Abstract
A circuit that includes a Darlington transistor pair having an input transistor and an output transistor configured to generate an output signal at an output node in response to an input signal received through an input node is disclosed. The circuit has a feedback coupling network coupled between the output node and the input node for feeding back to the input node a portion of an amplified version of the input signal that passes through the input transistor. The circuit further includes a bias feedback network that includes a bias transistor and a resistive network that consists of only resistive elements such that no inductors and no capacitors are provided within the bias feedback network.
Claims
1. A circuit comprising: a Darlington transistor pair comprising an input transistor and an output transistor wherein a gate of the input transistor is coupled to an input node, a drain of the output transistor is coupled to an output node, and a source of the input transistor is coupled to a gate of the output transistor to generate an output signal at the output node in response to an input signal received through the input node; a feedback coupling network having pair of split feedback resistors coupled between the output node and the input node for feeding back to the input node a portion of an amplified version of the input signal that passes through the input transistor; and a bias feedback network comprising a bias transistor and a resistive network that is coupled between a gate of the bias transistor, the gate of the output transistor, and the source of the input transistor, wherein a drain of the bias transistor is communicatively coupled to the input node through a tap point between the pair of split feedback resistors, and a source of the bias transistor is coupled to a fixed voltage node, and the resistive network consists of only resistive elements such that no inductors and no capacitors are provided between the gate of the bias transistor, the gate of the output transistor, and the source of the input transistor.
2. The circuit of claim 1 wherein the portion of the amplified version of the input signal that passes through the input transistor is phase and amplitude modulated.
3. The circuit of claim 1 wherein the output signal is fed back to the input node via the feedback coupling network.
4. The circuit of claim 1 wherein resistance of the resistive network is sized to attenuate by at least 0.5 dB an RF signal that passes through the resistive network.
5. The circuit of claim 1 wherein resistance of the resistive network is sized to attenuate voltage of an RF signal that passes through the resistive network by a ratio of at least 0.95.
6. The circuit of claim 1 wherein a third-order intercept point (IP3) for the output signal having a frequency of 0.5 GHz improves at least 2.8 dB relative to a bias feedback network having a shunt capacitor.
7. The circuit of claim 1 wherein an IP3 for output power for the output signal having a frequency of 1 GHz improves at least 2.5 dB relative to a bias feedback network having a shunt capacitor.
8. The circuit of claim 1 wherein an IP3 for output power for the output signal having a frequency of 2 GHz improves at least 1.6 dB relative to a bias feedback network having a shunt capacitor.
9. The circuit of claim 1 wherein an IP3 for output power for the output signal having a frequency of 2.7 GHz improves around 1.1 dB relative to a bias feedback network having a shunt capacitor.
10. The circuit of claim 1 wherein an IP3 for output power for the output signal having a frequency of 3 GHz improves around 1.0 dB relative to a bias feedback network having a shunt capacitor.
11. The circuit of claim 1 wherein an IP3 for output power for the output signal having a frequency of 2 GHz and ranging in output power from 30 dB to 10 dB improves at least 1.6 dB relative to a bias feedback network having a shunt capacitor.
12. The circuit of claim 1 wherein an improvement in composite triple beat (CTB) for output power for the output signal approaches 2 dB over a span of cable television (CATV) frequencies that range from 75 MHz to 575 MHz in comparison to bias feedback network with a shunt capacitor.
13. A mobile terminal comprising: an antenna; a duplexer/switch coupled to the antenna; amplifier circuitry selectively coupled to the antenna through the duplexer/switch, the amplifier circuitry comprising: a Darlington transistor pair comprising an input transistor and an output transistor wherein a gate of the input transistor is coupled to an input node, a drain of the output transistor is coupled to an output node, and a source of the input transistor is coupled to a gate of the output transistor to generate an output signal at the output node in response to an input signal received through the input node; a feedback coupling network having pair of split feedback resistors coupled between the output node and the input node for feeding back to the input node a portion of an amplified version of the input signal that passes through the input transistor; and a bias feedback network comprising a bias transistor and a resistive network that is coupled between a gate of the bias transistor, the gate of the output transistor, and the source of the input transistor, wherein a drain of the bias transistor is communicatively coupled to the input node through a tap point between the pair of split feedback resistors, and a source of the bias transistor is coupled to a fixed voltage node, and the resistive network consists of only resistive elements such that no inductors and no capacitors are provided between the gate of the bias transistor, the gate of the output transistor, and the source of the input transistor.
14. The circuit of claim 13 wherein the portion of the amplified version of the input signal that passes through the input transistor is phase and amplitude modulated.
15. The circuit of claim 13 wherein the output signal is fed back to the input node via the feedback coupling network.
16. The circuit of claim 13 wherein resistance of the resistive network is sized to attenuate by at least 0.5 dB an RF signal that passes through the resistive network.
17. The circuit of claim 13 wherein resistance of the resistive network is sized to attenuate voltage of an RF signal that passes through the resistive network by a ratio of at least 0.95.
18. The circuit of claim 13 wherein a third-order intercept point (IP3) for the output signal having a frequency of 0.5 GHz improves at least 2.8 dB relative to a bias feedback network having a shunt capacitor.
19. The circuit of claim 13 wherein an IP3 for output power for the output signal having a frequency of 1 GHz improves around 2.5 dB relative to a bias feedback network having a shunt capacitor.
20. The circuit of claim 13 wherein an IP3 for output power for the output signal having a frequency of 2 GHz improves around 1.6 dB relative to a bias feedback network having a shunt capacitor.
21. The circuit of claim 13 wherein an IP3 for output power for the output signal having a frequency of 2.7 GHz improves around 1.1 dB relative to a bias feedback network having a shunt capacitor.
22. The circuit of claim 13 wherein an IP3 for output power for the output signal having a frequency of 3 GHz improves around 1.0 dB relative to a bias feedback network having a shunt capacitor.
23. The circuit of claim 13 wherein an IP3 for an output power for the output signal having a frequency of 2 GHz and ranging in output power from 30 dB to 10 dB improves around 1.6 dB relative to a bias feedback network having a shunt capacitor.
24. The circuit of claim 13 wherein an improvement in composite triple beat (CTB) for an output power approaches 2 dB over a span of cable television (CATV) frequencies that range from 75 MHz to 575 MHz in comparison to bias feedback network with a shunt capacitor.
Description
BRIEF DESCRIPTION OF THE DRAWINGS
(1) The accompanying drawings incorporated in and forming a part of this specification illustrate several aspects of the disclosure, and together with the description, serve to explain the principles of the disclosure.
(2) FIG. 1 is a circuit diagram of a simple embodiment of a prior art self-biased Darlington feedback topology.
(3) FIG. 2 is a circuit diagram of a prior art enhancement mode (E-mode) pseudomorphic high electron mobility transistor (PHEMT) implementation of the self-biased Darlington feedback topology of FIG. 1.
(4) FIG. 3 is a graph that provides a third order intercept (IP3) comparison between an E-mode PHEMT Darlington circuit and an Indium Gallium Phosphide (InGaP) heterojunction bipolar transistor Darlington circuit.
(5) FIG. 4 is a circuit diagram of a prior art linearized Darlington amplifier having a common base cascode transistor including a frequency dependent voltage reference.
(6) FIG. 5 is a graph that provides a third order intercept point (IP3) comparison between a conventional Darlington amplifier (not shown) and the linearized Darlington amplifier having a common base cascode transistor of FIG. 4.
(7) FIG. 6 is a related art circuit diagram of a simplified related art linearized field effect transistor (FET) feedback amplifier.
(8) FIG. 7 is a related art circuit diagram of a simplified related art linear FET feedback amplifier including electronic tuning.
(9) FIG. 8 is a related art graph depicting nominal broadband performance of the E-mode PHEMT of FIG. 6.
(10) FIG. 9 is a related art graph showing IP3 simulations for a frequency bias feedback network of FIG. 6 having a fixed tuning capacitor value and swept resistance values for a tuning resistor.
(11) FIG. 10 is a related art graph showing IP3 simulations for the frequency bias feedback network of FIG. 6 having a fixed tuning resistor value and swept capacitance values for the tuning capacitor.
(12) FIG. 11 is a related art graph showing IP3 simulations for the frequency bias network of FIG. 7 having a fixed tuning capacitor value with varistor tuning.
(13) FIG. 12 is a related art graph depicting IP3 sensitivity to tone spacing over frequency for the linear FET feedback amplifier of FIG. 6.
(14) FIG. 13 is a related art graph depicting IP3 sensitivity to tone spacing and power level for the linear FET feedback amplifier of FIG. 6.
(15) FIG. 14 is a related art graph depicting detailed third order intermodulation (IM3) sensitivity to tone spacing and power level for the linear FET feedback amplifier of FIG. 6.
(16) FIG. 15 is a related art graph depicting IP3 sensitivity to temperature and power level for the linear FET feedback amplifier of FIG. 6.
(17) FIG. 16 is a related art graph depicting IM3 sensitivity to temperature and power level for the linear FET feedback amplifier of FIG. 6.
(18) FIG. 17 is a related art circuit diagram of a general embodiment of a linear FET feedback amplifier having a resistor-inductor-capacitor (RLC) type tuning network.
(19) FIG. 18 is a related art circuit diagram of a first embodiment of the linear FET feedback amplifier employing an RLC type tuning network with a tuning inductor.
(20) FIG. 19 is a related art graph showing IP3 simulations for the RLC type tuning network of FIG. 18 having swept resistance values for the tuning resistor.
(21) FIG. 20 is a related art graph showing IP3 simulations for the RLC type tuning network of FIG. 18 having swept inductance values for the tuning inductor.
(22) FIG. 21 is a related art graph showing IP3 simulations for the RLC type tuning network of FIG. 18 with operation swept over a temperature from 40 C. to 120 C.
(23) FIG. 22 is a related art graph showing IP3 response over a separation frequency of a two tone IP3 simulation for the RLC type tuning network of FIG. 18.
(24) FIG. 23 is a related art graph showing that the RLC type tuning network of FIG. 18 has a fundamental broadband linearization over a wide dynamic range.
(25) FIG. 24 is a related art graph showing preferred feedback resistor ratios for various desired IP3 values for given frequencies.
(26) FIG. 25 is a related art circuit diagram of a second embodiment of a linear FET feedback amplifier employing an RLC type tuning network with a choke inductor.
(27) FIG. 26 is a related art graph of IP3 performance between employing the tuning inductor of the RLC type tuning network of FIG. 18 and the choke inductor of the RLC type tuning network of FIG. 25.
(28) FIG. 27 is a related art graph depicting nominal S-parameter characteristics for the related art amplifiers.
(29) FIG. 28 depicts user equipment (UE) in the form of a mobile terminal that incorporates embodiments of the linear FET feedback amplifier of the present disclosure.
(30) FIG. 29 is a circuit diagram of the linear FET feedback amplifier that is included in the mobile terminal of FIG. 28.
(31) FIG. 30 is a graph of the IP3 for the linear FET feedback amplifier of FIG. 29 with and without a capacitor in a bias feedback network versus frequency of a signal input at the input node.
(32) FIG. 31 is a graph of measured IP3 for a first realized sample of the linear FET feedback amplifier of FIG. 29 with and without a capacitor in the bias feedback network versus signal power at the output node.
(33) FIG. 32 is a graph of measured IP3 in dBm versus frequency for the first realized sample of the linear FET feedback amplifier of FIG. 29 with and without a capacitor in the bias feedback network versus frequency of a signal input at the input node.
(34) FIG. 33 is a graph of measured IP3 in dBm versus output power (P.sub.OUT) in dBm for the first realized sample of the linear FET feedback amplifier of FIG. 29 with and without a capacitor in the bias feedback network.
(35) FIG. 34 is a graph of measured IP3 in dBm versus frequency for a second realized sample the linear FET feedback amplifier of FIG. 29 with and without a capacitor in the bias feedback network.
(36) FIG. 35 is a graph of measured IP3 in dBm versus output power in dBm for the second realized sample of the linear FET feedback amplifier of FIG. 29 with and without a capacitor in the bias feedback network.
(37) FIG. 36 is a graph of improvement in composite triple beat (CTB) in dB versus frequency for a signal input into the input node of the linear FET feedback amplifier of FIG. 29 with and without a capacitor in the bias feedback network.
DETAILED DESCRIPTION
(38) The embodiments set forth below represent the necessary information to enable those skilled in the art to practice the embodiments and illustrate the best mode of practicing the embodiments. Upon reading the following description in light of the accompanying drawing figures, those skilled in the art will understand the concepts of the disclosure and will recognize applications of these concepts not particularly addressed herein. It should be understood that these concepts and applications fall within the scope of the disclosure and the accompanying claims.
(39) While the prior art circuits depicted in FIGS. 1, 2 and 4 provide good results for many applications, embodiments of the present disclosure provide even greater improvements for IP3-BW over a wideband frequency range while operating under harsh conditions such as a wide temperature range that may be experienced during operation.
(40) FIG. 6 depicts a simplified embodiment of a related art linear FET feedback amplifier 28 that includes a Darlington transistor pair 30 having an input transistor M.sub.1 and an output transistor M.sub.2 configured to generate an output signal at an output node 32 in response to an input signal received through an input node 34. The linear FET feedback amplifier 28 also includes a frequency bias feedback network 36 that is communicatively coupled between the gate of the output transistor M.sub.2 and the input node 34 for providing biasing to the Darlington transistor pair 30. The frequency bias feedback network 36 is also used for adjusting a phase and amplitude of an amplified version of the input signal that passes through the input transistor M.sub.1 and into the frequency bias feedback network 36. A feedback coupling network 38 is coupled between the output node 32 and the input node 34 for feeding back to the input node 34 a portion of the amplified version of the input signal that passes through the input transistor M.sub.1. In particular, the frequency bias feedback network 36 passes portions of RF and/or intermediate frequency (IF) signals to the input node 34.
(41) In greater detail, a tuning resistor R.sub.TUNE combined with the gate to source (Cgs) capacitance of a bias transistor M.sub.3 makes up a low pass filter that is in cascade with a resistor-capacitor (R-C) network made up of a tuning capacitor C.sub.TUNE and a filter resistor R.sub.GM1. An adjustment of the resistance value of tuning resistor R.sub.TUNE and/or the capacitance value of the tuning capacitor C.sub.TUNE can change the phase and amplitude of the RF and IF signals originating from the source of the input transistor M.sub.1, and in turn are applied to the gate of the bias transistor M.sub.3. The bias transistor M.sub.3 inverts and amplifies the RF and IF signals as well as DC signals that are coupled to the input node 34 through a feedback coupling network 38. The phase of the RF and IF signals that are fed back to the input node 34 may be tuned by the tuning capacitor C.sub.TUNE and the tuning resistor R.sub.TUNE in order to produce cancelled intermodulation (IM) distortion at the output node 32. Moreover, the bias transistor M.sub.3 can create additional IM products with advantageous phase and amplitude characteristics that can reduce the IM distortion at the output node 32. The tuning resistor R.sub.TUNE and the tuning capacitor C.sub.TUNE are tunable to optimize the phase and amplitude of an RF spectrum that includes, but is not limited to desired tones, third order intermodulation (IM3) tones and beat tones that are coupled to the input node 34 and amplified by the bias transistor M.sub.3. An upper frequency band linearity of the linear FET feedback amplifier 28 can be increased by as much as 10 dB by tuning the resistor R.sub.TUNE to around 150. Comparatively, a lower to mid-band linearity of the linear FET feedback amplifier 28 may be improved by 3 dB to 5 dB by decreasing the value of capacitance for the capacitor C.sub.TUNE by about 1 pF.
(42) The feedback coupling network 38 includes a split feedback resistor tap point 40 for a pair of split feedback resistors made up of a first feedback resistor R.sub.FB1 and a second feedback resistor R.sub.FB2. A feedback capacitor C.sub.F blocks DC signals and the resistor RDC implement a DC bias set. The split feedback resistor tap point 40 is usable to adjust the amount of RF and IF signal being fed back to the input of the amplifier in order to optimize desired linearity cancellation at RF and IF frequencies. The total value of the first feedback resistor R.sub.FB1 and the second feedback resistor R.sub.RFB2 generally sets the RF gain-bandwidth of the linear FET feedback amplifier 28. By adjusting the first feedback resistor R.sub.FB1 to be proportionally larger than the second feedback resistor R.sub.FB2, more of RF-IF-DC feedback signals output from the bias transistor M.sub.3 will be fed back to the input node 34. If the first feedback resistor R.sub.FB1 is proportionally smaller than the second feedback resistor R.sub.FB2 a smaller amount of the RF-IF-DC feedback signals will be fed back will be fed back to the input node 34 from the bias transistor M3. As a result, more of the RF-IF-DC feedback signals will be directed toward the output node 32. Thus, the split feedback resistor tap point 40 provides another way to control the amplitude of the RF-IF-DC feedback signals.
(43) In one embodiment, the tuning resistor R.sub.TUNE, the tuning capacitor C.sub.TUNE, and the filter resistor R.sub.GM1 make up a resistor-capacitor-resistor (R-C-R) low pass network 42. The tuning resistor R.sub.TUNE, the tuning capacitor C.sub.TUNE, and the filter resistor R.sub.GM1 are each sized such that a pass pole allows a majority of the IF signal or beat tone (f1-f2) signal to pass through to the gate of the bias transistor M.sub.3. A pass frequency may be in the range of 5-10 MHz in order to allow the passage of IM resulting from wideband communication modulation with minimum attenuation. The RF and IF signals passing through the R-C-R low pass network 42 are inverted by the bias transistor M.sub.3 and coupled to the input node 34 through the feedback coupling network 38. In this way a negative feedback from a DC to an IF frequency for cancelling unwanted beat frequencies is generated. The RF and IF signals passing through the R-C-R low pass network 42 will experience amplitude changes and a phase shift that can be optimized to generate fundamental and IM products (f1, f2, 2f1-f2, 2f2-f1) through the bias transistor M.sub.3 where the net result will be the cancellation of distortion at the output node 32. The tuning resistor R.sub.TUNE is a primary component for tuning RF and IF signal phases introduced to the gate of the bias transistor M.sub.3. The tuning resistor R.sub.TUNE allows phase and amplitude tuning control since the value of resistance for the tuning resistor R.sub.TUNE creates a second low pass filter pole with the input capacitance (Cgs) of the bias transistor M.sub.3.
(44) As mentioned above, the R-C-R low pass network 42 is a preferred and low cost implementation of a more general function of phase and amplitude control. It should be appreciated that other passive components such as inductors (not shown) can be employed in place of one or more of the filter resistor R.sub.GM1, the tuning capacitor C.sub.TUNE, and the resistor R.sub.TUNE to achieve phase and amplitude control of the frequency bias feedback network 36. Moreover, the R-C-R low pass network 42 is simplified to a simple resistor capacitor (RC) network by combining the resistances of the filter resistor R.sub.GM1 and the resistor R.sub.TUNE.
(45) The resistor R.sub.TUNE and the tuning capacitor C.sub.TUNE are sized to provide a low pass filter (LPF) transfer function that allows the adjustment of the phase and amplitude of RF signals and IF signals being fed back to the input node 34. By reducing C.sub.TUNE and/or increasing R.sub.TUNE, the phase and amplitude can be adjusted at the RF frequency for optimizing the RF signal and IM3 signal phase and amplitude that are fed back to the input of the amplifier through the bias transistor M.sub.3, and the coupling network. In one preferred embodiment of the linear FET feedback amplifier 28 (FIGS. 6 and 7), the resistor R.sub.TUNE may be 100.
(46) In the prior art self-biased Darlington feedback amplifier 10 (FIG. 1), C.sub.BYP1 (FIG. 1) is a relatively large capacitor that bypasses, and effectively isolates RF signals from modulating the bias transistor M.sub.3 in order to prevent the generation of unwanted non-linearity. In contrast, the value of C.sub.TUNE and a non-zero value R.sub.TUNE of the linear FET feedback amplifier 28 effectively provides a phase and amplitude shift of the RF and IF signals presented to the base of the bias transistor M.sub.3 which generates RF products that have phase and amplitude characteristics which help reduce the IM distortion at the output of the amplifier.
(47) The R-C-R low pass network 42 comprised of the filter resistor R.sub.GM1, the tuning capacitor C.sub.TUNE, and the resistor R.sub.TUNE may be replaced by electronic devices that provide a phase shifter and amplitude attenuator. FIG. 7 depicts an adaptation of the linear FET feedback amplifier 28 that provides electronic tuning control for the frequency bias feedback network 36 in accordance with the present disclosure. In this particular embodiment, the resistor R.sub.TUNE is replaced with a FET variable resistor (varistor) S1 that has a variable resistance that is electronically controllable via a first control signal VC1. Moreover, in this embodiment the capacitor C.sub.TUNE has a variable capacitance that is electronically controllable via a second control signal VC2. In this way, the R-C-R low pass network 42 becomes an electronically controllable phase shifter and amplitude attenuator. The variable capacitance version of C.sub.TUNE may be, but is not limited to a variable capacitor in the form of a varactor diode and a switchable capacitor array. By adjusting the first control signal VC1 and/or the second control signal VC2, linearity versus frequency profile may be tuned for a desired frequency response. As such, the R-C-R low pass network 42 can be realized as an electronically controllable phase shifter and amplitude attenuator.
(48) It is important to note that other embodiments of linear FET feedback amplifiers may only include the FET varistor S1 or may only include the electronically controllable version of the capacitor C.sub.TUNE. In a case in which only the FET varistor S1 is used, the capacitor C.sub.TUNE will have a fixed capacitance value. Alternately, if the electronically controllable version of the capacitor C.sub.TUNE is used, the FET varistor S1 is replaced with the resistor R.sub.TUNE (FIG. 6), which has fixed resistance value.
(49) FIG. 8 is a graph depicting nominal broadband performance of the linear FET feedback amplifier 28 of FIG. 6. The left side vertical axis of the graph represents dB units for the magnitude data for the scattering parameters S(1,1), S(2,1) and S(2,2). The right side vertical axis of the graph represents dB units for noise figure (NF) data for the linear FET feedback amplifier 28.
(50) The scattering parameter S(2,1) representing forward gain shows about a 15.5 dB gain with a 3 dB bandwidth (BW) that is greater than 4 GHz, which is sufficient for many of the popular wireless frequency bands such as the long term evolution (LTE) Advanced frequency bands. As shown in the graph of FIG. 8, the NF data for the linear FET feedback amplifier 28 is relatively good, being less than 3 dB from about 0.2 GHz to about 3.6 GHz. Moreover, the input return-loss (i.e., S(1,1)) is excellent, being less than 15 dB across the entire frequency range of 0.1 GHz to 4.0 GHz. Further still, the output return-loss (i.e., S(2,2)) is relatively good, being below 10 dB from about 0.1 GHz to about 3.2 GHz.
(51) FIG. 9 is a graph showing IP3 simulations for the frequency bias feedback network 36 (FIGS. 6 and 7) having a fixed tuning capacitor value and swept resistance values for a tuning resistor. In particular, the graph of FIG. 9 shows simulated IP3 versus frequency response for the linear FET feedback amplifier 28 (FIG. 6). A fixed value of 10 pF for C.sub.TUNE was used for the simulation. Two tone IP3 measurements were performed with a difference frequency of 1.3 MHz with output tones of 0 dBm. The FIG. 9 graph shows the IP3 vs. frequency performance for various values of R.sub.TUNE and compares the various improvements to a baseline case where R.sub.TUNE is set to zero Ohms. At a frequency of 2.3 GHz, the optimum R.sub.TUNE value is 150, which achieves over 8 dB of IP3 improvement. At an R.sub.TUNE of 100, the IP3 improvement is a relatively high 6 dB.
(52) FIG. 10 is a graph showing IP3 simulations for the frequency bias feedback network 36 (FIGS. 6 and 7) having a fixed tuning resistor value and swept capacitance values for the tuning capacitor. The graph of FIG. 10 provides simulated IP3 vs. frequency response for the linear FET feedback amplifier 28 for a fixed R.sub.TUNE of 100 and swept capacitance values for the capacitor C.sub.TUNE. Two tone IP3 measurements were performed with a difference frequency of 1.3 MHz with output tones of 0 dBm. The graph shows the IP3 versus frequency performance for various values of capacitance for C.sub.TUNE and compares the various improvements to a baseline case where the capacitance of C.sub.TUNE is equal to 0 pf. A desired effect of reducing the size of C.sub.TUNE is to improve the lower frequency IP3 response. At an intermediate value of 1 pF, the response shows a broad IP3 frequency response with an average IP3 improvement of 3 dB to 4 dB from 200 M Hz up to 1.5 GHz. A nominal IP3 of 47 dBm is achieved from a 5V source while drawing 105 mA of source current.
(53) FIG. 11 is a graph showing IP3 simulations for the frequency bias network of FIG. 7 having a fixed tuning capacitor value with varistor tuning. The graph of FIG. 11 provides simulated IP3 versus frequency response of the linear FET feedback amplifier 28 having a fixed capacitance equal to 10 pF for the tuning capacitor C.sub.TUNE. Two tone IP3 measurements were performed with a difference frequency of 1.3 MHz with output tones of 0 dBm. The graph shows the IP3 versus frequency performance for various values of tuning voltage, which is swept from 1.24V to 1.40V. Results depicted on the graph show an improvement of as much as 5 dB at a frequency of 2.3 GHz. The simulations depicted on the graph include nonlinearities attributable to the FET varistor S1.
(54) FIG. 12 is a graph depicting IP3 sensitivity to tone spacing over frequency for the linear FET feedback amplifier 28 (FIG. 6). The graph of FIG. 12 shows the IP3 sensitivity to tone spacing over frequency in comparison to a baseline performance and illustrates that the linear FET feedback amplifier 28 can maintain an advantage of about 4 dB of IP3 improvement at 2.3 GHz over swept tone spacings from 10 kHz to 5 MHz. Moreover, the linear FET feedback amplifier 28 achieves at least a 3 dB improvement over the enhanced frequency range from 2 GHz to 2.7 GHz.
(55) FIG. 13 is a graph depicting IP3 sensitivity to tone spacing and power level for the linear FET feedback amplifier 28 of FIG. 6. FIG. 13 illustrates the IP3 improvements over swept tone spacing and input power levels. An IP3 benchmark is typically specified at Pout=0 dBm. Since there is 15.5 dB of gain, this would correspond to an input power of about 15 dBm. An IP3 improvement of greater than 6 dB is maintained over tone spacing at an RF input power of 15 dBm.
(56) FIG. 14 is a graph depicting detailed IM3 sensitivity to tone spacing and power level for the linear FET feedback amplifier 28 of FIG. 6. The detailed Pout and IM3 graphs show that the linear FET feedback amplifier 28 significantly improves the IM3 suppressing up to an RF input power of 10 dBm (Pout=5.5 dBm, exceeding the typical gain block spec of Pout=0 dBm) while maintaining the IM3 3:1 slope over power. This validates that the linearization is fundamentally sound and works over at least the 20 dB of dynamic range indicated in FIG. 14.
(57) FIG. 15 is a graph depicting IP3 sensitivity to temperature and power level for the linear FET feedback amplifier 28 of FIG. 6. An IP3 benchmark is typically specified for an output power (Pout) that is 0 dBm. Since there is 15.5 dB of gain, this would correspond to an input power of about 15 dBm. As illustrated graphically in FIG. 15, an IP3 improvement that is greater than 3 dB is maintained over temperature at an RF input power of 15 dBm.
(58) FIG. 16 is a graph depicting IM3 sensitivity to temperature and power level for the linear FET feedback amplifier 28 of FIG. 6. The graph depicting IM3 shows that the linear FET feedback amplifier 28 significantly improves the IM3 suppressing up to an RF input power of 10 dBm with a Pout of at least 5.5 dBm, thereby exceeding the typical gain block specification of a Pout of only 0 dBm while maintaining an IM3 3:1 slope over a desired power range. These results validate that the linearization provide by the linear FET feedback amplifier 28 is fundamentally sound and works over at least the 20 dB of dynamic range as indicated by the graph of FIG. 16.
(59) FIG. 17 is a circuit diagram of a general embodiment of a linear FET feedback amplifier 44 with a frequency bias feedback network 46 having a resistor-inductor-capacitor (RLC) type tuning network 48. The general embodiment of the linear FET feedback amplifier 44 is similar to the simple embodiment of the linear FET feedback amplifier 28 (FIG. 6) with the exception of the frequency bias feedback network 46. In this case, the frequency bias network includes an inductor to enhance IP3 performance.
(60) FIG. 18 is a circuit diagram of a first embodiment of the linear FET feedback amplifier 44 employing the RLC type tuning network 48. In this first embodiment the inductor is a tuning inductor L.sub.TUNE. The RLC type tuning network also include the tuning resistor R.sub.TUNE and the tuning capacitor C.sub.Tune that is also present in the R-C-R low network 42 of the linear FET feedback amplifier 28 (FIG. 6). In this first embodiment of the linear FET feedback amplifier 44, the tuning inductor L.sub.TUNE is coupled between the tuning resistor R.sub.TUNE and the filter resistor R.sub.GM1. The tuning capacitor C.sub.TUNE is coupled between a fixed voltage node such as ground and another node that connects the tuning inductor L.sub.TUNE to the filter resistor R.sub.GM1.
(61) FIG. 19 is a graph showing IP3 simulations for the RLC type tuning network 48 (FIG. 18) having swept resistance values for the tuning resistor R.sub.TUNE. In particular, the graph of FIG. 19 shows an improvement in IP3 performance as a result of adjusting tuning resistor R.sub.TUNE. Markers are illustrated at both 700 MHz and 2 GHz. Up to a point improvements in IP3 are achieved as the resistance value of R.sub.TUNE is increased. However, as can be seen from the graph, it is not necessarily the case that higher values of resistance for R.sub.TUNE yield higher IP3 performance. For instance in this exemplary case, a maximum IP3 performance occurs at a frequency of 2 GHz. Generally the largest improvement in IP3 performance is obtained at lower frequencies with a marginal improvement in IP3 performance occurring at higher frequencies.
(62) FIG. 20 is a graph showing IP3 simulations for the RLC type tuning network of FIG. 18 having swept inductance values for the tuning inductor L.sub.TUNE. The graph in FIG. 20 illustrates that by fixing the resistance of R.sub.TUNE to a predetermined value, and then sweeping L.sub.TUNE, the higher 2 GHz frequency IP3 may be improved significantly. In this exemplary case, an L.sub.TUNE having an inductance value of 12 nH may yield a higher performance than an inductance value of 16 nH. This is due to a more gradual IP3 roll off. Moreover, the lower inductance value would be less sensitive to process and temperature variations. Further still, in this exemplary case, the IP3 performance improvement over the linear FET feedback amplifier 28 (FIGS. 6 and 8) without L.sub.TUNE is as high as 6.5 dB.
(63) FIG. 21 is a graph showing IP3 simulations for the RLC type tuning network 48 (FIG. 18) with operation being swept over a temperature range that extends from 40 C. to 120 C. The graph of FIG. 21 shows that the enhanced IP3 response is well behaved over a wide temperature range in an exemplary case wherein the tuning resistor R.sub.TUNE is set to a resistance of 200 ohms and the tuning inductor L.sub.TUNE is set to 12 nH. Thus, a practical design using typical resistance and inductance values will have a variation in IP3 performance that is less than 2.5 dB over a 160 C. temperature range.
(64) FIG. 22 is a graph showing IP3 response over a separation frequency of a two tone IP3 simulation for the RLC type tuning network of FIG. 18. As a result, the RLC type tuning network 48 (FIG. 18) is suitable for improving the IP3 performance during the operation of complex modulation schemes such as wideband code division multiple access (WCDMA).
(65) FIG. 23 is a graph showing that the RLC type tuning network 48 (FIG. 18) has a fundamental broadband linearization over a wide dynamic range. Furthermore, in the exemplary case wherein the tuning resistor R.sub.TUNE is set to a resistance of 200 ohms and the tuning inductor L.sub.TUNE is set to 12 nH, the enhanced IP3 response is well behaved over a 10 dB power range. As such, the graph of FIG. 23 demonstrates that the RLC type tuning network 48 (FIG. 18) does not employ a narrow band and narrow power level cancellation effect, but instead produces a fundamental broadband linearization over a wide dynamic range.
(66) FIG. 24 is a graph showing preferred feedback resistor ratios for various desired IP3 values for given frequencies. The graph also illustrates a desired resistor ratio (R.sub.FB2/(R.sub.FB1+R.sub.FB2)) for the feedback coupling network 38, which sets the bias network feedback signal that passes back to the input, resulting in a desired IP3 response at a given frequency. A resistor ratio of 1 indicates that a relatively small amount of feedback signal will be passed back to the input of the Darlington transistor pair 30 (FIG. 18), thereby minimizing an improvement in linearity for the linear feedback FET amplifier 44 (FIG. 18). A ratio of 0.5 indicates that a relatively large amount feedback signal will be passed back to the input of the Darlington transistor pair 30. An exemplary operational frequency of 2 GHz and a resistor ratio of 0.7 will produce a desirable IP3 performance.
(67) FIG. 25 is a circuit diagram of a related art linear FET feedback amplifier employing an RLC type tuning network with a choke inductor. In this embodiment, an RLC type tuning network 50 includes a choke inductor L.sub.CHOKE2 that is in series with filter resistor R.sub.GM1. This is in contrast to having the inductor L.sub.TUNE in series with the tuning resistor R.sub.TUNE as employed in the RLC type tuning network 48 (FIG. 18). As a result, an improvement in IP3 performance by tuning L.sub.CHOKE2 is not as great as tuning L.sub.TUNE of the RLC type tuning network 48. This lower improvement in IP3 performance is illustrated in simulation results shown in FIG. 26 and FIG. 27.
(68) FIG. 26 is a graph of IP3 performance when employing the tuning inductor L.sub.TUNE of the RLC type tuning network of FIG. 18 and the choke inductor L.sub.CHOKE2 of the RLC type tuning network of FIG. 25. In particular, FIG. 26 illustrates that the tuning inductor L.sub.TUNE has a relatively greater impact on improving IP3 than the choke inductor L.sub.CHOKE2 does. Therefore, the placement of any inductor is critical to IP3 performance. In this exemplary case, the tuning inductor L.sub.TUNE in series with the tuning resistor R.sub.TUNE in the RLC type tuning network 48 (FIG. 18) provides a relatively large IP3 performance advantage over the RLC type tuning network 50 (FIG. 25), which has the choke inductor L.sub.CHOKE2 in series with the filter resistor R.sub.GM1.
(69) FIG. 27 is a graph depicting nominal S-parameter characteristics for the related art amplifiers. In particular, the graph of FIG. 27 shows nominal broadband S-parameter characteristics. Specifically, the gain and return-loss characteristics indicate that a multi-octave amplification capability exists for both the first and second embodiments of the linear FET feedback amplifier 44.
(70) FIG. 28 depicts the basic architecture of user equipment (UE) in the form of a mobile terminal 52 that incorporates an embodiment of a linear FET feedback amplifier 90 depicted in upcoming FIG. 29. In particular, the linear FET feedback amplifier 90 is usable in power amplifier circuitry 54 of the mobile terminal 52. The mobile terminal 52 may include a receiver front end 56, a radio frequency (RF) transmitter section 58, an antenna 60, a duplexer or switch 62, a baseband processor 64, control circuitry 66, a frequency synthesizer 68, and an interface 70. The receiver front end 56 receives information bearing radio frequency signals from one or more remote transmitters provided by a base station (not shown). The linear FET feedback amplifier 90 is also usable in a low noise amplifier (LNA) 72 that amplifies a received signal. A filter circuit 74 minimizes broadband interference in the received signal, while down conversion and digitization circuitry 76 down converts the filtered, received signal to an intermediate or baseband frequency signal, which is then digitized into one or more digital streams. The down conversion and digitization circuitry 76 may include an intermediate frequency (IF) amplifier comprised of the linear FET feedback amplifier 28. The receiver front end 56 typically uses one or more mixing frequencies generated by the frequency synthesizer 68. The baseband processor 64 processes the digitized received signal to extract the information or data bits conveyed in the received signal. This processing typically comprises demodulation, decoding, and error correction operations. As such, the baseband processor 64 is generally implemented in one or more digital signal processors (DSPs).
(71) On the transmit side, the baseband processor 64 receives digitized data, which may represent voice, data, or control information, from the control system 66, which it encodes for transmission. The encoded data is output to the RF transmitter section 58, where it is used by a modulator 78 to modulate a carrier signal that is at a desired transmit frequency. The power amplifier circuitry 54 amplifies the modulated carrier signal to a level appropriate for transmission, and delivers the amplified and modulated carrier signal to the antenna 60 through the duplexer or switch 62.
(72) A user may interact with the mobile terminal 52 via the interface 70, which may include interface circuitry 80 associated with a microphone 82, a speaker 84, a keypad 86, and a display 88. The interface circuitry 80 typically includes analog-to-digital converters, digital-to-analog converters, amplifiers, and the like. Additionally, it may include a voice encoder/decoder, in which case it may communicate directly with the baseband processor 64. The microphone 82 will typically convert audio input, such as the user's voice, into an electrical signal, which is then digitized and passed directly or indirectly to the baseband processor 64. Audio information encoded in the received signal is recovered by the baseband processor 64, and converted by the interface circuitry 80 into an analog signal suitable for driving the speaker 84. The keypad 86 and the display 88 enable the user to interact with the mobile terminal 52, input numbers to be dialed, address book information, or the like, as well as monitor call progress information.
(73) FIG. 29 is a circuit diagram of the linear FET feedback amplifier 90 that is included in the mobile terminal 52 of FIG. 28. In this newly disclosed embodiment, a bias feedback network 92 comprises the bias transistor M3 and a resistive network 94 that is coupled between a gate of the bias transistor M3, the gate of the output transistor M2, and the source of the input transistor M1. The resistive network 94 consists of only resistive elements such as the filter resistor R.sub.GM1. As such, no inductors and no capacitors are provided between the gate of the bias transistor M3, the gate of the output transistor M2, and the source of the input transistor M1. This is in contrast to having the tuning inductor L.sub.TUNE in series with the tuning resistor R.sub.TUNE and the tuning capacitor C.sub.TUNE in a shunt configuration as employed in the RLC type tuning network 48 (FIG. 18). The resistive network 94 is sized to attenuate by at least 0.5 dB an RF signal that passes through the resistive network 94. Alternatively, the resistive network is sized to attenuate voltage of an RF signal that passes through the resistive network 94 by a ratio of at least 0.95
(74) Unexpectedly, a substantial improvement in IP3 performance is realized by removing the tuning inductor L.sub.TUNE and the tuning capacitor C.sub.TUNE from the RLC type tuning network 48. This significant improvement in IP3 performance is illustrated in simulation results shown in upcoming FIG. 30 through FIG. 36.
(75) The circuit diagram of the linear FET feedback amplifier 90 resembles the schematic of FIG. 6 in U.S. Pat. No. 8,390,380 and U.S. Pat. No. 8,704,598 of related art with the exception of the bias feedback network 92 that omits the tuning capacitor C.sub.TUNE. The related art patents use the RF shunting function of the tuning capacitor to filter and suppress RF amplifier distortion from modulating the bias transistor M3 that functions as a current mirror.
(76) The present disclosure removes the tuning capacitor C.sub.TUNE which results in a linearity improvement, as the absence of a shunt filter capacitor on the gate of the bias transistor M3 causes the bias transistor M3 to be modulated by the entire signal frequency spectrum that passes from a common RF node between M1 and M2 of the RF Darlington transistor pair 30 and which is then introduced into the input of the RF amplifier as intermodulation distortion. Intermodulation distortion typically creates greater amplified intermodulation products at the output of the amplifier. In contrast, the linear FET feedback amplifier 90 of the present disclosure takes advantage of passing the full signal frequency spectrum through the bias feedback network 92, while amplifying and inverting a signal comprising the spectrum through bias transistor M3. The process continues by reintroducing a portion of the amplified and inverted signal back into the input node 34 by adjusting the ratio of a resistive divider feedback coupling network made up of the first feedback resistor R.sub.FB1 and the second feedback resistor R.sub.FB2 in order to improve IP3 at the output node 32. The filter resistor R.sub.GM1 is sized to substantially attenuate the amplitude of the RF signal applied to the gate of the bias transistor M3.
(77) FIG. 30 is a graph of IP3 for the linear FET feedback amplifier 90 with and without the tuning capacitor C.sub.TUNE versus frequency of a signal input at the input node 34. The IP3 was simulated for an output signal produced at the output node 32 in response to the input signal. Output power P.sub.OUT was held at 20 dB throughout the simulation. Notice that the IP3 for the linear FET feedback amplifier 90 remains improved from 0.5 GHz to around 4 GHz. The IP3 for embodiments having the resistance-only type bias feedback network 92 is improved over related art embodiments having tuning capacitor C.sub.TUNE by as much as 2.8 dB at a frequency of 0.5 GHz when C.sub.TUNE is held to a capacitance value of 10 pF. Further still, an improvement for IP3 is predicted to remain greater than 2 dB at 1 GHz and provide at least 1.6 dB improvement for IP3 at 2 GHz.
(78) FIG. 31 is a graph of measured IP3 for a first realized sample of the linear FET feedback amplifier 90 of FIG. 29 with and without a capacitor in the bias feedback network 92 versus signal power at the output node 32 at a 2 GHz output frequency. At a power level of 30 dB, an improvement of about 1.4 dB is available, whereas at a power level of 20 dB an improvement of about 1.43 dB is seen. At a power level of 16 dB, an improvement of at least 1.5 dB is provided.
(79) FIG. 32 is a graph of measured IP3 in dBm versus frequency for the first realized sample of the linear FET feedback amplifier 90 of FIG. 29 with and without a capacitor in the bias feedback network 92. A delta (A) of at least 3.6 dBm was measured at a frequency of 2 GHz as the frequency of the input signal was swept from 1.6 GHz to 2.5 GHz. An average of at least 2 dB was measured across a 1.6 GHz to 2.5 GHz bandwidth.
(80) FIG. 33 is a graph of measured IP3 in dBm versus output power in dBm for the first realized sample of the linear FET feedback amplifier 90 of FIG. 29 with and without a capacitor in the bias feedback network 92. A greater than 3 dB improvement is achieved over a 6 dB power range by incorporating the resistor-only bias feedback network 92 in place of bias networks having a capacitor.
(81) FIG. 34 is a graph of measured IP3 in dBm versus frequency for a second realized sample of the linear FET feedback amplifier 90 of FIG. 29 with and without a capacitor in the bias feedback network 92. In this case, the bias feedback network 92 provides a definitive improvement in IP3 that is achieved over an 8 dB power range in comparison to having a capacitor in the bias feedback network 92. While impressive, this improvement in IP3 is not as high as expected. The discrepancy between expected and realized results is attributed to difference in linearization of loop gain for the feedback coupling network 38 (FIG. 29). Adjustment of the resistance ratio between the first feedback resistor R.sub.FB1 and the second feedback resistor R.sub.FB2 sets the loop gain for the feedback coupling network 38. Thus, a desired improvement in IP3 can be realized by adjusting the resistance ratio between R.sub.FB1 and R.sub.FB2 in at least some cases.
(82) FIG. 35 is a graph of measured IP3 in dBm versus output power in dBm for the second realized sample of the linear FET feedback amplifier 90 of FIG. 29 with and without a capacitor in the bias feedback network 92. An average of around 1 dB in IP3 improvement was measured over a power level range of 0 dBm to round 8 dBm.
(83) FIG. 36 is a graph of improvement in composite triple beat (CTB) in dB versus frequency for a signal input into the input node 34 of the linear FET feedback amplifier 90 of FIG. 29 with and without a capacitor in the bias feedback network 92. An improvement in CTB approaches 2 dB over a span of cable television (CATV) frequencies that range from 75 MHz to 575 MHz using the bias feedback network 92 in comparison to the related art bias feedback networks with a 10 pF shunt capacitor.
(84) Unlike the related art embodiments disclosed in U.S. Pat. No. 8,390,380 and U.S. Pat. No. 8,704,598, the present disclosure provides a topology that omits the need for an inductor and a capacitor in the bias feedback network 92. As such, the linear FET feedback amplifier 90 has reduced size and cost in comparison to the related art. However, it should be noted that although there are linearity enhancements provided by the present disclosure, embodiments disclosed in U.S. Pat. No. 8,390,380 and U.S. Pat. No. 8,704,598 can obtain similar or even greater amounts of linearity enhancement in selected frequency bands and power ranges through the use of frequency selective resistive, inductive and/or capacitive elements as shown in the related art frequency bias feedback network 46 shown in FIG. 18. Nevertheless, embodiments of the present disclosure provide substantial linearity enhancement with a minimal impact to size and cost.
(85) Those skilled in the art will recognize improvements and modifications to the preferred embodiments of the present disclosure. All such improvements and modifications are considered within the scope of the concepts disclosed herein and the claims that follow.
