Design and Test Challenges for 8x8 MIMO Devices for LTE Advanced
Mobile network operators are continually upgrading their networks to match the demand for increased capacity and availability.
From less than two percent in 2011, mobile data will comprise 10 percent of all network traffic by 2016. The projected rise follows the forecasted increase in smartphones and tablet computers accessing data-intensive streaming video and enhanced social media applications, and a general increase in Internet activity from mobile devices — on-line retail and location-based services, for example.
By 2016, there are expected to be close to 20 billion mobile network devices, ranging from simple machine-to-machine (M2M) sensors to powerful PCs, and some of these will be in mission critical applications that require absolute network reliability. Mobile network operators are continually upgrading their networks to match this demand for increased capacity and availability.
In the mobile world, capacity gains come primarily from three variables: more spectrum, better modulation efficiency, and better frequency re-use through a progressively denser network topology with smaller cell sizes. The fourth generation networks currently being built use more frequency bands than previous generations and can use broader channel bandwidths. However, even with the increased spectrum made available from the shutdown of analog television services, network operators will struggle to meet long-term demand for capacity.
Spatial Multiplexing & MIMO
One solution to increased capacity without further additional spectrum is spatial multiplexing or multiple input multiple output (MIMO), which allows for the simultaneous transmission of more than one stream of data in the same time and frequency. To recap, 3GPP specifications describe the “channel” between transmitter and receiver in terms of inputs and outputs to and from it: a transmitter supplies input(s) to the channel, and a receiver takes output(s) from it.
First commercialized in Wireless LAN specification IEEE 802.11n with typically two spatial streams, both the downlink 3GPP specifications for LTE Advanced and 802.11ac for WLAN now specify up to 8x8 MIMO, or 8 separate streams. In both cases, the goal is to increase both overall capacity and the data rate that a single user can expect from the system.
LTE can use transmit diversity (multiple input, single output (MISO)) and receive diversity (single input, multiple output (SIMO)) as well as beamforming, either alone or in combination with MIMO. For LTE and LTE Advanced, successive 3GPP standards releases increase transmission complexity, beginning with release 8, which introduced transmission mode 7 (TM7), which supports single layer beamforming.
Release 9 added TM8, which supports dual layer beamforming (i.e. 2x2 MIMO with beamforming) and Release 10 adds TM9, which supports up to 8x8 MIMO with beamforming.
There are a total of nine different downlink transmission modes, each of which is suited to different channel and noise conditions.
Another option for network operators that is first specified in Release 10 and is enhanced in later releases is carrier aggregation (CA). While operators typically don’t have the contiguous spectrum to take advantage of LTE’s maximum transmission bandwidth of 100 MHz, they may have other spectrum available. CA allows them to “stitch together” a number of RF channels known as component carriers, to enhance downlink capacity.
Release 10 allows one additional component carrier, which can be either adjacent to the primary channel or in a different frequency band. Later releases allow a total of 5 downlink component carriers with a maximum of 20 MHz each for an aggregated 100 MHz bandwidth. These can be adjacent to the primary carrier, use a separate frequency in the same band, or be in a different frequency band. Each component carrier can have its own channel attributes: diversity, beamforming, and up to 8x8 MIMO, further compounding design and test complexity – particularly of the UE receiver.
The terms "codeword," "layer," "precoding," and "beamforming" have each been adopted specifically for LTE to refer to signals and their processing. The terms shown in Figure 1 are used in the following ways:
- Codeword: A codeword represents user data before it is formatted for transmission. In the most common case of single-user MIMO (SU-MIMO), up to two codewords are sent to a single handset or user equipment (UE).
- Layer (or stream): For MIMO, at least two layers must be used. Up to four for LTE and eight for LTE-A are allowed. The number of layers is always less than or equal to the number of antenna ports. The receiver needs at least as many antenna ports as the number of layers.
- Precoding: Precoding modifies the layer signals before transmission. This may be done for diversity, beamforming, or spatial multiplexing. The MIMO channel conditions may favor one layer (data stream) over another. If the spatial multiplexing is closed loop, the UE provides a precoding matrix indicator (PMI) so the eNB can cross-couple the streams to counteract the imbalance in the channel.
- Eigenbeamforming (beamforming): Beamforming modifies the transmit signals to give the best carrier-to-noise interference plus noise ratio (CINR) at the output of the channel, normally by maximizing antenna gain in the direction of a particular UE. It may also be set to minimize gain in the direction of a second UE which is managed by another base station (eNB). See Figure 2.
Circuit design starts with effective system simulation. Tools that can model the entire transmit/receive chain are important to shortening the design process and getting new products to market faster. A key element of simulation is the ability to comprehensively model the complete RF environment in a way that allows repetitive, deterministic design evaluation. The next step is to merge simulation and real-world system elements.
For the transmitter with multiple signals, the concern is with cross coupling within the block diagram. Testing needs to verify that the signals are coded and decoded correctly. Power amplifier RF performance must meet specifications because excess distortion will create signals that fall outside the channel and cause interference to other channels. New challenges include measuring amplifier performance under different operating conditions: envelope tracking reduces power consumption to match signal characteristics, reducing UE battery drain and increasing talk time. (For more on envelope tracking, see Agilent Application Notes 5991-1463EN and 5991-0797EN.)
Signal analyzers need to be able to pick out the correct signal in the spectrum and to demodulate and decode that signal. Transmitter measurements must be able to look at the spectrum in detail to see if interference is being generated. Agilent’s 89600 Vector Signal Analyzer (VSA) is an example of a measurement solution that gives an in-depth view of transmitter characteristics for many different wireless technologies, including both the FDD and TDD variants of LTE, and the aggregated carrier scenarios of LTE Advanced. Network analyzers give the required insight into the performance of power amplifiers and the quality of the RF chain components.
Receiver testing needs signal generation that can create a wide range of scenarios including fading and interference. The receiver needs to recover multiple simultaneous signals and adequately demodulate and decode them (Figure 3). Testing needs to ensure the performance of all the components within the block diagram, so the overall design meets the needs of real life conditions, including dynamic fading and interference.
Another important point that can be confusing is that MIMO in each of its transmission modes uses specific antenna ports. MIMO does require a minimum of two physical antennas, but a number of antenna ports may share the same physical antenna. Each antenna port has its own signal processing functionality, which separates the streams based on the UE reference signals defined for each transmission mode (Figure 4).
Looking at solutions, design systems such as Agilent’s SystemVue allow the export of simulated signals to a hardware block and importation of its real-world outputs to downstream simulation. For example, it’s possible to fully emulate an 8x8 MIMO channel and test a new receiver under the most stressful of scenarios. Using a fully-coded data source means testing can extend through the RF channel right to the receiver’s final decoding processes.
The system in Figure 5 combines SystemVue simulation tools with signal generators, multi-channel receiver, and vector signal analyzer to provide an overall system test solution for the evaluation of both transmitter and receiver characteristics for design and development of 8x8 MIMO devices.
Designing and testing LTE devices gets more complicated with each revision of the 3GPP standards. More frequency bands and more channel bandwidths, combined with MIMO, diversity, and beamforming techniques for each channel, make new device development a challenge. Test and measurement solutions that allow the deep insight required for device and system performance are key to reducing the time it takes to get new functionality into the hands of end-users.
This article originally appeared in the November/December print issue. Click here to view the full issue.