Although the move to MIMO offers benefits to network operators and users, its complexity sets test engineers a number of challenges, says Mark Elo
Today's radio devices use a SISO (single input single output) configuration with one transmitter and one receiver and information sent over a single channel. MIMO (multiple input multiple output) transmission transmits information over multiple radio channels, but only occupies the bandwidth of a single channel. MIMO presents one of the most significant changes to happen to radio architectures in recent history. This technology can now be used in a wide range of commercial communications devices including mobile phones, PDAs, and laptops and is an integral part of the 802.11n and
WiMAX Wave 2 Standards.
Along with the benefits of increased bandwidth, multi-signal transmission and reception adds more layers of complexity. For instance, the transmitters must be aligned in time and phase and have a high degree of isolation from each other. This presents a number of unique testing challenges when moving from SISO to MIMO based systems that test engineers must consider.
From SISO to MIMO
The SISO configuration is used in almost all contemporary radio designs. Sometimes there may be an extra antenna for spatial diversity that is constantly switched for the best signal path. However, this is still considered to be a SISO system because there is a single up converter and a single down converter, a single demodulator/modulator, and a single data stream in the higher levels of the product's communications stack.
Multi-path effects can degrade a SISO transmission. For example, a Bluetooth signal with a symbol rate of 1M symbols per second must receive a symbol within a window of one microsecond. If multi-path effect delays the signal by more than this, a significant symbol error will occur. MIMO systems, on the other hand, require multiple paths. If two signals are transmitted with known characteristics, for instance a header, at the receiver end, one can assume what the signal should look like and create a model of the channel effects. When the unknown signal comes, i.e. the data, subtracting the channel effects can solve for the transmitted symbols. The key to a MIMO system, and why it is different from SISO, is that the behavior of the channel is critical and must always be understood.
There are three ways to transmit data using a MIMO configuration. The spatial multiplexing technique transmits different data on each channel, thus increasing the throughput. Spatial diversity transmits the same data on each channel. This redundancy in effect increases the robustness of the signal and improves the transmission coverage. Lastly is a method called beam forming. This technique improves the throughput and coverage by controlling the directionality and the shape of the transmitted signal.
A typical MIMO configuration can range from a 2×2 system, containing two transmitters and two receivers, to a 4×4 system with four transmitters and four receivers. Many commercial WLAN devices today employ a 3×2 configuration of three transmitters and two receivers. In the future, beam forming based systems could have up to 8×8 configurations.
Testing Challenges
Perhaps the greatest testing challenge for MIMO systems involves synchronisation with good channel isolation in the transmitter and the receiver. Transmission of multiple signals requires accurate synchronization of multiple channels in phase and sampling alignment. This means that RF test equipment such as signal analysers and generators must have precise alignment and excellent isolation between channels in order to make accurate and repeatable measurements.
For most test engineers, a major challenge is the ability to transition smoothly from single-channel to multi-channel testing and therefore choosing instruments that provide a clear and easy upgrade path to MIMO. For example, moving from WiMAX SISO to the MIMO versions based on Matrix A, B, and even C, the highest 4X4 configuration, can significantly lower test costs. Test engineers should also consider whether or not there is a clear upgrade path beyond the 4X4 Matrix C configuration.
Another major concern is keeping the cost of test per channel low while maintaining good performance, especially with respect to maintaining excellent channel isolation. This is important because measuring the channel characteristics is fundamental to verifying any MIMO device. The test equipment should ideally have independent transmitters and receivers for the best channel isolation and at least 14-bit or better amplitude resolution for good dynamic range.
Bandwidth is another important consideration. For Mobile WiMAX, the sub-carrier spacing is fixed at 10.94 kHz. The standard allows for FFT sizes from 128 to 2048, which means that the maximum signal bandwidth will be in excess of 20 MHz – so test equipment needs to have at least 20 MHz of bandwidth. If working with WLAN, then 40 MHz of bandwidth is even better for the 802.11n MIMO standard.
Instrument usability, or its user friendliness, is an often overlooked but equally important consideration. Intuitive displays are essential for debugging complex radio systems, especially when dealing with multiple signals. Going beyond the constellation diagram, users need to see how modulation quality behaves over time and over sub-carriers.
The measurements for SISO are similar to MIMO. For example, EVM is a key metric for establishing the quality of any digital signal. In a MIMO system, it is still important to understand the EVM performance of the system, i.e. the composite EVM. However, as part of the design process, it is also important to be able to understand the EVM performance of each channel, while it is in the presence of all the other channels. Here lies a significant challenge. For instance, if one of the transmitters is generating an in-band spurious signal, then the composite EVM would be degraded. The next step is to check the EVM of each channel or stream. In so doing, the engineer would notice that one of the streams has a degradation in EVM. This performance could be attributed to either time domain or frequency domain effects. By then observing the EVM of each OFDM carrier over the frequency, it will quickly become apparent that some in-channel distortion is causing the radios performance to be degraded.
Test engineers also need to see how the radio responds to changes in the channel, especially with different multi-path models. Channel response shows how all the radio transmissions interact with each other in the channel. In a 2×2 system the interaction is between TX1 and RX1, TX2 and RX2, TX1 and RX2 and TX2 and RX2. As the channel or stream count increases, the number of channel interactions also increases. For example in, a 4×4 system, the measurement needs to process 16 streams or channel responses to determine how each channel interacts with the other.
Beam forming also presents many test challenges. Beam forming is a technique that helps increase receiver sensitivity to the desired signal and decreases the sensitivity to interference and noise. This is accomplished by creating a series of beams and nulls in the transmitted signal. Test equipment for beam forming should be capable of finite phase and amplitude adjustment to be able to effectively create and receive specific patterns of radiation from each antenna.
Mark Elo is RF Marketing Director, Keithley Instruments