The following sections include information about converting to and working with Touchstone format data, extrapolation, and using data files with noise simulation.
Some network analyzers use Citi format to store their measurement data. The NI AWR Design Environment suite does not directly import Citi data, but has scripts that can convert Citi files to Touchstone format, including multiple-parameter Citi files. See the NI AWR website for example Citi import scripts, or contact NI AWR Support to request these scripts.
There are several common problems you might encounter when using Touchstone data:
Having all of the data for one line. (The N-port matrix on one line of the data file for files bigger than 4 ports.) Remember that there is a maximum of 8 entries per line, so any file with more than 4 ports gets more complex because new lines are required after 8 entries are added. See “Touchstone File Format ” for an example of a 5-port data file with line wrapping. If your data is not line-wrapped this way, you can use the raw data format described in “Raw Data File Format ”, including a technique to convert data to properly formatted Touchstone files.
Improper line wraps in the data file (Touchstone v1.0 format only). There are many situations where this might occur. The solution is to use the raw data format described in “Raw Data File Format ”.
Duplicate frequencies in the data file. Duplicate frequencies produce the following error: "Problem with file format: Error reading line <x>: expecting 5 entries per line for noise data". Upon finding a duplicate frequency, the program thinks it has entered the noise data section of the file.
Having derived data (such as common mode rejection ratio) calculated from the raw network data. Some Vector Network Analyzers export Touchstone data files with derived data appended to each line. Contact NI AWR Support for scripts that help clean up this data.
You can generate one N-port Touchstone file from many M-port Touchstone files; where N > M (typically, M=2 and N=3 or 4). Chooseto run a Visual Basic script that performs this automatically.
When Touchstone, raw data, and MDIF files are used as subcircuits in a larger circuit simulation, problems arise when the simulation occurs at frequencies outside of the range of the data files frequency range. In this case, the software must extrapolate a response for the data file from the existing data.
Extrapolating to DC can cause common problems such as current flowing through blocking capacitors or transistors not biasing up properly. One method to check for problems is to turn on both current and voltage DC annotations so you can see these simulation values on the schematic. After identifying a problem, there are several things you can do to fix it:
Change the interpolation/extrapolation options. You can access and change these options for the entire project by choosing Interpolation/Passivity tab. You can also access and change these options for a single data file by selecting the data file under the Data Files node in the Project Browser, right-clicking and choosing Options, and then clicking the Interpolation/Passivity tab. You can try changing the interpolation method or the coordinate system. See “Options - Data File Dialog Box: Interpolation/Passivity Tab ” for more information on the options in this dialog box.and clicking the
Edit the data file directly and add the proper entries for DC. To do so you must know the proper entries in your files. This is more difficult for MDIF files because you need to edit each block of data.
Place your data file in a schematic and then use large inductors and capacitors to define the proper DC paths (use a capacitor to block DC and inductors between ports where DC current should flow). You can now use the schematic as a subcircuit, or export it as an output file in Touchstone format to generate a new Touchstone format with proper entries at DC. Be careful using the schematic with large capacitors and inductors when using transient simulations, as these components introduce very large time constants resulting in the need for many cycles to get to steady state.
Another problem is the behavior at the harmonics of the fundamental. For example, this can occur if you have a 2 to 3 GHz amplifier and data for some capacitors from 1 to 4 GHz, and you want to run harmonic balance analysis to get the compression characteristics of the amplifier. You are running 5 harmonics in the harmonic balance simulation, so the simulation needs to know the behavior of those caps at 15 GHz (3 GHz x 5 harmonics). 15 GHz is significantly outside the range of the data file, so the extrapolated data is most likely not accurate.
When you use Touchstone, raw data, and MDIF files with noise simulation, the program first determines if the data file is passive. If passive, the noise can be computed from the network parameters. If not passive, the data file expects to find noise parameters in the data file. Sometimes, data for passive structures can be slightly active (due to calibration errors or EM simulator numerical problems).
If this problem occurs, you can force the data file to be treated as passive for noise simulation. To do so, right-click the file in the Project Browser under the Data Files node, choose Options, click the Interpolation/Passivity tab, and then select the Consider Passive for Noise Simulation check box. For more information on the options in this dialog box, see “Options - Data File Dialog Box: Interpolation/Passivity Tab ”.
When you use S-parameters in a schematic, they are inserted as a subcircuit. You have three options for grounding types: normal, explicit ground node, and balanced ports. You can find these options on the Element Options: SUBCKT - Properties dialog box Ground tab (right-click the S-parameter subcircuit and choose .) Explicit ground node exposes the ground node so it is accessible in the schematic. Balanced ports adds a local “ground” port to each port of the S-parameter file.
You can view an exposed ground node as a local ground for the S-parameter file. It is important to understand that the same ground is used for all ports in the S-parameter file. This implies that, physically, the structure is electrically small, or has a very good (perfect) internal grounding system connecting the ports. Normally, exposing the ground node is used for transistor data, where a common ground node in the measurement is being exposed.
Balanced ports extend the exposed ground node concept by creating a local ground node for each port. Conceptually this is similar to attaching an ideal 1:1 transformer to each port, and using the exterior coil to create a local ground reference. It is possible to misuse this concept and obtain physically meaningless results.
To learn more about the different grounding types choose File > Open Example and search for "ground_node". See the Design Notes for the example for more information about different grounding types.