CN101589547A - System and method for on-chip im3 reduction over a broad range of operating powers - Google Patents

Abstract

Sets of power amplifier branches are power combined within each amplifier stage and each set of branches are biased in different classes of operation by bias circuits possessing different impedance characteristics such that the fundamental frequency components present at the output are in-phase with one another and the IMD3 components are anti-phase over a range of power levels. The RF input signal is provided by the output of the previous stage, each stage being formed by power combining sets of power amplifier branches each separately biased so the fundamental components are additive, while the IM3 components cancel partially or completely. Using a feed forward control loop to monitor the input power and appropriately adjusting the bias currents and impedance characteristics of the bias circuits feeding the individual branches can provide additional IM3 reduction or cancellation over a large range of output powers.

Description

The system and method that third-order intermodulation product reduces on the sheet on the wide operational power range

Priority request

The application requires the priority of No. the 60/850146th, the U.S. Provisional Application submitted on October 6th, 2006, and the complete of this application openly is incorporated in this as a reference.

Technical field

The present invention relates generally to be used for the RF amplifier of communication equipment, more specifically, relate to and a kind ofly be used on wide frequency ranges reducing third-order intermodulation product level (third order intermodulationproduct level) automatically so that reduce distorted signals and keep the system and method for the high data rate of communication system.

Background technology

In all cell phones, pass through cellular front end (frontend) before at modulation signal from transceiver, all use amplifier to amplify this signal, this cellular front end is typically formed by passive block with to the switch of antenna, and signal is sent to the base station at described antenna place.Fig. 1 illustrates the senior explosion views of the circuit that comprises in the cell phone of prior art.Usually, base band integrated chip 1 provides signal to transceiver 2.This transceiver is sent to power amplifier 3 with the RF signal, and this power amplifier 3 outputs to front-end module 4 with the RF signal, and by antenna 5 the RF signal is outputed in the air.All these assemblies are included in the typical cell phone 6.

Two key factors that promote the design of RF/ microwave power amplifier 3 are the linearity and efficient.As used herein, the linearity is meant the ability that equipment amplifies undistortedly, and efficient is meant that the least possible waste energy of equipment ground is the ability of RF/ microwave power with the DC power transfer.In traditional Designing power amplifier, the improvement in field can cause falling back of another field usually.The RF/ microwave amplifier has two operating areas: linear and non-linear.In the range of linearity, the input signal envelope is exaggerated, and does not have distortion at output.For big peak to peak incoming signal level, amplifier enters nonlinear area, and the output signal distortion that becomes.

Distortion in the RF/ microwave amplifier is normally by amplitude amplitude limit (amplitude clipping), cause as the phase change of the function of signal amplitude and intermodulation product.When peak to peak input signal envelope amplitude was extended to beyond the range of linearity of amplifier, the amplitude amplitude limit took place.When peak to peak input signal envelope amplitude is extended to outside the range of linearity of amplifier, also cause phase change along with signal amplitude.As the nonlinear result in the amplifier transfer function, intermodulation distortion (IMD) takes place, thereby cause at mix products (mixing products) input signal and that generate with the difference frequency place.

Third-order intermodulation product (IMD3) is very attractive, because this product is in close proximity to carrier signal on frequency spectrum.Because so near carrier signal, IMD3 is very difficult to be eliminated or even reduces and the IMD3 limiting factor of the linearity aspect of RF/ microwave amplifier normally.Output third order intermodulation intercept point (OIP3, output third-order intercept point) (being also referred to as OTOI (output third order intermodulation intercepting)) is defined as the crosspoint between 3: 1 slopes of extrapolation of 1: 1 slope of extrapolation of fundamental frequency power output and third-order intermodulation product.If extrapolate well in the range of linearity, then OIP3 (OTOI) becomes the useful standard of the linearity that is used for the predicted power amplifier.Therefore, OIP3 (OTOI) point is high more, and then power amplifier is linear more.As mentioned above, the IMD that reduces power amplifier has improved its linearity, and has therefore improved OIP3 (OTOI).

With reference to figure 2, power amplifier 3 comprises one or more grades of 7a, the 7b...7n that is connected in series by matching network 8a, 8b, 8c...8n, and described matching network 8a, 8b, 8c...8n are made up of the circuit layout of zero or a plurality of resistor, capacitor and inductor.Each grade 7a, 7b...7n are made up of the many 18a of branch, 18b...18n that are connected in parallel.Each branch 18 has one or more unit cells that are connected in parallel 20.A unit cell is made up of the one or more transistors in the circuit layout of zero or a plurality of resistor, capacitor and inductor.

Each power-amplifier stage 7 is biased the biasing of circuit (not shown) usually, and this biasing circuit provides suitable curtage so that operate for described branch in the single operation kind.As used herein, operation species is that the percentage of the input sinusoidal signal by unit cell conducting in each branch therein and conduction current is recently determined.For example, in kind A operation, branch all is biased, and makes them connect and conduction current for 360 degree of input sinusoidal signal.In kind B, branch connects and conduction current for 180 degree of input signal.Approximate (near) kind B bias condition is to be used for the connection of wherein said branch and to conduct angle approaching the still situation more than 180 degree.In kind AB, described branch typically connects and conduction current at 270 degree or about 270 degree, but can change between 180 degree and 360 degree.The restriction of each operation species is not strictly set, and is used to roughly to understand the purpose of the operating condition of amplifier herein.

Attempted the response that several methods is improved power amplifier.A kind of method is with lower power level operation RF/ microwave amplifier, remains in the range of linearity so that guarantee equipment.The defective of this method is when equipment during with the operation of lower power level, and its efficient than with higher power level operation the time is low.

The present invention seeks to improve by the OIP3 (OTOI) that reduces IM3 level and raising power amplifier the output response of RF/ microwave power amplifier.A solution of this problem is to reduce the distortion that exists in linear and the nonlinear area, is called possible nonlinear area thereby the operation of amplifier can be extended to higher power level.

Summary of the invention

The present invention recognizes and has solved the top consideration of the structure of prior art and method and other.

These and/or other purpose is achieved in a preferred embodiment of discrete amplifier, this discrete amplifier is used for offsetting at least one distortion component at the output of this discrete amplifier, comprise: the first order, at least have first branch and second branch that are connected in parallel to each other, in wherein said first branch and second branch each is formed by one or more transistor pins, and described one or more transistor pins have the one or more unit cells that are connected in parallel to each other; The second level, at least have the 3rd branch and the 4th branch that are connected in parallel to each other, in wherein said the 3rd branch and the 4th branch each is formed by one or more transistor pins, and described one or more transistor pins have the one or more unit cells that are connected in parallel to each other.This discrete amplifier further comprises: first biasing circuit, have first impedance and first bias level, and described first biasing circuit is operably connected to first branch of the described first order, in order to described first branch is biased to first operator scheme; Second biasing circuit has second impedance and second bias level, and described second biasing circuit is operably connected to second branch of the described first order, in order to described second branch is biased to second operator scheme; The 3rd biasing circuit has the 3rd impedance and the 3rd bias level, and described the 3rd biasing circuit is operably connected to described partial the 3rd branch, in order to described the 3rd branch is biased to the 3rd operator scheme; With the 4th biasing circuit, have the 4th impedance and the 4th bias level, described the 4th biasing circuit is operably connected to described partial the 4th branch, in order to described the 4th branch is biased to the 4th operator scheme.Select first operator scheme and second operator scheme, and in the 3rd operator scheme and the 4th operator scheme at least one, thereby at least one the corresponding output end place in the described first order and the described second level offsets at least one distortion component basically.

In another embodiment, first impedance of described first biasing circuit and first bias level are different from second impedance and second bias level of described second biasing circuit.In other embodiments, the described impedance of described first biasing circuit and described second biasing circuit is different with the described bias level of described first branch and described second branch, and the described impedance of described the 3rd biasing circuit and described the 4th biasing circuit is different with the described bias level of described the 3rd branch and described the 4th branch, and select the described impedance of described first biasing circuit and described second biasing circuit and the described bias level of described first branch and described second branch, and the described bias level of the described impedance of described the 3rd biasing circuit and described the 4th biasing circuit and described the 3rd branch and described the 4th branch, so that offset described at least one distortion component substantially in the described first order of described discrete amplifier and at least one the output in the described second level.

In other embodiments, described first branch, described second branch, described the 3rd branch and described the 4th branch are formed on the single integrated circuit chip, and described first biasing circuit, described second biasing circuit, described the 3rd biasing circuit and described the 4th biasing circuit physically are placed on the identical single integrated circuit chip minimizing electric ghost effect, thereby offset described at least one distortion component substantially.Other embodiment further comprise sensing circuit, be used to sense the input power of described discrete amplifier, and make in described first biasing circuit, described second biasing circuit, described the 3rd biasing circuit and described the 4th biasing circuit at least one regulate the operator scheme of its respective branch, so that offset described at least one distortion component substantially by at least one of regulating in described biasing circuit electric current and the impedance.

In certain embodiments, described first pattern, described second pattern, described three-mode and described four-mode are that from kind A, kind B, kind C and kind AB operator scheme selects, and regulate described first biasing circuit, described second biasing circuit, described the 3rd biasing circuit and described the 4th biasing circuit to change the impedance of described each biasing circuit, so that offset described at least one distortion component substantially.

In another embodiment, a kind of circuit that is used for offsetting at the output of circuit at least one distortion component comprises: the power dispenser, it has first input end mouth, first output port and second output port, wherein first signal is transfused to the first input end mouth of power dispenser, and is split into the 3rd signal at the second output port place of the secondary signal at the first output port place of power dispenser and power dispenser; Power combiner, it has first input end mouth, second input port and first output port.First amplifier has the input port of first output port that is couple to the power dispenser and is couple to the output port of the first input end mouth of power combiner.Second amplifier is in parallel with first amplifier, and this second amplifier has the input port of second output port that is couple to the power dispenser and is couple to the output port of second input port of power combiner.At least one biasing circuit operationally with first amplifier and second amplifier in one be connected, be used for another different operation species of first amplifier and second amplifier in described first amplifier of biasing and second amplifier described one, and at least one control logic circuit sensing first signal, and at least one changes based on first signal: one the size in first amplifier and second amplifier and the amplitude of one bias level in first amplifier and second amplifier in below making.The 4th signal makes to be reduced by at least a distortion component from first output port output of power combiner.

In other embodiments, second biasing circuit is operably connected to another in first amplifier and second amplifier, be used for a different operation species of first amplifier and second amplifier in described first amplifier of biasing and second amplifier described another, therefore in the 4th signal, be reduced by at least a distortion component.In certain embodiments, described at least one distortion component third harmonic that is described first signal.In other embodiments, described operator scheme is that from kind A, kind B, kind C and kind AB operator scheme one selects.In other embodiments, power dispenser and power combiner are orthogonal mixers.

The accompanying drawing of incorporating and form the part of this specification into illustrates one or more embodiment of the present invention, and is used for explaining principle of the present invention with describing.

Description of drawings

With reference to the accompanying drawings, in specification, set forth feasible the disclosing of complete sum of the present invention, the described best mode that openly comprises at those of ordinary skill in the art, wherein:

Fig. 1 illustrates the schematic diagram of the typical RF system in the prior art cell phone;

Fig. 2 illustrates the schematic diagram of the multistage RF/ microwave power amplifier of the prior art of using in the system of Fig. 1;

Fig. 3 illustrates the schematic diagram according to the power combination amplifier of the embodiment of the invention;

Fig. 4 illustrates the graph of a relation of the bias level of the amplifier among large-signal K3 curve and Fig. 3;

Fig. 5 illustrates the correlation of OIP3 for the bias point of the amplifier of Fig. 3;

Fig. 6 illustrates the raising that bias point when two amplifiers of Fig. 3 is chosen to be the OIP3 that makes that K3 equals-sees during K3 ';

Fig. 7 illustrates the schematic diagram according to the balance amplifier with control loop of the embodiment of the invention;

Fig. 8 shows the flow chart of the control logic of Fig. 7;

Fig. 9 illustrates the schematic diagram according to the power combination branch in the one-level of the power amplifier of the embodiment of the invention;

Figure 10 illustrates the schematic diagram according to the power combination branch in the one-level of the power amplifier with control loop of the embodiment of the invention; With

Figure 11 illustrates the flow chart of the control logic of Figure 10.

Reusing Reference numeral in this specification and accompanying drawing is to be used for representing identical or similar characteristics or element of the present invention.

Embodiment

Now will be in detail with reference to currently preferred embodiment of the present invention, one or more examples of the present invention are illustrated in the accompanying drawing.Provide each example with explanation the present invention, rather than restriction the present invention.In fact, those skilled in the art will be appreciated that, can make amendment in the present invention and change under the situation that does not deviate from scope and spirit of the present invention.For example, can use in another embodiment to produce another embodiment as the part diagram of an embodiment or the feature of describing.Therefore, the invention is intended to contain those modifications and variations as within the scope that falls into claims and equivalent thereof.

The present invention seeks one of assembly level (component level) at this amplifier and sub-component level or both and locates to reduce or offset significantly IM3 level in the power amplifier.In first embodiment shown in Figure 3, assembly level orthogonal balanced amplifier layout is made up of prime amplifier orthogonal mixer 110, amplifier stage and back amplifier orthogonal mixer 112 usually, and described amplifier stage is made up of two amplifiers 114 and 116 in parallel and that be biased in the different operation species.By form input V in the sine curve sum of two different frequencies _In, be expressed as follows:

Vin＝Asin(w ₁t)+Bsin(w ₂t)

Input signal V _InBe passed to orthogonal mixer 110, this orthogonal mixer 110 is positioned at before the amplifier stage.An example of suitable orthogonal mixer is the SHY55090 degree hybrid coupler of being made by Mid-Atlantic RF System Co., Ltd.Import to be divided into and have the same frequency component but two signals, one of them signal V with a part of input power _IBy the offset pi radian, and another signal V _QBy skew pi/2 radian, caused the relative phase difference of 90 degree:

$V_I=\frac{A}{\sqrt{2}}*\sin (w_{1}t-\mathrm{\&pi;})+\frac{\mathrm{<figure-callout\; id="2"\; label="B"\; filenames="A2007800449300022H1.png,A2007800449300023H1.png"\; state="\{\{state\}\}">B</figure-callout>}}{\sqrt{2}}*\sin (w_{2}t-\mathrm{\&pi;})$

$V_Q=\frac{A}{\sqrt{2}}*\sin (w_{1}t-\frac{\mathrm{\&pi;}}{2})+\frac{\mathrm{<figure-callout\; id="2"\; label="B"\; filenames="A2007800449300022H1.png,A2007800449300023H1.png"\; state="\{\{state\}\}">B</figure-callout>}}{\sqrt{2}}*\sin (w_{2}t-\frac{\mathrm{\&pi;}}{2})$

These signals are passed to amplifier 114 and 116 then respectively.An example of suitable amplifier is the interior mesh power FET of model EIC5359-810 volt that is made by Excelics Semiconductor Co., Ltd.Can use the modeling respectively of following approximated equation by the signal amplification that each amplifier is carried out:

V _A1＝K ₁*V _I+K ₂*V _I ²+K ₃*V _I ³

V _A2＝K ₁′*V _Q+K ₂′*V _Q ²+K ₃′*V _Q ³

K wherein _iThe coefficient of three rank voltage transfering functions of expression amplifier 114, K _iThe coefficient of three rank voltage transfering functions of ' expression amplifier 116.A key character of this Amplifier Design is the K in the amplifier 114 ₃K in item and the amplifier 116 ₃' equal and opposite in direction opposite in sign.Except this condition, the phase place of fundamental component is equal substantially, has caused the maximum power combination of fundamental component.

Following equation is checked one group of IMD3 item.Other IMD3 item is similar in form and follow identical mathematical operation.

An IM3 item from 114: $K_3\left[\frac{-3A^2\mathrm{<figure-callout\; id="2"\; label="B"\; filenames="A2007800449300022H1.png,A2007800449300023H1.png"\; state="\{\{state\}\}">B</figure-callout>}}{2\sqrt{2}}\right[{\sin }^2\left(w_{1}t\right)*\sin (w_{2}t\left]\right]$

Use identity: ${\mathrm{<figure-callout\; id="2"\; label="sin"\; filenames="A2007800449300022H1.png,A2007800449300023H1.png"\; state="\{\{state\}\}">sin</figure-callout>}}^2\mathrm{\&alpha;}=\frac{1}{2}(1-\cos \left(2\mathrm{\&alpha;}\right))$

$\&DoubleRightArrow;K_3\left[\frac{-3A^2\mathrm{<figure-callout\; id="4"\; label="B"\; filenames="A2007800449300022H1.png"\; state="\{\{state\}\}">B</figure-callout>}}{4\sqrt{2}}\right[\sin \left(w_{2}t\right)-\cos \left(2w_{1}t\right)*\sin \left(w_{2}t\right)]]$

Use identity $\cos \left(\mathrm{\&alpha;}\right)*\sin \left(\mathrm{\&beta;}\right)=\frac{1}{2}\sin (\mathrm{\&alpha;}+\mathrm{\&beta;})-\frac{1}{2}\sin (\mathrm{\&alpha;}-\mathrm{\&beta;})$

$\&DoubleRightArrow;K_3\left[\frac{-3A^2\mathrm{<figure-callout\; id="4"\; label="B"\; filenames="A2007800449300022H1.png"\; state="\{\{state\}\}">B</figure-callout>}}{4\sqrt{2}}\right[\sin \left(w_{2}t\right)-\frac{1}{2}\sin (2w_{1}t+w_{2}t)+\frac{1}{2}\sin ({2w}_{1}t-w_{2}t)\left]\right]$

An IM3 item from 116: ${{\mathrm{<figure-callout\; id="3"\; label="K"\; filenames="A2007800449300022H1.png,A2007800449300022H2.png"\; state="\{\{state\}\}">K</figure-callout>}}_3}^{\&prime;}\left[\frac{-3A^2\mathrm{<figure-callout\; id="2"\; label="B"\; filenames="A2007800449300022H1.png,A2007800449300023H1.png"\; state="\{\{state\}\}">B</figure-callout>}}{2\sqrt{2}}\right[{\cos }^2\left(w_{1}t\right)*\cos \left(w_{2}t\right)\left]\right]$

Use identity: ${\mathrm{<figure-callout\; id="2"\; label="cos"\; filenames="A2007800449300022H1.png,A2007800449300023H1.png"\; state="\{\{state\}\}">cos</figure-callout>}}^2\mathrm{\&alpha;}=\frac{1}{2}(1+\cos \left(2\mathrm{\&alpha;}\right))$

$\&DoubleRightArrow;{{\mathrm{<figure-callout\; id="3"\; label="K"\; filenames="A2007800449300022H1.png,A2007800449300022H2.png"\; state="\{\{state\}\}">K</figure-callout>}}_3}^{\&prime;}\left[\frac{-3A^2\mathrm{<figure-callout\; id="4"\; label="B"\; filenames="A2007800449300022H1.png"\; state="\{\{state\}\}">B</figure-callout>}}{4\sqrt{2}}\right[\cos \left(w_{2}t\right)+\cos \left(2w_{1}t\right)*\cos \left(w_{2}t\right)\left]\right]$

Use identity: $\cos \left(\mathrm{\&alpha;}\right)*\cos \left(\mathrm{\&beta;}\right)=\frac{1}{2}\cos (\mathrm{\&alpha;}-\mathrm{\&beta;})+\frac{1}{2}\cos (\mathrm{\&alpha;}+\mathrm{\&beta;})$

$\&DoubleRightArrow;{{\mathrm{<figure-callout\; id="3"\; label="K"\; filenames="A2007800449300022H1.png,A2007800449300022H2.png"\; state="\{\{state\}\}">K</figure-callout>}}_3}^{\&prime;}\left[\frac{-3A^2\mathrm{<figure-callout\; id="4"\; label="B"\; filenames="A2007800449300022H1.png"\; state="\{\{state\}\}">B</figure-callout>}}{4\sqrt{2}}\right[\cos \left(w_{2}t\right)+\frac{1}{2}\cos (2w_{1}t-w_{2}t)+\frac{1}{2}\cos ({2w}_{1}t+w_{2}t)\left]\right]$

The output of amplifier is passed to second blender 112 then, and this second blender 112 and first blender are same types, and is used to the synthetic RF output signal of RF sets of signals with the output existence of two amplifiers.A suitable examples of blender is the model SHY55090 degree hybrid coupler from Midatlantic RF System Co., Ltd.Blender 112 has two inputs (port one and 4) and two outputs (port 2 and 3)." top " output port, promptly port 2, by from port one, have a part of power and by the signal of offset pi radian with from port 4, have a part of power and formed by the addition of the signal of skew pi/2 radian." bottom " output port, promptly port 3, by from port 4, have a half-power and by the signal of offset pi radian with from port one, have a half-power and formed by the addition of the signal of skew pi/2 radian.The bottom output port is used as the output of equipment, and under ideal conditions, the IMD3 product will eliminate from frequency spectrum basically and fundamental frequency (fundamentals) still keeps.What has taken place in the IMD3 item VO2 that following equation is presented at the existence of " bottom " output port place.

$V_{\mathrm{<figure-callout\; id="2"\; label="O"\; filenames="A2007800449300022H1.png,A2007800449300023H1.png"\; state="\{\{state\}\}">O</figure-callout>}2}=-j\frac{V_{\mathrm{<figure-callout\; id="1"\; label="A"\; filenames="A2007800449300022H1.png"\; state="\{\{state\}\}">A</figure-callout>}1}}{\sqrt{2}}-\frac{{\mathrm{<figure-callout\; id="2"\; label="V\; A"\; filenames="A2007800449300022H1.png,A2007800449300023H1.png"\; state="\{\{state\}\}">V</figure-callout>}}_A}{\sqrt{2}}$

For one group of IMD3 item deriving above, we obtain following equation:

${\mathrm{<figure-callout\; id="2"\; label="V\; O"\; filenames="A2007800449300022H1.png,A2007800449300023H1.png"\; state="\{\{state\}\}">V</figure-callout>}}_O=K_3\left[\frac{-3A^{2}B}{8}\right[\sin (w_{2}t-\frac{\mathrm{\&pi;}}{2})-\frac{1}{2}\sin (2w_{1}t+w_{2}t-\frac{\mathrm{\&pi;}}{2})+\frac{1}{2}\sin ({2w}_{1}t-w_{2}t-\frac{\mathrm{\&pi;}}{2})\left]\right]$

$+{{\mathrm{<figure-callout\; id="3"\; label="K"\; filenames="A2007800449300022H1.png,A2007800449300022H2.png"\; state="\{\{state\}\}">K</figure-callout>}}_3}^{\&prime;}\left[\frac{{3A}^{2}B}{8}\right[\cos \left(w_{2}t\right)+\frac{1}{2}\cos ({2w}_{1}t-w_{2}t)+\frac{1}{2}\cos ({2w}_{1}t+w_{2}t)\left]\right]$

Use identity, $\sin (\mathrm{\&alpha;}-\frac{\mathrm{\&pi;}}{2})=-\cos \mathrm{\&alpha;}$

${\mathrm{<figure-callout\; id="2"\; label="V\; O"\; filenames="A2007800449300022H1.png,A2007800449300023H1.png"\; state="\{\{state\}\}">V</figure-callout>}}_O=-{\mathrm{<figure-callout\; id="3"\; label="K"\; filenames="A2007800449300022H1.png,A2007800449300022H2.png"\; state="\{\{state\}\}">K</figure-callout>}}_3\frac{{3A}^{2}B}{8}[-\cos \left(w_{2}t\right)+\frac{\cos ({2w}_{1}t+w_{2}t)}{2}-\frac{\cos ({2w}_{1}t-w_{2}t)}{2}]$

$+{{\mathrm{<figure-callout\; id="3"\; label="K"\; filenames="A2007800449300022H1.png,A2007800449300022H2.png"\; state="\{\{state\}\}">K</figure-callout>}}_3}^{\&prime;}\frac{{3A}^{2}B}{8}[\cos \left(w_{2}t\right)+\frac{\cos ({2w}_{1}t-w_{2}t)}{2}+\frac{\cos (2w_{1}t+w_{2}t)}{2}]$

As can be seen from the above, work as K ₃=-K ₃In ' time, will offset this group IMD3 item.

${\mathrm{<figure-callout\; id="2"\; label="V\; O"\; filenames="A2007800449300022H1.png,A2007800449300023H1.png"\; state="\{\{state\}\}">V</figure-callout>}}_O=-{\mathrm{<figure-callout\; id="3"\; label="K"\; filenames="A2007800449300022H1.png,A2007800449300022H2.png"\; state="\{\{state\}\}">K</figure-callout>}}_3\frac{{3A}^{2}B}{8}\left[\cos (2w_{1}t+w_{2}t)\right]$

Similarly, all other IMD3 items also will be offset, and this causes the raising of OIP3.Particularly, the minimizing of IMD3 level moves to more high power with the crosspoint between 3: 1 slopes of the extrapolation of 1: 1 slope of the extrapolation of fundamental frequency power output and third-order intermodulation product, thereby causes higher OIP3 value.Look at the single order item that exists in the amplifier,

Single order item from amplifier 114: $K_1[\frac{-\mathrm{<figure-callout\; id="2"\; label="A"\; filenames="A2007800449300022H1.png,A2007800449300023H1.png"\; state="\{\{state\}\}">A</figure-callout>}}{\sqrt{2}}\sin \left(w_{1}t\right)-\frac{\mathrm{<figure-callout\; id="2"\; label="B"\; filenames="A2007800449300022H1.png,A2007800449300023H1.png"\; state="\{\{state\}\}">B</figure-callout>}}{\sqrt{2}}\sin \left(w_{2}t\right)]$

Single order item from amplifier 116: ${{\mathrm{<figure-callout\; id="1"\; label="K"\; filenames="A2007800449300022H1.png"\; state="\{\{state\}\}">K</figure-callout>}}_1}^{\&prime;}[\frac{-\mathrm{<figure-callout\; id="2"\; label="A"\; filenames="A2007800449300022H1.png,A2007800449300023H1.png"\; state="\{\{state\}\}">A</figure-callout>}}{\sqrt{2}}\cos \left(w_{1}t\right)-\frac{\mathrm{<figure-callout\; id="2"\; label="B"\; filenames="A2007800449300022H1.png,A2007800449300023H1.png"\; state="\{\{state\}\}">B</figure-callout>}}{\sqrt{2}}\cos \left(w_{2}t\right)]$

As can be seen, because these, the output of orthogonal mixer 112 is as follows:

${\mathrm{<figure-callout\; id="2"\; label="V\; O"\; filenames="A2007800449300022H1.png,A2007800449300023H1.png"\; state="\{\{state\}\}">V</figure-callout>}}_O=K_1\left[\frac{j}{2}\right][A\sin \left(w_{1}t\right)+B\sin \left(w_{2}t\right)]+\left[\frac{{{\mathrm{<figure-callout\; id="1"\; label="K"\; filenames="A2007800449300022H1.png"\; state="\{\{state\}\}">K</figure-callout>}}_1}^{\&prime;}}{2}\right][A\cos \left(w_{1}t\right)+B\cos \left(w_{2}t\right)]$

This abbreviation is:

${\mathrm{<figure-callout\; id="2"\; label="V\; O"\; filenames="A2007800449300022H1.png,A2007800449300023H1.png"\; state="\{\{state\}\}">V</figure-callout>}}_O=\left[\frac{{\mathrm{<figure-callout\; id="1"\; label="K"\; filenames="A2007800449300022H1.png"\; state="\{\{state\}\}">K</figure-callout>}}_1+{{\mathrm{<figure-callout\; id="1"\; label="K"\; filenames="A2007800449300022H1.png"\; state="\{\{state\}\}">K</figure-callout>}}_1}^{\&prime;}}{2}\right][A\cos \left(w_{1}t\right)+B\cos \left(w_{2}t\right)]$

If K ₁Amplitude and symbol is basic and K ₁' amplitude identical with symbol, then will there be the fundamental frequency item in the output at orthogonal mixer 112.Make K ₁Amplitude and symbol is basic and K ₁' the phase place of the identical fundamental component of also having guaranteed two amplifiers of amplitude with symbol equate.

Fig. 4 shows K ₃Correlation for the bias level of amplifier.For example, in kind AB operator scheme, amplifier has and equals-K of α ₃Realize equaling the K of α for second amplifier in the parallel connection configuration ₃', the necessary bias level that changes this amplifier.Along with bias level descends, K ₃' item pass zero axle and become positive, and along with amplifier arrives α near kind B operator scheme.By from K wherein ₃Substantially equal-K ₃' figure in select bias point, if the IMD3 component that exists at output is not fully offset, also will greatly be reduced.

In yet another embodiment, by suppress operational amplifier in (compression) pattern or gain expansion (expansion) pattern in gain, can change K ₃Phase place.In the gain suppression mode, Amplifier Gain reduces along with the input power that increases, and in the gain mode of extension, amplifier gain increases along with the input power that increases.The power output at fundamental frequency place can be used down and list modeling: K ₁* A+K ₃* A ³+ K ₅* A ⁵, wherein A is the amplitude of input voltage signal, the K item represents to describe the Taylor series expansion coefficient of the output voltage of amplifier to the transmission characteristic of input voltage.Therefore even-order K item is not included in this equation the not contribution of IMD3 component.For the gain expansion, suppose K ₁For just, K then ₃And K ₅Xiang Weizheng suppresses K for gain ₃And K ₅Item is for negative.Three rank IMD3 components of two-tone excitation (two toneexcitation) also comprise K ₃And K ₅.This hint IMD3 component is a homophase with respect to fundamental frequency during the gain expansion, suppresses for gain, and be 180 degree out-phase with respect to fundamental frequency.In the RF amplifier, this can realize by changing the biasing that bias level and/or use have the bias circuit block resonance-amplifier of specific input impedance value.

In order to ensure K in amplifier 114 and 116 ₃The amplitude of item is equal substantially, must select correct size for each amplifier.The size of amplifier is directly connected to amplifier can produce for how much power.Amplifier itself is made by many single transistors that are connected in parallel.Each independent transistor has standard size according to manufacturing process.When transistor was connected in parallel, each transistor was also contributed the part in the gross output.When the number of parallel connected transistors increased, the gross output that amplifier provides also increased.Therefore, in this case, size is meant the product of transistorized number and each transistorized effective area (active area) amount.

When amplifier that design is used among the present invention, know that it is very important needing how many power outputs, can use the parallel transistor of right quantity thus.This transistorized sum is set the size of amplifier then.Yet, will produce different IMD3 level at two amplifiers of the same size of different some biasing.For example, the amplifier of a certain size of setovering in kind AB operator scheme is compared with the amplifier of the same size of setovering in approximate kind B operator scheme, will produce the IMD3 item of much less at identical input power levels place.Therefore, the size of approximate kind B pattern amplifier must be confirmed as making the input power levels place identical, and the IMD3 amplitude of this approximate kind B pattern amplifier equals the IMD3 amplitude of kind AB pattern amplifier substantially.A kind of method of determining size is to determine suitable IMD3 amplitude by a lot of different sizes of approximate kind B pattern amplifier test to(for) suitable input power levels.Just because the IMD3 amplitude of two bridging amplifiers separates, the raising of the OIP3 that sees at the output of output orthogonal blender 112 will not as the IMD3 level in basic equating on the amplitude with the same big on the contrary the time on phase place.And amplifier 114 and 116 IMD3 amplitude can be equal substantially at a power level place, but unequal at another power level place.

As shown in Figure 7, a kind of method that overcomes this problem is to use the correction signal that produces by the RF input signal that monitors blender 110 places, the size and the bias level of coming control amplifier 116.Just, sampling RF input signal V _InAnd be fed into control logic piece 118, this control logic piece 118 passes through input power levels V _InCompare with the level of storing in the look-up table and generate numeric word or correction signal, the level of storing in the look-up table is associated with the appropriate size and the bias level of approximate kind B amplifier 116.Correction signal is controlled at the number that is similar to effective parallel transistor in the kind B amplifier by some part that turns on and off approximate kind B amplifier, controls the size of approximate kind B amplifier 116.

More specifically,, control logic 118 should turn-off, then by bias level being reduced to zero some transistors that come in the cutoff approximation kind B amplifier if determining some part of approximate kind B amplifier.On the other hand, should connect if control logic piece 118 is determined some part of approximate kind B amplifier, then their bias level is that some zero transistors will rise to approximate kind B level.Biasing control logic 118 is arranged in the biasing circuit of approximate kind B amplifier 116.When approximate kind B amplifier size reduced, its IMD3 component also reduced.Similarly, when approximate kind B amplifier size increased, its consequent IMD3 component also increased.As the result of size adjustment ability, can on wide input power levels scope, see the raising of OIP3.

Fig. 8 illustrates the step that control logic is handled the change that pairing approximation kind B amplifier makes.At first in step 120 sensing input power levels, and in step 124, use this information to judge and to make what change by pairing approximation kind B amplifier.In step 126 and 128, increase making respectively or reduce the size of amplifier and be the decision that increases or reduce the biasing of amplifier, offset so that strengthen IMD3.In case make decision in step 126 and 128, make suitable change at step 130 pairing approximation kind B amplifier, so that the amplitude of the IMD3 component of two amplifiers is equal substantially, the phase place of IMD3 component relative to each other and 180 degree out-phase, and the phase place of fundamental frequency equates.

The experimental result that the assembly that then will can be purchased off the shelf by the commerce of using the 5.7GHz place obtains is described in an example of the effect of setovering in the layout of being advised.Fig. 5 shows the correlation of OIP3 for biasing.Fig. 6 show when select bias point make K3 equal substantially-K3 ' time the OIP3 that sees raising.Two experiments all are to use the amplifier (from mesh power FET in the EIC5359-8 10-volt of Excelics Semiconductor Co., Ltd) of the layout of being advised and same size to carry out.For input mixer and output blender, all select SHY55090 degree hybrid coupler from Midatlantic RF System Co., Ltd.In one case, two amplifiers are biased in same level and use kind A﹠amp in Fig. 5 in kind AB operator scheme; A represents.At this bias point, the amplitude of IMD3 component is equal substantially, but therefore IMD3 component homophase each other the counteracting of IMD3 component does not take place or reduce.

In second kind of situation, an amplifier as above is biased in kind AB operator scheme, and another amplifier is biased in approximate kind B operator scheme.Each amplifier is in different operation modes, and the IMD3 component relative to each other differs 180 degree on phase place, but their amplitude is equal substantially on short input power range.This experiment in accompanying drawing 5 by kind A﹠amp; B represents.Thus, there is the raising of OIP3 in two amplifiers in the biasing of variety classes place on little input power levels scope.About mathematical derivation, this experimental observation is the result of the correlation of K3 item and input power.

Fig. 6 exists nearly by diagram for little input power levels scope, and the raising of the OIP3 of 4dB further illustrates the raising shown in Fig. 5.If control loop shown in Figure 7 118 is incorporated in experiment into, will on wideer power bracket, there be OIP3 to improve.Therefore, although can obtain to improve by the amplifier of selecting similar size, the variable-sized of approximate kind B amplifier determined and setovers to cause the OIP3 on the broad power band to improve.Therefore, by regulating the size and/or the bias level of approximate kind B amplifier, its IMD3 amplitude can be adjusted to the IMD3 amplitude of mating kind AB amplifier on wide input power range better, thereby causes the raising of the OIP3 on the wideer input power range.

In another embodiment of the present invention, can on the sub-component level, greatly reduce or offset the IM3 level.More specifically, by using the biasing circuit of two separation, as described by operation species (that is, kind A, kind B, kind AB), two group of branches of bias power amplifier stage can greatly reduce or offset each level 7 of power amplifier 3 in different operation conduction angles _a, 7 _b... 7 _nThe IM3 level at place.Each biasing circuit will have distinct output impedance, and wherein the impedance at base band frequency place is noticeable especially.The scope of the modulating frequency of the input signal that base band frequency is defined as paying close attention to, its scope typically from kHz to tens MHz.The biasing parallel branch makes phase difference between the IM3 component of each branch group near 180 degree.Additionally, allow the additional phase shift of IM3 component, thereby can realize the poor of 180 degree between the IM3 component in the difference of the output impedance at the base band frequency place of biasing circuit.Regulate the quantity of the branch in every group, so that the amplitude of IM3 component is equal substantially.Just, because the amplification of branch group must equate substantially so that therefore maximizing performance regulates the number of the branch in every group so that difference is amplified in compensation.

By change the dc operating current in the branch with biasing circuit, come the operation conduction angle (for example, kind A, kind AB, kind B, approximate kind B) of regulatory work rate amplifier branch.Typically, in biasing circuit, change control voltage and/or Control current,, change operation conduction angle thus so that change the dc operation electric current.In one implementation, can use the resistor of connecting to revise the impedance of biasing circuit with biasing circuit.The value that can change resistor obtains different resistance values.

With reference to figure 9, the level of power amplifier 7a is made up of following three assemblies: power divider device 10 on the sheet, belong to two branch group 12 and 14 of a level of amplifier, wherein the biasing circuit (not shown) of Fen Liing setover each branch group, power combiner 16 on the sheet of following thereafter.Each branch group 12 and 14 comprises one or more branches 18 of unit 20 _a, 18 _b... 18 _n Unit 20 can be made up of one or more transistors and other passive electric circuit element.Input V ₁Sine curve sum by two different frequency places is formed:

V ₁＝Asin(w ₁t)+Bsin(w ₂t)

Input signal V ₁Be passed to power divider device 10 on the sheet.The power divider device example is a microstripline on the suitable sheet, and it has characteristic impedance Z _o, this circuit is divided into two and has characteristic impedance Z _o* the quarter-wave line of sqrt (2).Described input is divided into two signals, and these two signals have the same frequency component, but has a part of input power, one of them input signal V ₂With another input signal V ₃Do not have phase difference.

${\mathrm{<figure-callout\; id="2"\; label="V"\; filenames="A2007800449300022H1.png,A2007800449300023H1.png"\; state="\{\{state\}\}">V</figure-callout>}}_2=\frac{A}{\sqrt{2}}\sin \left(w_{1}t\right)+\frac{\mathrm{<figure-callout\; id="2"\; label="B"\; filenames="A2007800449300022H1.png,A2007800449300023H1.png"\; state="\{\{state\}\}">B</figure-callout>}}{\sqrt{2}}\sin \left(w_{2}t\right)$

${\mathrm{<figure-callout\; id="3"\; label="V"\; filenames="A2007800449300022H1.png,A2007800449300022H2.png"\; state="\{\{state\}\}">V</figure-callout>}}_3=\frac{A}{\sqrt{2}}\sin \left(w_{1}t\right)+\frac{\mathrm{<figure-callout\; id="2"\; label="B"\; filenames="A2007800449300022H1.png,A2007800449300023H1.png"\; state="\{\{state\}\}">B</figure-callout>}}{\sqrt{2}}\sin \left(w_{2}t\right)$

These signals are passed to branch group 12 and 14 subsequently respectively.According to the transfer function of amplifier, amplify the signal that each amplifier receives, described transfer function can use following each equation to come modeling approx:

${\mathrm{<figure-callout\; id="4"\; label="V"\; filenames="A2007800449300022H1.png"\; state="\{\{state\}\}">V</figure-callout>}}_4=K_{1A}\&times;{\mathrm{<figure-callout\; id="6"\; label="V"\; filenames="A2007800449300022H1.png"\; state="\{\{state\}\}">V</figure-callout>}}_6+K_{2A}\&times;{\mathrm{<figure-callout\; id="6"\; label="V"\; filenames="A2007800449300022H1.png"\; state="\{\{state\}\}">V</figure-callout>}}_6^2+K_{3A}\&times;{\mathrm{<figure-callout\; id="6"\; label="V"\; filenames="A2007800449300022H1.png"\; state="\{\{state\}\}">V</figure-callout>}}_6^3$

${\mathrm{<figure-callout\; id="5"\; label="V"\; filenames="A2007800449300022H1.png"\; state="\{\{state\}\}">V</figure-callout>}}_5=K_{1B}\&times;V_7+K_{2B}\&times;V_7^2+K_{3B}\&times;V_7^3$

Wherein, K _IAThe coefficient of three rank voltage transfering functions of expression branch group 12, and K _IBThe coefficient of three rank voltage transfering functions of expression branch group 14.A key character of this amplifier stage design is the K in the branch group 12 _3AWith branch group 14 in K _3BThe item size is basic to be equated and opposite in sign.Except this condition, the phase place of fundamental component is basic identical, thereby causes the maximum power combination of fundamental component.Following equation is checked one group of IMD3 item.Other IMD3 item is similar in form, and follows identical mathematical operation.

An IM3 item from 12: $K_{3A}\left[\frac{-3A^2\mathrm{<figure-callout\; id="2"\; label="B"\; filenames="A2007800449300022H1.png,A2007800449300023H1.png"\; state="\{\{state\}\}">B</figure-callout>}}{2\sqrt{2}}\right[{\sin }^2\left(w_{1}t\right)*\sin (w_{2}t\left]\right]$

Use identity ${\sin }^{}$ ${}^2\mathrm{\&alpha;}=\frac{1}{2}(1-\cos \left(2\mathrm{\&alpha;}\right))$

$\&DoubleRightArrow;K_{3A}\left[\frac{-3A^2\mathrm{<figure-callout\; id="4"\; label="B"\; filenames="A2007800449300022H1.png"\; state="\{\{state\}\}">B</figure-callout>}}{4\sqrt{2}}\right[\sin \left(w_{2}t\right)-\cos \left(2w_{1}t\right)*\sin \left(w_{2}t\right)]]$

Use identity $\cos \left(\mathrm{\&alpha;}\right)\&times;\sin \left(\mathrm{\&beta;}\right)=\frac{1}{2}\sin (\mathrm{\&alpha;}+\mathrm{\&beta;})-\frac{1}{2}\sin (\mathrm{\&alpha;}-\mathrm{\&beta;})$

$\&DoubleRightArrow;K_{3A}\left[\frac{-3A^2\mathrm{<figure-callout\; id="4"\; label="B"\; filenames="A2007800449300022H1.png"\; state="\{\{state\}\}">B</figure-callout>}}{4\sqrt{2}}\right[\sin \left(w_{2}t\right)-\frac{1}{2}\sin (2w_{1}t+w_{2}t)+\frac{1}{2}\sin ({2w}_{1}t-w_{2}t)\left]\right]$

An IM3 item from 14: $K_{3B}\left[\frac{-3A^2\mathrm{<figure-callout\; id="2"\; label="B"\; filenames="A2007800449300022H1.png,A2007800449300023H1.png"\; state="\{\{state\}\}">B</figure-callout>}}{2\sqrt{2}}\right[{\cos }^2\left(w_{1}t\right)\&times;\cos \left(w_{2}t\right)\left]\right]$

Use identity ${\cos }^{}$ ${}^2\mathrm{\&alpha;}=\frac{1}{2}(1+\cos \left(2\mathrm{\&alpha;}\right))$

$\&DoubleRightArrow;K_{3B}\left[\frac{-3A^2\mathrm{<figure-callout\; id="4"\; label="B"\; filenames="A2007800449300022H1.png"\; state="\{\{state\}\}">B</figure-callout>}}{4\sqrt{2}}\right[\cos \left(w_{2}t\right)+\cos \left(2w_{1}t\right)\&times;\cos \left(w_{2}t\right)\left]\right]$

Use identity $\cos \left(\mathrm{\&alpha;}\right)\&times;\cos \left(\mathrm{\&beta;}\right)=\frac{1}{2}\cos (\mathrm{\&alpha;}-\mathrm{\&beta;})+\frac{1}{2}\cos (\mathrm{\&alpha;}+\mathrm{\&beta;})$

$\&DoubleRightArrow;K_{3B}\left[\frac{-3A^2\mathrm{<figure-callout\; id="4"\; label="B"\; filenames="A2007800449300022H1.png"\; state="\{\{state\}\}">B</figure-callout>}}{4\sqrt{2}}\right[\cos \left(w_{2}t\right)+\frac{1}{2}\cos (2w_{1}t-w_{2}t)+\frac{1}{2}\cos ({2w}_{1}t+w_{2}t)\left]\right]$

Branch group 12 and 14 output are passed to second and go up power combiner 16, power combiner can be or can not be to go up the identical type of power divider device 10 with first on this sheet, and the RF signal combination that is used to be present in the output of two branch group is a RF output signal.As can be seen from the above, work as K _3A=-K _3BThe time, this group IMD3 item will be offset.Similarly, all other IMD3 items also will be offset, and cause the raising of OIP3.Particularly, the reducing of IMD3 level moves to higher power with the crosspoint between 3: 1 slopes of the extrapolation of 1: 1 slope of the extrapolation of fundamental frequency power output and third-order intermodulation product, thereby causes higher OIP3 value.Look to be present in the single order item in amplifier branch 12 and 14,

Single order item from 12: $K_{1A}[\frac{A}{\sqrt{2}}\sin \left(w_{1}t\right)+\frac{\mathrm{<figure-callout\; id="2"\; label="B"\; filenames="A2007800449300022H1.png,A2007800449300023H1.png"\; state="\{\{state\}\}">B</figure-callout>}}{\sqrt{2}}\sin \left(w_{2}t\right)]$

Single order item from 14: $K_{1B}[\frac{A}{\sqrt{2}}\sin \left(w_{1}t\right)+\frac{\mathrm{<figure-callout\; id="2"\; label="B"\; filenames="A2007800449300022H1.png,A2007800449300023H1.png"\; state="\{\{state\}\}">B</figure-callout>}}{\sqrt{2}}\sin \left(w_{2}t\right)]$

As can be seen, because these, the output of power combiner 16 is as follows on the sheet:

${\mathrm{<figure-callout\; id="6"\; label="V"\; filenames="A2007800449300022H1.png"\; state="\{\{state\}\}">V</figure-callout>}}_6=\frac{(K_{1A}+K_{1B})}{2}[A\&CenterDot;\sin \left(w_{1}t\right)+B\&CenterDot;\sin \left(w_{2}t\right)]$

If K _1ASubstantially equal K _1B, then will there be the fundamental frequency item in the output of power combiner 16 on sheet.Make K _1ASubstantially equal K _1BThe phase place of also guaranteeing the fundamental component of two branch group equates.

Amplitude in order to ensure the IMD3 component is equal substantially in parallel branch group 12 and 14, must select correct size for each branch group.The size of each branch group is directly connected to it can produce for how much power.Have equal number unit cell, will produce different IMD3 level in two branches of difference (that is, for each branch, different operation conduction angle) biasing.For example, at identical input power levels place, the branch with same size that the unit cell 20 of a certain number, the branching ratio that is biased in kind AB operator scheme be biased in approximate kind B operator scheme will produce the IMD3 of much less.Therefore, the size of approximate kind B pattern amplifier must be confirmed as making that at identical input power levels place, its IMD3 amplitude equals the IMD3 amplitude of kind AB pattern amplifier substantially.A kind of method of definite size is to determine suitable IMD3 amplitude by testing a lot of approximate kind B pattern branch group.Just because the IMD3 amplitude of two parallel branch groups separates, therefore the raising of the OIP3 that sees at the output of power output combiner 16 will not as the IMD3 level in basic equate and equally big on the contrary the time on phase place on the amplitude.And branch group 12 and 14 IMD3 amplitude can be equal substantially on a power level, but do not wait on another power level.

Except the size of each branch group, the quantity of the branch group in the specific amplifier stage also can change based on the application of amplifier.Therefore, if the level in the power amplifier has three branch group, then power divider device 10 and 16 must be designed to cut apart input V on the sheet on three paths ₁, and also the output of three branch group is combined into single output signal V ₆Therefore, the number that should be appreciated that the branch group of any one amplifier stage 7 can change according to the application of power amplifier 3.

In yet another embodiment, by operational amplifier in gain suppression mode or gain mode of extension, can change K ₃Phase place.In the gain suppression mode, the gain of branch or branch group reduces along with the input power that increases, and in the gain mode of extension, tap gain increases along with the input power that increases.The power output at fundamental frequency place can be used down and list modeling: K ₁* A+K ₃* A ³+ K ₅* A ⁵, wherein A is the amplitude of input voltage signal, the K item represents to describe the Taylor series expansion coefficient of the output voltage of branch to the transmission characteristic of input voltage.Therefore even-order K item is not included in this equation the not contribution of IMD3 component.

For the gain expansion, suppose K ₁For just, K then ₃And K ₅Xiang Weizheng suppresses K for gain ₃And K ₅Item is for negative.Three rank IMD components of two-tone excitation also comprise K ₃And K ₅.This hint IMD3 component is a homophase with respect to fundamental frequency during the gain expansion, and to suppress with respect to fundamental frequency for gain be anti-phase.In the RF amplifier, this can realize with the biasing of using the biasing circuit with different impedances to regulate one or more branches by changing bias level.

The scope of observing the power level of raising therein can be used in different semiconducter IC technology and changes along with this technology, and control loop incorporate the raising that can guarantee on broad power band into.A common problem is: the K relevant with the arithmetic differentiate ₃Item can be used as the function of RF input power and changes.Therefore, control loop can be used to compensate K by the bias level (promptly conducting angle) and the impedance of biasing circuit of regulating two power amplifier branch group ₃In variation.At K as the function of RF power ₃In this dynamic compensation of variation, allow IMD3 amplitude and phase place on wide input power range, to be conditioned for optimum IM3 counteracting/minimizing.

With reference to Figure 10, control loop 22 is shown as from input V ₁Be connected to the branch group 12 of amplifier stage 7a and 14 biasing circuit (not shown).By monitoring RF input signal V ₁Generate correction signal 24, so that as the bias level of the functions control branch group 12 of power and 14 and the impedance of biasing circuit (not shown).Just, sampling RF input signal V ₁And being fed into control logic piece 22, this control logic piece 22 generates suitable bias level at the RF input power that applies.These bias levels can generate by the means of arbitrary number, and for example (but being not limited to) is with input power levels V ₁Compare with the level of storing in the look-up table, perhaps by using circuit block directly to generate corresponding bias level at specific input power levels.Correction signal 24 is controlled the bias level and the bias circuit impedance of branch group 12 and 14 by the control voltage that changes the biasing circuit (not shown).More specifically, should be biased to more towards kind AB pattern, then will correspondingly regulate the control voltage of biasing circuit if control logic 22 is determined branch group 14.In one implementation, biasing control logic 22 can be arranged in the biasing circuit of branch group 14 and also can be arranged in branch group 12.And, use the switch (not shown) of the part turn on and off the resistors in parallel array can regulate the impedance of biasing circuit.As the result of the biasing/impedance adjustment performance of biasing circuit (not shown), the OIP3 that can obtain on the wide power level range improves.

With reference to Figure 11, show the flow chart of control logic 22.In step 26, sensing input power levels V at first ₁And, use described information to determine the change that to make branch group 12 and 14 by the value of storing in institute's sensed signal and (1) look-up table for example being compared (in step 28) or institute's sensed signal being converted to the corresponding biasing/impedance level of biasing circuit by (2).May changing of biasing circuit is: (1) changes the impedance of biasing circuit, step 34, and perhaps (2) increase/reduce the biasing of branch, step 32.In case make decision at control logic piece 22, just make suitable change in step 36 pair branch group 12 and 14, so that the amplitude of the IMD3 component of branch group 12 and 14 is basic identical, the phase place of IMD3 component relative to each other and 180 degree out-phase, and the phase place of fundamental frequency is near equating.

The above-mentioned amplifier that reduces at IM3 improves and has produced very wide band solution, and can fully realize on chip.Described technology is based on the power amplifier layout of uniqueness, and combines the biasing technique of the novelty on the amplifier branch level.Described technology is very flexible, and can be applicable to any IC technology (that is, and Si CMOS, SiGe BiCMOS, GaAs HBT, or the like).

Those of ordinary skill in the art will be understood that, under the situation that does not deviate from scope and spirit of the present invention, can carry out various modifications and variations in the present invention.This invention is intended to contain the described modifications and variations within the scope and spirit that fall into claims and equivalent thereof.

Claims (19)

1. a discrete amplifier is used for offsetting at least one distortion component at the output of this discrete amplifier, comprising:

A. the first order, at least have first branch and second branch that are connected in parallel to each other, in wherein said first branch and second branch each is formed by one or more transistor pins, and described one or more transistor pins have the one or more unit cells that are connected in parallel to each other;

B. the second level, at least have the 3rd branch and the 4th branch that are connected in parallel to each other, in wherein said the 3rd branch and the 4th branch each is formed by one or more transistor pins, and described one or more transistor pins have the one or more unit cells that are connected in parallel to each other;

C. first biasing circuit has first impedance and first bias level, and described first biasing circuit is operably connected to first branch of the described first order, in order to described first branch is biased to first operator scheme;

D. second biasing circuit has second impedance and second bias level, and described second biasing circuit is operably connected to second branch of the described first order, in order to described second branch is biased to second operator scheme;

E. the 3rd biasing circuit has the 3rd impedance and the 3rd bias level, and described the 3rd biasing circuit is operably connected to described partial the 3rd branch, in order to described the 3rd branch is biased to the 3rd operator scheme;

F. the 4th biasing circuit has the 4th impedance and the 4th bias level, and described the 4th biasing circuit is operably connected to described partial the 4th branch, in order to described the 4th branch is biased to the 4th operator scheme;

Wherein select at least one in following:

A. described first operator scheme and described second operator scheme and

B. described the 3rd operator scheme and described the 4th operator scheme,

Thereby the corresponding output end place of at least one in the described first order and the described second level offsets at least one distortion component basically.

2. discrete amplifier as claimed in claim 1, first impedance of wherein said first biasing circuit and first bias level are different from second impedance and second bias level of described second biasing circuit.

3. discrete amplifier as claimed in claim 1, wherein

The described impedance of a. described first biasing circuit and described second biasing circuit is different with the described bias level of described first branch and described second branch, and

The described impedance of b. described the 3rd biasing circuit and described the 4th biasing circuit is different with the described bias level of described the 3rd branch and described the 4th branch,

And select the described impedance of described first biasing circuit and described second biasing circuit and described bias level and the described impedance of described the 3rd biasing circuit and described the 4th biasing circuit and the described bias level of described the 3rd branch and described the 4th branch of described first branch and described second branch, so that, offset described at least one distortion component substantially at the described first order of described discrete amplifier and at least one the output in the described second level.

4. discrete amplifier as claimed in claim 1, wherein said first branch, described second branch, described the 3rd branch and described the 4th branch are formed on the single integrated circuit chip, and described first biasing circuit, described second biasing circuit, described the 3rd biasing circuit and described the 4th biasing circuit physically are placed on the described identical single integrated circuit chip minimizing electric ghost effect, thereby offset described at least one distortion component substantially.

5. discrete amplifier as claimed in claim 1, also comprise sensing circuit, be used to sense the input power of described discrete amplifier, and make in described first biasing circuit, described second biasing circuit, described the 3rd biasing circuit and described the 4th biasing circuit at least one regulate the operator scheme of its respective branch, so that offset described at least one distortion component substantially by at least one of regulating in described biasing circuit level and the impedance.

6. discrete amplifier as claimed in claim 1, wherein said first pattern, described second pattern, described three-mode and described four-mode are that from kind A, kind B, kind C and kind AB operator scheme selects.

7. discrete amplifier as claimed in claim 6, wherein regulate described first biasing circuit, described second biasing circuit, described the 3rd biasing circuit and described the 4th biasing circuit to change the impedance of described each biasing circuit, so that offset described at least one distortion component substantially.

8. discrete amplifier as claimed in claim 6, first pattern of wherein said first branch and second pattern of described second branch make at least one harmonic wave of basic neutralisation.

9. discrete amplifier as claimed in claim 6, the three-mode of wherein said the 3rd branch and the four-mode of described the 4th branch make basic neutralisation third order intermodulation distortion component.

10. discrete amplifier as claimed in claim 6, one in first branch of the wherein said first order and second branch of the described first order has a plurality of transistor pins, so that increase described one power output in second branch of first branch of the described first order and the described first order.

11. discrete amplifier as claimed in claim 1, the wherein said first order and the described second level are connected in series.

12. a discrete amplifier is used for offsetting at least one distortion component at the output of this discrete amplifier, comprising:

A. input port is used to receive first signal;

B. output port is used to export the secondary signal with described first signal correction;

C. first branch and second branch in parallel with described first branch, described first branch is operably connected to described input port and described output port, and described second branch is operably connected to described input port and described output port,

D. first biasing circuit is operably connected to described first branch, in order to described first branch is biased to first operator scheme;

E. second biasing circuit is operably connected to described second branch, in order to described second branch is biased to second operator scheme;

Wherein select described first operation and described second operator scheme, thereby reduce described at least one distortion component in the described secondary signal basically,

Described first branch, described second branch, described first biasing circuit and described second biasing circuit are placed on the single integrated circuit chip to minimize electric ghost effect, so that offset described at least one distortion component substantially.

13. discrete amplifier as claimed in claim 12, wherein said at least one distortion component are the triple-frequency harmonics of described first signal.

14. discrete amplifier as claimed in claim 12, the third order intermodulation distortion component that wherein said at least one distortion component is described first signal.

15. being from kind A, kind B, kind C and kind AB operator scheme, discrete amplifier as claimed in claim 12, wherein said operator scheme select.

16. discrete amplifier as claimed in claim 12, also comprise sensing circuit, be used for described first signal of sensing, and make in described first biasing circuit and described second biasing circuit at least one change the operator scheme of its respective branch, so that offset described at least one distortion component substantially by at least one of regulating in biasing circuit level and the biasing impedance.

17. discrete amplifier as claimed in claim 12, first impedance of wherein said first biasing circuit and first electric current are different from second impedance and second electric current of described second biasing circuit.

18. discrete amplifier as claimed in claim 16 is wherein regulated described first biasing circuit and described second biasing circuit changing the impedance of described each biasing circuit, thereby is offset described at least one distortion component substantially.

19. discrete amplifier as claimed in claim 12, one in wherein said first branch and described second branch has a plurality of transistor pins, so that increase described one power output in described first branch and described second branch.

Images