Go to www.awrcorp.com
Back to search page Click to download printable version of this guide.

Chapter 1. Simulation Basics

Visual System SimulatorTM (VSS) is the system level design component of the NI AWR Design Environment suite. With VSS software you can analyze a complete communications system, from data encoding, through transmission, reception and data decoding.

1.1. Additional Documentation

The NI AWR Design Environment suite includes the following additional documentation:

  • What's New in NI AWR Design Environment 14? presents the new features, user interface, elements, system blocks, and measurements for this release.

  • The Installation Guide (available on your Program Disk as install.pdf or downloadable from the Knowledge Base at NI AWR Knowledge Base) describes how to install the software and configure it for locked or floating licensing options. It also provides licensing configuration troubleshooting tips.

  • The Getting Started Guide familiarizes you with the NI AWR Design Environment suite through Microwave Office, VSS, Analog Office, AnalystTM, and Monolithic Microwave Integrated Circuit (MMIC) examples.

    Microwave Office example projects show how to design and analyze simple linear, nonlinear, and EM circuits, and how to create layouts. Visual System Simulator examples show how to design systems and perform simulations using predefined or customized transmitters and receivers. Analog Office examples show how to design circuits composed of schematics and electromagnetic (EM) structures from an extensive electrical model set, and then generate physical layouts of the designs. Analyst examples show how to create and simulate 3D EM structures from Microwave Office, and MMIC examples show MMIC features and designs.

    You can perform simulations using a number of simulators, and then display the output in a wide variety of graphical forms based on your analysis needs. You can also tune or optimize the designs, and your changes are automatically and immediately reflected in the layout.

  • The User Guide provides an overview of the NI AWR Design Environment suite including chapters on the NI AWR Design Environment user interface; using schematics/system diagrams; data files; netlists; graphs, measurements, and output files; and variables and equations in projects.

  • The Simulation and Analysis Guide discusses simulation basics such as swept parameter analysis, tuning/optimizing/yield, and simulation filters; and provides simulation details for DC, linear, AC, harmonic balance, transient, and EM simulation/extraction theory and methods.

  • The Dialog Box Reference provides a comprehensive reference of all NI AWR Design Environment dialog boxes with dialog box graphics, overviews, option details, and information on how to navigate to each dialog box.

  • The Microwave Office Layout Guide contains information on creating and viewing layouts for schematics and EM structures, including use of the Layout Manager, Layout Process File, artwork cell creation/editing/properties, Design Rule Checking, and other topics.

  • The VSS System Block Catalog provides complete reference information on all of the system blocks that you use to build systems.

  • The Microwave Office Element Catalog provides complete reference information on all of the electrical elements that you use to build schematics.

  • The Microwave Office Measurement Catalog provides complete reference information on the "measurements" (computed data such as gain, noise, power, or voltage) that you can choose as output for your simulations.

  • The VSS Measurement Catalog provides complete reference information on the measurements you can choose as output for your simulations.

  • The API Scripting Guide explains the basic concepts of NI AWR Design Environment scripting and provides coding examples. It also provides information on the most useful objects, properties, and methods for creating scripts in the NI AWR Script Development Environment (NI AWR SDE). In addition, this guide contains the NI AWR Design Environment Component API list.

  • The Quick Reference document lists keyboard shortcuts, mouse operations, and tips and tricks to optimize your use of the NI AWR Design Environment suite. This document is available within the NI AWR Design Environment suite by choosing Help > Quick Reference.

  • NI AWR Design Environment Known Issues lists the known issues for this release. This document is available on your program disk as KnownIssues.htm.

1.2. Typographical Conventions

This guide uses the following typographical conventions.

Item Convention
Anything that you select (or click on) in the NI AWR Design Environment suite, such as menus, submenus, menu items, dialog box options, button names, and tab names

Shown in a bold alternate type. Nested menu selections are shown with a ">" to indicate that you select the first menu item and then select the second menu item from the menu:

Choose File > New Project.

Text that you enter using the keyboard

Shown in bold type within quotes:

Enter "my_project" in Project Name.

Keys or key combinations that you press

Shown in a bold alternate type with initial capitals. Key combinations using a "+" indicate that you press and hold the first key while pressing the second key:

Press Alt+F1.

Filenames and directory paths

Shown in italics:

See the DEFAULTS.LPF file.

Contents of a file, fields within a file, command names, command switches/arguments, or output from a command at the command prompt

Shown in an alternate type:

Define this parameter in the $DEFAULT_VALUES field.

1.3. Basic Concepts

VSS is a system level simulator. A system level simulation models designs at the component level, such as amplifiers, mixers, or encoders.

VSS software models the effects of a system upon one or more signals which can be digital data, modulated analog signals, continuous wave (CW) tones, or other types. VSS designs start with a system diagram and are composed of interconnected blocks representing individual components in a design. The connections between the blocks describe the flow of data through the system.

Blocks typically have one or more ports, which serve as the connection points to other blocks. The connection point between ports is called a node.Due to historical reasons, ports are also often referred to as nodes.

There are two types of ports: input ports and output ports. Input ports are the entry point of data into a block, and receive data from an output port of another block. When a simulation runs, data flows from the output port of one block to one or more input ports of other blocks connected to the output port. When a block receives data it applies its behavior to the data and generates any appropriate output.

At least one block must be a source block, which is a block that generates a signal without requiring input from another block. Examples of source blocks are the modulated sources such as the QAM Modulated Signal block QAM_SRC, or the RF Tone Source block TONE.

To perform most measurements there must also be at least one 'meter' block with an input port connected to the system. Annotations, which are applied to an entire system diagram, are the exception.

The meter blocks are located in the Meters category of the Element Browser. The most commonly used meter is the Test Point block TP. TP lets you monitor any signal.

VSS software supports three types of simulations: Time Domain simulations, RF Budget Analysis simulations, and RF Inspector simulations.

1.3.1. RF Budget Analysis and RF Inspector Simulations

The RF Budget Analysis measurements, in the System > RF Budget Analysis measurement category, invoke the RF Budget Analysis (RFB) simulator. These measurements include cascaded noise figure, gain, and IP3 measurements.

The RF Inspector measurements, in the System > RF Inspector measurement category, invoke the RF Inspector (RFI) simulator.

Both RFB and RFI simulations are generally restricted to the RF components of a system design. Both simulators operate in the frequency domain and assume steady state behavior. Whereas Time Domain simulations are started and then run until stopped, RFB and RFI simulations occur in one step. This allows efficient use of the optimization and yield analysis features of the NI AWR Design Environment software.

RFB and RFI simulations are run similar to Microwave Office simulations. They are run when the Analyze command is chosen, and are only run if the system diagram has been modified. In addition, they are also run when the Run/Stop System Simulators command is chosen if the system diagram has been modified.

See the RF Modeling in VSS chapter for a more detailed description of RF Budget Analysis and RF Inspector simulations.

1.3.2. Time Domain Simulations

NOTE: Time Domain simulations require a VSS-250 or greater license. The Run/Stop System Simulators command, which is required to run Time Domain simulations, is not available without a VSS-250 license.

In a Time Domain simulation, data is represented as a stream of samples, with each sample representing the value of the signal at a specific point in time. The time difference between each sequential sample in a signal is called the time step. The inverse of the time step is the sampling frequency, commonly denoted as fs.

In VSS software the time step is fixed during the main simulation. However, different signals may have different time steps. For many blocks the sampling frequencies of the output signals are the same as the sampling frequencies of the input signals. Other blocks may modify the sampling frequency. For example, a digital-to-analog converter typically generates several analog output samples for each digital value input. If, for example, it generates 8 output samples for each digital value input, the output time step would have to be 1/8 of the input time step.

The time step and sampling frequency can be viewed at each output port in a system diagram using the SMPFRQ and TSTEP annotations.

When a simulation runs, the source blocks generate a sequence of data samples representing the time varying data. This sequence of samples is passed on to any connected blocks. These connected blocks in turn 'read' each incoming sample, processes each sample, and if necessary generates new samples for the blocks connected to their output ports.

Time Domain simulations typically require many thousands of samples be generated and processed before any significant measurement can be made. Because it takes real time to process all these samples, Time Domain simulations are normally started and then run in the background. While the simulation is running, the graphs update based on the samples available to the individual measurements.

VSS Time Domain simulations are started by choosing Simulate > Run System Simulators. A simulation runs until you pause or stop it, or it completes processing the required number of samples. Each time you choose Run System Simulators new simulations are started and you lose all measurement results from previous simulations.

1.3.3. Sweeping Simulations

It is often desirable to obtain simulation results for several values of a design parameter. For example, bit error rate (BER) measurements are typically made over several signal to noise ratio values, or adjacent channel power is measured for several signal power levels.

When one or more design parameters are modified during a simulation, the simulation is a swept simulation. Each sweep consists of running the simulation for a specific set of design parameter values. The design parameters are called the swept variables.

All VSS simulators support sweeping. Multiple swept variables are supported for performing multi-dimensional sweeps.

Each sweep can be viewed as a separate simulation. When a new sweep begins, most blocks reset their state. The exceptions are blocks that need to track inter-sweep information such as the BER block. For Time Domain simulations the sweep runs until a control block determines that the simulation has run long enough for that sweep.

There are several elements to a simulation sweep:

  • Design Parameters - Defining the design parameter(s) to sweep

  • Controlling Sweeps - Defining the criteria for starting a new sweep (Time Domain simulations)

  • Swept Measurements - Viewing the results

Design Parameters

Design parameters are the values that are varied from sweep to sweep. Having more than one design parameter results in a multi-dimensional sweep.

VSS software supports two types of design parameters. The first is an equation variable that is assigned to one or more block parameters. The second is a built-in CW power level generated by the Vector Network Analyzer (VNA).

In order for an equation variable to be a design parameter, the set of values to be assigned for each sweep must be defined. This is normally done using one of the following blocks:

  • BER, BER_EXT, FER_EXT, SER and SER_EXT - These are the BER meters found under Meters > BER.

  • SWPVAR - The swept variable control block, found under Simulation Control.

  • VSA - The vector signal analyzer block, found under Meters > Network Analyzers.

These block have a VARNAME and a VALUES parameter. The VARNAME parameter is set to the name of the variable to be swept, entered in double quotes. The VALUES parameter is set to a vector containing the value for each sweep.

The variable to be swept must be in the system diagram. It must also be assigned a constant as its value and not an equation or other variable.

The following illustrates using SWPVAR to define FREQ_Hz as a design parameter with the sweep values of 10e9, 10.1e9, and 10.2e9:

Swept Variable Block

Figure 1.1. Swept Variable Block


Note that the value for the VARNAME parameter, "FREQ_Hz", is enclosed in double-quotation marks as it is referring to the name of the variable, not the values of the variable.

While the equation variable must be assigned a constant value, the VALUES parameter may contain an equation, particularly equation functions. One of the more useful functions is stepped(). This function automatically creates a vector starting from a given value and containing equally spaced values up to a given end value. For example:

stepped(5,10,1)

is equivalent to the vector:

{5,6,7,8,9,10}

VNA can also be used to define an equation variable as a design parameter. The VARNAME parameter is used as above. The values for each sweep, however, are automatically set to the power levels generated by the block. For example, if PSTART is 5, PSTOP is 10, and PSTEP is 1, and the variable "PWR_dBm" were assigned to the VARNAME parameter, PWR_dBm would be a design parameter with the following values:

{5,6,7,8,9,10}

Although you can have more than one VNA in a system diagram, only the VNA block that has the largest number of power steps defines a design parameter. This VNA is the master VNA. Any other VNA block in the system diagram still generates a swept power signal. However once those blocks reach their PSTOP power level they stop participating in the remaining sweeps of the master VNA.

Once a design parameter has been defined, it can be used just like any other equation variable. In order for the design parameter to affect the simulation, however, at least one block's parameter must refer to that design parameter, either directly or indirectly through other equations.

NOTE: Previous versions of VSS software supported limited sweeping through the use of the sweep() equation function. This functionality is still supported. However, it is strongly recommended that the new sweep mechanism be used in place of the sweep() equation function. The new mechanism is required in order to use the generic swept measurement capabilities.

The most common use of the sweep() function was to define the swept Es/N0, Eb/N0 or SNR values for a BER simulation. As an example, suppose SNR were being swept. The following equation may have been defined:

SNR = sweep(stepped(0,4,1))

The variable might then have been assigned to the PWR parameter of the additive white Gaussian noise block AWGN.

To convert this to the current sweep mechanism, you do the following:

  • Cut 'sweep(stepped(0,4,1))' from the equation and replace it with "0":

SNR = 0
  • Paste 'sweep(stepped(0,4,1))' as the value of the VALUE parameter of the BER block and remove the sweep() function:

stepped(0,4,1)
  • Set the VARNAME parameter of the BER block to "SNR", including the double quotation marks.

Controlling Sweeps

When performing a swept Time Domain simulation, the simulator needs to be told when to start the next sweep. This is done through a sweep control block. Frequency analysis simulations do not need a sweep control block.

The same blocks that define design parameters are also used to control sweeps: the BER meters, SWPVAR, VSA, and VNA.

Only one block should control sweeping in any system diagram. The block should be located as far downstream from the sources as possible. This is to give all other blocks a chance to process samples before a new sweep is started.

The BER meters by default start a new sweep after a specified number of errors have been detected. This can be disabled by setting the DETACT secondary parameter of the BER block to "Do not sweep".

The SWPVAR, VSA and VNA blocks can be configured to start a new sweep by connecting the block's input to an output port and entering a value for either the block's SWPDUR or SWPCNT secondary parameter. The SWPCNT parameter specifies a minimum number of samples to receive before starting a new sweep. The SWPDUR parameter is similar, except a simulation stop time is specified. A new sweep is started after a minimum of SWPDUR·fs samples have been received, where fs is the sampling frequency of the input signal.

Swept Measurements

Most VSS measurements support displaying results from one or more sweeps. They typically allow you to view the results from either an individual sweep or to display the results of all the sweeps of a swept variable. When displaying the results of an individual sweep, the sweep can be the active sweep, a specific pre-selected sweep, or selected on the fly with the tuner.

When displaying the results of all the sweeps, the format depends upon the measurement. For measurements with a pre-defined x-axis, such as the waveform measurement WVFM (time is the x-axis) or the power spectrum measurement PWR_SPEC (frequency is the x-axis), the results of each sweep are displayed as separate traces overlaid on the same graph.

PWR_SPEC displaying all sweeps

Figure 1.2. PWR_SPEC displaying all sweeps


Measurements that inherently support sweeping, such as the adjacent channel power measurement ACPR or the AM to AM measurement AMtoAM_PS, support using the values of one of the swept variables for the x-axis. For example, by default S21_PS displays S21 versus measured input power. If the simulation swept frequency, S21_PS could be configured to display S21 versus frequency.

S21_PS displaying S21 vs. frequency vs. inputpower

Figure 1.3. S21_PS displaying S21 vs. frequency vs. inputpower


When a system diagram contains swept variables, the Add/Modify Measurement dialog box, opened by choosing Project > Add Measurement or by right-clicking an existing measurement in the Project Browser and choosing Properties, displays additional controls for configuring the sweeps to plot.

Modify Measurement dialog box with drop-downs for two swept variables

Figure 1.4. Modify Measurement dialog box with drop-downs for two swept variables


If there is a VNA configured to sweep, a Power Sweep (VNA) drop-down list box displays. For other swept variables, a drop-down list box displays with a title containing the name of the block defining the swept variable and the ID parameter of the block, for example "SWPVAR.SWP1".

The drop-down list(s) contains a number of options, including a list of all the values for the swept variable represented by the list, a "Plot all traces" option, and a "Select with tuner" option.

Measurements that have a pre-defined x-axis such as WVFM or PWR_SPEC also have a "Plot active sweep" option. Measurements that directly support sweeping such as ACPR or AMtoAM_PS also have a "Use for x-axis" option. Measurements that directly support sweeping typically have 'Swept' in their description. The BER measurements also directly support sweeping.

If a specific swept variable value is selected, the measurement displays only the results for that sweep value. If Select with tuner is selected, the measurement displays the results for one sweep value. The specific value can be changed by opening the Tuner through Simulate > Tune.

If Plot all traces is selected, the measurement displays the results for all the sweeps of that swept variable. For measurements with a pre-defined x-axis the results are overlaid as separate traces. Measurements that directly support sweeping plot their measured values using their 'built-in' x-axis. For the swept power measurements such as ACPR or AMtoAM_PS this is the measured input power from the VSA. For BER this is SNR, Es/N0 or Eb/N0.

If Plot active sweep is selected, only the results of the current sweep display. Note that for multi-dimensional sweeps, only the first drop-down list contains this option. Selecting it overrides the settings for the remaining drop-down lists.

1.4. Data Signals

Data signals form the means of passing simulation data from one block to another. There are several different types of data signals, for the different types of processing supported. They include:

  • Digital signals

  • Analog signals

  • Complex envelope signals

  • Generic signals

  • Fixed-point signals

Data signals have signal properties associated with them. Signal properties define static attributes of the signal, which are used by various blocks and measurements to aid in the simulation. Examples of signal properties include sampling frequency and data rate, estimated signal power, and estimated phase rotation.

Closely associated with signal properties is the concept of a primary data path. The primary data is important in blocks that have multiple main inputs, such as adders and combiners. The primary data path determines which input signal's properties are used for the output signal.

1.4.1. Digital Signals

Digital signals typically represent digital data to be transmitted or data that has been received.

Digital signals are signals with values that are restricted to a range of non-negative whole numbers. Every digital signal has an alphabet size property, commonly referred to as M, which defines the range of values. A digital signal with an alphabet size of M has values that fall within the following range:

{0,1,...,M-1} (1.1)

Each digital sample is also called a digital symbol. If M=2 then each sample represents a bit. Normally, when a digital signal is converted to an analog waveform such as in a digital-to-analog converter or a digital transmitter, the number of analog samples representing each digital symbol is maintained as a property of the analog signal. This is the samples per symbol property. It is also known as the oversampling rate. In the current release the samples per symbol property is always a positive whole number.

The time step of a digital signal is called the symbol period. The inverse of the time step is the data rate. The data rate and time step at every digital output port can be viewed on a system diagram using the DRATE and TSTEP annotations.

When working with digital sources, the data rate by default is automatically set. If the source is eventually connected to a block downstream that defines the data rate or sampling frequency, such as a transmitter, the data rate of the source is set to produce the proper data rate or sampling frequency downstream. The data rate automatically compensates for any intermediate blocks that may adjust the data rate, such as encoders.

If the data rate cannot be determined from the downstream blocks, it is set to the data rate specified on the Options dialog box Basic tab.

1.4.2. Analog Signals

Analog signals are used to represent analog waveforms such as voltage or current. In the current release almost all analog signals represent instantaneous total voltage seen at an output port.

Analog signals are usually oversampled, that is the sampling frequency is a multiple of the largest frequency of interest in the signal. To satisfy the Nyquist criterion the signal must be oversampled by at least 2. However, for modeling analog signals an oversampling rate of 8 or more should normally be used.

The oversampling rate is usually determined automatically. Blocks that specify the oversampling rate contain a SMPSYM parameter to let you explicitly set the oversampling rate. SMPSYM can be found in analog signal source blocks such as the TONE and SINE blocks, or in blocks that convert another signal to analog, such as a digital-to-analog converter or a QAM transmitter. (The parameter is named SMPSYM because it originally represented the number of samples per symbol used to represent a digital signal in a transmitter. To maintain consistency, the name was kept for pure analog blocks. Samples per symbol and oversampling rate have the same meaning within the VSS program.)

If the oversampling rate is not explicitly specified at a block that generates an analog signal, it is normally determined either automatically from the topology of the system diagram or from the Oversampling Rate specified on the Options dialog box Basic tab.

1.4.3. Complex Envelope Signals

Often an analog signal is a narrowband signal, where the frequency content of interest is within a narrow frequency band that is centered at a frequency (called the center frequency) much larger than the width of the narrow frequency band. An example is a modulated RF signal. These signals can often benefit from complex envelope representation.

In complex envelope, or CE, representation, the signal is modeled relative to the center frequency, and the sampling frequency only needs to accommodate the narrow frequency band. This can reduce simulation time significantly, especially if the bandwidth of interest is much smaller than the center frequency.

Take for example a 1 GHz carrier modulated by a 10 MHz data signal. If you wanted to monitor the IM3 and IM5 characteristics of the modulated signal near to the carrier, the frequency band from 950 to 1050 MHz is more than sufficient to represent those frequencies.

If that signal were represented as a real valued analog signal (or real signal), you would need a sampling frequency of at least 2100 MHz (twice the maximum frequency of interest, 1050 MHz) to model the 950 to 1050 MHz band. Using a complex envelope representation, however, the sampling frequency would only need to be 100 MHz with a center frequency of 1 GHz.

Mathematically, the relationship between a real signal and its complex envelope representation is:

(1.2)

where x(t) is the real signal, c(t) is the complex envelope representation, fc is the center frequency, and Re{a} is the real component of the complex value a. x(t) is also called the real passband signal.

As a practical interpretation of the above, consider a narrowband modulated signal centered about a high frequency sinusoidal carrier with frequency fc. Such a signal can be mathematically expressed as:

(1.3)

xc(t) and xs(t) are real valued lowpass or baseband signals representing the signal modulating the carrier. They have a bandwidth much smaller than the center frequency. xc(t) and xs(t) are the in-phase and quadrature (I and Q) components, respectively, of the real passband signal x(t).

If the complex envelope term of Equation 1.2 is represented as:

(1.4)

and multiplied by the ej2πfct term, the argument of Re{ } in Equation 1.2 becomes:

(1.5)

and Equation 1.2 becomes:

(1.6)

Equation 1.6 shows that for a narrowband modulated signal, the complex envelope representation is the same as the baseband modulating signal.

The following graph further illustrates the complex envelope representation of a modulated signal. In this graph, a QAM signal with a 100 MHz data rate is modulating a 500 MHz carrier. The blue curve is the real passband signal. The red and magenta curves are the real and imaginary components of the complex envelope signal. The green curve is the magnitude of the imaginary signal. The graph illustrates that the envelope of the modulated carrier signal is the same as the magnitude of the complex envelope signal at the carrier's center frequency.

Real and complex envelope waveforms for a QAM modulated signal

Figure 1.5. Real and complex envelope waveforms for a QAM modulated signal


Working with Complex Envelope Signals

Many VSS blocks support complex envelope signals directly. For example, the circuit filters in the Filters category, the transmitters, receivers, modulators and demodulators in the Modulation category, and many of the RF blocks in the RF Blocks category support complex envelope signals directly. Many of the RF blocks also support real signals.

Other blocks, such as many of the math operators in the Math Tools category or many of the signal processing blocks in Signal Processing, work with complex envelope signals, though they generally treat a complex envelope signal the same as a generic complex signal.

Internally, the VSS program represents a complex envelope signal using a complex signal with an associated center frequency tag. The value of the center frequency tag at all output ports with such a tag can be viewed on a system diagram using the CTRFRQ annotation.

Complex envelope signals can only be generated by blocks that satisfy two conditions:

  • The output signal must be complex.

  • A center frequency must be specified.

Blocks that generate complex envelope signals usually have a CTRFRQ parameter for specifying the center frequency.

Most of the modulators and transmitters in the Modulation category satisfy these conditions, as do the complex sources in Sources. Several other source blocks provisionally satisfy these conditions. They provide a CTRFRQ parameter for specifying a center frequency, but support real as well as complex signals. The TONE block in RF Blocks > Tones and the SINE block in Sources > Waveforms are such blocks.

In most cases, working with complex envelope signals in the VSS program simply involves specifying the center frequency at one or two blocks. After that, you can usually specify frequencies of interest using absolute frequency values. For example, cutoff frequencies for circuit filters are specified in absolute frequencies.

Most measurements also automatically detect and handle complex envelope signals. For example, when entering frequencies for the adjacent channel power measurement ACPR, the channel center is specified as an absolute frequency.

There may be cases, however, where you want to convert a complex envelope signal to the equivalent real passband signal. For example, if you want to view the modulated carrier signal itself, you need the real passband signal.

This conversion can be accomplished using the CE2R block in Converters > Complex Envelope. This block allows you to view and manipulate the modulated carrier signal rather than just the complex envelope. Of course, in order to properly represent the complex envelope signal as a real signal the sampling frequency of the generated real signal may be very large compared to the sampling frequency of the complex envelope signal.

A real passband signal can also be converted into a complex envelope signal. This is done using the R2CE block in Converters > Complex Envelope. Note that while converting a complex envelope signal to its real signal equivalent is fairly straightforward, generating a complex envelope representation of an arbitrary real signal is not. R2CE works by essentially down converting the real signal to baseband using the specified center frequency, lowpass filtering that result, and downsampling to reduce the sampling frequency.

Center Frequency Of Zero

The VSS program allows a complex signal with a center frequency of zero. However, when the center frequency is zero, the signal may have special characteristics depending on what the signal represents.

For CW signals such as those generated by TONE or SINE, a center frequency of zero is treated simply as fc=0.

Modulated signals treat a center frequency of zero differently. In this case the complex signal is treated as representing two separate real signals. The real component is the I channel of a baseband signal while the imaginary component is the Q channel. The signal is called a baseband I/Q channel signal.

Several RF blocks implement different behaviors for baseband I/Q channel signals and complex envelope signals. These blocks include the behavioral amplifiers such as AMP_B and the behavioral mixers such as MIXER_B. Figure 1.6 illustrates the equivalent diagram for an amplifier-mixer link.

There is one significant difference between the two layouts in Figure 1.6. The output of the mixer when using complex baseband I/Q channels is a complex envelope signal, while the equivalent layout using real signals results in a real signal. If the LO frequency is high relative to the data rate, a very large sampling frequency must be used for the real signal equivalent. The sampling frequency for the baseband I/Q channel based layout only requires the sampling frequency be adequate to represent the data rate.

The default behavior of the spectrum-based measurements is to display the full spectrum when the signal is a baseband I/Q channel signal. This is similar to what network analyzers do when displaying I/Q channel signals.

The circuit filter blocks also support baseband I/Q channel signals. The main difference between treating a complex signal as complex envelope versus a baseband I/Q channel signal is that the coupling between the real and imaginary components of the input signal.

In the complex envelope, complex math is applied directly, which results in coupling between the two components. For baseband I/Q channel signals, the signals are treated as two separate signals, so there is no coupling between the two components. Note that in many instances the coupling is minimal so no significant differences are observed.

Equivalent baseband I/Q channel amplifier and mixer

Figure 1.6. Equivalent baseband I/Q channel amplifier and mixer


Other Benefits and Limitations

One advantage of using complex envelope signals in a Time Domain simulation is the fact that phase information is present in the complex signal. The phase information does not have to be extracted from the real signal.

One of the big disadvantages of complex envelope signals is the modeling of wideband signals, particularly noise, when the center frequency is small enough that the complex envelope frequency band crosses DC. This occurs when the following is true:

fc<fs/2 (1.7)

For a complex envelope signal, any negative frequency content can be 'folded' across DC into an equivalent positive frequency content. The problem with a wideband or noise signal is any negative frequency content already has positive frequency content, and the net effect is that the signal is doubled over the overlapping frequency range. This topic is covered in more detail in the Negative Complex Envelope Frequencies section of this chapter.

Another limitation of complex envelope signals involves the modeling of nonlinear operations. If the intent of a nonlinear operation is to apply the nonlinearity to the equivalent real passband signal, the operation must take into account the complex envelope nature of the signal.

Many of the RF nonlinear blocks do apply their nonlinearities to complex envelope signals with this adjustment. In particular, the nonlinear amplifiers that incorporate the polynomial based nonlinearity found in AMP_B, which include AMP_BV, AMP_F, and VGA_F, along with the behavioral mixers MIXER_B, MIXER_F and MIXER_S, all apply their nonlinearities to the equivalent real passband signal. The AM/AM-AM/PM based amplifier models such as those found in NL_F and NL_S also operate properly with complex envelope signals.

Other RF nonlinear blocks, such as V_LIM, do not take into account the complex envelope nature of the signal, and must be used with caution with complex envelope signals. For these blocks, the safest, though slowest simulation time-wise, approach is to convert the complex envelope signal to the real passband equivalent using CE2R, apply the nonlinearities, then, if desired, convert the signal back to complex envelope form using R2CE.

RF Budget Analysis Simulations

Complex envelope signals are primarily a Time Domain simulation concept. However, the frequencies at which RF Budget Analysis simulations are typically performed are offsets from the center frequency. This is particularly true with modulated signals.

Tone sources do have the option to use the tone frequencies rather than the center frequency. This is controlled using the RFBSRC secondary parameter.

1.4.4. Sampling Frequencies, Data Rates and Oversampling

There are several key concepts related to sampling frequency in VSS Time Domain simulation signals. First and foremost is the sampling frequency itself. Every time domain signal has a sampling frequency. The sampling frequency is the inverse of the simulation time duration represented by each sample of the signal.

Signals also have a data rate, which represents the bandwidth of the 'interesting' portion of the signal. For digital signals the data rate and the sampling frequency are the same. Digital signals that have been converted to analog signals such as with an ADC or a transmitter usually have a sampling frequency that is a multiple of the data rate. This multiple is the samples per symbol, as each digital symbol is represented by that many analog samples. It is also called the oversampling rate.

For pure analog signals such as tones and other waveforms, the data rate does not have a direct interpretation. Instead, it is used to indicate the interesting portion of the signal to the simulator, and is called the signal bandwidth. The sampling frequency is the signal bandwidth multiplied by the oversampling rate. The use of the signal bandwidth in the VSS program is described in the Signal Bandwidth section.

The following equations show the relationships between these values:

(1.8)
(1.9)

In the VSS program you can specify the data rate, oversampling rate, and sampling frequency in several ways. The simplest method is to use the settings on the Basic tab of the Options or System Simulator Options dialog boxes, either for all system diagrams or for a specific system diagram. The settings in the dialog box are used as the default settings for the system diagram.

Source and transmitter blocks also allow you to specify these settings on an individual block level. Blocks that work with digital signals such as the QAM transmitter QAM_TX have a DRATE parameter for setting the data rate and a SMPSYM parameter for setting the samples per symbol/oversampling rate. More general purpose blocks such the RF tone TONE have a SMPFRQ parameter for setting the sampling frequency. They also have a SMPSYM parameter for setting the oversampling rate. (The parameter is SMPSYM for historical reasons and to illustrate that it has the same effect as the SMPSYM parameter for the digital transmitters.)

Note that the default data rate, sampling frequency and samples per symbol/oversampling rate values specified in the System Simulator Options dialog box are available as built-in variables in the system diagram. These variables are:

Sampling frequency _SMPFRQ
Data rate _DRATE
Samples per symbol/oversampling rate _SMPSYM

The data rate and sampling frequency at all the output ports in a system diagram can be viewed using the DRATE and SMPFRQ annotations. The FRQ_PROP measurement in the System > Tools category can be used to display the data rate, sampling frequency and center frequency for specific ports in a graph table.

The proper selection of sampling frequency is important in Time Domain simulations. The sampling frequency, and for complex envelope signals the center frequency as well, determines the range of frequencies that may successfully be simulated.

Sampling Theory and Aliasing

Sampling theory states that the minimum sampling frequency required to represent a given frequency is twice that frequency. For real signals the frequency range is:

(1.10)

For complex envelope signals, the sampling frequency band is:

(1.11)

Note the use of ≤ at the lower limit and < at the upper limit. This is because the frequency at -fs/2(or fc-fs/2) has the same value as fs/2 (or fc+fs/2)

Any frequency content outside these ranges is aliased into the sampling frequency band. The following diagram illustrates the effect of aliasing.

Aliasing of frequencies outside sampling frequency band

Figure 1.7. Aliasing of frequencies outside sampling frequency band


In this example, there are three tones: 7MHz, 12MHz and 33MHz. This analog signal is then represented as a complex envelope signal with sampling frequency of 10MHz and center frequency of 15MHz. The sampling frequency band is then 10MHz to 20MHz.

The 12MHz tone is sampled correctly since it is within the sampling frequency band. The 7MHz tone is outside the band, so it gets aliased into the sampling frequency band. The aliasing results in the 7MHz tone appearing at 17MHz. The 33MHz tone is also aliased, it appears at 13MHz.

Selecting A Sampling Frequency

In practice, the sampling frequency for analog signals should generally be several times larger than the minimum sampling frequency required. This is particularly important when using the circuit filters or RF blocks. The reason for this is many of these blocks incorporate filters as part of internal upsampling and downsampling. This resampling is a key element in the modeling performed by these blocks.

The filters are used to eliminate frequency content outside the original sampling frequency band when the input signal is upsampled, and again prior to downsampling. Ideally, the filters are bandpass filters with a passband matching the original sampling frequency band and infinitely sharp cutoffs at the edges of the band. However, ideal filters with infinitely sharp cutoffs are not possible, and the filters must exhibit a transition region at the band edges.

The end result is some roll-off near the edges of the sampling frequency band. The following figure illustrates the roll-off effect on a tone with white noise passed through the nonlinear behavioral amplifier AMP_B. The blue curve is the signal going into the amplifier, the pink curve is the signal output by the amplifier, with slight attenuation of the noise near the edges of the sampling frequency band.

Roll-off near sampling frequency band edges due to internal resampling

Figure 1.8. Roll-off near sampling frequency band edges due to internal resampling


Signal Bandwidth

The signal bandwidth, signal band, and oversampling rate for analog signals are used several different ways in the VSS program.

The circuit filter blocks such as BPFB or LIN_S use the signal bandwidth when in IIR mode to determine the frequency alignment points for the bilinear transform. The bilinear transform maps s-domain frequencies to the z-domain in a nonlinear fashion. The frequency alignment point is where the s-domain frequency exactly matches the z-domain frequency. The Filter Issues section in the RF Modeling in VSS chapter covers circuit filters in detail.

The RF behavioral amplifiers and mixers may also use the signal bandwidth. The RF blocks that support internal resampling can be configured to use the signal bandwidth in their determination of the amount of upsampling required.

Consider a very simplified example: suppose the sampling frequency band is 10GHz to 20GHz (sampling frequency of 10GHz and center frequency of 15GHz). If the signal were passed through a 5th order polynomial (which is what the behavioral amplifiers use if both P1dB and IP3 are specified), the maximum frequency that may be generated is 5 · 20 = 100GHz. The upsampling rate that is required for a given frequency is found from:

(1.12)

For this example, the upsampling rate required is 9. However, if the signal can be presumed to be entirely contained within the signal band, then the maximum frequency that needs to be accommodated is the maximum signal band frequency rather than the maximum sampling frequency band signal. If, for this example the sampling frequency band were only 1GHz, then the maximum frequency that may be generated is 5 · 15.5 = 77.5GHz, and the upsampling rate would only have to be 7.

Note that the default behavior for the amplifiers and mixers is to use the sampling frequency band. The SIGBW parameter must be cleared to use the signal frequency band. Also note that this is only of value when the center frequency is near the sampling frequency, where the ratio of the maximum signal band frequency to the maximum sampling frequency band frequency is largest.

Mixers can also use the signal frequency band to limit the spurs to be synthesized. In these cases the SIGBW parameter is also used to limit the frequency band used to determine which spurs to generate.

Many of the VSS measurements use the oversampling rate to determine the base number of samples to 'slide' their snapshot of the time domain samples. For example, the WVFM measurement advances the window of time displayed by oversampling rate samples. If the time step is 0.1 ns and the oversampling rate is 10, the WVFM measurement lines up the start of the time axis on 1 ns boundaries.

The spectrum measurements also use the oversampling rate in a similar manner. The point at which samples for the FFTs are taken starts on a multiple of the oversampling rate. This is one reason why a spectrum of a CW signal may appear to jump as the simulation runs. If the frequency of the CW signal relative to the sampling frequency and the oversampling rate are not related by a whole number, they may become out of sync, affecting the display of the spectrum.

1.4.5. Generic Signals

The term generic signals is used to classify signals that do not represent digital nor analog signals directly. For example, control signals are generic signals. The complex valued input to the I/Q Modulator block IQ_MOD is also a generic signal, as is the output of the QAM Mapper block QAM_MAP.

1.4.6. Fixed-Point Signals

Fixed-point signals are a special case of generic signals. Fixed-point signals are used to model the limited floating point precision commonly found in DSP applications. The resolution available for representing floating point values in DSP applications is generally much less than that available in the general purpose CPUs used in personal computers. The fixed-point library allows the modeling of those limitations.

The "Fixed-Point Simulations" chapter of the Getting Started Guide covers fixed-point simulations in detail.

1.4.7. Signal Properties

All signals in the VSS program have one or more properties associated with them. A signal property describes a particular characteristic of the signal during a simulation sweep. All signals have a time step property, which defines the time duration between each sample. The following table lists some of the more common properties:

Property Data Types Description
Time Step/Sampling Frequency All The time span between samples. The inverse of the time step is the sampling frequency.
Samples per Symbol/Oversampling Rate All The ratio of the data rate to the sampling frequency. The data rate is the sampling frequency divided by the samples per symbol.
Alphabet Size Digital The range of allowed values in a digital signal.
Center Frequency Complex Envelope The center frequency of a complex envelope signal.
Signal Power Analog The average power of the transmitted signal. This defines the operating point and is used for automatic gain control in receivers.
Generated Noise PSD Analog The average noise power spectral density in a time domain signal. This is used by the BER meters to automatically determine Eb/N0, Es/N0 or SNR.
Phase Rotation Analog The average phase rotation that has been applied. This is used to perform automatic phase rotation compensation.
Signal Delay All The amount of delay that has been imparted on the signal since its generation.

Many of the properties are used for automatic compensation. For example, most receivers and demodulators use the signal power, phase rotation and signal delay properties to scale and align the signal prior to demodulation and detection.

The signal properties are referred to as 'static' properties. They are static because they do not change during a simulation sweep. They are determined before any samples are generated, and are used to describe the expected state of the signal. Most of these properties can be viewed using either annotations or measurements found in the System > Tools category.

1.4.8. Primary Data Path

Signal properties are normally assigned starting from source blocks. Blocks connected to those blocks then assign properties to their output signals, and the process repeats until all blocks have assigned properties to their signals.

Most blocks in the VSS program have one input port and one output port for the main signal. Other input or output ports are generally secondary in nature. For these one input-one output main signal blocks, the signal properties from the input port are normally assigned to the output port unless the block changes a particular property.

However, blocks that combine multiple input signals, such as the adder block ADD or the combiner block COMBINER, have multiple main signals to choose from. To determine the signal properties for their output port, these blocks must select one of the input ports as the primary input port and use that port's signal properties as the basis for the output port. The signal path formed by this primary input port and the output port is called the primary data path.

The primary input port of blocks such as ADD or COMBINER can either be determined automatically or selected through a PRIMINP parameter. When automatic selection is used, the block chooses the primary input port based on the following rules:

  • Modulated signals have the highest priority.

  • With all else equal, the input port with the smallest node number is chosen.

There is no corresponding primary output port. A block with multiple output ports defines multiple primary data paths, one for each input port - output port path.

1.5. Automatic Configuration

The VSS program has the ability to configure many of its settings automatically. For example, when using a QAM transmitter block such as QAM_TX, you can use the general purpose receiver block RCVR to receive and detect the signal. The RCVR block automatically selects the appropriate demodulation based on the received signal's properties. You can change the number of symbol levels in the transmitter and RCVR is automatically reconfigured to use the new symbol levels. You can even change the transmitter to MPSK_TX and RCVR automatically reconfigures itself for PSK rather than QAM modulation.

At the same time, blocks that support automatic configuration normally have parameters that can be used to override the automatic configuration. These parameters are often secondary parameters since they are not frequently used. They also typically have no value assigned, which the blocks interpret as use automatic configuration.

Another example of automatic configuration is the ability to automatically set sampling and center frequencies at source blocks whose signals are eventually combined. In this case one block determines its sampling frequency and the other sources automatically set their sampling frequencies to be compatible. For example, if the LO input to a mixer does not have its sampling frequency explicitly set, it is assigned the sampling frequency of the input or output signal of the mixer.

The block setting the sampling frequency does not necessarily have to be a source block. For example, most of the transmitters contain a DRATE parameter that lets you define the data rate at the output of the transmitter. The input signal to the transmitter has its sampling frequency set to this data rate if it is specified, while the output signal has its sampling frequency set to the data rate times the samples per symbol/oversampling rate. If the input to the transmitter consisted of a digital source followed by a rate 1/3 convolutional encoder, the digital source is automatically configured to generate data at 1/3 the data rate of the transmitter.

When there are multiple data paths all set for automatic sampling frequency determination connected to a single block such as a combiner, the primary data path is used to determine which blocks define the sampling frequency.

1.6. Spectral Analysis in Time Domain Simulations

Performing spectral analysis in sampled Time Domain simulations requires taking into account several factors that may affect the results. These are primarily related to the limitations imposed by the need to represent continuous signals as a sampled data stream.

This section points out some of these factors without going into depth with the mathematical theory behind them. For more in-depth details, refer to general digital signal processing texts such as Proakis [1].

1.6.1. Fourier Transforms

Spectral analysis in Time Domain simulations is performed using the discrete Fourier transform (DFT), windowing and averaging.

The DFT converts N equally spaced time domain samples into N frequency domain samples at equally spaced frequencies. The general equation is:

(1.13)

The inverse DFT converts N equally spaced frequency domain samples into N equally spaced time domain samples according to:

(1.14)

The relationship between n and frequency is:

(1.15)

The two different equations result in one of the N frequencies always being at the center frequency.

These equations make a big assumption: that the signal is periodic over N samples. However, unless the signal is composed only of tones that fall exactly on one of the equally spaced frequencies, the frequency content is 'smeared', or leaked into nearby frequencies. This is most apparent when working with pure tone signals. When N results in the frequencies of the tones falling exactly on one of the N frequencies, the frequency spectrum is exact. However, increasing or decreasing N by 1 causes the spectrum to leak.

Voltage spectrum of an 11 GHz tone at center frequency 10 GHz

Figure 1.9. Voltage spectrum of an 11 GHz tone at center frequency 10 GHz


Figure 1.9 illustrates the leaking effect. The signal from both plots is the same: a single tone complex envelope signal at 11 GHz with center frequency of 10 GHz and sampling frequency of 10 GHz. The tone's amplitude is 1V, and the y-axis displays linear voltage.

The blue curve has N set to 10. This results in the following sampled frequencies:

5, 6, 7, 8, 9, 10, 11, 12, 13, 14 GHz

The pink curve has N set to 11. This results in the following sampled frequencies:

5.45, 6.36, 7.27, 8.18, 9.09, 10.00, 10.91, 11.82, 12.73, 13.64, 14.55 GHz

For N=1, 11 GHz is not one of the sampled frequencies. Therefore, the tone is spread into the available sampled frequencies.

1.6.2. Averaging, Windowing and Power Spectrum Estimation

An estimate of the power density spectrum called a periodogram can be made using Equation 1.13 by applying the following:

(1.16)

Unfortunately, the periodogram is not a consistent estimate of the true power density spectrum, particularly when working with non-periodic signals such as modulated signals.

One of the most common and simplest methods of improving the power spectrum estimate, and the approach taken by the VSS program, is the Welch method. This method involves computing the average of several overlapping periodograms with windowing applied to the time domain samples.

Effect of averaging and windowing on power spectrum estimates

Figure 1.10. Effect of averaging and windowing on power spectrum estimates


Figure 1.10 illustrates the smoothing effect of averaging and windowing on a QAM signal. The QAM signal is not pulse shaped, so the theoretical power spectrum is the sinc function squared. The black markers indicate the theoretical spectrum. The pink curve is windowed and averaged 10 times with 50% overlap. The blue curve is not windowed and is averaged 10 times with 50% overlap. The gray curve is not windowed, has no averaging and uses a DFT with 1000 points. The averaged spectrums use DFTs with 100 points, resulting in a total of 1000 time domain samples being used for all three cases.

Note that with windowing and averaging, the spectrum matches the theoretical spectrum closely in the main lobe -1 GHz to +1 GHz. However, it deviates towards the ends of the sampling frequency band.

In general, windowing and averaging should be used when working with non-CW signals, such as modulated signals, but not with CW signals. The automatic configuration feature of the VSS program attempts to select the proper settings for performing power spectrums based on the signal properties. If the signal is a modulated signal, it uses windowing and averaging and a fairly large N (1024 in the current release) for the DFT.

If the signal is CW and the tone spacing is a sub-multiple of the sampling frequency, it attempts to find an N that results in all tones in the signal falling on frequency sample points. If an N with a reasonable value is found, averaging and windowing is disabled.

The automatic configuration settings can be overridden. The measurements that rely on power spectrum computations include settings for selecting N, the number of averages, and the windowing settings. These settings are generally secondary settings in the measurements. The various test points and meters also include similar settings. By default the measurement settings use the settings from the test point/meter.

The default settings when averaging and windowing are applied are:

  • Number of FFT bins: 1024

  • Number of averages: Cumulative

  • Window type: Taylor

  • Windowing parameter: 4.5 for "Kaiser-Bessel" windowing, 0.0 for all others.

  • Sliding ratio: 0.5

1.6.3. Frequency Resolution and Video Bandwidth

Spectrum analyzer settings are normally in the form of frequency resolution and video bandwidth. There are equivalent settings for the DFT based spectrum measurements.

Frequency resolution is directly related to the number of samples used to compute the DFT:

(1.17)

Video bandwidth does not have such a direct equivalent. However, an effect similar to that of video bandwidth is produced by averaging. In a spectrum analyzer decreasing the video bandwidth has the effect of smoothing the spectrum. A similar behavior can be accomplished with DFTs by applying averaging.

VSS software uses the following relationship to convert from video bandwidth to number of averages:

(1.18)

Note that Equation 1.18 is only used to approximate the effects of changing the video bandwidth. It does not necessarily shape the spectrum to what you would see in a spectrum analyzer for the given video bandwidth.

1.6.4. Negative Complex Envelope Frequencies

One of the more confusing aspects of working with complex envelope signals is how to interpret negative frequencies. When possible, you should avoid working with complex envelope signals that contain negative frequencies. This is accomplished by ensuring Equation 1.19 is true:

(1.19)

However, this is not always possible.

There are three scenarios when a complex signal contains negative frequency content:

  • Complex baseband signal representing I and Q channels.

  • Complex envelope signal with center frequency of 0.

  • Complex envelope signal with center frequency other than 0.

The first two cases both have a center frequency of 0. The difference between the two is the interpretation of the complex values. In general, a signal is treated as separate I and Q channels when it represents the output of an I/Q modulator. CW signals are treated as complex envelope signals.

The second and third cases both treat the signal as complex envelope signals and result in the same behavior.

Complex baseband signals representing I and Q channels are covered in the Center Frequency Of Zero sub-section of the Complex Envelope Signals section.

For complex envelope signals, negative frequency content has an equivalent positive frequency content. The equivalent positive frequency content is the complex conjugate of the negative frequency content. This is derived from the complex envelope equation:

(1.20)

where S(f) is the spectrum of the real signal, SCE(f) is the complex envelope spectrum and a* indicates the complex conjugate of a.

Figure 1.11 illustrates this interpretation further. The upper diagram shows the complex envelope and the conjugate of the complex envelope when fc>fs/2. The lower diagram shows that as fc becomes less than fs/2, the complex envelope and its conjugate begin to overlap. The cross-hatched area is the negative frequency content of the complex envelope.

Negative frequencies in complex envelope signals

Figure 1.11. Negative frequencies in complex envelope signals


In most cases, when interpreting a spectrum any negative frequency content should be 'folded' into the equivalent positive frequencies. This results in a spectrum containing only non-negative frequencies, which is what is normally expected when working with real signals.

The default behavior of the spectrum-based measurements is to display negative frequencies folded into their equivalent positive frequencies. This presents the spectrum as you would see it on a spectrum analyzer.

One case where negative frequency folding does not work is the modeling of white noise or any other broad spectrum signal where the signal crosses DC. What happens in these cases is there is overlapping frequency content from both the positive and negative frequencies, and folding results in a doubling of the spectrum over the portion of the spectrum where the overlap occurs.

Negative frequency folding and broad spectrum signals

Figure 1.12. Negative frequency folding and broad spectrum signals


As Figure 1.12 illustrates, the resulting folded spectrum shows a sharp drop where the negative frequency overlap ends. This is most apparent when the signal contains white noise, since the white noise extends across the full sampling frequency band. Theoretically, the white noise should end at DC - there should not be any negative frequency content. However, in the current release white noise is always generated across the full sampling frequency band.

The Noise Modeling in VSS chapter of this guide contains more information on noise modeling within the VSS program.

1.6.5. Measuring Channel Power

Many VSS measurements use channel power in their computations. These include the power meter measurement PWR_MTR, the adjacent channel power measurement ACPR, and the swept AM to AM measurement AMtoAM_PS.

Channels are normally defined using a center frequency and a bandwidth. The channel center frequency should not be confused with the signal center frequency. The channel center frequency defines the center of the channel and the bandwidth defines the width. The channel frequencies are the frequencies that satisfy:

(1.21)

Channel power is computed by first obtaining a power spectrum. The use of the DFT divides the frequency spectrum into N frequency bins, where the value of each bin represents the power within the frequency range:

(1.22)

for n=0,1,...,N-1. Note that for n=0 the lower frequency bound is the lower edge of the sampling frequency band and for n=N-1 the upper frequency bound is the upper edge of the sampling frequency band.

Channel power is obtained by first summing the power spectrum values for the bins that fall entirely within the channel frequency band. If one of the edge frequencies does not fall exactly on a edge of a frequency bin, power proportional to the amount of the bin occupied by the edge frequency is added to the channel power. For example, if the lower frequency edge were 3.75 Hz and it fell in the frequency bin bounded by 4 Hz and 5 Hz, 25% of the power in that bin is included in the channel power.

When specifying channel bandwidth, the VSS program also supports using a single frequency bin. The "1 RBW (bin)" option for the channel bandwidth type, when available, does this.

1.7. Microwave Office Schematics: Parameterized Subcircuits and Sweeping

You can access parameterized Microwave Office subcircuits in the VSS program and use them in swept simulations. The RF blocks that support parameterized subcircuits are LIN_S, NL_S, OSC_S, and MIXER_S.

1.7.1. Passing Parameters from Microwave Office Circuits

In a circuit schematic you can define the parameters you want to pass to a VSS system diagram by using the x<<5 syntax, where x is the parameter to pass. The circuit schematic shown in the following figure has a passed parameter for DC voltage.

When the circuit is used in VSS system diagrams using the Simulation Based RF blocks (such as NL_S in this example), the passed parameter from the circuit schematic is also listed as one of the parameters inside the RF Block, as shown in the following figure. This parameter can be treated like any other parameter and used for swept simulation.

1.7.2. Using a Microwave Office Circuit Swept Parameter in a System Diagram

Parameters that are swept in a circuit schematic can also be used in VSS simulation. A sweep performed using the SWPVAR block or swept sources works the same when used in the VSS program. In the following figure notice that the circuit parameter "Vbias" is swept using the SWPVAR block.

When the circuit is used in VSS system diagrams using the Simulation Based RF Blocks (such as NL_S in this example), the swept parameter from the circuit schematic is also listed as one of the parameters inside the RF Block, as shown in the following figure. Note that the parameter name begins with 'SWP_'. This parameter can be treated like any other parameter and used for simulation.

The parameters passed to the VSS program in this manner can take on two sets of values.

  • Select one of the sweep values from the circuit schematic: In the corresponding RF block Element Options dialog box click Show Secondary. Notice the parameter IVARTYP as shown in the following figure. Set this parameter to Select from list and the passed parameter lists the sweep values in the drop-down list as shown in the following figure.

  • Enter any value and pin it to the nearest sweep value: Set the IVARTYP parameter to Allow any value for numeric, pin to nearest as shown in the following figure. The value entered is pinned to the nearest sweep value from the schematic.

1.8. References

[1] Proakis, J. and Manolakis, D., Digital Signal Processing

Please send email to awr.support@ni.com if you would like to provide feedback on this article. Please make sure to include the article link in the email.

Legal and Trademark Notice