Examples of AC simulation and S-parameter simulation of ADS simulation

This paper introduces the use of ADS2009 Simulation software for DC simulation , It is used to test the AC working characteristics and S Parameter simulation .
One 、 AC simulation
AC simulation can be used to analyze the small signal characteristics of circuits , It can also analyze the noise characteristics of the circuit , During small signal AC simulation , All systems need DC simulation of the circuit , Find the DC operating point of the nonlinear device . Through the small signal analysis of the circuit , You can get the voltage of the circuit / Current gain 、 Span resistance and other parameters .

Chart 1 AC Simulation schematic

Chart 2 Simulation results
Two 、S Parameter simulation
When designing RF or microwave circuits , Node circuit theory is no longer applicable , It is necessary to adopt the analysis method of distributed parameter circuit , Microwave network analysis is a popular method at present , For Microwave Networks ,S Parameter is a very important parameter to characterize the transmission characteristics of microwave network .
2.1 S General process of parameter simulation
1、 Select the device model and establish the schematic diagram ;
2、 determine S Parameter simulation IO port , And add terminal load control Term Respectively connected to IO port ;
3、 add to S Parameter emulation control , And set up S Parameters simulation parameters ;
4、 If you need to scan many variables , Need to add parameter sweep Control , At the same time, you can also add other controls for calculation as needed ;
5、 If you need to calculate noise or group delay , Need to be in noise Check... In the tab calculate noise and parameters Check... In the tab group delay;
6、 Run the simulation ;
7、 View the simulation results in the data window .
Here, a simple low-pass filter is used to introduce the general S Parameter simulation , The low-pass filter is 3 rank Chebyshev Type a filter , The ripple in the band is 0.5dB, The cut-off frequency is 3.0GHz, The schematic diagram after design is as follows ：

Chart 3 chebyshev Low pass filter simulation schematic
According to the schematic diagram above , stay ADS Establish the corresponding simulation schematic diagram ：
1、 Capacitance and inductance components can adopt lumped parameter components , stay Lumped-components Select inductance and capacitance in the panel and add them to the schematic ;
2、 stay Simulation-S para Add S Simulation control , And set the corresponding frequency scanning range 、 Scanning type and other parameters ;
3、 Add two Term terminal , Connect to input port and output port ; Impedance can be used by default 50 Ohm;
4、 You can omit it 2-3 step , Adopt menu insert->Template->S params To insert S Parameter simulation template ;
5、 Connect the schematic diagram according to the above figure and simulate ;
6、 View the simulation results in the data display window .
The simulation results of the above low-pass filter are as follows ：

Chart 4 Simulation results
2.2 tuning
For a preliminary design , There may be some differences between the simulation results and expectations , It is often necessary to make some minor changes to the circuit to meet the requirements , And in many cases , We need to know the influence trend of the change of a certain component on the whole result , This can be achieved by tuning .
For example, we designed a 5 rank chebyshev low pass filter , The ripple in the band is 0.5dB, The cut-off frequency is 3GHz, According to
The schematic diagram of the design is as follows ：

Chart 5 5 Tuning schematic diagram of first-order low-pass filter
And S Parameter simulation is similar , After establishing the simulation schematic , Conduct S Parameter simulation , The simulation results show that the cut-off frequency is 3GHz, Next, we can see the change trend of the filter through tuning ：
1、 Click... In the schematic , perhaps simulate->tuning, open tune parameter window ;
2、 Click the component to be tuned in the schematic L1, Check... In the pop-up dialog box L1, After confirmation, in tune parameter The window will show L1 Tunable state of , As shown in the figure below , Also select all components , Let it all enter the tunable range , here tune parameter The window shows all components to be adjusted ;
3、 stay tune parameter Click in the window enable/disable Button , Select all components , Means to enable all elements to be tunable ;
4、 stay tune parameter The tuning range of the element can be adjusted in the window 、 Step and tuning parameter change types ;
5、 Adjust the size of the data display window before tuning , And will history choice ON Pattern , In order to observe the change trend of the curve during tuning ;
6、 Moving the slider or arrow can change the value of the component , At the same time, you can see S21 Change trend of .
7、 After tuning, you can update To the schematic diagram .

Chart 6 Tuning window

Chart 7 tuning S21 Curve trend chart
From the simulation results, we can see , Increase capacitance C1 C2 C3 Value , The cut-off frequency of the filter decreases , The low-frequency passband bandwidth is correspondingly reduced , High frequency attenuation is larger than the original .
2.3 Optimize
Optimization is ADS Provides a method for calculating circuit parameters , Optimization requires setting a target value , Simultaneous setting ADS Available
Adjusted variables , By scanning the set variables , Calculate a set of variable parameters that meet the target value . The optimized control panel is Optim/Stat/Yield/DOE, The panel contains optimizations 、 Statistics 、 Yield and proprietary equipment simulation controls .

Chart 8 Optimize controls and panels
The optimization of a circuit often needs to include optimization controls 、GOAL Control ,GOAL Control is used to set a target state that the circuit wants to achieve . Below we take a 5 Order low-pass filter to illustrate the optimization process . The goal of optimization is to 5GHz Of S21 The maximum value is -40dB, stay 3GHz and 2.4GHz ±0.1GHz Within the scope of S21 The maximum value of is -1dB.
1、 stay 5.2 Section established 5 In the schematic diagram of the first-order low-pass filter , Replace the value of each component with a variable , As shown in the schematic diagram below , And add variable equations to the schematic , The variable named in this schematic diagram is c1~c3,l1、l2;
2、 double-click , Add above 5 A variable , Initial value , As shown in the figure below , And in Tune/opt/stat/DOE setup Set enable optimization options in the optimization tab 、 Variable change type and corresponding change range ;
3、 Add optimization control , Control settings generally follow the default settings , One thing that needs to be changed is the optimization method （Optimization Type）, Second, the number of optimization （Maxlters）. There are many optimization methods in the optimization method drop-down menu , They are based on different algorithms . Generally, if you are uncertain about the initial value of the variable, you can use Random, It's random , Suitable for large-scale search ; If the result is close to the expectation , Then it can be changed to Gradient, It will change the variable value in a small range . One more Discrete, Its value is discrete , Suitable for such as capacitors 、 resistance 、 Actual models of inductors and other devices .
4、 Add target control , Set the corresponding value as required .
5、 After optimization, pass simulate->update optimization values Update the optimized value .

Chart 9 Set optimization variable window
Schematic diagram after setting ：

Chart 10 Optimize schematic diagram
The simulation results are shown in the figure below ：

Chart 11 Optimization results
2.4 tolerance / Yield analysis
The reason for tolerance analysis , It is because there are certain errors in any electronic components , Such as inductance 、 Accuracy of capacitance, etc . For example, a nominal is 2.0nH±0.1nH The inductance of , Products that stand for 99.74% The probability of falls on 2.0nH±0.1nH Within the scope of , The meet 6σ,σ Is the standard deviation or variance , When the difference between the product random variable value and the average value is 6σ when , The yield of the product is 99.74%, This is the field of Statistics .
2.4.1 Tolerance analysis
The tolerance analysis of circuit modules can be done by Monte Carlo （Monte carlo） Analysis is carried out , Through tolerance analysis, we can know the influence of component error on circuit performance , A bandpass filter is taken as an example to illustrate the method of tolerance analysis . The following is a band-pass filter designed by optimization , Ask for in 2.4GHz-2.5GHz Passband interpolation loss less than 2dB, stay 3.2GHz The minimum attenuation value at is 20dB, stay 1.6GHz The minimum attenuation value at is 15dB, The design schematic diagram is as follows ：

Chart 12 3 Schematic diagram of step bandpass filter
Without adding Monte Carlo simulation control, the results are as follows

Chart 13 3 Simulation results of order bandpass filter
Tolerance analysis can be performed on several or one component , But at least 1 Variables of components with errors , Here to C1 and C3 The influence on the whole filter from the change of , The specific process is as follows ：
1、 Set the component value that needs tolerance analysis as a variable , And add variable equations VAR;
2、 stay VAR Set the initial value of the variable in , And in Tune/opt/stat/DOE setup Medium statistics Tab enable statistics status, At the same time, set the distribution type and error of variables . Set up here c1 and c3 Respectively ±0.25 and 5%.
3、 add to MONTE CARLO Control , Set up SimInstanceName and Numitem, That is, select the embedded simulator and sampling times , This example sets 10 Time .
4、 Set up OK Post run simulation , Check the simulation results as follows .

Chart 14 Tolerance analysis results
The blue curve is the template that the filter needs to conform to , Simulation can see C1 and C3 Balance the attenuation in the passband of the filter , In some cases 2.5GHz The attenuation value of has exceeded 2dB Insertion loss of . Therefore, such filter parameters can not deal with the influence of component error well .
2.4.2 Yield analysis
Yield analysis is used to analyze the ratio of the number of design circuits passing a given standard to the total number , But for a circuit design , The total amount of design that may exist cannot be estimated , Therefore, the yield analysis is carried out by a limited number of tests , When the number of tests is more , The closer it is to the truth .
The above bandpass filter optimization is not perfect , Let's analyze a previously designed 3 The yield of order low-pass filter , Definition 3 Order low pass filter spec by 0-3GHz The interpolation loss is 1dB, The minimum return loss is 15dB,4.8GHz-6GHz The minimum insertion loss is 10dB, Establish schematic diagram according to the above requirements ：
1、 Add... To the schematic Yield Analyzed controls and Yield SPEC Control , And set the corresponding Yield Of SimItem The number of
1000, And in parameters Check save data all for all trials To maintain all test data . Set up
Yield SPEC Control , Define the reference value of yield analysis .
2、 Yield analysis requires at least one variable , Here we analyze the impact of three components on the yield . The schematic diagram of setting is as follows ：

Chart 15 Schematic diagram of yield analysis
The simulation results are as follows ：

Chart 16 Yield analysis results
It can be seen that the yield of the filter is 76.5%, Change the device C1 and C2 The accuracy of the , The yield rate is increased to 81.7%.

Chart 17 Improve the influence result of accuracy and yield
Added senshist Control , It is used to count the simulation results , The influence degree of a certain variable on the yield , For example, in the schematic diagram, the control is set to sensHist1=histogram_sens(dB(S(1,1)),l1,-15,2.4GHz,2.5GHz), Representative means l1 This variable , stay 2.4GHz-2.5GHz In the frequency range , Yes S11 Less than -15dB The degree of influence , From the two simulation results , When l1 If the value of is too large , Yes S11 Less impact , So without change C1 and C2 In the case of accuracy , Set the value of inductance to 2.9nH Change it to 3.0nH when , The yield is 93%, The results are shown in the following figure , It can be seen that without changing the accuracy of the device , increase L1 The value of is of great help to the yield .

Chart 18 modify L1 Impact on yield