Next year in China, China Mobile will build a pre-commercial LTE network and, according to In-Stat, is expected to be the first operator to commercially launch LTE services. In Spain, Telefónica has conducted the first live test on an LTE network installed in Madrid. This test consisted of VoIP calls and a video call with data and image downloads at speeds exceeding 140 Mbps, which is approximately ten times faster than what can be achieved with conventional 3G networks using HSPA technology. But the transition to LTE is not as simple as switching or adding more base stations. There are complex problems to solve, especially regarding new operating bands, interference issues, and high data transmission speeds in a complete IP (Internet Protocol) mobile system. To achieve this, system designers must revise the criteria for devices, and device manufacturers must find ways to meet the demand for advanced features.
As the number of wireless operators committed to LTE networks continues to grow (26 in the latest count, according to the GSA, the Global Mobile Suppliers Association (iv)), numerous technological challenges arise. Primarily, we can expect faster devices with low power consumption, high linearity, and smaller form factors to enable LTE. The advantages of low-power and easily integrated RF CMOS technology are well known. Building on this foundation, UltraCMOS™ technology is a CMOS process in which a 50–100 nm silicon film is formed directly onto a sapphire substrate. This results in completely empty devices with little to no body charge under the gate. Consequently, the UltraCMOS process delivers faster devices with lower power loss, excellent linearity, and high isolation, making it ideally suited to the demands of mobile devices and LTE base stations.
What makes LTE unique?
The move to LTE is driven by the demand for broadcast data from customers on the go. Having grown accustomed to exceptional user experiences in their offices, consumers now demand the same from their mobile devices: fast upload and download speeds and long battery life, all in a compact form factor. To succeed, operators need to meet these demands and find more efficient uses of available spectrum while reducing capital and operating expenses to remain competitive.
Fortunately, LTE offers download speeds of 100 Mbps and upload speeds of 50 Mbps for every 20 MHz of allocated spectrum; these are the kinds of data speeds that will give consumers what they're looking for with the spectral efficiency that operators need. Even higher data speeds are possible, such as up to 326.3 Mbps download when using multiple antennas. Since mobile customers expect their service to work anywhere, it is important that, despite being optimized for speeds of 0-15 km/h, LTE can support high-performance mobility even when the device is moving at 120-250 km/h.
Another unique feature of LTE is its use of two different access techniques: orthogonal frequency-division multiplexing (OFDM) for downlink (base station to mobile) and single-carrier frequency-division multiple access (SC-FDMA) for uplink (mobile to base station), with an adaptive modulation technique ranging from QSPK to 64QAM. The use of different access techniques improves power amplifier (PA) efficiency on the mobile side (resulting in longer battery life) and enhances spectral efficiency on the base station side. It is also important to note that some LTE mobile devices will support FDD and TDD duplexing techniques, allowing users to dynamically adapt to different systems across countries. Furthermore, LTE operates over a scalable bandwidth of 1.4 to 20 MHz for both uplink and downlink, supporting both mobile and existing frequency bands.
Impact on System Design:
In the implementation of OFDM within LTE, several closely spaced orthogonal subcarriers form a resource block (RB). Depending on the bandwidth of the particular system, the number of RBs will vary (as a rule, 1RB = 12 subcarriers of 15 kHz each). This technique presents many challenges to receiver design in terms of adjacent channel selectivity, since the LTE specification indicates much higher interference in the localized band at just 1 RB. As a result, the choice of system architecture (direct conversion or IF sampling) and signal processing technique (analog vs. digital signal processing) has a significant impact on meeting LTE system requirements. Generally speaking, devices with high linearity and isolation will help overcome this challenge.
Initial LTE base stations will employ 2x2 antenna technology and will likely quickly adopt 2x4 antenna technology. This increases market pressure for greater integration in order to control the number of components and minimize the bill of materials and design complexity. For example, in a 2x2 configuration, there will be a digital pass-through attenuator (DSA) on each of the Tx and Rx paths and one on the digital pre-distortion feedback path, meaning there will be 15 DSAs per base station. Clearly, the need to reduce dimensions and increase device performance is a critical factor in system design.
Since LTE network support extends beyond current bandwidth deployments, bandwidth expansion is also crucial. All devices in LTE mobile and base station networks, including switches, mixers, and DSAs, will need extremely high bandwidth to accommodate the additional frequencies. The switching capacity, in particular, must be sufficient to achieve an unprecedented number of contact positions—up to 12 or even more in a single-pole device (SP12T). Why so high? The additional states require an advanced serial interface to reduce I/O and improve functionality. Furthermore, some LTE deployments will incorporate TDD, which increases the requirements for the number of switching contact positions. This increased switching capacity must be achieved with fast setup times and in the smallest possible form factor, with compact routing to accommodate the increased functionality within the device.
Finally, LTE requires very low losses to achieve the signal-to-noise ratios (SNR) necessary for high data rates. This is particularly complicated due to LTE's additional operating bands, which place a significant demand on the antenna. In addition to the remaining requirements, active antenna tuning will likely be necessary to achieve the desired LTE device performance (v).
Meeting the challenges
Fortunately, the UltraCMOS process is widely available to achieve high speeds, low power consumption, high linearity, and greater integration of switches, mixers, and DSAs into the LTE signal chain.
High Speeds:
A fast switching speed is imperative to protect the receive path from damage when strong blocking signals are present, and this is also key to base station gain control. As the number of LTE contact positions increases, this specification becomes even more important. Higher switching speeds and shorter settling times lead to more reliable and accurate performance, and UltraCMOS inherently offers these advantages. For example, the UltraCMOS PE43204 is a DSA with a typical switching speed of 30 ns (see Figure 1), while maintaining an input third-order intercept point (IIP3) of +61 dBm (typical), insertion loss of 0.6 dB, and an electrostatic discharge (ESD) of 2 kV. In comparison, a GaAs DSA exhibits a typical switching speed of 130ns, which is more than four times slower than the UltraCMOS DSA.
Low power consumption
In the LTE specification, SC-FDMA was chosen for the uplink to reduce power consumption, but lower power consumption within mobile devices will allow for longer battery life, a key consideration for consumers. An SP9T switch like the UltraCMOS PE42692 is well-suited for LTE applications as it offers a typical Idd supply current of 120 µA.
High linearity
is not a new requirement for mobile devices. In fact, the RF input section module has long been the most linear element in a mobile device. However, as LTE systems become more complex, it is difficult to have a process that can integrate more functions while still achieving high linearity. Essentially, although data rates increase with LTE, the linearity requirements remain the same despite the additional semiconductor content. As a result, devices in an LTE system must offer better linearity than those used in previous generations of mobile devices.
Due to the isolation gate of CMOS technology and its inherent ability to incorporate mixed-signal design techniques, UltraCMOS ICs can meet the linearity performance requirements that demand a monolithic solution. In fact, UltraCMOS is currently being used in the design and manufacture of devices that exhibit high levels of linearity and isolation. The SP9T PE42692 switch, for example, exhibits an input third-order intercept point (IIP3) of +71 dBm (see Figure 2) with an insertion loss (IL) of 0.6 dB and Tx-Rx isolation of 43 dB (900 MHz). Specifications such as these enable higher data rates through the system and improve interference resistance, as well as better performance across the entire spectrum.
Linearity is closely linked to high isolation, which improves signal quality in the presence of interfering signals, and high isolation is necessary to meet the demanding performance requirements of the duplexer. As LTE introduces new operating frequency bands, isolation becomes even more critical. At the base station, digital pre-distortion (DPD) is a critical factor in improving access point (AP) efficiency. Essentially, the system samples the transmission path, corrects the signal, and feeds it back to enhance AP efficiency. Using components with high linearity/high isolation in the DPD feedback path is important to avoid introducing distortion and thus degrading the AP efficiency that is being improved. Operating between DC and 3000 MHz, the PE4257 UltraCMOS SPDT switch, for example, is well-suited to the base station feedback loop design requirements with 64 dB of isolation at 1000 MHz (see Figure 3); an isolation specification resulting from the very high isolation properties of the sapphire substrate in UltraCMOS.
Smaller Size/Greater Integration
Simply put, LTE services require devices to handle large volumes of data at high speeds over a wide bandwidth. This additional functionality will require more integration to maintain a compact footprint and meet power consumption forecasts. Because it is a CMOS process, UltraCMOS offers support for high levels of integration. UltraCMOS switches, for example, have an integrated decoder, so they do not require additional control signals such as GaAs. Furthermore, blocking capacitors are eliminated because the switches integrate a negative voltage generator to cut off the FETs. In an effort to further increase integration, Peregrine Semiconductor engineers have developed the MultiSwitch™ (see Figure 4), which incorporates four independent, high-performance, multi-position RF switches in a monolithic flip-chip IC controlled by a unique integrated CMOS controller, providing a size reduction of over 85% compared to other solutions. For example, the MultiSwitch measures 1.6 x 1.93 mm, and the SP9T UltraCMOS with integrated decoder, voltage generator, and ESD protection measures 1.36 x 1.28 mm. By comparison, the GaAs SP9T-based implementation measures 3.0 x 3.5 mm and requires 29 interconnects in a custom multi-chip package.
Unlike any other available device, the MultiSwitch RF IC provides +71 dBm IIP3 linearity and over 70 dB of isolation on critical paths. The device integrates key elements that would normally be outside the GaAs circuitry, including three control lines on 12 independent paths.
By 2011, ABI Research expects that approximately 34 million users worldwide will subscribe to LTE (vi), which promises consumers speeds on their mobile devices that will rival those available via cable or DSL. With the DSAs, mixers, and SP9T switches currently available in high volume (and with plans for SP12T and above), UltraCMOS is well-suited to support mobile device and LTE base station designs.
(i) ABI Research, July 2009. http://www.abiresearch.com/research/1003359. Accessed August 25, 2009.
(ii) In-Stat Research, July 2009. http://www.instat.com/press.asp?ID=2577&sku=IN0904599CCM
. Accessed August 25, 2009.
(iii) Telefonica conducts 1st tests of LTE in Spain http://www.cellular-news.com/story/36835.php
(iv) GSA, March 2009. http://www.gsacom.com/news/gsa_265.php. Accessed August 25, 2009.
(v) Ranta, Tero and Rodd Novak. “Antenna Tuning Approach Aids Cellular Handsets,” Microwaves & RF, November 2008. http://www.psemi.com/articles/2008/2008_ar_1.pdf. Accessed August 25, 2009.
(vi) ABI Research, July 2009. http://www.abiresearch.com/research/1003359. Accessed August 25, 2009
Authors:
Dylan Kelly and Mark Schrepferman, Peregrine Semiconductor Corp.
More information or a quote