The Cadence® AWR® Network Synthesis networks synthesis wizard is a tool for creating optimized two-port matching networks composed of discrete and distributed components. You specify the maximum number of sections and the types of components to include in the search space. The wizard searches for the best circuit topologies and optimizes the component parameter values.
The optimization goals are specified in the wizard using a dedicated set of synthesis measurements, much like optimization goals are normally defined in the AWR Design Environment software. Specialized measurements are provided for input noise matching, amplifier output power matching, and interstage matching. The optimum reflection coefficients are specified over frequency and can be provided in the form of load pull data, network parameter data files, or circuit schematics.
Options allow you to specify DC constraints on the networks and optimally attach a user-provided bias-feed network to the circuit. Also, the component parameter values may be confined within minimum and maximum limits and additionally may be constrained to discrete values.
For information on using the wizard via the AWR Design Environment platform API, see “Network Synthesis Wizard”.
At the bottom left of the Synthesis Definition tab a block diagram defines the terminology used in the wizard. The rectangle in the middle of the diagram represents the network that the wizard will synthesize. Attached to Port A of that network is a block that represents impedance A. That impedance might be the source impedance of an LNA or the output load for a PA. Connected to Port B of the matching network is impedance B, which for an LNA is the input impedance of the active device and for a PA is the output impedance of the active device. The wizard creates a network to optimize the match between Port B and the block labeled Impedance B.
After the wizard synthesizes the matching networks, the selected networks are drawn in project schematics when you click the To MWO button. By default, Port A in the network is port "1" and Port B is port "2". You can reverse this by selecting P=2 as the Matching network port numbering.
Some of the synthesis measurements allow you to specify impedance A as a parameter for the measurement. If this is not specified, impedance A is taken from Default A impedance.
This tab includes two name fields. Synthesis session name specifies the name for the wizard saved state, listed under the Network Synthesis wizard node in the Project Browser. Synthesis results name is used as a base name for the generated networks and schematics and for the user folder in which the schematics are saved.
The Frequencies to match list specifies where the measurements are made for calculating the cost of a network during optimization. This list is initialized with the project frequencies. Click Edit Frequencies to customize the list.
The Components tab provides options for tailoring the set of network topologies over which the wizard searches.
Select the desired components in the shunt and series component lists. The components at the top of the lists are ideal, lumped components, followed by TLines, which are ideal transmission lines. The wizard replaces the TLines with microstrip lines in an additional optimization step if Replace TLines with MLINs is selected. Microstrip tees are also added where needed, and microstrip step junctions are added if MSTEPs is selected.
By default when MLINs are used, the wizard gets the substrate definition from the first MSUB element it finds when looking through the Global Definitions windows in the project. If there are multiple MSUBs in the project, you should click the MSUB button to specify the correct substrate definition.
At the end of the component lists are the vendor library component categories: Vendor Lib Inductor, Vendor Lib Capacitor, and Vendor Lib Resistor. Select these to have the wizard choose between non-ideal components available in vendor component libraries. (The wizard does not vary the values of parameters on these components–if the components have parameters, the default values are used.) Click the button to display the “Select Vendor Library Components for Network Synthesis Dialog Box” where you can edit the lists of allowed library components.
Valid topologies are determined by the types of components selected and the value specified for Maximum number of sections. Each section is either a series component or a shunt component. The wizard considers topologies having the maximum number of sections "N", and with fewer, down to N-3 sections. In its synthesis algorithm, any number of series components may be connected in series, but shunt components must have at least one series component between them.
In a typical RF/microwave circuit, some minimum length of transmission line is needed between the active device and any components used in a matching network. The First comp (Port A side) and Last comp (Port B side) drop-down lists provide a way to specify that a series TLine is required. Alternatively, the first and/or last components can be forced to any other type of component. On the Parameter Limits tab, you can specify separate limits for the parameters of these first/last components.
The Search space size display is for informational purposes only. It provides an indication of how the size of the search space (number of circuit topologies) increases as a function of the number of sections in the networks and the selection of series and shunt component types.
To access this dialog box, click the Components tab.button on the
On the left of the dialog box a copy of the AWR Design Environment platform Elements Browser Libraries node displays. To select components, browse to a folder in the tree and then choose from the components in that folder listed in the pane on the right. Shift- and Ctrl-clicking are supported for multi-selection. Click on the button to copy the selected item(s) into the Synthesis Components pane on the right. This list of components is available to the wizard in the category (for example, Capacitors) specified in the drop-down list. The check boxes before each component name allow you to select which components can be considered for the synthesized networks.
Theand buttons affect these check boxes. The button deletes selected components from the list. When only a single component in the list is selected, the button is enabled. Click this button to view the category in the Elements Browser from which the component was added.
You can export the list of components on the right into a comma-separated values (CSV) file for later use by clicking thebutton. Click the button to reimport the list into another wizard session. You can easily edit these CSV files outside of the wizard with a spreadsheet program or text editor software.
The Parameter Limits tab includes a table that lists each type of component selected on the Components tab.
Each component parameter shows the minimum and maximum values allowed when the component is used in either a series or a shunt configuration.
If a specific component type is selected for the First comp or Last comp on the Components tab, a separate component type displays in this table, designated with "F/L" to indicate that it is a First or Last component in the network. This allows these First/Last components to have different limits than components of the same type used elsewhere in the networks.
When Replace TLines with MLINs and MSTEPs are selected on the Components tab, a constraint on the ratio of adjacent MLIN widths, labeled as "W2/W1", also displays, with default min and max values set to the hard limits imposed by the MSTEP model.
To edit any of the constraints on a parameter, select it and click the Edit Limits button or simply double-click the parameter. A dialog box containing the various settings for the parameter constraints displays.
By default the Upper, Lower, and Initial value values for a parameter are the same for series and shunt components. To set the values independently, first clear the Use same limits for shunt and series elements check box.
Click the Calc Init Values button to compute the initial value based on the upper and lower limits. Normally the geometric mean of the limits is used, but if the lower limit is set to zero, the arithmetic mean is used instead.
There are four options available for constraining the parameter to discrete values (if Continuous values is selected (the default), Use discrete value list is disabled):
Select Round to, then in the text box enter the
precision in the displayed units to round to a specified precision. For
example, to round the inductance parameter shown in the previous dialog
box to the nearest tenth of a nH, enter
the text box.
Select Round to # sig. digits, then in the text box enter the specified number of significant digits.
Select Use table of significant digits to enable the Table name drop-down and constrain the three most significant digits to those provided in a table. The names of built-in tables in the list are for the “E-series” system of preferred numbers, a standard (IEC 60063) that was created for use with electronic components. The three-digit values in the selected table display below the option and represent the values that are allowed for the three most significant digits of the parameter. User-defined tables are also supported. Click the Add button to create a table from scratch, and click the Copy button to create a new table using the values of an existing table as a starting point. User-defined tables are not defined only for a specific component parameter; they can be used for any parameter.
Select Use discrete value list and then select a List name from the drop-down list to constrain to a list of discrete values. Click the Add button to create a new list of allowed parameter values specified in base units, and click the Copy button to create a new list using the values of an existing list. The controls behave similar to when Use table of significant digits is selected, although there are no built-in lists of values.
The DC & Bias Feed tab includes two sections. The top section has options for restricting the topologies to those that have certain DC characteristics (open or short from a port to ground or between ports). By default no DC constraints are specified. The bottom section provides options for attaching a bias feed network to the matching network.
When Attach bias injection network to matching network is selected, the Bias network source document drop-down displays a list of the project documents (schematics and data files) that you can use for the feed network. The Port # option specifies which port of the feed network is connected to the matching network. For each matching network topology the wizard chooses a location to attach the feed network. If there is a DC constraint on the matching networks specifying that an open circuit is required between the two ports, you can also indicate whether the bias feed network should be located on the A side or the B side of the matching network.
The fitness (or cost) of a matching network is evaluated by taking a measurement at each of the specified frequencies and summing comparisons of the measurements versus the goal. The definitions of the measurements are listed at the top of the Goals tab, and the goals for each measurement display in the bottom half of the dialog box.
In the Measurements section of the dialog box, click the Add or Edit buttons to display a Synthesis Measurements dialog box that lists the available synthesis measurements.
For synthesizing output matching networks for power amplifiers, the HarmAreaMatch and LoadPull measurements are useful. For low-noise amplifier input matching use the NoiseMatch measurement. The NetMatch measurement is good for input or output matching of a linear amplifier, and the NetMatch2 measurement provides a way to create an interstage match between two devices. (Note that the NetMatch and NetMatch2 measurements compute mismatch loss, not return loss.) You can use the NetGp measurement in conjunction with the other measurements to place a constraint on the amount of loss introduced in the matching network.
The HarmAreaMatch measurement provides a flexible way to directly specify a region (annular sector) of reflection coefficients to match into, at a specified harmonic. To aid in visualizing the region defined by the measurement parameters, when a HarmAreaMatch measurement is selected, the View Region button on the Goals tab is enabled. Click this button to display a Smith Chart with arcs drawn to show the boundaries of the region.
The CompCount measurement allows you to specify constraints on the number of components. When its Filter Type parameter is set to Unique Vendor Lib Components CompCount returns the number of different types of vendor library components used in the matching network. This can be helpful for minimizing the number of line items on a bill of materials. The other Filter Type setting is Lumped Element Components. When this mode is selected, the measurement returns the count of components in the network that are not transmission lines.
After defining the measurements, the goals are specified on the lower half of the dialog box. Click the Add button to display the New/Edit Optimization Goal dialog box to select a measurement and define a goal for it. The dialog box displays the formula used to compute the cost from the measurement and goal values. The values of the constants used in the formula are adjustable and you can alter the range of frequencies for the goal from the default MIN and MAX values, which correspond to the minimum and maximum frequencies listed on the Synthesis Definition tab.
Note that you can create multiple goals for each measurement, which means for example that the frequency range can be split into multiple bands with different goals for each.
The Search Options tab provides advanced settings for refining how network topologies are created and optimized. Descriptions of each option follow.
Maximum number of topologies to search per section: When the wizard generates the network topologies, it begins by creating all the possible topologies with a single component. The number of such topologies is normally equal to the number of selected check boxes on the Components tab. To create the 2-section topologies, it takes each 1-section topology and goes through the list of components that are allowed to follow the first one. The process repeats, producing an exponential growth in the number of topologies as a function of the number of sections. This setting (referred to as "M" and with a default of 1000) provides a way to constrain the exponential growth, by limiting the number of N-section topologies used to create the topologies with N+1 sections. Only the "M" best topologies are propagated.
Search depth: There are a variety of hard-coded control values in the algorithm used for optimizing the component values. These constants were determined empirically to provide a good tradeoff between covering the whole space of allowed values and limiting the optimization for speed and memory usage. This setting gives you some control over these optimization parameters.
Frequencies for initial search: During the initial phase of the synthesis process, when the full set of topologies is being evaluated for pruning, it can be helpful to use only a subset of the frequencies to speed up the evaluations. The list of frequencies on this tab is the subset used during this initial phase. The wizard chooses a default subset which you can override by selecting the Customize check box and clicking the Edit Frequencies button to specify alternate frequencies.
The outcome of the synthesis process is summarized on the Results tab.
Listed under Results are all of the synthesized networks that have a cost less than or equal to the specified maximum cost, ordered by number of sections and cost. Click a column header to change the sort order. Networks with a lower cost are closer to meeting the goals. A network with zero cost achieves all of the goals– the synthesis procedure terminates if a network is found that attains a cost of zero. The Sens. column provides an indication of how sensitive the cost is to variations in the component parameters.
Click on a network name to see a schematic representation drawn in the Simplified topology area. If Attach bias injection network to matching network is selected on the DC & Bias Feed tab, a green arrow shows where the bias injection network is attached.
Click the To MWO button to export into project schematics those networks with a check mark before their name. By default, a certain number of networks (specified in Maximum networks) with the best costs are auto-selected.
In addition, if one or both of the options under Result display options is selected, an Output Equations data display document is created in the project. This window will contain a schematic view and one or two graphs with measurements pre-populated. An S_TERM_Z measurement is used on the Reflection Coefficient graph, with the value for its Z0 (real) parameter specified in Z0, real. The network displayed in the schematic, which is also the data source for the measurements, is selected by clicking on the name of the schematic under the User Folders node in the Project Browser.
If there is more than one goal specified, the Cost Breakdown button is enabled. Click this button to display a Cost Breakdown dialog box that shows for each network how much each goal contributed to the total cost. For each network, the goal that contributed most to the total cost is highlighted in red and the goal that contributed least is highlighted in green.