Maxim's adjustable RF predistortion ICs are ideal for wireless and basestation applications. Our closed loop RF predistorters use the "RFPAL" technology for fully adaptive distortion correction where the highest performance is needed. Our open loop RF predistorters provide a basic 'set and forget' manual adjustment for those applications where only a few dB of correction is needed.
Many factors determine the overall performance of any linearization solution. Comparing Maxim's RF predistortion system implementation, RFPAL, against solutions operating in backoff, customers can realize the greatest performance benefits. By "spending" 0.4W of added power consumption for RFPAL, customers can get in return up to a 4X improvement in efficiency, thus enabling proportional decreases in the system power consumption. Additionally, linearization enables operation of the PA closer to its PSAT operation point allowing the use of smaller and less expensive power transistors to achieve a desired output-power goal. A direct benefit of the improved power consumption is a reduced annual operating cost (electricity), which, at antenna power levels above 5W, can offset the initial cost of the RFPAL solution in a relatively short time. In the example below (Table 1), the system power consumption and system efficiency advantages are clear: a 3X improvement in efficiency, a system power reduction of over 67W and a yearly operating cost saving of almost $30. Though not every system will operate at maximum output power all the time, the power supply capacity and system cooling requirements must be designed for worst case. An often overlooked benefit of predistortion, compared to operation in backoff, is the dramatic reduction in size/volume and costs associated with the power supplies and cooling elements (heat-sinks, fans, etc.).
|Desired Max Antenna Pout (avg)||37||37||37||dBm|
|Component IL Between PA & antenna||2||2||2||dB|
|Final stage PA output power (Max antenna power - IL)||39||39||39||dBm|
|Final stage PA efficiency at 39 dBm (avg)||25.0%||27.0%||8.0%|
|Linearizer power consumption||0.4||2.1||0||W|
|External DAC, ADC, up/down-converters, etc. power consumption||1||3.3||1|
|Pre-amp + driver power consumption||1||1||1||W|
|Total power consumption (w/o) final stage PA)||2.4||6.4||2||W|
|System effeciency at MAX antenna POUT = 37 dBm||14.7%||14.0%||4.9%|
|System power consumption at Max antenna POUT = 37 dBm||34.2||35.8||101.3||W|
|Cost of energy (cost per kWh)||$0.05||$0.05||$0.05|
|Yearly cost of operation per system||$14.98||$15.70||$44.40|
|Added yearly cost of operation compared to RFPAL||$0.72||$29.42|
Table 1. Comparison of system power consumption, efficiency and annual operating cost
Figure 9. Comparison of system power consumption, efficiency and annual operating cost
Comparing RFPAL against DPD, an important difference between the two solutions is revealed. Even when DPD provides better final stage PA efficiency (2% better in Table 1 but the difference in most cases is much smaller) than RFPAL, the overall system efficiency is worse due the DPD's higher power consumption. As antenna output power levels drop below 10 W and approach 500mW, the improvement in system efficiency grows progressively larger for RFPAL due to the DPD power consumption becoming a larger portion of the overall system power consumption. With similar correction performance to DPD, but improved power consumption and efficiency, reduced complexity, lower system cost and a small footprint, RFPAL is an ideal choice for small cell designs which are part of heterogeneous deployments.
Finally, there exists an entire class of applications that DPD cannot address due to costs and, by its very nature, the requirement for a digital processor. These markets include analog remote radio heads, power amplifier modules, repeaters and microwave backhaul systems. Maxim's RFPAL is an adaptive RFIN / RFOUT and standalone linearization solution that requires no external digital control. The digital port shown in Figure 4a is optional and used in applications where reporting functions are desired and not needed for the markets above. RFPAL is simple to integrate and requires no expertise in predistortion algorithms making it today the only alternative to operation in backoff.
Maxim's RFPAL devices are ideally suited to cellular applications requiring operation from 168 to 3600MHz, PAR of up to 10dB, static average power waveforms like CDMA and WCDMA, dynamic average power signals like WiMAX, HSDPA and LTE, and with a wide range of PAs including Class A/B or Doherty, in different process technologies including LDMOS, GaN, GaAs, etc. Maxim’s RFPAL can be operated over a wide case temperature range of -40°C to +100°C.
Looking beyond cellular infrastructure applications, Maxim also has products for the microwave radio and broadcast infrastructure markets. For example, in the broadcast market, the RFPAL has been integrated into DVB-T, ATSC and CMMB transmitters operating at greater than 600W of average output power. Additionally, any application requiring linear amplification of RF signals can benefit from Maxim's solutions, including white space, military and public safety, etc.
RF Predistortion (RFPD) vs. Digital Predistortion (DPD)
Maxim's RF predistortion technology (also known as analog predistortion) shares similarities with DPD in that both compensate for amplitude-modulation-to-amplitude-modulation (AM-AM) and amplitude-modulation-to-phase-modulation (AM-PM) distortion, intermodulations and PA memory effects, and both employ feedback information to compensate for impairments due to temperature variations and PA aging. Though both approaches share underlying theoretical similarities, the similarities end with their circuit design and system implementations. As there are many articles and application notes written describing the main functional blocks which comprise a digital predistortion system, the theory of operation of digital predistortion will not be described here. This section will focus on the key differences between the two architectures and the theory of operation of RF predistortion.
Before delving into the details and inner working of Maxim's RF Power Linearizer (RFPAL) family, a basic explanation is required for context: Figure 4a below shows a high level block diagram of a PA system utilizing RFPAL. Of note, this RFPD-based linearizer is an adaptive RFIN / RFOUT system enabling standalone operation in remote radio heads, PA modules and any application without direct access to a digital processor. Directional couplers are used to drive the linearizer's RF inputs (RFIN and RFFB). The correction signal (RFOUT) is then combined with the PA input signal using a directional coupler as well. Looking deeper into the inner workings, the linearizer uses the PA output signal to adaptively determine the nonlinear characteristics of the PA at a given average and peak power level, center frequency and signal bandwidth. This feedback signal (RFFB) from the PA output is analyzed in the frequency domain to generate a spectrally resolved linearity metric used for the adaptation cost function.
Figure 4a. Block diagram of RFPAL system implementation
As shown in Figure 4b, the entire linearizer system (including all components within dashed lines of Figure 4a) can be realized in a compact PCB footprint measuring less than 6.5cm2 and a low bill of materials.
Figure 4b: Photograph of RFPAL system implementation on printed circuit board
With a baseline established for RFPD's basic operation, its system implementation can be described and compared to digital predistortion (DPD). Figure 5 illustrates how DPD expands the bandwidth (adds the predistortion correction signal to the desired signal) at the earliest point in the signal chain – the digital baseband. That bandwidth expansion is then propagated through the entire transmitter chain and back again through the feedback path to the digital baseband. The bandwidth expansion burdens the entire system with greatly increased clock speeds and expanded bandwidth requirements leading to high system power consumption. Added complexities include (but not limited to) challenging clock generator requirements, including jitter performance, multi-pole high frequency reconstruction filter and wideband linear up-mixer.
Figure 5. Digital predistortion system implementation (30MHz BW, WCDMA, 14-b converters)
With DPD systems, the frequency response of the filter after the up-converter has to be wide enough to accommodate the wanted signal plus the BW expansion required for the PA predistortion. Unfortunately, any noise generated by the DACs, up-converters, etc., within the filter's bandpass will also be amplified by the PA. In the majority of applications, the only way to filter the noise falling in the receive band is at the output of the PA. This approach requires the use of filters whose size, cost and insertion loss varies based on the design requirements. There is also a potential increase in the cost of the filter to meet the more stringent rejection requirements. Any added insertion loss due to this filter degrades efficiency and requires the PA to be driven harder to achieve the same output power at the antenna as originally desired. The filter partially negates the benefits that were achieved through the use of digital predistortion. Alternatively, lower noise DACs and up-converters may be used to remove the requirements for a post-PA filter, but at higher cost and current consumption than their higher noise counterparts. Note, the power consumption is estimated based on an integrated DPD/DSP ASIC and external ADCs, DACs, down-converter, clock generator and power detectors. The power consumption estimates do not include DUC, CFR and PA as they are present in both DPD and RFPD implementations.
As discussed previously, by leveraging a standalone RFIN/RFOUT architecture and adaptive RF predistortion technology, Maxim's integrated approach allows the correction signal to be injected only at the point it is needed – the PA's input. The benefits of this implementation can be seen in Figure 6. The requirements on the clock generator, reconstruction filter and up-mixer are all relaxed while all the components in the transmitter chain from the digital baseband up to the PAs can operate at 1X signal bandwidth. However, the linearizer can operate with a signal bandwidth greater than 5X with no system design or power consumption penalty since far out residual intermodulation products can be easily filtered. The total predistortion BW of RFPAL devices is about 250 MHz enabling compensation up to 11th order IMs for an instantaneous BW (wanted signal) of 20MHz, or up to 5th order IMs for an instantaneous BW of 50 MHz. Additionally, the RFPD-based system needs only a narrow band filter before the PA, hence relaxing DAC and up-converter noise requirements, and avoiding costly filtering at the PA output. Though not required for a RFPD implementation, the RFPAL devices also integrate the entire RFFB feedback path, thus greatly simplifying the overall system design and limiting the active components affected by the BW expansion to just the PA and the linearizer itself. These benefits lead to very low power consumption and a greatly simplified, lower cost transmitter and baseband architecture.
Figure 6. RF predistortion system implementation (30MHz BW, WCDMA, 14-b converters)
In the example given, the RFPD implementation consumes 4W less than the DPD implementation. Though the power difference may not be a critical issue in macrocell designs, the reduced power consumption, lowered system cost, and smaller footprint of RFPD-enabled designs become important factors in micro, pico and enterprise femtocells designs. Finally, Maxim’s RFPAL devices include optional features which provide measurement functionality for forward and reverse power and monitoring capabilities of temperature and spectral mask thus simplifying the system implementation even further.
Predistortion Techniques vs. Operation in Backoff
When a power amplifier (PA) is presented with a multi-tone signal at its input, it will amplify the desired signal as well as generate unwanted intermodulation (IM) terms (Figure 1a). This non-linear distortion increases as the PA approaches its saturation point and will vary in nature based on operating conditions and from PA to PA. To achieve the desired linearity at the PA output (without predistortion), the PA must be operated with significant backoff from its saturation point (PSAT(3dB) in Figure 2a). Operation in backoff means that the PA's maximum output power level must be reduced so that the entire signal is within the linear region of the PA transfer curve. However, the PA's efficiency (PA's ability to convert DC supply power into RF energy) decreases as the PA's operating point is lowered further away from its saturation point (Figure 1b). Efficiencies of 8% or less for a Class AB PAs are not unusual in order to accommodate the signal’s peak-to-average ratio (PAR) and the additional backoff required to meet the system linearity requirements.
Figure 1a. Intermodulation terms generated by PA
Figure 1b. Relationship between output power, efficiency and distortion
Considering that the most popular linearization method by far for Class A/AB PAs transmitting 20W average power and below is operation in backoff, for these applications active linearization can provide very compelling benefits. Active linearization techniques, including digital predistortion (DPD) or RF predistortion (RFPD), allow the transmitter to operate close or even slightly above its PSAT - PAR operating point (Figure 2b). Both use predistortion techniques where a correction signal is injected at the PA’s input in order to reduce the overall distortion at the output of the PA (Figure 3a and Figure 3b).
Sidebar: Note that no predistortion solution can correct signals whose peaks extend much past the PA's saturation point as the information becomes more difficult, or even impossible, to recover as the amount of signal clipping increases. Pushing the PA past its saturation point is ultimately a system design decision that is based on many factors including margin to the adjacent channel leakage ratio (ACLR) specification, spectral emissions mask (SEM) specification and/or error vector magnitude (EVM) requirements for example.
Figure 2a. Unlinearized performance of PA with no predistortion (operation in backoff)
Figure 2b. Linearized performance of PA with predistortion enabled
Figure 3a. PA output characteristics without linearization
Figure 3b. PA output characteristics with predistortion linearization
Using active linearization, a Class AB PA can achieve 3 dB to 6 dB of additional (linearized) output power and improve its efficiency by 2X to 4X. Compared to operation in backoff, active linearization enables the final stage amplifier, power supply, cooling elements and operating costs to be reduced by half or more. Maxim's RF PA linearizer efficiency calculator clearly demonstrates that for all but the lowest average output power Class A/AB PAs, the cost of the predistortion solution can be easily recovered in less than 2 years of operation just due to lowered electricity usage. Considering the added benefits of lowered PA device cost, lowered power supply cost and lowered cooling costs, predistortion offers a very compelling value proposition.
In systems requiring wide signal bandwidth like LTE, or in wideband multi-carrier/multi-protocol systems, active linearization is the only option if the PA cannot reach the desired linearity regardless of the amount of backoff applied. In these systems, active linearization is required to pass regulatory radiated emission testing and meet project requirements.
Finally, systems requiring an improvement in efficiency (beyond what is achievable using a Class A/AB PA) will use more advanced PA topologies like a Doherty configuration. The advanced topologies depend on predistortion solutions to meet their system linearity requirements.
RF Predistortion (RFPD) Theory of Operation
Maxim's approach to RF power amplifier linearization, called RF predistortion, repartitions portions of the predistortion algorithm from the digital domain to the analog/RF domain. This implementation is sometimes called analog predistortion. The text below will identify the key architectural blocks which comprise Maxim's RF Power Linearizer (RFPAL) and describe their function.
Nearly the entire Correction Processor block shown in Figure 7 is implemented using RF/analog circuits resulting in very low power consumption, wide bandwidth performance and compact circuits compared to the equivalent digital implementation.
Figure 7. RFPAL architecture and functional block diagram
Correction Processor block, the RFIN signal passes through a quadrature phase shifter (QPS) to create an I and Q signal [RFIN(I), RFIN(Q)], which is used in multiple locations. The envelope power of RFIN(I) and RFIN(Q) is also used in the Volterra Series Generator block to create the even order IM terms by applying a non-linear transformation. In order to compensate for PA memory effects, four different sets of coefficients are created based on delay terms (τ1 to τ4) ranging from 0ns to 300ns (Figure 8a). All coefficients are individually controlled and generated by the digital controller running a proprietary adaptation algorithm. For each of the memory terms, the even order correction functions are summed and then multiplied with their corresponding RFIN(I) and RFIN(Q) signals generated by the QPS. This final multiply converts the even order terms into odd orders terms. The I and Q correction signals are then summed creating the RFOUT correction signal. The Correction Processor uses a full 360° modulator enabling it to correct IM terms of any phase and magnitude. The digital controller adapts the coefficients based on the information derived from the RFFB feedback signal and applies them to the Correction Processor until an optimal set of coefficients is found that minimizes the cost function (error metric).
Figure 8a. Volterra series generator block diagram
Figure 8b. Volterra series equation
The Monitor block is implemented largely in the digital domain, as functions like FFTs and error metric generation are better suited to an implementation that uses digital signal processing (DSP). As seen in Figure 7, the Monitor inputs include the down-converters and ADCs required to provide the spectrally resolved data used by the DSP. The integration of the RFFB ADC is a significant difference compared to DPD which relies on external down-converters and ADCs. Maxim's unique partitioning approach results in a monolithic and highly integrated solution that maintains the flexibility of digital approaches while offering the simplicity and low power consumption of analog approaches.
System Efficiency and Power Consumption Calculator
|Select linearization method(s) to compare with RFPAL:|
|Max Pout (Antenna)|
|Min Pout (Antenna)|
|Component IL between PA and antenna||dB|
|PA (final stage) Efficiency at||%|
|PA (final stage) Efficiency at||%|
|Linearization Power Consumption||W|
|Support components power consumption||W|
|Pre-Amp + Driver Power Consumption||W|
|Total Power Consumption (w/o final stage PA)||W|
|System Efficiency at Max Pout (Antenna) =||%|
|System Power Consumption at Max Pout (Antenna) =||W|
|Cost of energy (cost per kWhour)||$|
|Annual operation cost at||$|
|Incremental annual cost of operation compared to RFPAL||$|
Efficiency & Power Consumption Results
---- Efficiency __ Power Consumption
Cellular service providers are moving to deploy heterogeneous networks in an attempt to keep up with the surging demand for data services. They are transitioning their infrastructure from voice-driven macro base stations to small cell, distributed antenna, and MIMO architectures with smaller coverage areas and low power transmit architectures. As the volume of these types of cells increases, solutions are necessary to constrain the hardware, deployment and operating (power consumption) costs of these units. Furthermore, the transition to 4G services with multi-mode, multi-carrier and wide bandwidth signals increases the linearity requirements of transmitters even at the lowest output power levels. However, no commercially available solution previously existed that could cost-effectively linearize power amplifiers (PA) from 60W down to 500mW (average output power) while also meeting operator's and OEM's stringent performance and system requirements.
Maxim's RF Power Linearizers (RFPAL) fill the gap left unaddressed by digital predistortion (DPD) solutions – especially when used with PAs operating below 60 W average output power where system power consumption, solution cost and footprint are as important as final stage efficiency and correction performance. For PAs operating below 10W, where Class A/AB amplifiers dominate, Maxim’s solution offers a very compelling alternative to operating the PA in backoff by improving system efficiency by 2X to 4X.
- Single and multi-carrier: WCDMA, CDMA
- Single carrier, multi-carrier/standard: EVDO, TD-SCDMA, WiMAX, HSDPA, LTE & TD-LTE
- Traditional in-cabinet BTS amplifier, Remote Radio Unit (RRU), tower mounted power amplifiers, microwave repeaters, booster amplifiers, microcells, picocells, enterprise femtocells, Distributed Antenna System (DAS), Adaptive Antenna System (AAS) and MIMO systems.
Key product specifications:
- Amplifier: Class A/AB, Doherty
- Average PA output power: 500 mW to 60 W
- Process: LDMOS, GaAs, GaN and InGaP
Microwave Point-to-Point (P2P)
With the rapid proliferation of smartphones, tablets and other data-hungry devices, operators are working to quickly transition from voice-centric networks to data-centric networks. Keeping up with the surging demand in data services and applications is a continuous challenge. Even newly deployed equipment is showing signs of strain including the wireless/microwave backhaul link due to the increased data rate and bandwidth requirements. The growing mobile backhaul bottleneck can be addressed with microwave point-to-point (P2P) equipment that adapts to these changing demands. As operators increasingly deploy microwave P2P equipment, they will also expect carrier class reliability, low CAPEX cost and low OPEX cost.
Maxim provides RFPAL devices that are able to meet the stringent requirements of both the equipment manufacturer and the operator. It can operate across a wide range of intermediate frequencies, modulations, operating conditions and signal characteristics and can effectively linearize microwave amplifiers operating at nearly any output frequency and deliver the full benefits of predistortion including increased efficiency and linearity.
Key product specifications:
- Modulation: QPSK, 64-QAM to 1024-QAM, CDMA, OFDM, and others
- Instantaneous signal bandwidth: up to 60 MHz
- Intermediate Frequencies (IF): 225-3800 MHz
- Peak-to-Average Ratios (PAR): 4-11 dB
General RF PA Linearization
Modern communication standards place unique requirements on the system design of transmitters. Application requirements including wide bandwidth, high efficiency, multi-carrier, multi-standard coupled with specific signal requirements including Spectral Emissions Mask (SEM), Bit Error Rate (BER) and Adjacent Channel Leakage Ratio (ACLR) or Adjacent channel power ratio (ACPR) to name a few, can create unwanted system design tradeoffs.
Maxim's RFPAL family addresses a wide range of applications including small niche applications that have previously been unable to take advantage of predistortion solutions due to complexity and/or cost. With its unique RFIN / RFOUT architecture, RFPAL can be easily integrated into most any transmitter system with high linearity requirements. By using RFPAL, a design engineer can deliver impressive gains in efficiency while maintaining or even improving the linearity performance of most any amplifier operating in backoff – in other words when replacing an existing Class A/AB amplifier running in backoff with one using RFPAL there are no tradeoffs. In fact, by using RFPAL it is even possible to convert a poor efficiency Class A/AB transmitter into a Doherty configuration and achieve even higher efficiency results.
- Any application requiring PA linearization
- Software Defined Radio (SDR), HMS/Mobile military communications and white space
- Public safety mobile and portable transmitters
- Customer Premises Equipment
Key product specifications:
- Wide range of PAs and output power levels:
- Amplifier: Class A/AB, Doherty
- Average PA output power: 500 mW to 600+ W
- Process: LDMOS, GaAs, GaN and InGaP