Conducting Measurements on LTE Transmitters
Engineers preparing to measure long-term evolution (LTE) face many challenges. For one, LTE introduces orthogonal frequency division multiplexing (OFDM) into the cellular world. It also brings bandwidths of up to 20 MHz, significantly wider than those of earlier technologies such as GSM and WCDMA. Understanding the tests defined in the specification, knowing the issues associated with particular tests, ensuring that the equipment is in place, and training personnel are some of the considerations.
However, the specification is not complete. Test equipment selected now must be flexible enough to accommodate new tests as they are defined, and test equipment vendors must show how their products will keep pace with the evolving LTE specification.
One solution to testing LTE is to upgrade existing instruments by adding expensive software and hardware options. While this may seem cost-effective, the existing equipment must have the processing power to handle LTE analysis.
OFDM splits the channel bandwidth into multiple subcarriers transmitted simultaneously to the receiver. The FFT analysis required to analyze the incoming LTE signal is computationally intense. If an instrument is more than two to three years old, the upgrades may make the instrument LTE capable, but the new software actually could slow measurement time. If speed is critical, upgrading equipment designed for 2G and 3G may not deliver the desired performance.
While OFDM has distinct advantages—easy and efficient handling of multipath interference and robustness against narrowband interference—there are a few disadvantages. Two potentially critical concerns are its sensitivity to frequency offset and phase noise and a high peak-to-average power ratio (PAPR) problem that reduces the power efficiency of the RF amplifier at the transmitter.
For the uplink, 3GPP selected single-carrier frequency division multiple access (SC-FDMA), a variant of OFDM that reduces the PAPR. As a result, the uplink and downlink have different modulation schemes. All of this makes selecting the proper test instrument essential to ensure performance.
An LTE frame has dimensions for frequency and time. In time, the LTE frame is 10 ms long and contains 10 subframes, each of which contains two slots. The resource block is a transmission unit exactly one slot wide in time and 12 subcarriers wide in frequency. The slot or resource block contains six or seven symbols depending on the cyclic prefix (CP) parameter. A normal CP resource block has seven symbols while an extended CP resource block has six symbols.
Resource blocks are composed of resource elements (RE). The REs are the smallest unit in an LTE frame with dimensions of one subcarrier by one symbol. Measurements for conformance typically will be made at the subframe level; however, measurement instrumentation needs resource block, subframe, and frame resolution. When evaluating test equipment, the resolution of the various measurements is an important consideration.
LTE UE Power Measurement Overview
For any transmitter, output power is the fundamental metric. However, test cases and measurement equipment must ensure that all LTE user equipment (UE) developers measure power identically. As defined by the 3GPP LTE specifications, the proof of power conformance involves transmit power and output power dynamics.
The metrics evaluated under the transmit power category include maximum output power (MOP), maximum power reduction (MPR), and UE additional maximum power reduction (A-MPR). Output power dynamics evaluations apply to power control, minimum output power, and transmit off power. For the LTE uplink, output power is not a simple metric with a single maximum value for each UE power class.
In actual use, a UE cannot transmit excess power because it has the potential to interfere with other UEs and adjacent systems. MOP defines the maximum transmitted power in the channel bandwidth for all transmission bandwidths. LTE measures transmission bandwidth in units of resource blocks.
The channel bandwidth determines the maximum number of resource blocks in an LTE slot, as shown in Table 1. The MOP applies to a particular UE power class for a particular channel bandwidth and transmission bandwidth as shown in Table 2 using Power Class 3 as the example. When all resource blocks in a slot are used, MPR is applied.
Table 1. Channel Bandwidth vs. Transmission Bandwidth
Table 2. Transmission Bandwidth Configurations for MOP Specification
Source: 3GPP TS36.803 V1.1.0
MPR is a power reduction value used to control the adjacent channel leakage power ratio (ACLR) associated with the various modulation schemes and the transmission bandwidth. An adjacent channel may be either another Evolved Universal Terrestrial Radio Access (E-UTRA) channel or an UTRA channel. Different ACLR specifications apply to either scenario.
Using UE Power Class 3 as the example, Table 3 shows the power reduction value for the various modulation types and the transmission bandwidth. Comparing Table 2 and Table 3, some specification harmonization still is required to ensure that the transmission bandwidths for the MOP and MPR are consistent.
Table 3. MPR for Power Class 3
Source: 3GPP TS36.521-1 V2.0.0
When the situation warrants it, the LTE network can indicate to the UE that additional spectral emission control is necessary. A-MPR is not power control but power reduction due to specific regulatory or deployment constraints.
The specification indicates that the application of A-MPR should be the exception rather than the rule with the admonishment that additional spectrum emission requirements should only be used in a restricted set of transmission bandwidth configurations and deployment scenarios. A set of network signaled values is defined based on the E-UTRA band, the channel bandwidth, and the number of resource blocks in use.
Modern cellular communications systems use power control to ensure that each UE device transmits only as much power as it needs to achieve a successful communications link. Various schemes exist to determine the amount of power the UE should use for transmission. Whatever the method used to convey power control information, the UE must respond appropriately and set its power output to the desired value accurately.
For Power Class 3 UE devices, the specified power range is +23 dBm to -40 dBm. Using power control, the UE must be able to set its output power accurately across this range. When commanded to the minimum power control value, the UE must transmit at or below the specified -40 dBm. Excess minimum output power, like excess maximum power, can adversely influence the coverage area and performance of other UEs accessing the system.
As the circuitry that drives the UE is always active, residual output power is present when the transmitter is not actively sending data. The residual power transmitted during off periods must be ≤-50 dBm.
With the introduction of LTE and its dynamic bandwidth, the transmission bandwidth or number of resource blocks in use is a critical parameter for LTE measurements. For example, when the channel bandwidth is 10 MHz, if the number of resource blocks required is less than the maximum number of 50, the transmission bandwidth can be significantly less than the channel bandwidth.
As mentioned previously, the transmission bandwidth determines the maximum power that the UE can transmit in any given subframe. Power reduction is required when the transmission bandwidth approaches maximum.
Figure 1 shows various permutations of a 10-MHz, 64-QAM uplink signal using a low-power laboratory amplifier. The ACLR offsets assume adjacent E-UTRA channels. The waveforms were created using Anritsu's IQproducer for LTE application software, and each waveform consists of a single frame.
Figure 1. Various Power Levels and 64-QAM Transmission Bandwidth Configurations
The blue trace represents a fully occupied transmission bandwidth. Each slot in the frame contains 50 resource blocks. The reported values for ACLR shown below the trace are for the blue trace.
The pink trace was captured when the power was reduced 2 dB, the recommended MPR for 64-QAM in the specification (Table 3). The yellow trace demonstrates transmission bandwidth effects. Reducing the number of resource blocks from 50 to 16 created the signal defined in Table 2 for the 10-MHz MOP. Both the yellow trace, showing 16 resource blocks, and the blue trace, showing 50 resource blocks, used the same input power to the amplifier. Notice the impact of these various configurations on the ACLR, particularly L1 and U1.
In Figure 2, the waveform changes from 64-QAM to 16-QAM. The ACLR mask has changed from the 10-MHz offsets used for adjacent E-UTRA channels to 5-MHz offsets for UTRA adjacent channels.
Figure 2. Various Power Levels and 16-QAM Transmission Bandwidth Configurations
Some signal analyzers have predefined ACLR masks for all the various permutations of ACLR defined in the LTE specification. Moving from 10-MHz offsets to 5-MHz offsets is as simple as making a selection from a drop-down menu on the instrument.
The traces in Figure 2 again show the impact of power levels and the transmission bandwidth on ACLR. The yellow trace represents a signal having maximum transmission bandwidth. For a 10-MHz signal, each slot is carrying 50 resource blocks. The ACLR values shown below the trace window are for this trace.
For the pink trace, the power into the amplifier was reduced by 1 dB, the MPR for 16-QAM as shown in Table 3. The 1-dB drop in power in-band reduces the out-of-band power by about 10 dB. The green trace shows the results when the resource blocks are centered in the transmission bandwidth. The 16 resource blocks are placed in the center of the channel bandwidth beginning with resource block position 17. This balanced reduction in resource blocks gives the best ACLR performance.
The input power level for the pink trace and the green trace are identical. The blue trace shows the signal without the amplifier to show any residual effects in the measurement. When comparing the results in Figure 2 to Figure 1, note that the span in Figure 1 is 50 MHz while the span in Figure 2 is 30 MHz.
Analysis of waveforms having various numbers of resource blocks and various positions for those resource blocks helps the developer understand LTE outside of simulations. The trade-offs among output power, the transmission bandwidth, and the ACLR for both adjacent E-UTRA and UTRA remain under investigation. The interaction of these parameters will drive UE design, development, and test.
When looking into TS 36.521-1: Evolved Universal Terrestrial Radio Access (E-UTRA), it is interesting to note the items that still must be determined. The messaging used to control the UEs is not finalized so the actual procedures cannot be completed. For all of the power test cases, the reference measurement channel must
In the waveforms shown in this article, the uplink signal contained physical uplink shared channel (PUSCH) and the reference signal but did not contain a physical uplink control channel (PUCCH). The maximum and minimum output power test cases require the yet undefined fixed power allocation for resource blocks.
Although not yet final, 3GPP TS 36.101 V8.3.0, TR 36.803V1.1.0, and TS 36.521-1 V2.0.0 are the resources LTE UE and component developers use to define and design their products. While conformance test cases are not yet complete, the key metrics are available.
Transmission bandwidth, measured in resource blocks, is a new parameter that comes into play whenever output power is measured. Power measurements take on two forms: measuring the absolute power level produced for a particular set of parameters or measuring the change in power level in response to control messages and transmission bandwidth.
Making power measurements is not difficult and, typically with the correct setup, power measurements are fast, single-button tests on most analyzers. As the specifications evolve, test and measurement vendors are keeping a close eye on the developing test cases to ensure that conformance testing is accurate, fast, and efficient.
About the Author
Lynne Patterson is a business development manager for wireless infrastructure test and measurement products at Anritsu. She previously held positions at the Georgia Tech Research Institute, Nortel, and several small telecommunication firms. Ms. Patterson received her B.E.E. from Georgia Tech in 1987. Anritsu, Americas Sales Region Headquarters, 1155 E. Collins Blvd., Richardson, TX 75081, e-mail: Lynne.Patterson@anritsu.com
FOR MORE INFORMATION
3GPP TS 36.101 V8.3.0,
and TS 36.521-1 V2.0.0