CLAIM OF PRIORITY

This patent application claims the benefit of priority to U.S. Application Serial No. 17/402,476, filed August 13, 2021, which is incorporated by reference herein in its entirety.

BACKGROUND OF THE INVENTION

Field of the Invention

This invention relates to Monolithic Microwave Integrated Circuit (MMIC) Power Amplifiers, and more particularly to dual-band MMIC Power Amplifiers.

Description of the Related Art

A typical RF transmitter may include an RF IC, a MMIC power amplifier and an antenna. The RF IC modulates a data signal onto an RF carrier signal. The data signal is typically at a relatively low rate e.g. ~ 1 GHz or ~ 10% of the center frequency of the RF carrier signal. Typically the RF IC can vary the RF carrier signal over a narrow frequency band e.g., 12-15 GHz or 42-45 GHz. The MMIC amplifies the modulated RF carrier signal to drive the antenna, which transmits the RF signal over the air.

A MMIC power ampifier is a distributed amplifier that is fabricated on a single chip using, for example, GaN, GaAs or SiGe (bipolar transistors or RF CMOS). The distributed amplifier includes a plurality of amplification stages connected in a chain. Each ampification stage includes a transistor biased to provide gain and a matching network (e.g. a lumped element inductor (L)/capacitor (C) or distributed circuit) to allow the amplified RF signal to flow from one amplification stage to the next. The peak-to-peak voltage (a proxy for RF power assuming current is constant) of the signal driven into each amplification stage determines whether that stage operates in its linear or compressed regions. Linear operation provides less distortion of the input signal but less amplified power.

In certain applications, it may be required that the RF transmitter have the capability to selectively transmit in multiple, typically dual, RF bands e.g. 16-18 GHz (K _u band) and 33-50 GHz (Q-band). As shown in FIG. 1, a network of satellites 10 communicate with each other in, for example, a 45 GHz band 12 and with ground data links 14 in, for example, a 12 GHz band 16. The lower frequency Ku-band being used to communicate through the atmosphere. The higher frequency Q-band providing higher bandwidth utilized for inter-satellite communication. In such an application the satellites require an RF transmitter that can amplify and transmit RF signals in both the Ku and Q bands.

One approach to realizing a dual band MMIC power amplifier is to have two channels, one MMIC channel that operates in the lower RF band and another MMIC channel that operate in the upper RF band to drive a dual-band antenna. Another approach is to design extremely wideband MMIC power amplifier that spans both the lower and upper bands.

Another approach is to have a MMIC that operates at the lower band and integrate a diode ring and a broadband local oscillator (LO) on the MMIC power amplifier. With this approach the data signal (with a waveform frequency denoted Fi) is mixed with the LO (with a waveform frequency denoted F2) generating harmonics using a process called heterodyning. The mixing process generates high and low frequencies F 1 - F2 and Fi + F2 (major lowest order mixing components). In most RF devices the low frequency mixed signal Fi - F2 is filtered. The high frequency mixed signal Fi + F2 is then passed to the amplifier and is eventualy routed to the antenna where the waveform is radiated into space as the transmission frequency. To vary the transmission frequency Fi + F2 the LO frequency F2 is altered. The frequency of the local oscillator is typically varied by tuning a voltage in the local oscillator denoted VLO. Consequently, the output frequency is a function of a voltage in the local oscillator denoted F2(VLO). Since it is possible to vary the input voltage into the local oscillator VLO the output MMIC frequency can be varied thereby effectively creating a dual (or multi) band power amplifier MMIC by the following formula Fi + F2(VLO).

SUMMARY OF THE INVENTION

The following is a summary of the invention in order to provide a basic understanding of some aspects of the invention. This summary is not intended to identify key or critical elements of the invention or to delineate the scope of the invention. Its sole purpose is to present some concepts of the invention in a simplified form as a prelude to the more detailed description and the defining claims that are presented later.

The present invention provides a dual -band MMIC power amplifier configured to receive RF input signals within a narrow input frequency band, amplify and upconvert at least one of the frequencies via a compressed non-linear response of the last amplification stage to output amplified signals in two different frequency bands. This allows for the use of a conventional narrowband RF IC to drive the MMIC and does not require additional circuitry (e.g., a local oscillator) on the MMIC power amplifier.

This is accomplished by modifying the matching network of the last amplification stage and controlling the RF frequencies fi and fi and amplitude of the input signals that are driven into the MMIC power amplifier. The matching networks of the first L amplification stages are configured to pass fi and fi and to reflect at least the 2 ^nd and 3 ^rd harmonics thereof. The matching network of the last (M ¹¹¹ ) amplification stage is configured to pass fi (or a harmonic thereof, namely N*fi where N = 1, 2 or 3) and block fi, to pass a P ^th harmonic of fi where P is typically 2 or 3 and to reflect any unused 1 ^st , 2 ^nd or 3 ^rd order harmonics of fi or fi. When the RF IC generates an input signal at fi, the MMIC power amplifier amplifies and outputs a signal at fi (or a harmonic thereof). When the RF IC generates an input signal at fi at sufficient RF power, the last amplification stage operates in compression. When a MMIC power amplifier operates in compression it has a non-linear transfer function. This results in the generation of higher-order harmonics of the input signal (f2 + 2fi + 3f2 + . . .). The amplifier’s matching network is designed such that only one harmonic, denoted the P ^th harmonic with a frequency Pfi is amplified and is the signal output from the MMIC. To summarize, in the example above when the MMIC is driven at a frequency fi, an amplified fi signal is produced by the MMIC. When the MMIC is driven at a frequency fi, an amplified Pfi is produced by the MMIC.

The RF IC is configured to generate the RF input signal at frequencies fi and fi that span a narrow input frequency band. The bandwidth choice is dependent upon the value of the frequencies chosen for amplification and the design of the matching network. At a minimum however, the bandwidth for fi denoted 2Afi (see: fi - Afi < fi < fi + Afi) and bandwidth of fi denoted 2Af2 (see: fi - Afi < fi < fi + Af2) needs to be such that fi + Afi < f2 - Afi. If the previous inequality is not satisfied, then signal leakage will occur across the bands and the matching networks in the device will attenuate or multiply undesired frequencies resulting in signal distortion. Moreover, the choice of frequencies fi and fi and associated bandwidths 2Afi and 2Af2 need to be selected so the harmonics correctly pass through the matching networks.

The RF IC may generate the RF input signal at frequency fi either at low RF power such that the last amplification stage (and all preceding stages) operate in the linear region or at high RF power such that the last amplification stage (and possibly some or all preceding stages) operate in the compressed region. If operating in the linear region, the MMIC power amplifier will output the N=1 harmonic of fi (i.e., the fundamental frequency). If operating in the compressed region, the MMIC power amplifier may be configured to output N*fi where N = 1, 2 or 3. In this later case, the unused harmonics are reflected back into the transistor to improve power added efficiency.

The matching network of the final amplification stage may be further configured to pass a Q ^th harmonic of a third input frequency fs where Q is 2 or 3. The frequencies fi, fi and f? are selected and the passbands of the matching network configured such that any unused 1 ^st , 2 ^nd or 3 ^rd order harmonics of fi, fi, or fi are reflected back into the amplification stage. 4 ^th or higher order harmonics are naturally attenuated to a point that they do not have to be reflected and add little to the recycled power.

The matching network of the next to last (M - 1 ^th ) amplification stage may be configured to pass fi (or a harmonic thereof), reflect fi and pass an R ^th harmonic of fi where R is 2 or 3 and to reflect any unused 1 ^st , 2 ^nd or 3 ^rd order harmonics of fi or f2 in order to cascade the higher order harmonics of the M - 1 ^th and M ^th amplification stages to output an amplified RF signal at a frequency of R*P*f2. Cascading of two or more stages allows the MIMIC power amplifier to reach much higher frequencies.

These and other features and advantages of the invention will be apparent to those skilled in the art from the following detailed description of preferred embodiments, taken together with the accompanying drawings, in which:

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1, as described above, illustrates communication between multiple satellites and ground data links over different RF bands in which the satellites communicate between themselves in a higher RF band and with the ground data links in a lower RF band;

FIG. 2 is a block diagram of an RF transmitter including an RF IC and a dual- band MMIC power amplifier in which the non-hnear properties of the transistor in the final stage are used to convert closely spaced frequencies into two amplified output bands at disparate frequencies that are fed into a dual-band antenna;

FIG. 3 is a schematic diagram of an embodiment of a 4-stage dual-band MMIC power amplifier in which the matched network of the final amplification stage is configured such that a frequency fi is amplified and a frequency fi is amplified with the final amplification stage in compression to amplify a harmonic P*f2;

FIGs. 4A-4C are a block diagram and power spectrum plot illustrating the amplification of frequency fi in both compression and linear regions of operation;

FIGs. 5A-5B are a block diagram and power spectrum plot illustrating the amplification and upconversion of frequency fi to P*f2;

FIG. 6 is a power spectrum plot for an alternative dual-band configuration;

FIG. 7 is a power spectrum plot for a three-band configuration; and

FIG. 8 is a schematic diagram of an embodiment of a 4-stage dual -band MMIC power amplifier in which the matched networks of the next to last and last amplification stages are configured to amplify a frequency fi and with both stages in compression to cascade the R ^th and P ^th harmonics of the stages to amplify a frequency R*P*f2.

DETAILED DESCRIPTION OF THE INVENTION

The present invention provides a dual-band MMIC power amplifier configured to receive input signals within a narrow input frequency band, amplify and upconvert at least one of the frequencies via a compressed non-linear response of the last amplification stage to output amplified signals in two different frequency bands. This allows for the use of a conventional narrowband RF IC to drive the MMIC and does not require additional circuitry (e.g., a LO) on the MMIC power amplifier. For example, an RF IC that can generate signals in the K _u band can be configured to drive the dual-band MMIC power amplifier to generate amplified signals in both the K _u and Q bands without modifying the RF IC or adding circuitry to the MMIC.

The difference between the linear and the compressed region depends on the relationship between the input and output voltage from the transistor amplifier. If the relationship between input and output voltage is linear, namely V _ou t = GVin where G is the amplifier gain, then the frequency input is the same as the frequency output if Vin = Vosin(27tfint). Now if the relationship between the input and output voltage is nonlinear, namely expressed as a finite or infinite polynomial series V _ou t = G7Vi _n + G2Vin + GjVm ³ + . . . then passing in a sinusoidal signal through the device as Vin = Vosm(27tfi _n t) and using trigonometric product identities (ex: sin ² (a) = - (l/2)cos(2a)) results in V _ou t composed of a sum of sinusoidal function with frequencies that are integer multiples of the input frequency fin, 2fi _n , 3fi _n , .... The power in each of the tones (fin, 2fi _n , 3fi _n , . . .) is dependent on the values of the A ¹¹¹ power coefficients Gk. For most real transistors the Gk coefficients decay in magnitude for higher order tones. Typically 4fi _n and higher order tones are not significant contributors to the total output waveform power. Therefore, driving a MMIC with a single tone waveform at frequency fi _n in the nonlinear region generates an output signal that is composed of a sum of tones that are integer multiples of the input drive frequency fin.

For most transistors driven at small RF input powers the voltage input to output response is linear. As the input RF power becomes progressively larger the voltage input to output response curves transition from a linear to a non-linear response (creating what is known as harmonic distortion). This process continues until in the transistor the compression is so extreme that no more gain is possible and the transistor cannot output any more power. Power amplifiers operate at maximum power added efficiency in compression, where some harmonic distortion is generated.

As used herein, when a transistor is described as being driven in the “linear region” it implies that the voltage input to output relationship is approximately linear as the transistor is driven at low power and no spurious tones are generated. When a transistor is described as being driven in “compression” it means the transistor is being driven at high input RF power and the resulting voltage input to output relationship is non-linear. This non-linear input to output voltage relationship generates multiple integer multiples of the input signal frequency in the output signal.

Referring now to FIGs. 2 and 3, in an embodiment an RF transmitter 20 includes an RF IC 22, a dual -band MMIC power amplifier 24 and a dual-band antenna 26.

A conventional off-the-shelf RF IC 22 (e.g, SiGe or SiGe CMOS) includes a frequency mixer 28 that modulates a carrier signal V _cs (fcs) 30 with a data signal Vdata(fd) 32 to produce a modulated carrier signal V _s (f _cs ) 34 referred to as the RF input signal. A filter 35 removes unwanted frequency products generated by mixer 28. The frequency fd of the data signal is much lower than the frequency of the carrier signal f _cs . A control voltage V _c is applied to a voltage controlled oscillator (VCO) 36 to generate carrier signal V _cs (fcs) 30. By varying control voltage V _c , the carrier frequency f _cs can be varied over an input frequency band to produce different but closely spaced frequencies fi and 12.

The RF IC is configured to generate the RF input signal at frequencies fi and fi that span a narrow input frequency band. The bandwidth choice is dependent upon the value of the frequencies chosen for amplification and the design of the matching network. At a minimum however, the bandwidth for fi denoted 2Afi (see: fi - Afi < fi < fi + Afi) and bandwidth of fi denoted 2Af2 (see: fi - Afi < fi < fi + Af2) needs to be such that fi + Afi < f2 - Afi. If the previous inequality is not satisfied, then signal leakage will occur across the bands and the matching networks in the device will attenuate or multiply undesired frequencies resulting in signal distortion. Moreover, the choice of frequencies fi and fi and associated bandwidths 2Afi and 2Af2 need to be selected so the harmonics correctly pass through the matching networks.

Dual-band MMIC power amplifier 24 (e.g. GaN, GaAs or SiGe (bipolar junction or RF CMOS devices)) is a distributed amplifier that includes a plurality of M amplification stages 40 operatively coupled in a chain to amplify the RF input signal 34. Each amplification stage 40 includes a transistor 42 biased to provide gain and a matching network 44 to allow the amplified RF input signal to flow from one amplification stage to the next. The matching network is shown as a simple LC circuit but is actually a plurality of lumped parallel-connected LC circuits or distributed matching networks designed to provide certain passband and rejection band characteristics.

The matching networks 44 of the first L<M amplification stages are configured with a passband 46 at the RF input frequency band and a rejection band 48 to reflect at least 2 ^nd and 3 ^rd order harmonics of the RF input signal. Accordingly, the RF input signal at either fi or fi is amplified and flows from one amplification stage to the next. If these intermediate stages are operated in compression mode, the higher order harmonics are reflected back into each stage and the power recycled to increase the power at the fundamental frequency fi or f2 thereby boosting MMIC power added efficiency.

In accordance with the present invention, the matching network of the M ^th (last) amplification stage is configured with a first passband 50 at an N ^th harmonic of fi where N is 1, 2 or 3 that rejects fi, a second passband 52 at a P ^th harmonic of fi where P is 2 or 3, and a rejection band 54 to reflect any unused 1 ^st , 2 ^nd or 3 ^rd order harmonics of fi or f2.lt is critical that frequencies fi and fi are different frequencies with a certain minimum spacing to avoid overlap of the filter/matching network passbands, fi and fi are suitably closely spaced so that they can be generated by an off-the-shelf (OTS) RF IC.

Dual-band antenna 26 suitably includes first and second antenna elements configure to transmit RF signals 56 at N*fi and P*f2. For example, the antenna may include a pair of patch antennas configured to resonate at N*fi and P*f2.

For dual-band operation of the RF transmitter, the RF IC 22 selectively generates the RF input signal 34 at frequencies fi and fi within the input frequency band.

At frequency fi, the RF input signal 34 is amplified by the transistor 42 at each amplification stage and flows from one stage to the next via matched network 44 through passband 46. At the last or M ^th amplification stage, the RF input signal is amplified, passes through passband 50 and is output at the N ¹¹¹ harmonic N*fi. If the peak-to-peak voltage of the RF input signal applied to the last stage is such that the transistor operates in the linear region than N=1. If the peak-to-peak voltage of the RF input signal (e.g., the RF power) applied to the last stage is such that the transistor operates in the compressed region, the transistor will generate multiple higher order harmonics 60 of the input signal such that N may equal 1, 2 or 3. The MMIC may be configured and driven to operate in compression to maximize the RF transmit power at fi (i.e. operate at the point of maximum power added efficiency). Alternately, the MMIC may be configured and driven to operate in the linear region to minimize signal distortion. Either is possible at drive frequency fi.

At frequency fi, the RF input signal 34 is amplified by the transistor 42 at each amplification stage and flows from one stage to the next via matched network 44 through passband 46. At the last or M ^th amplification stage, the RF input signal must have sufficient peak-to-peak voltage (RF power) to compress the transistor and generate the higher order harmonics 60. One of the higher order harmonics (P = 2 or 3) passes through passband 52 and is output at the P ^th harmonic P*fi. The fundamental frequency fi and the other unused harmonic are reflected by rejection band 54. For the case of N = 1, passband 50 is narrower than passband 46 in the preceding stages such that fi is passed but fi is rejected.

When driven into compression, the transistor generates the harmonics 60 of the RF input signal. For most transistors used in power amplification, the power in each successive harmonic exhibits a natural decay that is dependent on the transfer function of the specific transistor. To optimize RF output power, it is important that the matching networks are designed to reflect the power of any of the unused harmonics (including the fundamental) back into the transistor. Some of that power will recombine and increase the overall power added efficiency of the MMIC thereby increasing the overall output power of the selected fundamental or harmonic frequency. Any remaining power will be waste heat. However, the natural decay of the power spectrum means reflecting the 4 ^th harmonic and higher order tones yields a minimal improvement in device power added efficiency (as there simply isn’t that much power in these tones). Therefore, the MMIC power amplifier is limited to generating the fundamental and 2 ^nd or 3 ^rd harmonics. Simply put, 4 ^th and higher harmonics have so little power that engineering the matching network to reflect these harmonics is not necessary.

The dual-band MMIC power amplifier as described is a fixed or ‘dumb’ device, it merely acts on the RF input signals according to their frequency and power. The design of the dual-band MMIC power amplifier, and particularly the matching network for the last stage, is intimately tied to the selection of the specific input frequencies fl and fi. Given the complex matching network requirements of the dual-band MMIC, compared to single-band power amplifier MMIC, a narrower input frequency bandwidth around fi and fi will be a design compromise. To illustrate how this dual band MMIC architecture will result in reduced bandwidth around fi and fi consider what happens as fi and fi are varied slightly about their respective center frequencies. Varying fi may cause the fundamental to be rejected by the last stage instead of amplified and output or may place an unwanted harmonic of fi in the second passband resulting in distortion of the output signal. Varying fi may cause the desired P ^th harmonic to lie outside the second passband resulting in a rejection of the fundamental and harmonics and no output signal. Depending on the tightness of specifications for the passbands and rejection band some small variation in fi and fi may be tolerated. Therefore, the design of the matching/filter networks will place constraints on the allowable bandwidth around fi and fi (in addition to the natural constraints associated with the performance of the transistors themselves).

Referring now to FIGs. 4A-4C and 5A-B, a dual-band MMIC power amplifier 100 is configured to amplify an RF input signal at frequency fi and to amplify and upconvert an RF input signal at frequency fi to 3*fi using the non-linear properties of the compressed final amplification stage. The matching networks for the first M-l amplification stages are designed to exhibit a passband 102 that passes fi and fi within the input frequency band. The matching network for the last or amplification stage is designed to exhibit a passband 104 that passes fi while rejecting fi and a passband 106 that passes 3*fi.

As shown in FIG. 4B, for frequency fi, if the RF input power is sufficient to compress the final stage, harmonics at fi, 2fi and 3 fi are generated, fi is passed through passband 104 while 2fi and 3fi are rejected and recycled. It is critical that passband 106 does not pass either of these unwanted harmonics. This is why f2 must be different than fi and the frequencies carefully selected to avoid overlap and unwanted filtering and reflection of various frequencies.

As shown in FIG. 4C, for frequency fi, if the RF input power operates the final stage in its linear region than only the fundamental fi is generated and passed through passband 104.

As shown in FIGs. 5A-5B, for frequency fi, the RF input power must be sufficient to compress the final stage and generate harmonics at fi, 2fi, 3fi ... Although fi passed through the passband 102 in the preceding M-l amplification stages it lies outside passband 104 in the final stage and is recycled along with the 2 ^nd harmonic 2*fi. The 3 ^rd harmonic 3*fs passes through passband 106 and is output as the amplified RF signal. Although the 3 ^rd harmonic is naturally attenuated, the transitor gain is increased due to the recycled power from the fundamental and second harmonics.

Consider an example using real frequencies: if the input frequency band is the K _u band from 12-18 GHz the RF IC may be configured to generate fi = 12 GHz and fi = 15 GHz. Passband 102 in the first M-l stages is configured to pass frequencies between 11 Ghz to 16 GHz. In the final stage passband 104 is configured to pass frequencies from 11 GHz to 13 GHz and passband 106 is configured to pass frequencies from 44 GHz to 46 GHz. Assuming the final stage is compressed for both fi and fi, fi will generate harmonics at 12, 24, 36 and 48 GHz and fi will generate harmonics at 15, 30, 45 and 60 GHz. At fi, the RF input signal is amplified through all M stages producing an amplified RF signal at 12 GHz. At fi, the RF input signal is amplified at 12 GHz through the first M-l stages. At the final stage, fi is rejected and recyled into the final stage while the 3 ^rd harmonic is amplified and produces an amplified RF signal at 45 GHz.

With only a modification to the matching network of the final amplification stage and careful selection of fi and fi in the K _u band, the dual-band MMIC power amplifier can generate an amplified signal in the K _u band and the V band. This is achieved without requiring duplicative channels at the K _u and V band, without requiring an RF IC that supports frequncies across both bands and without additional circuitry on the MMIC.

Referring now to FIG. 6, the RF IC and dual-band MMIC power amplifier are configured to amplify and output RF signals at 2*fi and 3*fi. As shown, both fi and fi pass through a passband 200 in the first M-l amplification stages. The matching network for the final amplification stage includes a passband 202 at the 2 ^nd harmonic of fi (2*fi) and a passband 204 at the 3 ^rd harmonic of fi (3*fi). In both cases, the RF power of the input signal must be sufficient to compress the transistor in the final ampification stage to generate the higher order harmonics. Furthermore frequencies fi and fi must be selected such that the fundamental frequencies and, more particularly, the unused harmonics do not overlap and conflict with passband 204. The unused fundamental and 2 ^nd or 3 ^rd order harmonics are reflected and recycled to increase the output power of the transistor in the final stage at the desired harmonic of either fi or f ₂ .

For example, if the input frequency band is the K _u band from 12-18 GHz the RF IC may be configured to generate fi = 12 GHz and f ₂ = 15 GHz. Passband 200 in the first M-l stages is configured to pass frequencies between 11 Ghz to 16 GHz. In the final stage passband 202 is configured to pass frequencies from 23 GHz to 25 GHz and passband 204 is configured to pass frequencies from 44 GHz to 46 GHz. The final stage is compressed for both fi and f ₂ . fi will generate harmonics at 12, 24, 36 and 48 GHz and f ₂ will generate harmonics at 15, 30, 45 and 60 GHz. At fi, the RF input signal at 12 GHz is amplified through the first M-l stages. At the final stage, the fundamental at 12 GHz is reflected (and thus recycled) while the 2 ^nd harmonic is amplified and pass through passband 202 producing an amplified RF signal at 24 GHz. At f ₂ , the RF input signal is amplified at 12 GHz through the first M-l stages. At the final stage, f2 is reflected and recyled into the final stage while the 3 ^rd harmonic is amplified and produces an amplified RF signal at 45 GHz. In this configurations, the fundamental frequencies fi, f ₂ , the 3 ^rd harmonic of fi and the 2 ^nd harmonic of f ₂ are reflected and recycled.

Referring now to FIG. 7, the RF IC and dual-band MMIC power amplifier are configured to amplify and output RF signals at fi, 2*f ₂ and 3*fi. As shown, each of fi, f ₂ and f? pass through a passband 300 in the preceding M-l amplification stages. The matching network for the final amplification stage includes a passband 302 at fi that rejects f and f , a passband 304 at the 2 ^nd harmonic of fi (2*fi) and a passband 306 at the 3 ^rd harmonic of fs (3*fs). The last amplification stage must operate in compression for f2 and fi. Furthermore, frequencies fi, fi and fs must be selected such that the fundamental frequencies and, more particularly, the unused harmonics do not overlap and conflict with passband 304 and 306. The unused fundamental frequencies and 2 ^nd or 3 ^rd order harmonics are reflected and recycled to increase the output power of the transistor in the final stage at the desired fundamental or harmonic of either fi, fi or fs. The addition of a 3 ^rd output frequency makes selection of the input frequencies fi, fi and f? and the placement/characteristics of the passbands and bandwidth available for the 3 drive tones more demanding than the dual band case.

To illustrate a 3 band MMIC consider it being driven by the RF IC at the following frequencies in K _u band: fi = 13 GHz, fi = 15 GHz and fs = 17 GHz. Passband 300 in the firstM-1 stages is configured to pass frequencies between 12 Ghz to 18 GHz. In the final stage passband 302 is configured to pass frequencies from 12 GHz to 14 GHz, passband 304 is configured to pass frequencies from 29 GHz to 32 GHz and passband 306 is configured to pass frequencies from 50 GHz to 52 GHz. The final stage is compressed for at least fi and fi. If the MMIC is in compression, fi will generate harmonics at 13, 26, 39 and 52 GHz, fi will generate harmonics at 15, 30, 45 and 60 GHz and fi will generate harmoics at 17, 34, 51 and 68 GHz. At fi, the RF input signal at 13 GHz is amplified through all M stages to output an amplified RF signal at 13 GHz. At fi, the RF input signal is amplified at 15 GHz through the first M-l stages. At the final stage, fi is reflected and recyled into the final stage while the 2nd harmonic is amplified and produces an amplified RF signal at 30 GHz. At fs, the RF input signal is amplified at 17 GHz through the first M-l stages. At the final stage, fs is reflected and recycled into the final stage, while the 3 ^rd harmonic is amplified producing 51 GHz RF signal. Note, the 4 ^th harmonic of fi is at 52 GHz, which may partial pass through passband 306. However, the 4 ^th harmonic is naturally attenuated to such an extent that this is unlikely to be a problem. If it is a problem, fi could be operated in the linear region to avoid the generation of the overlapping 4 ^th harmonic.

Referring now to FIG. 8, the RF IC and dual-band MMIC power amplifier 400 are configured to amplify and output RF signals at fi and R*P*f2 where R=2 and P=3. The matching network 402 in the first M-2 stages is configured to exhibit a passband 404 that spans both input frequencies fi and fi. The matching network 402 for the M- 1 stage is configured to exhibit a passband 406 that passes fi and reflects fi and a passband 408 that pass R*f2 where R=2. The matching network 402 for the final stage is configured to exhibit a passband 410 that passes fi and reflects fi and a passband 412 that passes R*P*f2 where P=3. At fi, the RF input signal is amplified at each stage flowing from one stage to the next through passbands 404, 406 and 410 until being output as the amplified RF signal at fi. At fi, the RF input signal is amplified at fi by the first M-2 stages. The compressed M-l stage generates harmonics 414, of which the 2 ^nd harmonic is passed through passband 408 and input to the final stage. The compressed final M ^th stage generates harmonics 416, of which the 3 ^rd harmonic at 6*f2 is passed through passband 410 and output as the amplified RF signal at 6*fi. Cascading the non-linear effects of the last two stages greatly extends the reach of the second band of the dual-band MMIC power amplifier. The natural attenuation of cascading two harmonics will reduce the available output power, a portion of which can be recaptured by recyling the unused harmonics.

For example, if the input frequency band is the C and X band from 6-10 GHz the RF IC may be configured to generate f 1 = 6 GHz and f2 = 10 GHz. Passband 400 in the first M-2 stages is configured to pass frequencies between 6 Ghz to 10 GHz. In the next to last stage, passband 406 is configured to pass frequencies between 6 GHz and 9 GHz and passband 408 is configured to pass frequencies from 19-21 GHz. In the final stage, passband 410 is configured to pass frequencies between 6-9 GHz and passband 412 is configured to pass frequencies between 59-61 GHz. The final two stages are both compressed for fi. fi will generate harmonics at 10, 20, 30 and 40 GHz at the M-l stage and 20, 40, 60 and 80 GHz at the final M ^th stage. At the M-l stage, the 2 ^nd harmonic at 20 GHz passes through passband 408 to drive the final stage. In turn, the 3 ^rd harmonic at 60 GHz passes through passband 412 and is output as the amplified RF signal at 60 GHz.

While several illustrative embodiments of the invention have been shown and described, numerous variations and alternate embodiments will occur to those skilled in the art. Such variations and alternate embodiments are contemplated, and can be made without departing from the spirit and scope of the invention as defined in the appended claims.