TD-LTE: testing times for MIMO beamforming
At a high level, TD-LTE is simply LTE with the uplink and downlink multiplexed in the time domain rather than the frequency domain. TD-LTE’s ability to be realized in a single frequency band (as opposed to the paired bands required for FDD-based LTE) makes it an attractive option in an age where the pressure on spectrum is rapidly increasing. In addition, since the downlink in a cellular system typically carries more data than the uplink, TD-LTE has a desirable ability to change uplink and downlink capacity, often dynamically, by altering the number of time slots allocated to each direction.
A key part of making TD-LTE an effective solution for operators is another wireless innovation: Multiple-Input Multiple-Output (MIMO) antenna techniques. These offer faster data rates and better system capacity than Single Input Single Output (SISO) systems, without expending resources in the time or frequency domains. In an m × n MIMO system (one using m transmitting antenna elements and n receiving antenna elements), theoretical maximum data rates are limited by the smaller of {m, n}.
MIMO beamforming
When MIMO antenna techniques are married with TD-LTE, an interesting and useful attribute results: since the uplink and downlink share a single frequency band, the channel can be regarded as reciprocal. As a result, channel estimation of the uplink can be used to make reasonable assumptions regarding downlink channel characteristics. Channel reciprocity in a single uplink/downlink frequency provides lends itself to a way of improving both coverage and system quality: MIMO beamforming.
Beamforming uses multiple antennas to “steer” signals towards areas where subscriber density is high. Two or more antennas will transmit the same signal, but with carefully controlled phase/amplitude characteristics. The result is constructive interference in areas of desired reception and destructive interference where reception is unimportant. While beamforming has other uses (for example, it can be used as with receiving antennas to “null out” a nearby interferer), it is most commonly applied to optimize both system capacity and the quality of the user experience.
Early TD-LTE deployments plan to combine MIMO and beamforming, offering the advantages of higher data rates as well as capacity and quality improvement. A typical MIMO beamforming configuration can be thought of as a 2×2 MIMO system, except that each of the two transmitted “layers” is actually a steered beam formed by four transmitting antenna elements. This has led to growth in the study of 8 × n systems, where each base station is equipped with eight antenna elements, as a cost-efficient, spectrally-efficient alternative to the addition of cell sites or additional carriers.
All of this creates an incredibly complex RF environment that must be tested before deployment, and the stakes are high. It is no secret that the world’s largest network operator, China Mobile, is currently deploying TD-LTE trial systems in six of China’s most populated cities. It’s also a safe bet that most if not all of the spectrum awarded in 2010’s Indian BWA auction will be used for TD-LTE as well, since that auction awarded single (“unpaired”) 20-MHz wide bands. It is true that this spectrum could also be used for WiMAX deployments, but Indian operators have made their intentions clear… in one case canceling a planned WiMAX deployment across two major cities. While it will be years before the majority of these Chinese and Indian subscribers use LTE service, it is worth noting that the operators involved represent billions (literally) of today’s mobile phone users.
There’s no question that the stakes are high, and there’s little doubt that combining antenna techniques creates an order-of-magnitude increase in the complexity of the radio link. It is difficult enough to optimize MIMO or beamforming when one of those techniques is deployed in isolation, but balancing the requirements of both adds new depth to both the design and testing processes.
Equipment being used to emulate the radio channel for testing must now be able to accurately account for a number of parameters that were less critical in earlier technologies. One relatively arcane but significant example is in the phase relationship between uplink and downlink channels. In frequency domain duplex scenarios such as FDD-LTE, where uplink and downlink exist in separate frequency bands, there is no expectation of any particular phase relationship between the uplink and downlink. TD-LTE, however, is different. Uplink and downlink share a single band and are separated only by negligible slices of time. The technology itself relies on channel reciprocity, including nearly identical phase relationships between the two links.
This poses a risk in the receiver testing process. In order for test equipment to provide the proper level of control to the engineer, the equipment used to emulate the radio channel internally disaggregates uplink and downlink channels. Unless the equipment being used has been designed with phase accuracy in mind, there is no reason to expect that an emulated channel will accurately represent the reciprocal TD-LTE channel. To make matters even more complicated, the sheer number of transmitted signals (e.g. eight transmitted RF signals in planned TD-LTE deployments) must be replicated within this phase-controlled environment.
This issue has been recognized and resolved with modern channel emulation equipment. As an example, Spirent’s MB5 MIMO Beamforming Test System implements automated phase calibration specifically for this type of testing. Any system set up intended to properly emulate the TD-LTE channel must not only maintain the proper correlation between radio links in both directions, but it must offer a realistic phase relationship between uplink and downlink in order to maintain the realistically reciprocal properties of a TD-LTE system. The MB5, for example, maintains this relationship through the use of hardware that automatically performs phase calibration on a per-band basis.
Conclusion
There is no question that TD-LTE technology will, over the next several years, become a significant global force in the deployment of mobile subscriber services. The promises of TD-LTE are tightly coupled to the deployment of a combination of complex radio antenna techniques. With the complexity of the problem comes complexity and potential risk in the testing process as well.
Proper lab testing requires very careful control and, above all, repeatability. While a live TD-LTE radio channel is reciprocal by nature, it is neither a controllable nor a repeatable scenario, and can’t be created on a lab bench located near a design engineer. Controlling the RF scenario means disaggregating the uplink and downlink channels, which might a potential issue when replicating the reciprocal nature of the TD-LTE radio channel. New methods of implementing radio channel emulation mitigate that issue and the associated risk of receiver failures in the field.
The author Nigel Wright is in charge of Wireless Product Marketing and Corporate Marketing at Spirent Communications.
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
