Changes

Analyzing Circuits Using RF.Spice A/D Tests

9,728 bytes added, 06:01, 2 August 2015
[[File:b2MAN_Fig203.png|thumb|200px|A typical DC operating point table.]]
===DC Bias Test===
The DC Bias Test, also known as the Operating Point Analysis, calculates all the node voltages and branch currents at the steady (quiescent) state of your circuit. In its most common form, the input can be a constant voltage source or a constant current source. During a DC Bias Test, all capacitors are opened, all inductors are shorted, and all voltage and current sources except for the input source are set to their initial values.
===DC Sweep Test===
A DC Sweep Test is a set of DC Bias [[Tests]] performed successively, during which the value of a parameter (called the sweep variable) is varied. At the end of the simulation, a set of data are generated that correspond to your circuit's response to the different values of the sweep variable.
[[File:b2MAN_Fig231.png|thumb|200px|The output of a DC Sensitivity Test.]]
===DC Sensitivity Test===
A DC Sensitivity Test calculates the sensitivity of the output port to all the device values and model [[parameters]] at the circuit's operating point. Sensitivity is defined as the derivative an observable like a node voltage with respect to a parameter value such as the resistance of a resistor. In other words, it gives a measure of the variation of the voltage due to a small variation in the value of a part parameter. The output port for sensitivity calculation is specified by a pair of positive and negative (reference) nodes.
The results of a DC sensitivity test are displayed in a table window. The left column of the table lists the part [[parameters]] and the right columns shows the derivative of the output voltage with respect to the respective [[parameters]]. For example, consider the simple voltage divider circuit of Tutorial lesson No. 1, which consists of a 1V DC voltage source in series with a 1k resistor and another 2k resistor. The output voltage is designated as the voltage across the 2k resistor. The results of the DC sensitivity test for this circuit indicate that the output voltage is increased by 666.66mV for every 1-volt increase in the input source voltage, it is decreased by -1.111 mV for every 1 Ohm change in the first resistor's value, and it is increased by 555.55 microvolts for every 1 Ohm change in the second resistor's value. Note that these are all small-signal operating point results.
===Device Output Parameters Test===
This test shows the operating point details for each of the primitive devices in your circuit. [[B2.Spice A/D]] displays the results for all the devices in the circuit in a table window.
===Model Output Parameters Test===
This test shows the operating point details for each of the process models present in your circuit. [[B2.Spice A/D]] displays the results for all the process models present in the circuit in a table window..
[[File:b2MAN_Fig51.png|thumb|200px|AC Sweep Test Settings.]]
===AC Frequency Sweep Test===
The AC Frequency Sweep Test is used to examine your circuit's behavior at different input frequencies. For example, a low-pass filter circuit could be analyzed to determine the cut off frequency of the filter, that is, the frequency above which the magnitude of the output is negligible.
[[File:b2MAN_Fig52.png|thumb|200px|AC Sensitivity Settings.]]
===AC Sensitivity Test===
The AC Sensitivity Test calculates the small-signal sensitivity of an output port to all device values and model [[parameters]] over a range of frequencies. The output port for sensitivity calculation is specified by a pair of positive and negative (reference) nodes. You can also prespecify whether the plots use decibels or magnitude, or degrees or radians.
===Distortion Test===
The Distortion Test computes the steady-state harmonic products of your nonlinear devices for small-signal inputs. In a distortion analysis, a multidimensional Volterra series analysis is performed using multidimensional Taylor series to represent the nonlinearities at the (DC bias) operating point. Terms of up to the third order are used in the Volterra series expansion. The Distortion test is supported for the following nonlinear devices: diode, BJT, JFET, MOSFET (all levels), MESFET as well as all linear devices. If there are switches present in your circuit, the analysis continues to be accurate provided the switches don not change state under the small-signal excitations used for distortion analysis.
[[File:b2MAN_Fig232.png|thumb|470px|The output of a Small-Signal Transfer Function Test.]]
===Small-Signal Transfer Function Test===
The Transfer Function Test calculates the small-signal transfer function of your circuit by linearizing it around its DC operating point. It generates the input impedance, output impedance and voltage gain seen across your circuit from the input port to the output port. All you need to do is specify the the input source and the output port.
===Pole-Zero Test===
The Pole-Zero Test generates a list of small-signal poles and zeros of the transfer function of your circuit given the input and output nodes.
|-
|}
 
[[File:b2MAN_Fig61.png|thumb|200px|Noise Test Settings.]]
===Noise Test===
The Noise Test analyzes the device-generated noise over a range of frequencies. For this test you provide an input source, an output port and a range of frequencies. [[B2.Spice A/D]] calculates the noise contributions of each device (and each noise generator within each device) to the output port voltage. The program also calculates the equivalent to the output noise referred back to the specified input source. This is what is meant by the "Input Noise". The calculated value of the noise over the specified range of frequencies corresponds to the spectral density of the signal viewed as the square root of a stationary Gaussian stochastic process. After calculating the spectral densities, the simulator integrates these values over the frequency range to arrive at the total noise voltage/current over this frequency range. This calculated value corresponds to the variance of the circuit variable viewed as the square root of a stationary gaussian process.
 
As for the specific noise sources, they include the "shot" noise associated with the DC currents in semiconductor devices and the "thermal" noise associated with resistance. Semiconductors also display "flicker" (1/f) noise. However, due to the lack of a unified model, [[B2.Spice A/D]] handles this type of noise on a "case by case" basis. There are flicker noise [[parameters]] available for transistors in the list of model [[parameters]]. The program can also provide the noise gain associated with the 1/f source.
 
 
The [[parameters]] of the noise test are specified using the setup dialog of the Test Panel. The input noise source can be either a voltage or current source, which you can select from the drop-down list labeled "Noise Source". The output port must be specified as a pair of node numbers, the "Output Node" and the Reference Node". The rest of the [[parameters]] specify the frequency interval over which the noise analysis will be performed. They are similar to an AC analysis. If you check the checkbox labeled "Include Noise Generator Contributions", then noise contributions of all the individual devices will be shown as well. If not checked, the input noise and output noise will be shown only.
 
 
As an example, consider the RTL inverter circuit of Tutorial Lesson 3, which is shown here in the opposite figure. The voltage source Vs is designated as the input noise source. The output port is set at node 3, i.e. the collector of the bipolar junction transistor. The noise is calculated over the frequency range from 1kHz to 10MHz on a decade scale with 10 steps per interval. The figures below show the graph of the noise spectral density over the specified frequency ranges as well as the tabulated results of the noise analysis. The latter are the integrated noise results, representing the total input and output noises as a result of contributions from the discrete devices in the circuit (namely, the resistors and the BJT). The results shown below are the square root of what Berkeley SPICE generates. This is because Berkeley SPICE gives values that are proportional to the noise power rather than the noise voltage.
 
<table>
<tr>
<td> [[File:b2MAN_Fig240.png|thumb|300px|The RTL inverter circuit of Tutorial Lesson 3.]]
</td>
<td> [[File:b2MAN_Fig241.png|thumb|500px|Plot of input and output noise spectral densities.]]
</td>
<td> [[File:b2MAN_Fig242.png|thumb|370px|The integrated input and output noise results.]]
</td>
</tr>
</table>
 
 
== AC Frequency Sweep of RF Circuits ==
 
The simplest RF circuit analysis type in [[RF.Spice]] is the "AC Frequency Sweep" Test. As mentioned earlier, this is identical to the AC frequency sweep test of [[B2.Spice A/D]]. The only difference here is that the frequency-domain models of [[Multiport Networks|multiport networks]] and transmission line segments or components are added to the analog or mixed-mode simulation of your circuit. Just as in [[B2.Spice A/D]], the AC Test is run from the Test Panel of the Toolbox.
 
 
==Defining Sources and Loads==
 
Just as in low frequency circuits, RF circuits must be excited using a voltage or current source. However, in order to ensure a successful RF circuit simulation, your source must be a single-frequency (sinusoidal) AC source. Remember that in [[B2.Spice A/D]] circuits, you must specifically designate a source to operate as an AC source for all [[tests]] of the AC type. The RF Menu of [[RF.Spice]] provides two additional types sources: the AC voltage source with the keyboard shortcut "Alt+V" and the AC current source with the keyboard shortcut "Alt+I". These sources are identical to the regular voltage and current sources of [[B2.Spice A/D]], whose "Use" checkboxes in the AC section of their source property dialog are automatically checked.
 
[[File:RFAC1.png|thumb|400px| AC voltage and current sources with internal series or shunt impedances.]]
 
{{Note | For AC-type RF circuit analysis, you can only use AC voltage or current sources with a single common frequency.}}
 
 
In most RF circuits, the sources are modeled to have an internal source impedance typically denoted by Z<sub>s</sub>. This source impedance is usually real-valued and typically has a value of 50 Ohms. To model the source impedance, you can simply use a resistor in series with the AC voltage source or a resistor in parallel with the AC current source as shown in the opposite figure. If you need a complex-valued source impedance, you can use a "Complex Impedance" and connected it either in series or in parallel with the AC voltage or current source, respectively.
 
 
[[File:RFAC2.png|thumb|400px| A simple RF circuit driven by a voltage source and with a resistive load.]]
 
Similarly, most RF circuits have a load impedance typically denoted by Z<sub>L</sub>. This load impedance, too, is usually real-valued and typically has a value of 50 Ohms. To model the load impedance, you can simply use a resistor at the output port of your RF circuit. The opposite figure shows a simple RF circuit consisting of a two-port network N1 connected to an AC voltage source with a 50&Omega; internal resistance and terminated at a 100&Omega; resistive load. Note how the negative input and output pins of the two-port device have been grounded.
 
Your load impedance can also be a combination of resistors, capacitors or inductors to model capacitive or inductive loading. Note that in that case you will have a complex-valued load impedance that varies with the operational frequency. In some other cases, you may prefer a user-defined "Complex Impedance" as your load, which cannot be simply modeled as a combination of RLC elements. A resonant antenna load is a good example of this case. The port characteristic data for the antenna structure can be generated by an electromagnetic simulator like [[EM.Cube]] and then imported to [[RF.Spice]]. If you have the input impedance values as a function of frequency, then you should define a complex impedance load. If you have the return loss (s11) data as a function of frequency (as is usually the case), then you can define a one-port as your load.
 
 
[[File:RFAC3.png|thumb|600px| A more realistic version of the previous RF circuit including connecting transmission line segments.]]
 
==Using Transmission Lines for Connecting Parts==
 
You can use Generic T-Line segments or physical transmission line types to connect the various RF parts and devices in your circuit. The opposite figure shows the same simple RF circuit of the previous figure, but containing two T-line segments, one connecting the source to Port 1 of two-port N1 and the other connecting Port 2 of N1 to the resistive load. Note how the negative input and output pins of both T-line segments have been grounded. Lossless transmission line segments cause a phase shift of the propagating signal, while lossy [[Transmission Lines|transmission lines]] also cause additional signal attenuation.
 
 
[[File:RFAC4.png|thumb|200px| Setting the parameters in the AC Frequency Sweep Test Panel.]]
 
==Defining RF Circuit Observables==
 
Before you run an AC frequency sweep test of your RF circuit, you need to set the sweep [[parameters]] and define the output data for plotting or tabulation. The frequency sweep [[parameters]] are set from the test panel of the Toolbox. You set the start and stop frequencies, the frequency interval type (typically linear) and the frequency step. In the lower part of the test panel, you define your simulation output data. Click on the buttons labeled "Preset Graph Plots..." or "Present Table Plots..." to open the Edit Plot List dialog. The dialog gives a list of all the node voltages and currents. You can choose different complex data formats such as Mag/Phase, dB/Phase or Real/Imag.
 
 
{{Note | Typically the input and output voltage, currents and powers are of primary interest. These are measured at the input port (between the source impedance and input transmission line segment) and output port (between the output transmission line segment and the termination load).}}
 
 
In the previous RF circuit, the input port has node index 2 and the output port has node index 5. Therefore, the voltage v(2) and v(5) are designated as the simulation output data. The figure below shows a plot of the computed input and output voltages over the frequency range 1-10 GHz. A frequency step size of 10MHz has been set for the frequency sweep.
 
 
<table>
<tr>
<td>
[[File:RFAC5.png|thumb|420px| Defining the input and output voltages for the previous RF circuit.]]
</td>
<td>
[[File:RFAC6.png|thumb|550px| Graph of the computed input and output voltages of the previous RF circuit over the frequency range 1-10 GHz.]]
</td>
</tr>
</table>
==Network Analysis==
The Network Analysis Test characterizes a circuit with one or two ports. It is used to determine the circuit’s behavior as seen through its port(s) and generates data in the form of Z (impedance) [[parameters]], Y (admittance) [[parameters]], S (scattering) [[parameters]] , and H [[parameters]]. In the case of a two-port circuit, the four Z [[parameters]] are Z11 (the input impedance), Z22 (the output impedance), and Z12 and Z21 (the cross-impedances). Network analysis is often used for characterization of circuits that operate at very high frequencies. The components operating at these frequencies are often modeled with tables of S [[parameters]].
</tr>
</table>
 
Next, we consider the same RLC circuit as a two-port network. The first port is left intact at the voltage source (between Node 1 and the ground), while Port 2 is set up across the 50-Ohm resistor between Node 2 and the ground. The figures below show the Cartesian plot of the magnitude of the S11, S12, S21 an S22 [[parameters]] as well as the Smith chart for the two-port network.
</table>
=== Network Analysis of RF Circuits ===
RF circuits are typically characterized as [[File:b2MAN_Fig61.pngMultiport Networks|thumb|200px|Noise Test Settings.multiport networks]]==Noise Test== The Noise Test analyzes the device(usually one-generated noise over a range port or two-port). In many practical cases, rather than computing the input or output voltages or currents, you might be more interested in the port characteristics of frequenciesyour RF circuit. For this test one-port circuits, you provide would designate an input sourceport (Port 1) and would like to calculate its return loss or input impedance. For two-port circuits, you would specify an input port (Port 1) and an output port (Port 2) and a range of frequencieswould be interested in finding its insertion loss or gain. The most commonly used set of [[B2.Spice A/Dparameters]] calculates for RF circuit characterization are the noise contributions of each device scattering (and each noise generator within each deviceS) to the output port voltage[[parameters]]. The program also calculates the equivalent to the output noise referred back to the specified input source. This is what is meant by the "Input NoiseNetwork Analysis Test". The calculated value is one of the noise over the specified range AC-type [[tests]] of frequencies corresponds to the spectral density of the signal viewed as the square root of a stationary Gaussian stochastic process[[B2. After calculating the spectral densitiesSpice A/D]], the simulator integrates these values over the frequency range which is of particular importance to arrive at the total noise voltage/current over this frequency range[[RF. This calculated value corresponds to the variance Spice]]. Network analysis computes four sets of [[parameters]]: S, Z, Y and H. Of these, S-[[parameters]] and the circuit variable viewed as the square root "Smith Chart" are of a stationary gaussian processprimary interest, although Z-[[parameters]] are also frequently sought.
As for the specific noise sources, they include the "shot" noise associated with the DC currents in semiconductor devices and the "thermal" noise associated with resistance. Semiconductors also display "flicker" (1/f) noise. However, due to the lack of a unified model, [[B2.Spice A/D]] handles this type of noise on a "case by case" basis. There are flicker noise [[parameters]] available for transistors in the list of model [[parameters]]. The program can also provide the noise gain associated with the 1/f source.
To run a network analysis of your RF circuit, open the Test Panel of the Toolbox. Check the checkbox labelled "Network Analysis" and then open the corresponding Settings Dialog. The top part of this dialog has three separate tabs: Connections, Sweep and Output, as shown below. In the Connections tab, you set the input port of the circuit as well as the output port, if it is a two-port network. The ports are defined by specifying their positive and negative pins. You also have to specify the port reference impedance (Z0). The default value of Z0 is 50 Ohms. The Sweep tab of the dialog is identical to the sweep section of AC Sweep Test Settings dialog. Here you set the start and stop frequencies and the step size. In the Output tab, you specify which port characteristics to compute at the end of the network analysis. You can choose only one of the four parameter sets: S, Z, Y or H. All [[parameters]] can be plotted on cartesian graphs with three data formats: Amp Only, Amp/Phase or Real/Imag. The magnitude data can be plotted on either linear or dB scales. S-[[parameters]] are the only option that can generate either a Smith chart or a polar graph.
The [[parameters]] of the noise test are specified using the setup dialog of the Test Panel. The input noise source can be either a voltage or current source, which you can select from the drop-down list labeled "Noise Source". The output port must be specified as a pair of node numbers, the "Output Node" and the Reference Node". The rest of the [[parameters]] specify the frequency interval over which the noise analysis will be performed. They are similar to an AC analysis. If you check the checkbox labeled "Include Noise Generator Contributions", then noise contributions of all the individual devices will be shown as well. If not checked, the input noise and output noise will be shown only.
<table>
<tr>
 
<td> [[File:Net1.png|thumb|200px|Network Analysis Settings: Connections Tab.]]
</td>
<td> [[File:Net2.png|thumb|200px|Network Analysis Settings: Sweep Tab.]]
</td>
<td> [[File:Net3.png|thumb|200px|Network Analysis Settings: Output Tab.]]
</td>
 
</tr>
</table>
 
 
[[File:RFAC3.png|thumb|600px| An RF circuit consisting of a two-port network and connecting transmission line segments.]]
 
As an example, consider the RF circuit shown in the opposite figure, which was earlier examined in the discussion of AC frequency sweep test. This circuit can be treated as a one-port network with its input port defined between nodes 2 and 0, i.e. between the source's internal resistor and the input T-line segment. As a one-port, the circuit has a single s11 and a single z11 parameter. The first figure below shows the Smith chart for the return loss (s11) over the frequency range 1-10 GHz at larger steps of 100MHz. The second figure below shows Cartesian plots of the real and imaginary parts of z11 over the same frequency range both with finer steps of 10MHz.
As an example, consider the RTL inverter circuit of Tutorial Lesson 3, which is shown here in the opposite figure. The voltage source Vs is designated as the input noise source. The output port is set at node 3, i.e. the collector of the bipolar junction transistor. The noise is calculated over the frequency range from 1kHz to 10MHz on a decade scale with 10 steps per interval. The figures below show the graph of the noise spectral density over the specified frequency ranges as well as the tabulated results of the noise analysis. The latter are the integrated noise results, representing the total input and output noises as a result of contributions from the discrete devices in the circuit (namely, the resistors and the BJT). The results shown below are the square root of what Berkeley SPICE generates. This is because Berkeley SPICE gives values that are proportional to the noise power rather than the noise voltage.
<table>
<tr>
<td> [[File:b2MAN_Fig240Net4.png|thumb|300px640px|The RTL inverter circuit of Tutorial Lesson 3Smith Chart.]]
</td>
<td/tr> [[File:b2MAN_Fig241.png|thumb|500px|Plot of input and output noise spectral densities.]]</tdtr><td> [[File:b2MAN_Fig242Net5.png|thumb|370px640px|The integrated input and output noise resultsCartesian graph of Z-parameters.]]
</td>
</tr>
</table>
 
<p>&nbsp;</p>
[[Image:Back_icon.png|40px]] '''[[RF.Spice_A/D | Back to RF.Spice A/D Wiki Gateway]]'''
Kazem Sabet
28,333
edits