement digital signals, respectively, of a digital signal pair corresponding to certain values of the N-bit LO signal as will be described further below, and interface circuit provides one pair for each converter cell.
Capacitor 142 has a first terminal connected to the drains of transistors 162 and 164, and a second terminal connected to the ground power supply voltage terminal. Capacitor 144 has a first terminal connected to the drains of transistors 166 and 168, and a second terminal connected to the ground power supply voltage terminal. Capacitors 142 and 144 serve as filter capacitors and are shared between all cells.
The RF input signal is represented as a differential voltage signal between VRF+ and VRF−. VBIAS is a bias voltage selected to keep transistors 154 and 158 operating in the triode (linear) region of their voltage-current characteristic. Thus as VRF+ and VRF− vary, they modulate the voltage at the sources of transistors 152 and 156, forming a differential current signal on the drains of transistors 152 and 156. The bits of the digital local oscillator signal LO cause transistors 162, 164, 166, and 168 to switch the current of current cell 150 from side to side. These currents are switched at the output speed of DDFS 130, fCLOCK, which is constrained by Nyquist's theorem to be greater than twice the maximum fLO. It is noted that the transistors 154 and 158 could also be operated in the saturation region of their voltage-current characteristics, if desired, and the region of operation for these transistors 154 and 158 could be dynamically selected during operation depending upon the nature of the received signal or based upon one or more other operational parameters.
The resolution (and hence the number of bits) required by DAC 120 can be determined by considering the worst-case energy in the undesired channels, since quantization noise will be mixed by the undesired channels into the desired band at the output of the mixer. A terrestrial television receiver may need to tune a relatively weak desired channel when the receiver is close to the transmitter of a relatively strong undesired channel. For example assume the desired channel has a signal strength of −83 dBm (where dBm represents a decibel power level with reference to a power level of 1 milliwatt dissipated across a 75 ohm load), an undesired channel has a signal strength of −15 dBm, and the minimum signal to noise ratio (SNR) required at the output of the mixer is 15 dB. The integrated quantization noise for the LO signal in a 6 MHz band for a 10-bit DAC clocked at 2 gigahertz (2 GHz) is −84 dBc (decibel level with respect to carrier frequency fLO). This quantization noise appears in every 6 MHz band from DC to fCLOCK/2 (1 GHz) and is mixed by the −15 dBm undesired channel into the desired channel's band at a −99 dBm level (−15 dBm+(−84 dBc)). The resulting SNR is thus −83 dBm−(−99 dBm)=16 dB, which is greater than the minimum required SNR of 15 dB. Thus a 10-bit DAC yields barely acceptable results while a 9-bit DAC would not.
The switching speed of the DAC, which determines the maximum fLO which can be created because the clock of the DAC must be greater than twice the maximum fLO, is determined by the on resistance of transistors 162, 164, 166, and 168 in
Interface circuit 170 converts the N-bit LO signal into pairs of true and complementary bits so that converter cell 160 can switch the currents differentially. In one embodiment, each of the D and DB signals as well as the current and converter cells are binarily weighted; thus the most significant bit pairs switch currents that are twice the currents switched by the second most significant bit pairs, the second most significant bit pairs switch currents that are twice the currents switched by the third most significant bit pairs, and so on. In this case interface circuit 170 provides 2N output signals consisting of N pairs of D and DB signals.
Preferably to achieve better performance, however, the less significant bits are binarily weighted as just described while the more significant bits are thermometer encoded. In a thermometer encoding scheme, the binary values will be switched using a corresponding number of equally-weighted currents. Thus M thermometer encoded bits switch 2M−1 equally-weighted current cells. The number of most significant bits that are thermometer encoded will vary depending on the desired performance, and the number of output pairs generated by interface circuit 170 will also vary accordingly.
The output signal is also preferably a differential signal formed between IOUT+ and IOUT−. In an alternative embodiment, however, if the drains of transistors 166 and 168 were connected to a reference voltage terminal, such as an analog ground terminal, the drains of transistors 162 and 164 would form a single-ended output signal.
It is noted that the embodiment for a tuner 100 as depicted in
Looking now to
The DDFS 130 must typically be clocked at a frequency greater than twice the LO frequency range to satisfy the Nyquist criteria. This implies that the DDFS 130 must be clocked at greater than 2-times the frequency range of interest (e.g., 2×(107.9-88.1 MHz) for FM radio signals). The DAC 620 will typically be clocked at an even higher rate. The DDFS 130 will generate a signal, which will get used by the DAC 620 to mix the RF signal to IF. The DDFS 130 generates these signals in discrete steps, where the step size is equal to the channel spacing in the broadcast spectrum (e.g., 200 KHz for FM broadcasts in the US). It is noted that the actual frequencies created by the DDFS can be arbitrary because the signals can be mixed up to the desired output frequency (e.g., FM broadcast spectrum frequency) in the DAC 620. And these DDFS frequencies can be chosen near DC to minimize power dissipation within the DDFS 130. It is also noted that for FM broadcasts where a fixed reference oscillator is used, the reference oscillator could provide a reference signal between about 10-20 MHz, and the output clock signal (fCLK) 652 of the clock circuit (CLK) 650 could be a digital clock signal at about 220 MHz.
In operation for the embodiment depicted in
It is noted that the total quantization noise is increased as M gets smaller while the quantization noise must reside from DC to fCLK/2. Lowering the quantization noise within the broadcast band of interest increases the noise outside that band. The in-band noise must meet a minimum requirement based upon an undesired channel signal mixing the quantization noise into the desired output IF frequency. For example, the LO signal 604 at the output of the modulator 610 contains signal and quantization noise. The LO signal 604 will mix the desired channel to the desired IF. An undesired channel Δf away from the desired channel will mix quantization noise Δf away from the LO into the desired IF frequency. In a situation where the undesired channel is much stronger than the desired channel, the quantization noise will need to be much lower than the LO signal in order to obtain the needed signal to noise ratio out of the receiver. Energy outside of the broadcast band is preferably attenuated by the frequency response of the RF input amplifier. Therefore, quantization noise that is outside of the broadcast band is allowed to rise to some extent. The order of the modulator 610, the coefficients of the modulator 610 and the number of bits the modulator 610 quantizes is chosen to get the appropriate signal to noise output. The digital M-bit signal 604 out of the modulator 610 feeds into the switching network 602. And the signal 604 can be an M-bit binary weighted output. The current output of the switching network 602 is the multiplication of the analog RF input signal and the digitally generated LO signal and, therefore, contains the IF output.
In one implementation, there can be M binary weighted transconductors that make up the transconductor block 110. These transconductors can convert the RF input voltage into M binary weighted currents. The currents are weighted, if desired, to match the bit weighting of the digital M-bit signal 604 from the modulator 610. The switches of the switching network 602 can then be matched up such that bit M of the LO signal 604 switches in or out transconductor M; bit M-1 of the LO signal 604 switches in or out transconductor M-1, etc. In another implementation, the M-bit LO signal 604 can be converted into a thermometer code of 2M−1 bits and correspondingly there are 2M−1 equivalent transconductors for transconductor 110. Still further, a portion of the M-bit LO signal 604 and the transconductors can both be converted to thermometer coding. In another implementation the thermometer coded transconductors for transconductor 110 can be scrambled with respect to the thermometer coded LO bits 604. By using scrambling, the transconductors that could show up as non-linearities could be made to appear as random noise.
The architecture of the present invention provides a number of advantages. For example, using the tuner of the present invention, the requirements of oscillator phase noise and spur performance are relaxed. Oscillator phase noise requirements in traditional architectures, such as
Another advantage of the architecture of the present invention is that the reference oscillator can be fixed instead of programmable. LC oscillators often require a tuning range that corresponds to a variability of the LC product of up to 50%. The programmability and variability of the LC product adds area and noise as compared to the fixed frequency oscillator that can be utilized with the architecture of the present invention.
In addition, as shown in the examples below, direct conversion (IF is zero) is possible with the architecture of the present invention. In the conventional architectures of
Another advantage of this architecture is that LO leakage is minimized. Just as there is no direct mechanism for the LO to couple to the RF path in the architecture of the present invention, there is no direct mechanism for the LO to radiate out through the input.
Linear mixing may also be utilized with the architecture of the present invention. In conventional mixers, the oscillator tends to have large non-linearities that manifest themselves as spurs in the frequency spectrum at multiples of the oscillator frequency. These non-linearities can mix energy in the RF spectrum to the IF frequency. The present invention allows for linear mixing where the non-linearities are much smaller than in the conventional architecture. In fact, the non-linearities can be made arbitrarily small through the use of scrambling or calibration.
It is noted that the embodiment for a tuner 100 as depicted in
Television Implementation Example
The majority of television tuners used today are discrete single conversion tuners. Typical television tuners have inputs for receiving a radio frequency (RF) signals from an antenna or cable source having channels in the range of 48 megahertz (MHz) to 870 MHz. A tracking bandpass filter is often used to receive the RF INPUT signal and to attenuate undesired channel energy in order to provide a filtered signal to an input of a low noise amplifier (LNA). An RF synthesizer is often used to control a variable local oscillator (LO) provide a mixing signal in the range of 95 to 911 MHz. The mixing signal is combined with the output of the LNA in a mixer, which mixes the desired channel to an intermediate frequency (IF) of 44 MHz. The output of the mixer is then amplified in a programmable gain amplifier (PGA) and is filtered in an IF filter having a center frequency at the conventional IF of 44 MHz and having a passband of 6 MHz. Thus the IF filter provides an output signal containing the desired channel and having frequency content primarily from 41 MHz to 47 MHz.
Discrete single conversion tuners suffer several disadvantages. These Tuners require large amounts of circuit board space due to the large number of discrete components. They also require RF expertise to lay out the circuit board to avoid undesirable signal cross coupling and interference. Tracking bandpass filters often need manual calibration, increasing production cost. Also the performance of these tuners varies significantly over temperature.
It has long been thought that a silicon-based television tuner could be manufactured cheaper and with more stable performance than a discrete tuner and that silicon-based television tuners would ultimately replace discrete television tuners. Unfortunately, existing silicon-based television tuners do not perform as well as discrete tuners and have not become significant in the marketplace.
Existing silicon-based television tuners have used a so-called “up/down” or double conversion architecture. A tuner in such an architecture can include an LNA, an up conversion mixer, an RF synthesizer, a local oscillator, a surface acoustic wave (SAW) filter, a PGA, a down conversion mixer, a local oscillator, and an IF filter. The LNA typically has an input for receiving the RF input signal from an antenna or cable source. The up conversion mixer typically has a first input connected to the output of LNA and a second input for receiving a first LO mixing signal generated using the RF synthesizer and the oscillator. The SAW filter is typically external to the integrated circuit and has an input connected to the output of this up conversion mixer. The down conversion mixer will have a first input connected to the output of the external SAW filter, which has been passed through the PGA, and will have a second input for receiving a second LO mixing signal. The output of the down conversion mixer is often passed through an IF filter.
In operation of the double conversion receiver, the up conversion mixer receives a first mixing signal LO at a frequency chosen to mix the selected channel to a frequency band centered around 1100 MHz. The external SAW filter separates the desired channel, centered around 1100 MHz, from the unwanted channels. The down conversion mixer mixes the signals down to the desired IF frequency of 44 MHz using the second LO mixing signal at 1056 MHz.
While the up/down or double conversion architecture does not require manual calibration during manufacturing and is stable over temperature, it has many deficiencies that make its performance inferior to the discrete tuner. These tuners use two high frequency oscillators. Because they are high frequency it is possible to implement them in silicon using inductor-capacitor (LC) oscillators. However, LC-based oscillators have many drawbacks that reduce their desirability. First, they are susceptible to electric and magnetic interference, which can create spurs (or tones) and noise and lower overall performance. Second, two oscillators which are close in frequency tend to lock to one another. To avoid locking, there needs to be a lot of isolation between the two oscillators, which is difficult to achieve. Third, the first oscillator's range is nearly 100% of its frequency which means that the LC product must vary by about a 4:1 ratio to successfully tune over this range (since frequency is proportional to the square root of one over the LC product). However this range of values is difficult to achieve in silicon. Such an oscillator would usually be implemented as many selectable LC oscillators but this approach requires a lot of integrated circuit area. Fourth, having multiple LC oscillators adds phase noise that can degrade performance for digital television applications.
Another disadvantage relates to external SAW filter. The SAW filter is required because undesired channels need to be attenuated by a large amount and only SAW filters have the desired transfer characteristic at such high frequencies. However, SAW filters are expensive. They need to be driven with a matched impedance, which increases power dissipation substantially. SAW filters are lossy. Also while SAW filters have good attenuation they have poor frequency selectivity and pass more than just the desired channel.
Another disadvantage relates to the mixing process in different signal environments. Cable television tuning requirements are very different from terrestrial television tuning requirements because of the difference in energy levels between a desired channel and undesired channels at adjacent frequencies. A cable head-end drives all channels with similar power levels and therefore a cable television tuner receives the desired and undesired channels at similar power levels. A terrestrial television receiver could be much closer to undesired channels' transmitters than to the desired channel's transmitter, leading to the undesired channels having much more signal energy than the desired channel. The tracking filter of a discrete tuner helps filter the undesired channels. However, since there is no tracking filter in a double conversion architecture and since the SAW filter passes more than the desired channel, the mixers see the large energy difference between the desired and the undesired channel. And this energy difference is very problematic since any spur or noise in the oscillator or non-linearity in the mixing process can mix the large undesired channel or channels into the desired channel and destroy the reception of the desired channel. The result is that the double conversion tuner has sufficient performance for some cable television applications, in which signal strength of all channels is nearly uniform, but poor performance as a terrestrial television tuner.
Mixer 220 has a first input terminal connected to the output terminal of tracking bandpass filter 204, a second input terminal for receiving signal LO1, and an output terminal for providing an in-phase baseband signal labeled “I”. Mixer 220 includes a transconductance amplifier 222 and a DAC 224. Transconductance amplifier 222 has an input terminal connected to the output terminal of filter 204, and an output terminal. DAC 224 has a first input terminal connected to the output terminal of transconductance amplifier 222, a second input terminal for receiving signal LO1, and an output terminal for providing signal I. Filter 226 has an input terminal connected to the output terminal of mixer 220, and an output terminal. PGA 228 has an input terminal connected to the output terminal of filter 226, and an output terminal.
Mixer 230 has a first input terminal connected to the output terminal of tracking bandpass filter 204, a second input terminal for receiving signal LO1-90°, and an output terminal for providing a quadrature baseband signal labeled “Q”. Mixer 230 includes a transconductance amplifier 232 and a DAC 234. Transconductance amplifier 232 has an input terminal connected to the output terminal of filter 204, and an output terminal. DAC 234 has a first input terminal connected to the output terminal of transconductance amplifier 232, a second input terminal for receiving signal LO1-90°, and an output terminal for providing signal Q. Filter 236 has an input terminal connected to the output terminal of mixer 230, and an output terminal. PGA 238 has an input terminal connected to the output terminal of filter 236, and an output terminal.
Up conversion mixer 240 includes an analog-to-digital converter (ADC) 242, a lowpass filter 244, a mixer 246, an ADC 248, a lowpass filter 250, a mixer 252, and a summing device 254. ADC 242 has an input terminal connected to the output terminal of PGA 228, and an output terminal. Lowpass filter 244 has an input terminal connected to the output terminal of ADC 242, and an output terminal. Mixer 246 has a first input terminal connected to the output terminal of lowpass filter 244, a second input terminal for receiving signal LO2, and an output terminal. ADC 248 has an input terminal connected to the output terminal of PGA 238, and an output terminal. Lowpass filter 250 has an input terminal connected to the output terminal of ADC 248, and an output terminal. Mixer 252 has a first input terminal connected to the output terminal of lowpass filter 250, a second input terminal for receiving signal LO2-90°, and an output terminal. Summing device 254 has a positive input terminal connected to the output terminal of mixer 246, a negative input terminal connected to the output terminal of mixer 252, and an output terminal. DAC 260 has an input terminal connected to the output terminal of summing device 254, and an output terminal. IF filter 262 has an input terminal connected to the output terminal of DAC 260, and an output terminal for providing an output signal of tuner 200 labeled “IF OUTPUT”.
In operation tuner 200 illustrates one example of an RF receiver in which the circuitry described in
The tuned output signal, which includes the desired channel plus attenuated undesired channels, is mixed to baseband as follows. The output of filter 204 is converted into a current signal by highly linear transconductance amplifier 222. The current signal is then mixed in DAC 224, which functions as a current multiplying DAC using LO1 as a mixing signal, to provide the in-phase baseband signal I. Likewise the output of filter 204 is converted into a current signal by a separate highly linear transconductance amplifier 232, and is mixed in DAC 234 using a phase-shifted version of LO1, namely LO1-90°, as a mixing signal to develop the quadrature baseband signal Q. DDFS 206, transconductance amplifiers 222 and 232, and DACs 224 and 234 are implemented as described above with reference to
Depending on which channel is chosen as the DESIRED CHANNEL, DDFS 206 creates an appropriate sine wave to mix the desired channel down to baseband (i.e. DC). For example if the desired channel is centered at 500 MHz, DDFS 206 creates a 500 MHz sine wave for LO1 as well as a 90° shifted sine wave for LO1-90°. In an alternate embodiment LO1-270° could be used as the phase-shifted version of LO1 in which case the Q signal would be an inverted version of the Q signal as shown. Note that according to Nyquist's theorem DDFS 206 and current multiplying DACs 224 and 234 need to be clocked at greater than twice the highest frequency sine wave. Thus FCLOCK>2*870=1.740 GHz and is preferably 2 GHz. The operation of the various circuits at those speeds is possible using currently-available CMOS or bipolar-CMOS (BICMOS) integrated circuit manufacturing technology.
The outputs of mixers 220 and 230 include the desired channel information spectrum and all other energy is filtered by filters 226 and 236. These outputs are then further processed at baseband. Filters 226 and 236 are anti-alias lowpass filters having a cutoff frequency of about 3 MHz. The outputs of filters 226 and 236 are amplified by PGAs 228 and 238 in order to increase the amplitude of small signals and to minimize the dynamic range required of ADCs 242 and 248 in up conversion mixer 240.
In up-conversion mixer 240 the clock rate of ADCs 242 and 248 is preferably below the frequency of any received signal (i.e., 48 MHz) to minimize interference created from the switching that gets reflected back into the analog signal, but is also as high as possible to minimize the order of filters 226 and 236. For use in a television receiver having an IF center frequency of 44 MHz, a clock rate of 40 MHz was chosen. Lowpass filters 244 and 250 provide additional attenuation for undesired channels and are implemented in the digital domain. In an alternative embodiment, up conversion mixer 240 could be implemented in the analog domain and in that case ADCs 242 and 248 would not be necessary. DDFS 208 is preferably clocked at 100 MHz to satisfy the Nyquist criterion for generating a 44 MHz mixing signal.
Even though a baseband digital signal can be used by most televisions available today with simple modifications, tuner 200 preferably provides the analog IF OUTPUT at a standard IF of 44 MHz, although any other desirable IF such as 38 MHz may be used as well. Thus it is necessary for tuner 200 to re-combine the baseband I and Q signals to reconstruct the full 6 MHz spectrum in the IF signal. Up conversion mixer 240 converts the filtered, gain-adjusted I and Q signals into the digital domain using ADCs 242 and 246. Conversion of these signals into the digital domain avoids generating local oscillator signals that can creates spurs or tones, allows a relaxation of the specifications of analog filters 226 and 236, and makes it easier to extract audio signals. Up conversion mixer 240 mixes the outputs of ADCs 242 and 246 (the digital I and Q signals) to IF using digital local oscillator signals LO2 and LO2-90° before combining them in summing device 254. The output of summing device 254 is converted back to analog using IF DAC 260 and filtered in IF filter 262 for driving off-chip. In other embodiments which interface to televisions at baseband, up conversion mixer 240, DAC 260, and IF filter 262 may be omitted.
Tuner 200 uses the circuitry described in
It should be apparent that the local oscillator and mixer circuit described in
FM Radio Implementation Example
In the United States, FM audio signals are broadcast in 200 KHz channels in the frequency band from 88.1 MHz to 107.9 MHz. Europe, Japan and other countries have frequency plans similar to the United States. FM receivers for broadcast audio convert the signal received at the antenna into audio signals that drive a speaker. These FM receivers can typically be divided into a RF tuner section and an IF section. The typical RF tuner filters out undesired channels and amplifies the RF spectrum at the desired channel frequency. Also, within the RF tuner, a programmable or tunable local oscillator (LO) creates a signal which is an IF frequency away from the desired channel. The desired channel and LO signal are then multiplied together in the mixer which creates sum and difference frequency components thereby translating the desired channel from RF to IF. The typical IF section contains an IF filter which filters out channels near the desired channel (these channels are often not adequately filtered by the RF tuner). The IF section also typically contains an IF amplifier and a limiter to remove amplitude information, an FM modulator to convert the FM modulated IF signal to an audio signal, and an audio amplifier for driving a speaker.
One advantage of the FM architectures utilizing a tuner 100 of the present invention is the IF filter utilized in traditional architectures can be removed. In conventional FM tuners, the output IF (typically 10.7 MHz) is fed into an external ceramic or SAW filter to remove adjacent channel signals prior to FM demodulation. This filter can be expensive. Because the present invention allows direct conversion (IF=zero) or low-If conversion (IF near zero) architectures, it is possible to remove the adjacent channel signals with high order analog or digital filters. The desired channel can then be mixed up in the analog or digital domain to the standard IF.
Multiple Tuner Implementation Examples
The tuner 100 of the present invention also allows multiple tuners to be implemented on the same integrated circuit. In conventional tuners, reception of multiple radio frequency bands (e.g., AM, FM, weather) requires multiple oscillators. Utilizing the present invention, only a single oscillator is required to receive multiple radio frequency bands because all LO frequencies can be digitally generated. Also, for conventional tuners it is extremely difficult to put two tuners that receive the same band on the same die and have them simultaneously tune to channels that are near to each other. This difficulty is due to the tendency for two oscillators to lock to one another (i.e., oscillate at the same frequency) if there is inadequate isolation between them. The architecture of the present invention does not suffer from this problem because the digital LOs for each tuner are generated from the same oscillator. The use of digital LO signals and mixing DACs, therefore, allows multiple tuners to be included on the same integrated circuit without causing interference problems caused by different analog LO frequencies that would exist if multiple tuners were included on a single integrated circuit using the prior art architectures of
With respect to
Further modifications and alternative embodiments of this invention will be apparent to those skilled in the art in view of this description. It will be recognized, therefore, that the present invention is not limited by these example arrangements. Accordingly, this description is to be construed as illustrative only and is for the purpose of teaching those skilled in the art the manner of carrying out the invention. It is to be understood that the forms of the invention herein shown and described are to be taken as the presently preferred embodiments. Various changes may be made in the implementations and architectures. For example, equivalent elements may be substituted for those illustrated and described herein, and certain features of the invention may be utilized independently of the use of other features, all as would be apparent to one skilled in the art after having the benefit of this description of the invention.
This application is a continuation-in-part application of the following applications: application Ser. No. 10/375,967 that is entitled “Local Oscillator and Mixer for Radio Frequency Receiver and Related Method,” which was filed on Feb. 28, 2003 now U.S. Pat. No. 6,778,117; and application Ser. No. 10/377,573 that is entitled “Tuner Suitable for Integration and Method For Tuning a Radio Frequency Signal,” which was filed Feb. 28, 2003 now U.S. Pat. No. 7,447,493, the entire text and all contents for both of which are hereby expressly incorporated by reference in their entirety.
Number | Name | Date | Kind |
---|---|---|---|
4346477 | Gordy | Aug 1982 | A |
4361906 | Sakamoto | Nov 1982 | A |
4464770 | Maurer et al. | Aug 1984 | A |
4626803 | Holm | Dec 1986 | A |
4653117 | Heck | Mar 1987 | A |
4803700 | Dewey et al. | Feb 1989 | A |
4817167 | Gassmann | Mar 1989 | A |
4866261 | Pace | Sep 1989 | A |
4926130 | Weaver | May 1990 | A |
4944025 | Gehring | Jul 1990 | A |
4975699 | Frey | Dec 1990 | A |
5251218 | Stone et al. | Oct 1993 | A |
5263194 | Ragan | Nov 1993 | A |
5440587 | Ishikawa et al. | Aug 1995 | A |
5442352 | Jackson | Aug 1995 | A |
5495512 | Kovacs et al. | Feb 1996 | A |
5544200 | An | Aug 1996 | A |
5640698 | Shen et al. | Jun 1997 | A |
5737035 | Rotzoll | Apr 1998 | A |
5826180 | Golan | Oct 1998 | A |
5828955 | Lipowski et al. | Oct 1998 | A |
5867535 | Phillips et al. | Feb 1999 | A |
6154640 | Itoh et al. | Nov 2000 | A |
6172569 | McCall et al. | Jan 2001 | B1 |
6177964 | Birleson et al. | Jan 2001 | B1 |
6222891 | Liu et al. | Apr 2001 | B1 |
6249179 | Maalej et al. | Jun 2001 | B1 |
6263195 | Niu et al. | Jul 2001 | B1 |
6278391 | Walker | Aug 2001 | B1 |
6321075 | Butterfield | Nov 2001 | B1 |
6330290 | Glas | Dec 2001 | B1 |
6377315 | Carr et al. | Apr 2002 | B1 |
6593805 | Kunieda et al. | Jul 2003 | B2 |
6600373 | Bailey et al. | Jul 2003 | B1 |
6639534 | Khoini-Poorfard et al. | Oct 2003 | B2 |
6711149 | Yano et al. | Mar 2004 | B1 |
6778117 | Johnson | Aug 2004 | B1 |
6785345 | Blazo | Aug 2004 | B2 |
6785529 | Ciccarelli et al. | Aug 2004 | B2 |
6799020 | Heidmann et al. | Sep 2004 | B1 |
6829311 | Riley | Dec 2004 | B1 |
6867693 | Radin | Mar 2005 | B1 |
6950047 | Piasecki et al. | Sep 2005 | B1 |
6963734 | Sorrells et al. | Nov 2005 | B2 |
7035595 | Kim et al. | Apr 2006 | B1 |
7139329 | Takahiko | Nov 2006 | B2 |
7183958 | Quilligan et al. | Feb 2007 | B2 |
7199740 | Ferguson et al. | Apr 2007 | B1 |
7227438 | Song et al. | Jun 2007 | B2 |
20010041532 | Tomasz et al. | Nov 2001 | A1 |
20020168951 | Paulus et al. | Nov 2002 | A1 |
20020177423 | Cowley | Nov 2002 | A1 |
20030083033 | Staszewski et al. | May 2003 | A1 |
20030112893 | Lee et al. | Jun 2003 | A1 |
20030199264 | Holenstein et al. | Oct 2003 | A1 |
20030223525 | Momtaz et al. | Dec 2003 | A1 |
20040169773 | Johnson | Sep 2004 | A1 |
20050117071 | Johnson | Jun 2005 | A1 |
20050239499 | Oosawa et al. | Oct 2005 | A1 |
20050266818 | Johnson et al. | Dec 2005 | A1 |
20060073800 | Johnson et al. | Apr 2006 | A1 |
20060083320 | Feher | Apr 2006 | A1 |
20060141973 | Behrens et al. | Jun 2006 | A1 |
20070085719 | Maxim et al. | Apr 2007 | A1 |
20080181337 | Maxim | Jul 2008 | A1 |
Number | Date | Country |
---|---|---|
WO 8701531 | Mar 1987 | WO |
Number | Date | Country | |
---|---|---|---|
20040176058 A1 | Sep 2004 | US |
Number | Date | Country | |
---|---|---|---|
Parent | 10375967 | Feb 2003 | US |
Child | 10784838 | US | |
Parent | 10377573 | Feb 2003 | US |
Child | 10375967 | US |