How to effectively use S-parameters in RF design

Figure 3: S-parameter result of the inductor (a) and S-parameter result of the SAW filter (b).

When using touchstone files, design engineers also need to consider the situation where the simulation frequency is occasionally outside the range of the data file. The software can provide a mechanism to extrapolate data or stop the simulation, because there is no valid data in part or all of the frequency range. Two popular extrapolation techniques are: (1) using the closest known data point and constant extrapolation; (2) using the two closest known data points for linear extrapolation. The results obtained by these two extrapolation techniques are approximate. If the data points are not selected properly, there will be a large error, and even the simulation will not converge. Figure 3 shows two examples.

Figure 4: (a) AmpliferP2D power scan results; (b) large signal frequency scan results.

In the example in Figure 3, both data files have data in the 50MHz to 6GHz frequency range. In the S-parameter results of the inductor, linear extrapolation produces an inflection point at 50MHz, which is not consistent with the expected inductor characteristics. When a linear extrapolator is used, the surface acoustic wave filter exhibits a "gain" mutation at about 8.5 GHz, which is not possible with passive components. The user needs to know the useful frequency range in the touchstone file and select the appropriate extrapolator. Another way to solve this problem is to manually add the DC point or high frequency value to the touchstone file. This method can eliminate the risk of incorrect extrapolation.

A very important feature of the S-parameter data file is that the simulator should also have interpolation capabilities within the data range, but the premise may be that very close data samples are required within a specific response range, otherwise the important characteristics of the response may be completely lost ( For example, the resonance of a high-Q circuit). In this case, the design engineer should first make broadband measurements, then focus on the narrowbands of interest to them, make a second measurement with a higher frequency resolution, and finally combine these files manually. These files are ASCII data files and can be edited with any text editor.

Figure 5: Simulation results of the PA and dual multiplexer circuit.

Large signal S parameter

The previous discussion defined the default S-parameter to be linear by definition. For CAD users, this means that the device is either a passive device or is assumed to work within the linear range of the simulator's response. Therefore, for S-parameter analysis, even if measured under the excitation of a large signal, the output will not generate harmonics. So, how effective are these large-signal S-parameters in the design process? When no other models are available, S-parameters provide a black-box circuit description, even for large-signal analysis.

One advantage of using S-parameters in large-signal analysis is that you can obtain a data file containing S-parameter measurements, which are a function of power and frequency (also called large-signal S-parameters). In the ADS emulator, it is completed by P2D files. The P2D file is called by a special emulator component called Amplifier P2D. P2D files are better than large signal touchstone files, because by definition, touchstone files are used for small signal power levels, while P2D files provide S-parameters at different frequencies and powers. The software can be used to easily create measured P2D files. In other words, the software can control the instrument and generate P2D files. P2D files can also be created from simulation results.

The following analysis examples illustrate the effectiveness of the P2D data model. First, we analyze the relationship between the output power of the power amplifier (PA) and the input power at a fixed input frequency, as shown in Figure 4a. P2D files can also be used to simulate the relationship between output power and frequency, as shown in Figure 4b (fixed input power).

Another great use of P2D and touchstone files is system-level analysis. For example, sometimes you need to look at the output after connecting the above PA and the dual multiplexer (represented by a touchstone file). The circuit diagram and analysis results are shown in Figure 5.

These data files can also be used by the modulation source in the circuit envelope simulation. The P2D file is a short description of the performance of the device under test (DUT) under single tone excitation, so it can be considered as a tuning model. This means that when design engineers use Amplifer P2D components in ADS, they only get the S parameters at the frequency defined by the component (remember: there are no harmonics at the output). Therefore, a very wide modulation bandwidth may cause large errors, especially when adjacent channel measurement results are also useful. For example, Figure 6 is the output of a modulated carrier that is fed to a P2D model and a discrete circuit model. It should be noted that within the modulation bandwidth there are two signals that appear to be the same. As the signal moves away from the carrier frequency, the spectral density of the two signals becomes different.

Figure 6: P2D file simulation results using modulated signals.

Another very important benefit of using Touchstone files and P2D files is the ability to understand the design performance earlier. In the early stages of the project, the system engineer may need to estimate the link budget tolerance and understand the PA performance. A typical method is to use spreadsheet-type parameters such as gain, 1dB compression point, IP3, etc. This has always been the mainstream method. But now, with just one file, the actual measured performance can be obtained and applied to the analysis process. A good example of this is that P2D files can provide higher accuracy than spreadsheets. Another method of use may involve the protection of intellectual property rights. P2D files and touchstone files can be generated directly from the emulator. Therefore, during circuit design, IC manufacturers or design teams can also provide design data to the system design team or customers to determine whether the proposed design meets the requirements of the entire system without having to wait until the actual hardware is produced.

The accuracy of the data model file is the key to the success of this method. The measurement-based file has the same accuracy as the measurement result, and should simulate the actual system design conditions (bias, temperature, etc.) as much as possible. Please note that if the design is a time-varying circuit (such as a power control loop), or there is a strong memory effect, they cannot be recorded or included in the P2D model.

Understanding how to use touchstone and P2D files is very useful for RF circuit design. In addition, P2D files can also increase simulation speed. In very complex design simulations, the P2D look-up table makes designs using P2D models several orders of magnitude faster than discrete designs.

Author: Wilfredo Rivas-Torres; EEsof EDA Application Engineer department; Email:; Agilent Technologies