White Paper Small Cells Call for Scalable Architecture By Jean-Christophe Nanan and Barry Stern
freescale.com Small Cells Call for Scalable Architecture
Introduction If 2011 was the year of the heterogeneous reasonable to think that when pico and femto Small cells seem to have a bright future network (HetNet) concept1, Mobile cells are focusing on 3G and 4G, a high-end because they are considered the most World Congress 2012 in Barcelona was portion of small cell equipment will evolve cost-effective solution to add capacity to characterized by the emergence of “small toward broader capabilities (2G, 3G and 4G). existing networks. However, some important cells” as a solution to help operators optimize For example, microcell equipment may be challenges remain in order to transform this their network architecture and face the rapidly a downscale version of the multi-standard promise into reality. First, the backhaul. There growing demand for coverage and capacity: macrocell base station supporting all types of are solutions, such as fixed fiber, line-of-sight currently there are close to 5.3 billion mobile transmission. point-to-point radio links at 60 GHz or non- subscribers texting, talking and downloading. line-of-sight radio below 6 GHz, each of them Small cell deployment is expected to follow By 2015, mobile subscriptions are expected being particularly well adapted to a different a similar trend with first applications in the to reach 6.4 billion and mobile data will scenario, but there is no one-size-fits-all spectrum commonly allocated to 3G (2.1 exceed 30 times that of current levels. solution, and this lack of consistency is an GHz) and 4G (2.6 GHz). These spectrums are issue for operators. The second challenge Reducing distance between the user and particularly well adapted to short coverage is managing a network with a significantly the base station and reducing the number in urban environments where the greatest larger number of cells. The self organizing of instantaneous users seeks to improve the need exists for high capacity. However, network (SON) concept defines the capability signal quality and allow a higher data rate in leveraging the new spectrum freed by the of equipment to self-optimize its settings for both up and down links. The names “femto,” digital TV broadcast at 0.7 GHz, we may efficient and consistent transmission, but “pico” and “micro” are now complementary see small cell deployment in this frequency here again the exact translation in technical to the well-known “macrocell” terminology as well. Considering further extension into specifications depends on the equipment and define a wide range of options for the GSM bands, this means that the RF vendor. Finally, small cell equipment must be network deployment2. They have been transmitter will target wideband operations in compact for acceptance in the commercial considered recently by the standardization the near future. However, due to the limited market and by regulating authorities for bodies under the names “home base station,” number of subscribers that small cells must positive public opinion. A base station with “local area base station,” “medium range support, the instantaneous signal bandwidth MIMO 2 x 2 should fit in a volume lower base station” and “wide area base station” is significantly smaller than macro base station than 10 liters for a weight lower than 10 Kg, respectively3,4. Additionally, we have the term transmissions. including backhaul solutions. “metrocell,” which defines high capacity, Along with the complex modulation schemes compact equipment deployed in urban areas. Based on these considerations, this paper used by 3G and 4G, advanced antenna Depending on the deployment scenario, analyzes the challenges and opportunities schemes are a key technology employed to metrocell equipment can be a picocell or created by new small cell equipment and achieve the high data throughput required microcell. shows how careful system analysis enables by mobile subscribers. The multi input, multi new types of semiconductor solutions for Future networks must offer the best service output (MIMO) antenna scheme is now part of efficient and scalable designs. using the most advanced communication the specifications3,4, and more sophisticated standards while maintaining 2G (GSM) service active antenna systems are considered also whose legacy will be continued for some to further improve network performances6. time. How both types of communications On the other hand, most of the small cells are managed simultaneously by the network consider only one or two sectors. poses an interesting question. One scenario All of the above considerations are is that 2G communications are managed summarized in table 1. by the macro base stations while the small cells focus on providing the most advanced standards (W-CDMA and LTE). This approach Table 1: Small Cell Family, in Contrast with Macrocell Application means that small cell equipment should be Max Cell Max RF Signal optimized for the fastest data rate to avoid RF Spectrum Cell Subscribers Radius Power Standard Bandwidth Sector MMO Band the tight requirements of signal integrity linked (Km) (dBm) (MHz) to the multi-carrier GSM protocol. A second Femto Indoor 4–16 0.01 20 3G/4G/Wi-Fi® 10 1, 3, 7, 34, 1 2 x 2 scenario takes into consideration countries 38, 40... with a solid 2G legacy. Many of these Pico In/outdoor 32–100 0.2 24 3G/4G 20 1, 3, 7, 34, 1 2 x 2 countries will transition directly to 4G without 38, 40... having deployed 3G networks. Moving from Micro/Metro Outdoor 200 2 38 2G/3G/4G 20,40 1, 2, 3, 5, 6, 8, 1–2 4 x 4 2G to 4G requires equipment designed to 7, 13... manage 2G and 4G standards simultaneously. Macro Outdoor 200–1000 10 50 2G/3G/4G 60–75 1, 2, 3, 5, 6, 3 4 x 4 Based on those considerations, it seems 8, 7, 13...
freescale.com 2 White Paper Small Cells Call for Scalable Architecture
FigureSingle-sector 1: Single-Sectormulti-standard downlink Multi-Standard architecture Downlink Architecture
Baseband Digital Front End Analog IF/RFRF PA
L2 LTE
DAC
01 ΣΣ L3 L2 WCDMA Σ CFR DPD -90 2 DAC
GSM/ L2 EDGE DAC 0 1 Adapt ΣΣ DUC -90 2 DAC
System Architecture DigitalFigure Baseband 2: LTE Processing Processing Elements in in eNodeB/eNBLTE eNodeB (eNB) (LTE Base StationBase Station) The level of performance required for a femtocell is not the same as the eNB requirements for a macrocell; neither are Inter Cell RRM the constraints on cost and manufacturing RB Control volume. Femtocell equipment can be Connection Mobility Cont. MME plugged into a domestic AC power supply or Radio Admission Control Ethernet cable, with dedicated constraints eNB Measurement NAS Security on DC power consumption, whereas Configuration and Provision Idle State Mobility macrocell equipment must be able to work Dynamic Resource Handling on a battery, justifying an architecture Allocation (Scheduler) EPS Bearer Control optimized for power efficiency. Together RRC with the criteria we listed in the first section PDCP S-GW P-GW (channel bandwidth, number of sectors), RLC these examples help define the best Mobililty UE IP Address MAC Anchoring Allocation architecture for each cell type. S1 PHY Packet Filtering Internet A typical one-sector transmission link is depicted in figure 1 with the important E-UTRAN EPC interfaces identified. Source: 3GPP TS 36.300 V8.12.0
Digital baseband processing in an LTE base station (eNB) is divided into several layers (figure 2). L2 and L3 layers can be PHYFigure (L1) Physical 3: PHY Layer Layer 1 Functions decomposed in three sub layers: medium access control (MAC), radio link control MAC Layer PHY Layer Downlink Processing Functions (RLC) and packet data convergence protocol (PDCP). Typically, these are implemented by a Layer Scrambling Mapping general-purpose processor (GPP) device. CRC Turbo Rate and Pre-Coding IFFT Attach Encoding Matching Modulation and Resource The charts in figure 3 depict the chain of Mapping functions building the PHY (L1) layer in an LTE base station. Typically, PHY functions are implemented in the digital signal processing (DSP) device by the DSP cores and baseband MAC Layer accelerators. PHY Layer Uplink Processing Functions
Finally, digital radio front end logic typically Rate- Freq. Dematching, Transport in an ASIC, FPGA or off-the-shelf transceiver Channel MIMO De-Modulation CRC FFT IDFT Offset De-Interleaving HARQ Block Estimation Equalizer Descrambling Check prepares the signal to be sent to the RF Compensation Combining, CRC Turbo Decoding amplifier.
freescale.com 3 White Paper Small Cells Call for Scalable Architecture
Digital Baseband Processing in single mode and in multimode operations, As mentioned previously, there is a need to Elements in LTE eNodeB (eNB) handling: support multiple standards concurrently as users slowly migrate to LTE. It is especially Base Station • Fourier transform processing element: Used primarily in LTE for FFT and DFT important that small cells covering a given, As wireless networks evolve, support for Fourier transforms operations as well as relatively limited cell radius and number of LTE and WCDMA standards and multimode RACH operations. It also can be used users continue to support multimode while operation with both technologies running in WCDMA for frequency domain search providing an upgrade path for handling more simultaneously are becoming requisite. and RACH operations. The ability to advanced technologies. In order to handle Given the inherent differences between these perform additional vector post and pre- multimode operation, the DSP cores are wireless standards, a number of technical multiplier operations makes this unit also fully programmable and can implement any challenges have to be solved on various levels very suitable for correlation and filtering standard. The MAPLE hardware block was of the processing stacks. operations. designed to enable multimode operation such as turbo and Viterbi decoding; • Turbo/Viterbi decoding processing On the L1 physical layer, the 3GPP turbo encoding and FFT/DFT can operate element: Used for forward error correction standards for third-generation WCDMA and concurrently on both standards in terms of (FEC) deploying turbo and Viterbi next-generation LTE have taken different the algorithms’ processing and capacity. decoding algorithms in both in LTE/LTE-A approaches to modulate and map the and WCDMA. Other functions like CRC data onto the physical medium. As the The layer 2 and layer 3 algorithms use a mix calculation, rate de-matching operations ® name indicates, WCDMA is based on of Power Architecture general-purpose, and HARQ combining are also covered. code division multiple access and typically high-performance cores and security requires processing resources to efficiently • Downlink encoding processing element: acceleration. Most of this processing is perform spreading/despreading, scrambling/ Used for FEC deploying turbo encoding done on programmable cores where any descrambling and combining operations. algorithms for both LTE/LTE-A and standard including multimode operation can These are the main functions needed in the WCDMA and rate matching operations. be implemented efficiently. The commonality RAKE receiver approach typically used in • Chip rate processing element: Used to between WCDMA and LTE is the requirement WCDMA. The L1 operations in WCDMA are a accelerate downlink (DL) and uplink (UL) for secure backhaul processing. The mix of streaming and batch type operations, spreading/despreading and scrambling/ bulk of this is Ethernet, QoS, IPSec and which the baseband architecture must descrambling operations for both data WCDMA frame protocol processing, which process efficiently. and control channels. This block is used is offloaded to hardware acceleration and exclusively for WCDMA and CDMA2K/ leaves software flexibility for the actual L2 In contrast, LTE uses a mix of OFDMA for EV-DO standards. stacks of both standards. downlink and SC-FDMA modulation for uplink. This multicarrier approach follows • Equalization processing element: the principle of modulation for orthogonal Performs the MIMO equalization Meeting the Latency Budget subcarriers to maximize the spectrum operations based on minimum mean square error (MMSE), interference rejection To ensure continued competitiveness to density. The predominant operations in 3G technology, the 3GPP standards body OFDMA/SC-FDMA are the discrete Fourier combining (IRC), successive interference cancellation (SIC) or maximum likelihood based LTE on orthogonal frequency division transforms (DFT) in the form of FFT or DFT multiplexing (OFDM) and MIMO antenna and multiply-accumulate operations. (ML) approaches, with internal algorithms and outputs done and generated in techniques. The major performance goals are The nature of data organization and floating point mathematics. A number significantly increased data rates, reduced subframe structure in LTE allows the of configurable operation modes allow latencies and improved spectrum efficiencies. L1 processing steps to be scheduled adaptation of the equalization process Latency is a key network metric and has sequentially according to the available to the user characteristics and channel a major influence on a user’s experience subframe user and allocation information. conditions. These equalization algorithms both in voice calls and data transactions The key challenge is meeting the tight are quite complex, requiring substantial such as video and Internet applications. The latency budgets of the physical layer computation resources. Hence, Freescale key challenge is meeting the tight latency processing to maximize the available time selected to implement the algorithms in budgets of the PHY processing to maximize budget in the MAC layer scheduler. hardware acceleration, which is adaptable the available time budget for the rest of the to different nuances and frees them from PHY processing and MAC layer scheduler Baseband Acceleration and the DSP cores, leaving these for other tasks. The LTE standard defines the end- tasks in the processing chain. Addressing the Multimode user roundtrip latency at less than 5 ms, • Physical downlink processing element: which requires the latency within the base Challenges Performs an encoding of the physical station to be significantly lower (less than downlink shared channel (PDSCH) starting 0.5 ms in DL and less than 1 ms in UL). With Freescale devices, the physical layer from the user information bits up to the (PHY) on the DSP is implemented using MIMO equalization/detection and FEC are cyclic prefix (CP) insertion and antenna a mix of StarCore SC3850 or SC3900FP heavily used in newer, high bit-rate wireless interface handshake. Including DL-MIMO high-performance DSP cores and a communication standards such as LTE and precoding and layer mapping operation. baseband accelerator platform called MAPLE WiMAX. The MIMO equalizer and turbo • (multi accelerators platform). The MAPLE Physical uplink processing element: coding error correction algorithms both in Performs decoding of physical uplink accelerators provide highly efficient hardware UL and DL are the major influencers on base shared channel (PUSCH) resulting in implementation of the standardized building station throughput and latency. blocks for each of the air interface standards decoded information bits.
freescale.com 4 White Paper Small Cells Call for Scalable Architecture
Figure 4: Baseband Integration Figure 4: Baseband Integration
Antenna Antenna Backhaul PHY 1 Gb/s Layer 2/3 GE SoC
Layer 1 Transport and Control PHY 1 Gb/s CPRI CPRI
CPRI Maint. COST AND POWER I2C Multicore Multicore DSP MPU UART
Flash CPRI SPI
Flash DDR3
Freescale Technology Optional In typical macro and micro base stations, Figure 5: LTE 5: 20MHzLTE 20 Spectrum MHz andSpectrum CCDS Efficiency and CCDF the baseband channel card is composed of a single GPP device and multiple DSP -50 devices to handle scalable and variable -100 Before CFR After CFR numbers of sectors, the number of users -150 and throughputs based on the specific -200 -250 deployment requirements. Alternately, femto, -300 pico and metrocell base stations typically -350 handle a single sector and a given number of Spectrum (dB) -400 users and data throughputs. The traditional -450 single GPP device and single DSP discrete -500 device paradigm is evolving into a single -550 -80 -60 -40 -20 020406080 unified system-on-chip (SoC) solution Frequency (MHz) (see figure 4). 10x Before CFR Digital Front End (DFE) 10x After CFR
The DFE receives the different signals created 10x by the baseband processing board and prepares them to be physically transmitted 10x
to the RF transmitter. In a standard 3G radio CDDF lineup, its first function is multiplexing and 10x combining the different carriers, the second is peak-to-average power ratio (PAPR) limiting 10x 042 6910121416 (i.e. crest factor reduction) and the third is RF PAPR (dB) hardware linearization (i.e. digital predistortion).
Digital Up Converter (DUC): Receives the different carriers created by layer 1, which can and transceiver chips may offset the gain excess of 30+10=40 dBm. Then, the PAPR be of different natures (for example: one carrier obtained in the RF power amplifier. The dictates the minimum back-off from Psat LTE + one carrier WCDMA), pulse shapes and main features involved to increase RF lineup (OBO) at which the PA operates most of the sums them according to the carrier specified efficiency are crest factor reduction (CFR) time. This has a direct consequence on the pattern. This operation involves oversampling and digital pre-distortion (DPD). amplifier efficiency as this parameter lowers when OBO increases (figure 6). Consequently, and filtering functions. Crest Factor Reduction: Complex reducing the PAPR helps to reduce both the modulated signals vary in both amplitude and In a macro base station, the RF power RF PA device size (as well as cost) and its phase. When they can be described in the amplifier contributes the most to the overall power consumption. Due to its impact on frequency domain by their power spectrum, power consumption and it is imperative to the equipment performance, CFR is a clear we can characterize the amplitude in the allocate baseband processing resource to differentiator among competitive solutions7. increase the RF power amplifier efficiency. time domain by its statistical distribution In small cell equipment, it makes sense to (figure 5). This analysis extracts the PAPR, Digital Pre-Distortion: Amplitude modulated carefully investigate the overall digital and which is used to size the RF power amplifier signals create challenges for the amplifier as RF front-end architecture, as the power devices; if one needs to transmit 30 dBm its non-linear behavior creates in-band (EVM consumed in the baseband processing average with a signal PAPR=10 dB, the impacting) and out-of-band (spectral regrowth) RF PA saturated power (Psat) should be in distortion. To meet linearity specifications and
freescale.com 5 White Paper Small Cells Call for Scalable Architecture
increasing RF PA efficiency requirements, the Doherty amplifiers exhibit significantly larger envelope tracking PA, which is a competitive PA must be linearized and DPD is the most memory effects (and higher efficiency). Then, architecture with Doherty to increase average popular linearizer scheme in wireless cellular the burden put on DPD to meet linearity efficiency, the DFE also includes the drain infrastructure applications today (see figure 7). specifications depends on the PA architecture. envelope signal generation together with its synchronization with the RF path. A key step in DPD consists in approximating The simplest DPD scheme is an open- the PA behavior by a behavioral model. The loop correction without memory effects There are different possibilities to implement level of the correction depends on the model compensation, based on the static AM/AM the DFE functionalities in practical hardware: accuracy and its capacity to follow the PA and AM/PM PA behavior. The more complex FPGA, ASIC and DSP. Each has its strength behavior variations. Such variations can be DPD architecture is closed loop with memory and weakness in terms of flexibility, cost and “slow” compared to the signal envelope correction (based, for example, on Volterra power consumption and those aspects have to variations. For example, this is the case of series) and requires a demodulation path to be taken into consideration when defining the temperature, carrier frequency or output sample the output signal and compare it to overall system architecture. power variations. PA behavior variation can the desired transmitted one. also be fast (comparable to the envelope On the Rx path, the DFE gets the multi- speed) and can depend on the previous PA RF/IF Transceiver carrier uplink signal, extracts and filters each state. This dependency to the past is called The RF/IF transceiver design depends upon carrier to be transferred to the PHY layer. “memory effects.” The lower the contribution the transmitted signal; GSM or LTE require a The functional blocks described above of memory effects, the easier the correction. different resolution and dynamic range for the typically fall under the umbrella of the DFE. DAC and ADC. Also, CFR and/or DPD may Memory effect content depends on the PA Depending on the transmitter architecture, impact this resolution marginally. architecture: when Class A or Class AB PA some of those can be eliminated or others show low memory effect (but low efficiency), can be added. For example, in the case of an Conversion speed depends not only on the signal bandwidth, but also on the presence or absence of DPD, because the pre-distorted Figure 6: Single-Stage Class AB PA Gain and Efficiency Figure 6: 1-stage Class AB PA Gain & Efficiency signal should be able to include up to fifth order intermodulation products to allow good Gain (dB) Eff (%) 20 No CFR 100 correction. Gain 19 Eff 90 with CFR DPD may also require an additional return path to sample the output signal back to the 18 90 DFE for adapting the pre-distorted signal. 17 70 This added complexity has a cost, which should be justified or offset by the system 16 60 performance improvements. 15 50 14 40 RF PA Transmitter 13 30 We have shown how DFE features are
12 20 linked with the RF front end architecture and both should be optimized together for 11 10 optimum system performances. Compared to
10 0 macrocells, small cells present an additional -20 -18 -16 -11 -12 -10 -6 -6 -4 -2 02
Figure 7: Final Stage PA, Simulated ACP with and without DPD Figure 7: Final Stage PA, simulated ACP with and without DPD
-30 50 -30 50 ACP_noDPD ACP_noDPD -35 ACP_Volterra 45 -35 ACP_Volterra 45 ACP_no memory ACP_no memory -40 Eff 40 -40 Eff 40
-45 35 -45 35
-50 30 -50 30
-66 25 -66 25
-60 20 -60 20
-65 15 -65 15
-70 10 -70 10
-75 5 -75 5
-80 0 -80 0 -17 -16 -15 -14 -13 -12 -11 -10 -9 -8 -7 -17 -16 -15 -14 -13 -12 -11 -10 -9 -8 -7 Class AB PA: ACP vs. OBO, LTE 20 MHz, PAR=9 dB Doherty PA: ACP vs. OBO, LTE 20 MHz, PAR=9 dB freescale.com 6 White Paper Small Cells Call for Scalable Architecture
Figure 8: DC Power Budget for the RF Amplifier that provides performance, power and Figure 8: Small Cell PA Power Consumption cost benefits. Being relatively independent 50 from external IP provider’s next-generation AB technologies and timelines enables Freescale AB + CFR to drive a roadmap of devices that helps AB + CFR + memoryless DPD 40 Doherty + CFR + memory DPD OEMs meet targets for performance and timelines for their next-generation wireless technologies.
30 The key processing elements in any device for mobile wireless infrastructures are the programmable cores, hardware accelerators, 20 DC Power (Watts) internal interconnects and high-speed interfaces. Freescale has long been an embedded processing leader. The market- 10 proven Power Architecture core is at the heart of our strength and has been used by leading wireless OEMs worldwide for many 0 years. Significantly enhanced from generation 13 15 17 19 21 23 25 27 29 31 33 35 37 39 to generation, it comes with a rich ecosystem Power at Antenna Connector (dBm) to provide customers with a seamless migration from their current products to higher performance products. challenge in that the DC power consumed by This analysis highlights how system the baseband signal processing is no longer performance can be improved with CFR The StarCore DSP core has been enhanced negligible compared to the power consumed and DPD, together with an optimized RF by Freescale from generation to generation by the RF power amplifier and a careful amplifier architecture, but equivalent analysis for more than a decade and is known for analysis should be done to define the best measuring DC power consumption and its high performance and programmability. system architecture. estimating the cost associated with the signal The SC3850 is used today in DSP processing features implemented by the devices deployed by many of the wireless Based on realistic performances of RF power DFE is needed to achieve optimum system manufacturers in LTE, WCDMA and WiMAX transistors (5 V GaAs, 28 V LDMOS) and definition, keeping in mind that figure 8 is deployments and has earned leading results RF power amplifier architectures (class AB, based on a single RF transmitter, when usual from top benchmarking firms. Doherty), an analysis has been performed to equipment includes at least two, sometimes estimate a budget for the DC power required four, RF power amplifiers for enabling MIMO Other important components are the by the RF power amplifier for the different transmission. internal fabric, accelerator throughputs and small cell types and frequencies. Figure 8 standards compliance. The internal fabric is compares the performances for the different Still, there are side effects that need to be a component that connects all processing scenarios. These figures apply for >50 dB considered to get the full picture: improving elements and memories within the device; gain lineup, and signals showing PAR=10 dB power amplifier efficiency is not only “green” it must enable high throughputs and low without CFR, PAR=7 dB with CFR (@0.01 from a pure power point of view, but also latencies for data movement throughout the percent probability). Depending on the power dictates the need for managing the dissipated SoC as well as preventing stalling of any of amplifier architecture and DPD scheme, power. For example, it impacts the RF device the elements attached to it for processing margin for linearity varies from 4 to 1 dB. The mounting (bolt down, PCB with copper coin, data. Both the internal fabric and the figures consider 3 dB losses between the true SMD mounting on PCB with via holes) accelerators are proven to be highly efficient output of the power amplifier and the antenna and the cooling principle chosen (forced and were field deployed by Freescale connector to account for the isolator and convection with fan, free air convection). customers. filters blocks. Consequently, equipment size and cost are directly impacted by the DFE and RF front From figure 8, we can extract the following Device Architectures and Capacities end choices. assessments: for low-power transmitters Leveraging the StarCore and Power (<15 dBm), there is no need for CFR or DPD. Architecture legacies, Freescale has developed For medium power (15
freescale.com 7 White Paper Small Cells Call for Scalable Architecture
Figure 9: Freescale QorIQ Qonverge SoC for Base Stations
B4420 Micro/Metro SoC B4860 Macro SoC • Hundreds of users • Very high throughputs • High throughputs • Thousands of users • Multistandard • Multisector and multimode • Multistandard and multimode • One to two sectors • Full compliance to 3GPPRel. 10 • Full compliance with 3GPPRel. 10 • New advanced cores and acceleration • Ready HetNet deployments technologies • Pin compatible with B4860 • Ready for cloud and HetNet deployments
BSC9131 Femto SoC BSC9132 Pico SoC • Eight to 16 users (LTE (FDD, • 32 to 100 users (LTE (FDD, TDD), TDD), WCDMA, CDMAx) WCDMA) and multimode and multimode • Two MPU + two DSP + wireless • One MPU + one DSP + accelerators wireless accelerators
To address the high-end portion of the FigureRF Lineup 10: for 5 Freescale W @ 2.6 GHz RF Lineup for 5 W @ 2.6 GHz wireless infrastructure market, Freescale introduced the B4420 for microcells and the MMG3008NT1 B4860 for macrocells in 2012, both built MMZ25332B AFT26HW050GS around the latest DSP and Power Architecture cores (SC3900FP and e6500 respectively) (see figure 9).
RF Front End
Freescale is a clear leader in RF power Gtyp=14 dB Gtyp=26 dB Gtyp=15 dB products for cellular infrastructure P1dB=15 dBm P1dB=33 dBm P3dB=46.8 dBm applications and holds a solid number one market share position with the broadest product portfolio in the industry from femto Pout=37 dBm to macro cells8. Under our new Airfast brand, Freescale offers innovative RF solutions Lineup efficiency @ 37 dBm=36% Gain=55 dB integrating the latest developments in die technologies, package concepts and system architecture expertise.
Leveraging these advantages, Freescale metro cell LTE applications between 1.8–2.7 GHz. To complement the lineup, provides optimized solutions for small 2500–2700 MHz and exhibits 48 percent Freescale provides the MMG3008, a cells. Recently, the company released the drain efficiency at 8 dB output backoff from broadband general-purpose amplifier in AFT26HW050GS, an in-package single- its saturated power, the highest reported the SOT89 SMD low-cost package stage Doherty for 2.6 GHz LTE applications in the industry at the 2.6 GHz. To drive the (see figure 11 on next page). based on the latest 28 V LDMOS generation. final stage, Freescale offers the MMZ25332, The AFT26HW050GS is designed specifically two-stage class AB 5 V InGaP HBT MMIC, for wide instantaneous bandwidth micro/ exhibiting 25 dB linear gain and 33 dBm output power at 1 dB compression over
freescale.com 8 White Paper Small Cells Call for Scalable Architecture
Figure 11: Dual-Band RF Front End for Femto Base Station RF Module Block Diagram
RX SAW MML20211H
Freescale LNAs MML09211H
sp2t MMZ09312B Duplexer To Antenna 2 TX SAW Freescale Power Amplifiers Duplexer LTE-FDD/ MMZ25332B TDD and WCDMA sp2t sp3t Transceiver MMZ09312B Duplexer To Antenna 1 TX SAW Freescale sp2t Power Amplifiers Duplexer MMZ25332B RX SAW MML20211H
Freescale LNAs MML09211H GSM Sniff
sp2t
The MMZ25332 is not only an excellent driver, Conclusion it can be used in a pico base station as final stage with an output power up to 24 dBm, This paper analyzes the new requirements where its embedded output power detector relating to the emergence of small cells and opens the door for power control and alarm identifies paths for system optimization. information. On the receive path, Freescale Particularly, it demonstrates how the offers a state-of-the-art, single-stage, low- baseband processing features and the RF noise amplifier built on E-Phemt 5 V GaAs transmitter architecture should be considered technologies. together to achieve optimum cost and performances. As an example illustrating the effectiveness of a clever system approach, the block diagram Freescale is approaching this market in figure 11 describes the RF front end of a by providing scalable baseband signal femto dual-band reference board designed to processors and RF solutions. Associated with work with the BSC913X SOC. a very broad device portfolio, a set of design tools completes the ecosystem allowing customers to design in a consistent and stable environment with reduced time to market.
freescale.com 9 White Paper How to Reach Us: References:
1 “Heterogeneous networks–increasing cellular capacity”, Ericsson review, 1-2011 Home Page: 2 freescale.com “Small cells–what’s the big idea?” white paper, Small Cell Forum, February 2012 3 “LTE Base Station (BS) radio transmission and reception, ETSI TS 136 104 V9.11.0”, ETSI organization, RF Portfolio Information: March 2012 freescale.com/RFpower 4 “UMTS Base Station (BS) radio transmission and reception (FDD), ETSI TS 125 104 V10.5.0”, ETSI organization, March 2012
Email: 5 “LightRadio Portfolio: White paper 1, technical overview”, Alcatel-Lucent [email protected] 6 “Active Antenna Systems: a step-change in base station site performance”, NSN White paper, January 2012
7 “Digital modulators with crest factor reduction techniques”, Olli Väänänen, PhD thesis, Helsinki University Information in this document is provided solely to enable of Technology, March 2006 system and software implementers to use Freescale products. There are no express or implied copyright licenses granted 8 ABI report “RF Power Amplifiers, Equipment and RF Power Device Analysis for Cellular and Mobile hereunder to design or fabricate any integrated circuits based Wireless Infrastructure Markets”, Dec 2011 on the information in this document.
Freescale reserves the right to make changes without further notice to any products herein. Freescale makes no warranty, representation, or guarantee regarding the suitability of its products for any particular purpose, nor does Freescale assume any liability arising out of the application or use of any product or circuit, and specifically disclaims any and all liability, including without limitation consequential or incidental damages. “Typical” parameters that may be provided in Freescale data sheets and/or specifications can and do vary in different applications, and actual performance may vary over time. All operating parameters, including “typicals,” must be validated for each customer application by customer’s technical experts. Freescale does not convey any license under its patent rights nor the rights of others. Freescale sells products pursuant to standard terms and conditions of sale, which can be found at the following address: freescale.com/SalesTermsandConditions.
For more information, please visit freescale.com/QorIQ
Freescale, the Freescale logo and StarCore are trademarks of Freescale Semiconductor, Inc., Reg. U.S. Pat. & Tm. Off. Airfast and QorIQ Qonverge are trademarks of Freescale Semiconductor, Inc. The Power Architecture and Power.org word marks and the Power and Power.org logos and related marks are trademarks and service marks licensed by Power.org. All other product or service names are the property of their respective owners. © 2012, 2013 Freescale Semiconductor, Inc.
Document Number: SMCELLRFWP REV 3