Measuring the Impact of RF Impairments on an LTE System
Communication systems performance places a high demand on RF components such as low noise amplifiers (LNAs), power amplifiers (PAs), or mixers. The constraints dictated by these RF components directly affect baseband design. Two famous examples include the choice of GMSK modulation for GSM or, more recently, SC-FDMA for the uplink of LTE. Therefore, it is desirable to assess the impact of such RF components on a complete system early in the design process. This requires setting up a simulation with both accurate baseband modeling and realistic RF effects.
LTE System Toolbox™ offers easy access to detailed baseband modeling and standard measurements such as EVM and ACLR. It also integrates with SimRF™ (Figure 1), which provides fast simulation of RF systems and insight into intermodulation distortion, image rejection, and phase noise, among other effects.
This example shows how to assess the impact on EVM of an RF front end. The imperfections we want to take into account include:
- LNA: Nonlinearity (IP3), frequency-dependent response (S-parameters), noise
- Mixer/demodulator: I/Q imbalance, phase noise, input isolation, nonlinearity (IP2), LO leakage
- VGAs: Nonlinearity (IP3, 1dB compression point)
RF Receiver Set Up
Figure 2 shows the model of the RF receiver. The input signal goes through an LNA, and then it travels through a direct-conversion demodulator and down to baseband. The I/Q components are amplified with a VGA and combined into a complex output signal for further processing and analysis
Figure 3 compares the measured RMS EVM for an ideal RF front end with a realistic RF front end with impairments. This figure clearly shows the influence of impairments in the RF front end. By leveraging such simulations early in the design, we can fine-tune RF component parameters to verify that an EVM target is satisfied.
Details of the Transmit/Receive Chain
We use LTE System Toolbox in MATLAB® to generate a subframe of standard-compliant LTE data. The output variable,
tx, represents the baseband output of the OFDM modulator after cyclic prefix insertion. As shown below, LTE System Toolbox includes functions such as
lteRMCDLTool that generate standard-compliant test signals .
% Create eNodeB transmission with fixed PDSCH data rmc = lteRMCDL(‘R.6’); data = randi([0 1], sum(rmc.PDSCH.TrBlkSizes),1); [tx, ~, info] = lteRMCDLTool(rmc, data);
sim command, the signal generated with LTE System Toolbox is imported into Simulink®.
% SimRF testbench sim(model, time(end)); xInitial = xFinal;
As illustrated in Figure 4, the imported signal (orange) undergoes filtering and free-space, and then enters the receiver, where white Gaussian noise is added. The received signal then enters the RF front end shown in blue in Figure 4 and discussed in Figure 2. This is where the RF simulation engine from SimRF comes into play.
As an example of RF impairment, the mixer includes phase noise. Figure 5 shows the profile of the phase noise around the LO used in the mixer in dB versus a log of the frequency offset.
Other effects include nonlinearities, which result in a DC offset at the output of the RF front end due to harmonics mixed back to DC through the direct-conversion process. To compensate for this DC offset, the signal is passed to an ADC and a DC offset compensation algorithm (not shown).
SimRF uses circuit envelope technology to achieve fast simulation of the RF signals without loss of accuracy. Circuit envelope technology enables easy integration of the RF model with digital baseband algorithms, and it provides a natural framework where RF effects and imperfections can be easily described.
Figure 6 shows the spectrum of the signal at the following stages of the chain:
- Output of the transmitter (yellow)
- Input of the receiver (blue), after attenuation and noise addition
- Output of the RF front-end (red), with an increased power but a DC offset
- Output of the DC offset compensator (green)
The DC offset compensation algorithm successfully removed the DC offset, represented by the red peak at the center frequency.
Finally, we measure EVM signals in MATLAB with LTE System Toolbox providing the EVM as per 3GPP TS 36.101 .
% Compute EVM measurements evmmeas = hPDSCHEVM(rmc, struct(‘PilotAverage’,’TestEVM’),rx); evmpeak(n) = evmmeas.Peak; evmrms(n) = evmmeas.RMS;
LTE System Toolbox integrates with SimRF to simulate RF components in the context of an LTE system. LTE System Toolbox provides accurate LTE baseband modeling while SimRF, with its RF components and core simulator, lets you simulate the RF section of the design transparently.
By assessing the performance and influence of RF components early in the design process with LTE System Toolbox and SimRF, we can reduce risk and issues that might appear at a later stage.
 Generating LTE Waveforms
 3GPP TS 36.101 – Base Station Radio Transmission and Reception
 PDSCH Error Vector Magnitude (EVM) Measurement
Published 2015 - 80710v00