High frequency technology

Telecommunication project

There are only limited numbers of frequency bands for mobile communication services. The required channel bandwidth for the data transmission and the data rate are key factors which characterize the effectiveness of a transmission system. The goal is to maximize the data rate within a frequency band. There are several methods which allow a higher data rate for the same channel bandwidth and thus supporting a more efficient data transmission.

A critical factor of a mobile phone is the transmission characteristic of the RF-transmitting amplifier regarding the transmitted EDGE or UMTS RF-signals. In contrast to the GMSK modulation the 3π/8-8PSK modulation and the QPSK modulation are both phase and amplitude modulated signals. The result is a spectral broadened output signal after the non-linear power amplifier or a significant distortion of the transmitted signal.

This leads to an increase in bit error rate (BER) in the uplink with the same received field strength. In order to minimize these distortions, the use of a linear power amplifier is required. However the efficiency of linear amplifiers is about 35%, which is clearly below the potential of non-linear power amplifiers. They can operate at an efficiency of over 60%. The high power consumption of the system due to the low efficiency of the components is contrary to the desire to achieve a long operation period of the mobile phone. In order to be able to use non-linear power amplifiers with high efficiencies, suitable linearization procedures must be used to recover the output signal.

Project: Linearization of a nonlinear power amplifier for mobile communication applications



The following procedure for linearization of a nonlinear power amplifier can be used:

Transceiver Design

  • Feed forward linearization technique
  • Feedback linearization technique
  • Cartesian loop linearization technique
  • Polar loop linearization technique
  • Linear amplification using nonlinear components
  • Digital pre-distortion technique
  • Analog pre-distortion technique

Depending on the level of abstraction simulation tools as Matlab/Simulink, ADS and CST Microwave Studio can be used.


  • Matlab/Simulink Simulation
  • ADS (Advance Design System) Simulation
  • CST (Microwave Studio) Simulation
  • CAD-Design

Hardware Design

  • Transmitter Hardware Design for 4G Mobile Communication Systems
  • Receiver Hardware Design for 4G Mobile Communication Systems
  • Antenna Design for 4G Mobile Communication Systems

Figure 1: Constellation diagrams of the 3π/8-8PSK modulator before and after the approximated Gaussian filter


Polar-Loop method

Figure 2 shows the basic structure of the polar-loop method. The key element is the non-linear power amplifier (PA), which is controlled by using control loops. Thus the output signal of the PA is an amplified copy of the input signal. A PLL (Phase Locked Loop) control loop is used to correct the phase angle between a reference signal and the output signal. Additionally an ALL-control loop (Amplitude Locked Loop) for correcting the amplitude of the output signal has to be used. Because there are two control loops, mutual adjustment of the control loops are extremely important for the quality of the output signal.

Figure 2: Schematic of a Polar-Loop method


Figure 3: FFT-Power spectrum of a modulated EDGE-Signal before adjusting the bandwidth


The FFT power spectrum of the modulated EDGE signal is shown in Figure 3. The original EDGE generator signal is shown in red. This is a simulation before the adjustment of the bandwidth of the ALL-control loop (Amplitude Locked Loop) and the PLL-control loop.


Figure 4: FFT-Power spectrum of a modulated EDGE-Signal after adjusting the bandwidth

Figure 4 shows the simulation results after adjusting the bandwidth of the ALL-control loop and the bandwidth of the PLL-control loop. Compared to figure 3, a significant improvement in the reconstruction of the original EDGE generator signal can be achieved.

Basic structure of other linearization techniques

In order to give a complete overview, the following principal structures of other linearization methods will be shown.

The linearization procedure of the Cartesian-loop method is done in the baseband frequency area. In Figure 5 the operating principle of the Cartesian-loop process is shown. As you can see a part of the output signal is mixed down to the baseband frequency. Then the difference between the mixed-down output signal and the undistorted input signal can be used to generate an error signal. This error signal controls the IQ modulator and the power output stage.

Figure 5: Linearization by baseband feedback (Cartesian-Loop)


Figure 6 shows the principle of feedforward method. The principle is based on a cancellation of non-linear effects by using an additional amplifier (Error Amplifier).

Figure 6: Basic structure of a feedforward amplifier

Another linearization method is shown in Figure 7. It is the LINC method (Linear Amplification using Nonlinear Component) without any feedback.

Figure 7: Linearization by LINC (Linear Amplification using Nonlinear Component)

Circuit implementation of the Polar-Loop method

Figure 8 shows the circuit implementation of the polar-loop method.


Figure 8: Realization of the Polar-Loop Tx-circuit for 1700 MHz


The results in Figure 9 and 10 confirm the simulation results before and after adjusting the bandwidth of the control loops. By adjusting the bandwidth a significantly improved reproduction of the output signal can be obtained.


Figure 9: Measurement of the output signal


a.) Signal generator SMIQ 06B
b.) Spectrum of the GSM-EDGE-Signal after the
Polar-Loop Tx-circuit (Bandwidth of the control loops has not been adjusted)


Bild 10: Figure 10: Measurement of the output signal


a.) Signal generator SMIQ 06B
b.) Spectrum of the GSM-EDGE-Signal after the
Polar-Loop Tx-circuit (Bandwidth of the control loops has been adjusted)



Antenna design: Splinehorn 75-100 GHz

The task of an antenna in the transmission case is the transformation of a guided electromagnetic wave into free space. During the receiving case the antenna is converting a plane wave into a guided wave.

For quasi-optical applications antennas with a homogeneous Gaussian beam characteristic are preferably used. The goal is a low-loss transformation of a waveguide mode into a Gaussian beam fundamental mode. In order to qualify the antenna the correlation/similarity of the electric field of the antenna and the electric field of the Gaussian beam fundamental mode is used. In the literature you can find different antenna shapes with different Gaussian coupling coefficients and bandwidth requirements. The maximum coupling coefficient and the corresponding optimal beam parameters can be calculated by solving the coupling integral. Thus the coupling of the field at the antenna aperture to the field of a Gaussian beam fundamental mode is determined.

The most important antenna shapes with the corresponding parameters are listed in the table below.



Design of an optimized smooth-walled spline-profile Horn

In order to realize a quasi-optical diplexer it is necessary to design an antenna with a homogeneous beam profile and a high coupling coefficient. It is known that a smooth-walled spline-profile Horn meets these requirements.


Figure: CST Microwave Studio Simulation of a Splinehorn at 75-100 GHz


The MWS model of the optimized spline horn is shown in the figure. The spline horn consists of 71 different circular waveguide segments.

Figure: Simulation result of the directivity of the optimized Splinehorn at 87.5 GHz


Figure: Simulation result of the gain of the Splinehorn at 87.5 GHz


Figure: Measured reflection coefficient of the Splinehorn at 75-100 GHz


It can be noted that there is a very good match between the beam profiles of the simulation and the measurement.

Figure: Measured Spot of the Splinehorn averaged over all frequency points 75-100 GHz


Figure: Realization of the Splinehorn at 75-100 GHz