Testing TD-LTE with real world test technologies
However, as broadband data services have rapidly expanded, the demand on the network has become asymmetrical. In other words, the downlink and uplink loads on the network are no longer balanced as subscribers typically request (and download) much more content than they upload. With a symmetrically provisioned data service, the asymmetrical data demand will rapidly cause the uplink to become underutilized as downlink capacity is reached.
Time Division Duplexed (TD) protocols have an advantage in this situation since the relative allocation of bandwidth between uplink and downlink can be adjusted by adjusting the scheduling of uplink and downlink transmissions. With effective scheduling, carriers can operate their networks at a higher level of utilization than one which was built on a symmetrical model. TD-LTE allows bandwidth allocations to be dynamically modified and uplink and downlink transmission schedules within a single channel can be updated, depending on the particular needs of the network. This enables carriers to operate LTE networks at a higher level of utilization.
The key to a robust and reliable TD-LTE test methodology is to ensure that test equipment supports several key requirements. In the real world, the uplink and downlink coexist in the same spectrum, so effective test equipment designed to support successful TD-LTE deployments must also provide these same features. In particular, test equipment must support bidirectional testing and have phase and amplitude balanced uplink and downlink channels.
Advanced wireless test equipment such as a channel emulator can provide the real world environment necessary for TD-LTE Test. Integration of channel emulators within test solutions that support device and infrastructure testing provides test results that more closely reflect what happens in the real world. With the growing demand for bandwidth and the development of MIMO based protocols, channel emulators that can also support both the RF requirements for TD-LTE testing as well as implement the correlation models typical of real world activity are essential to effective testing.
Requirements for real world testing of
TD-LTE systems and devices
MIMO protocols such as LTE are now impacted more than ever by the changing nature of the “real world” radio environment and are significantly affected by the degree of correlation exhibited by the channel models used during test. While basic conducted testing uses standard channel models, more advanced test solutions can enable test of TD-LTE solutions both over the air and with replicated radio field conditions in the lab.
Standard laboratory testing completed using wired connections produces repeatable results but lacks the “real world” and “through the antenna” aspect of over-the-air testing. While over the air tests (such as drive tests) represent the real world, testing such as drive testing lacks repeatability. This is because there are many variables which impact performance during a real world test, for example, channel conditions could change depending upon seasonality and network loading; real world testing such as drive testing is also expensive to perform.
To bridge this gap between laboratory and real world over-the-air testing, channel emulators are deployed in the laboratory testbed. Channel emulators replicate real world channel propagation conditions in a controllable and repeatable fashion through the use of complex channel models and multiple, programmable parameters (figure 1). Sophisticated channel emulators offer bi-directional operation (simultaneous activity in downlink and uplink directions) with independent programmability of channel characteristics in both directions. By using channel emulators, radio designs and performance can be verified, test coverage can be improved, test cycles decreased and higher quality products can be introduced to the market in a shorter period of time.
Data communications technologies, as employed in TD-LTE, demand high system dynamic range and excellent RF fidelity. These radio systems often employ advanced digital modulation technologies to increase capacity. An example is 64QAM (Quadrature Amplitude Modulation) that carries six bits per symbol per OFDM subcarrier. In addition, techniques like OFDMA improve the operation of the system and allow for scalable capacity. These techniques, coupled with multiple antenna technologies, MIMO, ultimately result in system operation that can provide scalable, reliable capacity to mobile stations with aggregate downstream data rates in excess of 100 Mbps and upstream data rates in excess of 50 Mbps.
But these increases do not come without some cost. Higher order modulations demand high dynamic range and linearity. A 64QAM signal may need in excess of 20 dB SNR to achieve better than desired maximum block error rate. OFDM systems transmit many small subcarriers which introduces wide changes in instantaneous power level; a peak to average power ratio (PAPR) greater than 10 dB may not be uncommon. And, with frequency selective fading environments typical of mobile communications, certain OFDM subcarriers may be deeply faded while others may not, further increasing the demand on dynamic range. The TD-LTE standard currently implements the uplink with SC-FDMA and was specifically designed to reduce the impact of deeply fades and therefore reduce the power consumption of the UE.
Channel emulator input dynamic range
There are several considerations associated with input power when choosing a channel emulator for use with a 3GPP TD-LTE device. These considerations include input power range, peak power and signal-to-noise margin.
The transmitted signal from the 3GPP LTE device can have a very wide dynamic power range. Although the average power may have some maximum value, when OFDM is employed, the PAPR can be greater than 10 dB, and hence the system must accommodate this maximum. Even with SC-FDMA as employed in transmitters for 3GPP TD-LTE UE devices, PAPR is still present and may be 8 dB or more. Mobile devices also implement transmit power control to vary their output power, typically as a function of their distance from the eNodeB. TD-LTE transmit power control may result in 63 dB or more of actual power change. Furthermore, when the device is transmitting a higher order modulation, such as 64QAM, an adequate signal to noise ratio (SNR) must be maintained. When interfacing with a 3GPP TD-LTE UE, a channel emulator that can allow for direct connection of devices that transmit from +23 to -40 dBm, and still allows for sufficient PAPR and SNR margin, provides for a robust and efficient test configuration.
Bidirectionality and phase balance
A time division duplexed signal poses a unique challenge to test equipment design. The signal path through test equipment used for FDD protocols isn’t required to support phase balance when conducting bidirectional testing since the uplink and downlink both work in different spectrum and follow their own pilot.
A TD protocol however requires that both uplink and downlink paths be balanced in order to correctly emulate a bidirectional connection. This is especially important as the base station is able to use information from the uplink to control the downlink transmission.
Fading and noise floor
A fading channel emulator is employed to provide realistic fast fading conditions; ideally, the emulated fading will match that observed by a subscriber using the devices on the service provider’s network.
With an OFDM signal, as used in 3GPP TD-LTE, certain subcarriers may be faded, or momentarily reduced in amplitude by 20 dB or more due to the frequency selective fading. As each subcarrier is a modulated signal, with modulation up to 64QAM, this momentary drop in signal amplitude must be considered relative to the noise floor of the channel emulator equipment.
For example, if a signal with an average output power of -40 dBm is momentarily reduced by 20 dB due to fading, the amplitude will be -60 dBm. To maintain an adequate SNR for 64QAM at 25 dB, the test equipment noise floor should be no more than -85 dBm. Noise floor is often expressed by test equipment vendors as a noise power spectral density. Assuming 25º C and a 10 MHz wide signal as would be typical for 3GPP LTE, the noise power spectral density of the test equipment would need to be less than -155 dBm/Hz to insure that the signal fidelity was maintained even during fading conditions. If the noise floor was greater than this, it is possible that as the emulator provides the fading channel, the emulator could also introduce a noise level that will cause demodulation errors in the receiver as a direct result of the channel emulator noise floor and not of the device under test.
Implications of beamforming
In the TD-LTE environment, many service providers and equipment vendors have been considering beamforming in deployments. By focusing the transmission energy in beams:
• Higher range can be reached;
• Less energy can be used for the same range;
• Interference can be mitigated;
• Network capacity can be increased;
• Overall improvements in system performance.
Beamforming algorithms are distinguished on the basis of the algorithm used to select the “beamformer.” With regards to TD-LTE, one must consider that “air” is reciprocal in nature; i.e. the downlink wireless path looks exactly like the wireless path in the uplink. There may also be multipath reflections, wireless channel variations, phase changes etc. as the signal traverses the wireless environment. In general, beamforming algorithms utilize the characteristics of the air interface such as channel variations and reciprocity.
Channel emulation provides a methodology to reproduce over-the-air conditions in the lab for testing and benchmarking different devices. Channel emulation can be used to validate improvements and performance gains due to beamforming algorithms. However, robust test of a beamforming device in the lab will require the use of a channel emulator that is bidirectional and has reciprocal and balanced paths in a cabled lab environment (figure 2).
This is because beamforming algorithms depend on uplink phase and amplitude information to steer the downlink antenna field pattern. Channel estimates and other signaling information are exchanged continuously between the BS and the MS and hence a bi-directional connection needs to be provided in the lab. Reciprocity means that transfer functions for each path in a MIMO system look exactly the same in both directions and the impulse responses hij(t) must identical for both directions (figure 3).
In practical terms, the test equipment used for beamforming must ensure that the phase of the downlink channel should be calibrated to be equal to the phase of the uplink channel i.e. ΦDL needs to be equal to ΦUL, the balance should be end to end i.e. right from the point of the connection of the antenna port of the eNodeB to the point of the connection of the antenna port of the MS and similar amplitude balancing may be required for UL and DL paths (figure 4).
At the edge of a cell, a terminal device typically raises it’s transmit power to the maximum permissible in order to ensure continued robust communications. However, this quickly leads to the situation where the terminal device is only able to transmit one resource block in any given time period. Any attempt to transmit more than one resource block would necessarily spread all available transmit power across more resource blocks and reduce max range.
However, TD-LTE will in reality deliver cell edge performance comparable to FDD performance, since almost all cells will typically have more than one user and thus each device would be limited to a single resource block per unit time irrespective of the max capacity of TD-LTE or FDD LTE.
Verification of this performance can be accomplished using a variable AWGN noise source. AWGN noise is a good approximation of the noise seen by a device due to other sectors/cells in the vicinity of the terminal device.
The rising demand for wireless broadband data is driving the adoption of TD-LTE around the globe. With TD-LTE, the allocation of bandwidth can be modified by adjusting the scheduling of uplink and downlink transmissions on a single channel depending on the particular needs of the network. This enables carriers to operate LTE networks at a higher level of utilization, but the protocol poses a unique challenge to equipment design as it requires that both the uplink and downlink paths be balanced in order to correctly emulate a bidirectional connection.
TD-LTE testing must consider test equipment dynamic range, phase and amplitude balancing, and bidirectionality in order to model TD-LTE deployment scenarios including beamforming. To ensure excellent “real world” test of TD-LTE devices and systems, channel emulators can be chosen that both meet the aforementioned requirements and which also provide appropriate automation and channel models that help bridge the field and the lab and enable effective replication of real world conditions in the lab. One good example of such a channel emulator is Azimuth Systems’ ACE MX MIMO Channel Emulator. The ACE MX channel emulator is designed to easily support bidirectional testing and because it include internal isolators and circulators, is delivered to customers with phase and amplitude balanced channels — which ensures that regardless of the customer’s testbed, Azimuth’s real world solutions deliver excellent performance for real world TD-LTE testing.