Microwave Journal

Real-world Stimulus-response Testing for LTE Transmitter RF Components

March 4, 2010

The 3GPP Long Term Evolution (LTE) specifications present some new challenges for manufacturers of components and equipment for LTE systems. The standards include multiple channel bandwidths, different transmission schemes for the downlink and uplink, both frequency and time domain duplexing (FDD and TDD) transmission modes, and use of multiple-input, multiple-output (MIMO) techniques. LTE will co-exist with current 2G and 3G cellular systems, so potential interference is an important issue.

This article describes some of the RF stimulus-response measurements required to ensure transmitter RF components perform as required to meet system specifications. Complete characterization will require additional measurements such as gain compression, intermodulation distortion, noise figure and power consumption, but those traditional tests are not covered here.

The physical layer for LTE utilizes orthogonal frequency division multiple access (OFDMA) for the downlink, and single-carrier frequency division multiple access (SC-FDMA) for the uplink. Six different channel bandwidths from 1.4 to 20 MHz may be used; with a fixed subcarrier spacing of 15 kHz, or 7.5 kHz for multimedia broadcast multicast service (MBMS).

A simplified high-level block diagram of the transmitter and receiver in an LTE device is shown in Figure 1 below. The lower-level components are determined by the type of architecture used in the design, and may include additional components such as filters and variable-gain amplifiers. LTE uses MIMO technology, so basestations will include multiple transmit chains, each with its own data source.

Figure 1. Simplified block diagram of transmitter and receiver in an LTE device

RFICs in user equipment (UE) may include different portions of this block diagram, with some containing the transmitter or receiver only and others combining transmitter and receiver in one component. Traditionally RFICs have used analog baseband IQ inputs and outputs. However, digital IQ interfaces are becoming more common. The Mobile Industry Processor Interface (MIPI) Alliance developed the DigRF standard, which describes the high-speed digital baseband to RFIC interface for new-generation radios, and includes embedded control protocols for the RFIC.

Amplifier testing will require stimulus-response measurements in which a signal generator is used to create an RF LTE signal that is input to the device under test, and the output signal from the DUT is measured using a signal analyzer. Testing the transmitter chain will require using a digital or analog baseband signal from the signal generator to drive the I/Q inputs and a signal analyzer to measure the RF output.

Downlink and Uplink Signal Generation

Agilent’s Signal Studio software is a family of PC-based applications that provide waveform files for download and playback in a variety of Agilent instruments, including the N5182A MXG and E4438C ESG signal generators. The MXG provides the industry’s best adjacent channel leakage ratio (ACLR) performance for power amplifier testing. The ESG can also be used with a digital signal interface module, which converts the digital IQ data to an appropriate format for the device under test. Signal Studio can also be used with Agilent’s 16800/16900 series logic analyzers as part of the Agilent RDX solution for DigRF, to create digital IQ signals.

The N7624B and N7625B Signal Studio applications for 3GPP LTE FDD and TDD respectively provide preconfigured signals for many of the test models that are used in the base station and user equipment conformance tests, and these signals can be used for component test. The software also provides highly flexible custom signal configuration capabilities, including the creation of multi-carrier signals.

Downlink and Uplink Signal Analysis

Signal analyzers such as the Agilent X-Series (PXA, MXA, EXA, and CXA) and PSA provide signal analysis with detailed modulation analysis, and can be used to make RF power measurements on LTE signals. The MXA and PXA can also measure analog IQ signals. Optional measurement applications provide convenient one-button power measurements and modulation analysis of LTE downlink and uplink signals.

The 89600 vector signal analysis (VSA) software is a PC-based application with options for LTE FDD and TDD that provide flexible and comprehensive modulation and MIMO signaling analysis. The same software can be used for testing at RF, analog IQ, and digital IQ interfaces in the block diagram, and works with a variety of signal analyzers, oscilloscopes, logic analyzers, and the N5344A DigRF analyzer module.

Output Power Measurements

As with other wireless systems, a requirement of LTE is the fundamental measurement of both broad- and narrow-band power. Channel power indicates the mean power within the appropriate integrated channel bandwidth. Occupied bandwidth measures the bandwidth of the LTE signal that contains 99 percent of the channel power. Due to the nature of the downlink and uplink signal characteristics, verifying LTE amplifier performance will also involve power measurements all the way down to the Resource Element (RE) level, which is one OFDMA or SC-FDMA symbol lasting 66.7 μs on one subcarrier. For such measurements, a vector signal analyzer (VSA) is essential and power measurements associated with specific portions of the signal require the VSA digital demodulation capabilities.

Note that for TDD signals, a time-gated measurement is necessary to ensure that data is captured during the period when the burst power is on. The standard does not specify the time period for the measurement, but typically it is performed during the period in which burst is completely on, not including the ramp up or ramp down time. Figure 2 below shows the channel power measurement using an MXA for a TDD signal with the gate view turned on.

Figure 2. Channel power measurement using an MXA for a TDD signal with gate view turned on

Frequency Error, Error Vector Magnitude, and IQ Parameters

Frequency error is the offset of the signal’s center frequency from the desired center frequency. Error vector magnitude (EVM) is a key test of modulation quality for a transmitter and indicates the amount of distortion in the signal. IQ error measurements can reveal some of the possible sources for the distortion. The UE transmitter conformance tests require the measurement of the IQ offset, which can be an indicator of the carrier feedthrough or a DC offset in the baseband signal. The frequency error, EVM, and IQ parameters are reported in the Error Summary trace in the X-Series applications or VSA software, as shown in Figure 3 below.

Figure 3. Frequency error, EVM, and IQ parameters are reported in the Error Summary trace in the X-Series applications or VSA software

Complementary Cumulative Distribution Function (CCDF)

The X-Series LTE applications and 89600 VSA both offer CCDF measurements. For TDD signals, the measurement should be limited to a time period in which the RF burst is turned on, since measuring the power during the time when the burst is off will result in an incorrect average power value and cause errors in the CCDF measurement. A typical TDD measurement is shown in Figure 4 below.

Figure 4. Typical CCDF measurement for a TDD signal

Adjacent Channel Leakage power Ratio (ACLR) or Adjacent Channel Power Ratio (ACPR)

ACLR is a key characteristic of power amplifiers, since power amplifiers are a key contributor of distortion in the transmit chain. LTE systems have to co-exist with legacy 2 and 3G systems in the same frequency bands, so the LTE RF conformance tests include cases in which the adjacent channels can be either an LTE signal or a legacy signal. All LTE channels are measured using a square filter (essentially the same as no filter), while W-CDMA channels are measured using a root-raised-cosine filter with a rolloff factor of 0.22 and a bandwidth equal to the chip rate (e.g., 3.84 MHz). For both base station (downlink) and UE (uplink) components, the tests are performed using test signals specified in the standards. For details, see the latest information at www.3gpp.org.

Setup of the ACLR measurement includes configuration of the carriers, offset frequencies, integration bandwidths, resolution and video bandwidths, measurement filters, and limit test. To minimize clipping and prevent overload, the X-Series signal analyzer automatically selects an attenuation value based on the current measured signal level at the input mixer. Using noise correction can provide a substantial improvement in the ACLR measurement, particularly when the distortion is close to the noise floor of the analyzer. With noise correction, the analyzer measures its internal noise floor, then subtracts that noise floor from the measurement data.

Spectrum Emission Mask

The spectrum emission mask (SEM) measurement covers unwanted emissions in the operating band from the 3GPP specifications. The configuration for the SEM measurement is similar to that for ACLR: as with ACLR, the X-Series LTE application provides predefined limit masks for the SEM measurement, as well as allowing manual configuration of the parameters. Figure 5 below shows an example SEM measurement with limit lines set for a complete transmitter. Typically, tighter limits would be used for component testing to ensure that the system performance meets the specification.

Figure 5. SEM measurement with limit lines set for a complete transmitter

Spectrum Flatness (Uplink)

Spectrum flatness is a useful test in addition to EVM, because EVM measurements are typically performed after the analyzer has estimated and removed the amplitude and phase errors over frequency through the equalization process, as defined in the standard. It measures the relative power variation across the subcarriers of the allocated resource block over one slot in the time domain, for 20 measurements in a frame.

The spectrum flatness measurement can be made by examining the equalizer channel frequency response results that come from the equalization process. This data is available in the 89600 VSA software and the modulation analysis measurement in the X-Series application. Both applications also provide a display that shows the equalizer channel frequency response per slot, as shown in Figure 6 below.

Figure 6. Equalizer channel frequency response per slot

In-Band Emissions for Non-allocated RB (Uplink)

A UE occupies only a portion of the channel bandwidth, since the uplink channel is shared with other UEs. The in-band emissions for non-allocated resource block (RB) test is a measure of the amount of power that the UE may transmit into non-allocated RBs. This test applies only to UE transmitters and components.

Figure 7 below shows a measurement of a 5 MHz uplink signal. The signal has 8 allocated RBs with the stipulated QPSK modulation. There are 25 available RBs (numbered RB #0 to RB #24) in the 5 MHz bandwidth.

Figure 7. Measurement of in-band emissions for non-allocated RB for a 5 MHz uplink signal

Spurious Emissions

The conformance tests measure spurious out-of-band emissions from 9 kHz to 12.75 GHz (excluding the frequency range that is covered by the SEM test) to ensure protection of other radio systems that may be operating in the same geographical area. They can be caused by harmonics, intermodulation products, and frequency conversion products in the power amplifier or other transmit chain components. The conformance tests contain test limits for specific situations such as co-existence with other systems such as GSM900, DCS1800, PCS1900, PHS, and public safety radio systems, which are dependent on regional regulations and the operating band of the LTE equipment.

Summary

This article described testing considerations for LTE components and provided information on creating LTE stimulus test signals for components, and analyzing LTE signals from components and transmitters. Additional information about Agilent’s LTE solutions is available online at www.agilent.com/find/lte.

Further reading

The following Agilent publications describe other key measurements on amplifiers and transmitters:

Stimulus-Response Testing for LTE Components –Application Note p/n 5990-5149EN

Improved Methods for Measuring Distortion in Broadband Devices – Application Note p/n 5989-9880EN

Fundamentals of RF and Microwave Noise Figure – Application Note p/n 5952-8255E

Amplifier CW and Swept IMD Measurements – Application Note p/n 5988-9474EN