Efficient envelope tracking power supply for radio frequency or other power amplifiers

[0001] This disclosure relates generally to power amplifiers. More specifically, this disclosure relates to an efficient envelope tracking power supply for radio frequency or other power amplifiers.

[0002] Wireless base stations are routinely used to support wireless communications with various wireless devices. As base station technology has evolved, base stations have often required the use of much more complex transmission schemes. These transmission schemes usually employ complex modulation techniques that often require linear power amplification. Linear power amplification is often performed by one or more power amplifiers (PAs).

[0003] Power amplifiers typically consume a significant amount of power in base stations. For example, in the majority of cases, this may account for more than half of the total power consumed by a base station. This typically increases the cost of operating the base station. Moreover, linear power amplification may require one or more radio frequency (RF) power amplifiers to operate in a “backed off” state. This state decreases a base station's overall efficiency because RF power amplifiers are much less efficient when in the backed off state.

[0009] FIGS. 1 through 7, discussed below, and the various embodiments used to describe the principles of the present invention in this patent document are by way of illustration only and should not be construed in any way to limit the scope of the invention. Those skilled in the art will understand that the principles of the present invention may be implemented in any type of suitably arranged device or system.

[0010] FIG. 1 illustrates an example linear-assisted architecture 100 for power amplifier (PA) envelope tracking according to this disclosure. In the context of a PA, envelope tracking (ET) refers to a process in which a dynamically changing voltage is applied to a PA to ensure that the PA remains in a linear operating region.

[0011] In the example embodiment shown in FIG. 1, the efficiency of a PA 102 is improved using an envelope tracking power supply 120. The PA 102 can operate in a linear mode, and its supply voltage V_dd is provided by the envelope tracking power supply 120. The power supply 120 dynamically adjusts the supply voltage V_dd to increase or maximize the efficiency of the PA 102 while maintaining a sufficient linearity of the PA 102 in the linear mode. The supply voltage V_dd can be proportional to the envelope of an input signal coming into the PA 102 with a minimum supply voltage being clamped so that the gain variation of the PA 102 is tolerable. By continuously operating the PA 102 close to its saturation (but not too close to lose linearity), high efficiency can be maintained across all power levels. In the linear-assisted architecture 100 of FIG. 1, the power supply 120 is current controlled, DC power is provided to the PA 102 by a switching converter (“switcher”) in the power supply 120, and AC power of the envelope is provided to the PA 102 by a linear amplifier in the power supply 120.

[0012] The PA 102 represents any suitable structure for amplifying an input signal. In this example, the PA 102 represents a radio frequency (RF) power amplifier capable of receiving an input signal RF_in and generating an output signal RF_out. The power supply 120 represents any suitable structure for supplying power to a PA and implementing envelope tracking. Note that the embodiment of the envelope tracking power supply 120 shown in FIG. 1 is for illustration only. Other embodiments of the envelope tracking power supply 120 could be used without departing from the scope of this disclosure.

[0013] In the example shown in FIG. 1, the power supply 120 receives a power supply voltage V_dc and an input voltage V_in. The power supply voltage V_dc may represent a voltage received from a power supply. The input voltage V_in may represent a voltage used to control the supply voltage V_dd generated by the power supply 120. A linear amplifier 104 receives the input voltage V_in as an input and is powered by the power supply voltage V_dc. The amplifier 104 amplifies the input voltage V_in and operates to regulate the supply voltage V_dd sent to the PA 102. The amplifier 104 includes any suitable structure for amplifying an input voltage in a substantially linear manner.

[0014] An output of the amplifier 104 is coupled to a resistor 106, which is coupled to a comparator 108. The resistor 106 represents any suitable resistive structure having any suitable resistance. A voltage drop across the resistor 106 is compared by the comparator 108, which could determine whether the voltage drop is greater than some threshold value. The comparator 108 represents any suitable structure for comparing inputs.

[0015] An output of the comparator 108 is coupled to a gate driver 110, which drives two transistors 112-114. The transistor 112 is coupled to the power supply voltage V_dc, and the transistor 114 is coupled to ground. By turning the transistors 112-114 on and off, the gate driver 110 can control the voltage provided to an inductor 118, which generates the supply voltage V_dd for the PA 102. The inductor 118 is also coupled to the resistor 106. The gate driver 110 represents any suitable structure for driving one or more transistors. The transistors 112-114 represent any suitable switching devices, such as NMOS transistors. The inductor 118 represents any suitable inductive device having any suitable inductance.

[0016] Although FIG. 1 illustrates an example linear-assisted architecture 100 for PA envelope tracking, various changes may be made to FIG. 1. For example, while described as being used to amplify RF signals, the architecture 100 could be used to amplify any other suitable signals. Also, other embodiments of each circuit within FIG. 1 could be used without departing from the scope of this disclosure.

[0017] FIGS. 2 through 5 illustrate example embodiments of a linear amplifier used in the architecture of FIG. 1 according to this disclosure. In particular, FIGS. 2 through 4 illustrate example embodiments of the linear amplifier 104 in FIG. 1.

[0018] As shown in FIG. 2, an input voltage V_in is supplied to an amplifier 202 and to level shifters 208-210. An amplifier 206 receives a shifted input voltage from the level shifter 210 and outputs a signal to an amplifier 204. The amplifier 204 receives a shifted input voltage from the level shifter 208 and outputs a signal to the amplifier 202. The amplifier 202 receives the input voltage V_in and generates a supply voltage V_dd. It is understood that the output from amplifier 206 is used as the lower supply rail for the amplifier 204. The output from amplifier 204 is used as the lower supply rail for the amplifier 202.

[0019] The input voltage for each amplifier 204-206 is DC shifted by the level shifters 208-210, and the supply voltage V_dd is generated by the amplifier 202. The amplifier 206 has input (from the level shifter 210 denoted V_ls1), with supply rails V_sup and V_min that may be governed by the following equations when the supply headroom is ignored:
V_sup=(V_max−V_min)/N+V_min  (1)
V_ls,i=(V_max−V_min)/N*(N−i)  (2)

[0020] Amplifier 204 works similarly to Amplifier 206. Here, V_sup denotes the supply voltage for the amplifier 206, i denotes the level, Vmin and Vmax are the min and max levels of V_dd, and N refers to the stacked levels of the amplifiers. It is understood that the offsets from level shifters 208, 210 are preferably different. If an amplifier output is used as the supply rail, it is considered stacked. For example, V1 is the output of amplifier 206, and is being used as the lower supply rail for amplifier 204. V2 is the output of amplifier 204, and is being used as the lower supply rail for amplifier 202. Note that any number of additional amplifiers may be added to the linear amplifier 200 consistent with FIG. 2.

[0021] In FIG. 3, a linear amplifier 300 is similar to the linear amplifier 200 in FIG. 2. Here, the linear amplifier 300 includes the three amplifiers 202-206 and the two level shifters 208-210. In addition, the linear amplifier 300 includes diodes 302-304 and capacitors 306-308. The supply voltage V_sup is coupled to the diodes 302-304. The diode 302 is coupled to the amplifier 202 and the capacitor 306, and the capacitor 306 is connected to the output of the amplifier 204. The diode 304 is coupled to the amplifier 204 and the capacitor 308, and the capacitor 308 is connected to the output of the amplifier 206.

[0022] This architecture may provide various benefits depending on the implementation. For example, the architecture in FIG. 3 may be more efficient as the loss would be effectively 1/N of the single amplifier case if the dominant energy is provided by the amplifier 202. This is because the supply voltage difference of the amplifier 202 can be 1/N of the full rail V_max−V_min. Also, the architecture in FIG. 3 may allow the use of lower breakdown voltage processes, which could have higher bandwidths (due to less capacitance) and smaller areas for the same function. Further, the architecture in FIG. 3 could deliver more output power with higher output voltages. In addition, the architecture in FIG. 3 may increase the output power capability of the amplifiers 202-206.

[0023] In some embodiments, the voltage difference between the two rails of each level may be constant ((V_max−V_min)/N). It is understood that the voltage difference of the two rails in each level could be set differently with a different clamp voltage in each level during the capacitor charging phase.

[0024] A high output voltage V_dd can be provided with the use of all three stacked amplifiers 202-206. When a high output voltage V_dd is not present, the output of the amplifier 206 could be near its lower rail (near V_min) because its input is near V_min due to the DC shift performed by the level shifter 210.

[0025] It is understood that all amplifiers may be in use during the operation of the circuit. The lower amplifier (204 and/or 206) may be in a clamp mode (e.g., a capacitor is being charged). The amplifier can be in the amplification mode, i.e., it is the supply for the higher level through the capacitor (instead of the higher level own supply through a diode).

[0026] In particular embodiments, the sizes of the amplifiers 202-206 can be similar. This may be done so that the amplifiers 202-206 provide the same peak current in a peak power condition during stacking. Also, in particular embodiments, the diodes 302-304 are of a high-voltage type.

[0027] In the example shown in FIG. 3, the floating supplies in the high levels are being generated by the charge pumps. In particular embodiments, the floating supply is generated similarly to a class-H amplifier. The voltage difference between the two supply rails may be constant, which is different from class-H supplies for digital subscriber line (DSL) applications where a lower rail is moved lower when an upper rail is moved higher. The capacitors 306-308 are charged most of the time through the diodes 302-304, respectively, using the supply voltage V_sup when the lower rail of each level is low. When the input voltage V_in is higher, the lower rail can be moved upwards to follow the input. During the delivery of low power levels, both capacitors 306-308 may be charged. During the delivery of medium power levels, the capacitor 308 may be charged, while the capacitor 306 may be discharged to supply the output voltage. During the delivery of high power levels, the capacitors 306-308 may be discharged to supply the output voltage.

[0028] In FIG. 4, a linear amplifier 400 includes a diode 402, a capacitor 404, a level shifter 406, and amplifiers 408-410. This arrangement is similar to that shown in FIG. 3, but the arrangement in FIG. 4 has two levels instead of three. Here, the level shifter 406 generates a voltage V_il provided to the amplifier 410, which generates a voltage V_l (such as 10-25V). Supply voltage V_sup2 is from V_sup through the diode 402 when the capacitor is in the charging phase, and V_sup2 is generated by the lower amplifier 410 through the capacitor 404. The amplifier 408 generates the output voltage V_dd (such as 10-40V). The lower rail for the amplifier 410 could be, for example, 10V. The gain from V_in to V_dd is 1 shown in the waveforms of FIG. 4. However, it is understood that the gain could be much higher for a low V_in.

[0029] Instead of using a single power buffer, two levels are stacked up in FIG. 3. The power loss from the linear amplifier 400 is effectively cut in half in the first order. This is because the supply voltage of the main amplifier 408 can be reduced to half, while the current consumption can remain the same. Most of the time, the amplifier 408 is the one that delivers the output power. During peak power levels, the capacitor 404 could supply approximately half of the peak power, and the amplifier 410 could supply approximately the other half of the peak power. In particular embodiments, the amplifier 410 delivers the same current as the amplifier 408 in the peak power case, so its chip area can be doubled, but the linear amplifier's power loss is cut by half. In further embodiments, when the input power is higher, the lower rail is moving upward following the input, the output power is being delivered partly by the lower level amplifier through the capacitor, and partly by the stored energy in the capacitor.

[0030] When a charge pump is used, it may require peak power occurring infrequently so that capacitors can be replenished before being depleted too much for intolerable voltage drops. This is normally the case for envelope signals with large peak-to-average ratios.

[0031] In the above description, it has been assumed that all amplifiers have rail-to-rail outputs. This, however, is not necessary. In many situations, a Darlington source-follower can be used for high output currents, so the output voltage swing may be within ±3V of its supply rails. In order to have an output with the required swing, the supply headroom can be increased to accommodate the output, and the DC level shifting offsets can be optimized, as well.

[0032] The efficiency of a linear amplifier can be further improved with the consideration that its main amplifier output source-follower supplies do not have to be the same as the amplifier's drivers. This is because the drivers may need a higher voltage for their upper rails and a lower voltage for their lower rail than the output devices. By optimizing the rails of the driver and output devices, the efficiency is further improved. This may be considered a special case of a class-G amplifier. When the output swing is small, the efficiency can be significant. This is particularly true for the multi-level architecture in FIG. 2 since the full swing is divided over multiple levels. The multi-level architecture of amplifiers permits the operation of the power supply without a full swing.

[0033] FIG. 5 may be an implementation of the efficient envelope tracking power supply using a class-G driver supply (i.e., the output rails having less headroom than the drivers). In FIG. 5, a linear amplifier 500, a level shifter 506, and amplifiers 508-510 are in the same arrangement as shown in FIG. 4. In this example, however, the amplifier 510 includes four power supply inputs, including V_sup, V_sup+, V_min− and V_min. The V_sup+ could represent the V_sup voltage plus an additional voltage (such as 3V). The V_min− could represent the V_min voltage minus the additional voltage.

[0034] The linear amplifier 500 also includes a diode 512 coupled to the supply voltage V_sup and a capacitor 514 coupled to the diode 512. The linear amplifier 500 can further include a diode 518 coupled to the amplifier 508 and a capacitor 520 coupled across the supply voltages V_sup+ and V₁− of the amplifier 508. The output from the amplifier 510 is fed into a node connected to the amplifier 508 and to the capacitor 514. In this embodiment, dropout losses can be reduced to a secondary level by reducing headroom. Also shown in FIG. 5 are inputs for amplifier 508 output device and the amplifier 508 driver referred to as V₁ and V₁−. V₁− and V₁ have a relationship such that V₁−=V₁−V_headroom, where V_headroom could be a voltage similar to 3V. Once the headroom is reduced, the quiescent current loss is dominant over other losses, except for switcher loss. A more complicated implementation with additional levels may be used to remove substantially all of the headroom.

[0035] In the embodiment shown in FIG. 5, the voltages on the driver (upper) rails of the amplifiers 508-510 can be generated using one or more charge pump DC/DCs or one or more inductive floating supplies. When the top amplifier driver supply is implemented by a charge pump similar to the main supply, it needs a separate smaller amplifier with a bigger offset. This separate smaller amplifier should be very small and require low power as no significant power is required for the driver rail.

[0036] Although FIGS. 2 through 5 illustrate example embodiments of a linear amplifier used in the architecture of FIG. 1, various changes may be made to FIGS. 2 through 5. For example, other linear amplifiers having any number of levels could be used.

[0037] FIG. 6 illustrates an example method 600 for envelope tracking according to this disclosure. In step 602, an input voltage is received. In step 604, the input voltage is split into two signals. One signal is transmitted to a voltage shift device, and another signal is transmitted to a first amplifier. In step 606, an output from the voltage shift device is transmitted to a second amplifier. In step 608, an output from the second amplifier is transmitted to an input of the first amplifier. In step 610, an output signal is generated using the first amplifier that maintains a PA in a linear mode of operation.

[0038] Although FIG. 6 illustrates an example method 600 for envelope tracking, various changes may be made to FIG. 6. For example, while shown as a series of steps, various steps in FIG. 6 could overlap, occur in parallel, or occur multiple times.

[0039] FIG. 7 is an example of a plurality of waveforms 700 that may be present within the linear-assisted architecture 100 for power amplifier (PA) envelope tracking according to this disclosure. The signal names in the waveforms may be referring to the signals illustrated in FIG. 4. As shown in FIG. 7, when V_in increases from 25V to 40V, V_sup2 increases from 25V to 40V at approximately the same time. In addition, V_il/V_l and V_dd also increase at approximately the same time as the increase in Vin and reflect the change in V_in. The level shifter 406 in FIG. 4 has a built-in clamp, as demonstrated from the V_il waveform. With the clamp, the capacitor voltage will be maintained relatively constant independent of the input signal V_in.

[0040] It may be advantageous to set forth definitions of certain words and phrases that have been used within this patent document. The term “couple” and its derivatives refer to any direct or indirect communication between two or more components, whether or not those components are in physical contact with one another. The terms “include” and “comprise,” as well as derivatives thereof, mean inclusion without limitation. The term “or” is inclusive, meaning and/or. The phrases “associated with” and “associated therewith,” as well as derivatives thereof, may mean to include, be included within, interconnect with, contain, be contained within, connect to or with, couple to or with, be communicable with, cooperate with, interleave, juxtapose, be proximate to, be bound to or with, have, have a property of, have a relationship to or with, or the like.

[0041] While this disclosure has described certain embodiments and generally associated methods, alterations and permutations of these embodiments and methods will be apparent to those skilled in the art. Accordingly, the above description of example embodiments does not define or constrain this disclosure. Other changes, substitutions, and alterations are also possible without departing from the spirit and scope of this disclosure, as defined by the following claims.