***性和非线性器件模型？

对于直流电源：1。请取出R3。
你不需要它。
实际上它耗散了18mW的测量直流功率，而你却没有得到任何东西。
没有必要设置直流偏置。
你可以获得大约60mA的漏极电流，R3 = 0（或去掉），R1设置在29k欧姆左右。
您可以根据需要调整值以设置所需的电流。
2.就结果而言，连接输入/输出负载是无关紧要的。
您已连接DC模块，因此DC模拟将无法看到它们或受其影响。
如果DC分析单独完成，您可以使用Terms，P1_tone源，P2_tone源或开路。
这没关系。
如果DC分析是作为谐波平衡仿真的一部分完成的，则DC分析首先作为该仿真的一部分自动运行，以设置HB仿真的起始条件，那么在这种情况下，您需要适当的HB1_Tone源/负载
或想要运行的HB2_Tone模拟。
在这种情况下，源或负载的选择不会改变DC结果。
3.为了更准确地测量直流功耗，I_Probe或用于测量总电流的第二个I_Probe应在R1之前直接连接到V_DC源。
您没有测量通过R1 / R2到地的电流或任何栅极电流。
或者，您也可以使用来自V_DC源的总电流SRC1.i而不是IDS.i来计算电路的直流功耗。



     以下为原文

  For DC Power: 

1. PLEASE TAKE OUT R3. You do not need it. In fact it is dissipating 18mW of your measured DC power and you are getting nothing for it. It is unnecessary for setting the DC bias. You can get exactly the same Drain current of about 60mA with R3 = 0 (or removed) and R1 set somewhere around 29kOhms. You can tune the value as desired to set the desired current.

2. It does not matter as far as the results are concerned what input/output loads are connected. You have DC blocks connected so DC simulations will not see them or be affected by them. If the DC analysis is done on its own you can have Terms, P1_tone sources, P2_tone source or open circuits. It really does not matter. If the DC analysis is done as part of a Harmonic Balance simulation, a DC Analysis is automatically run first as part of that simulation to set the starting conditions for the HB simulation, then in that case you need the appropriate sources/loads for the HB1_Tone or HB2_Tone simulation you want to run. The choice of source or load in this case will not alter the DC results. 

3. To more accurately measure the DC power dissipation the I_Probe, or a second I_Probe used to measure the total current, should be connected directly to the V_DC source before R1. You are not measuring the current through R1/R2 to ground or any gate current. Alternatively you could also use SRC1.i, the total current from the V_DC source, instead of IDS.i to calculate the DC power dissipation by the circuit.

模型的选择：在所有情况下，这个问题都没有明确的答案。
正如您在使用线性s参数模拟进行某些分析之前所提到的，一些使用非线性谐波平衡模拟。
每种类型的模型有时可以在一种样式中使用，有时在另一种样式中使用，在某些情况下可以在两者中使用
例如，IP3和P1dB压缩是大信号测量，电路的响应取决于输入信号的特性，即幅度，DC偏移等，并且这些可以使用谐波平衡模拟来完成。
这必须使用非线性样式模型。
噪声系数，群延迟（品质因数）和稳定性通常是小信号测量，其中输入信号的特性（频率除外）在计算中不予考虑，这是使用线性（S参数）完成的
模拟。
对于这些，可以使用样式模型，线性或非线性。
最常使用的是线性模型。
增益和带宽是不同的，因为它们可以使用谐波平衡和线性模拟进行测量。
对于线性模拟，增益为dB（S21），这可以作为频率的函数来改变，以定义带宽但是独立的信号幅度，这可以使用线性模型或非线性模型来完成。
对于谐波平衡，增益可以测量为Pout / Pin，但可以定义为不同的频率（谐波），也可以表示为输入信号幅度的函数，这显然是P1dB测量的一部分，但需要进行任何非线性模拟。
模拟功率相关性能必须使用非线性样式模型。



     以下为原文

  Choice of Models:

There is no definite answer to this question in all cases. As has been mentioned to you before some analysis are performed using linear s-parameter simulations some using nonlinear harmonic balance simulations. Each type of model can be used sometimes in one style, sometimes in the other and in certain cases in both.

For example IP3 and P1dB compression are large signal measurements, the response of the circuit depends on the characteristic of the input signal(s), i.e. amplitude, DC offset etc., and these can be done using Harmonic Balance simulation. This must use a nonlinear style model.

Noise Figure, Group Delay (Figure of Merit) and stability are typically small signal measurements, where the characteristics of the input signal(s), except frequency, are not considered in the calculations, and this is done using Linear (S-Parameter) simulation. For these is is possible to use either style model, linear or nonlinear. Most often though linear model are used.

Gain and Bandwidth are different as they can be measured using both Harmonic Balance and Linear simulations. For Linear simulation the gain is dB(S21) and this can change as a function of frequency to define the bandwidth but is independant signal amplitude and this can be done using either the linear model or the nonlinear model. For Harmonic Balance the gain can be measured as Pout/Pin but can be defined at different frequencies (harmonics) and can also be expressed as a function of input signal amplitude and this is obviously part of the P1dB measurement but any nonlinear simulation that required to simulate power dependant performance must use a nonlinear style model.

> {quote：title = sdgrant写道：} {quote}>模型的选择：>>在所有情况下都没有明确的答案。
正如您在使用线性s参数模拟进行某些分析之前所提到的，一些使用非线性谐波平衡模拟。
每种类型的模型有时可以在一种样式中使用，有时在另一种样式中使用，在某些情况下可以在两者中使用
例如，IP3和P1dB压缩是大信号测量，电路的响应取决于输入信号的特性，即幅度，DC偏移等，并且这些可以使用谐波平衡模拟来完成。
这必须使用非线性样式模型。
>>噪声系数，群延迟（品质因数）和稳定性通常是小信号测量，其中输入信号的特性（频率除外）在计算中不予考虑，这是使用线性（S-）完成的。
参数）模拟。
对于这些，可以使用样式模型，线性或非线性。
最常使用的是线性模型。
>>增益和带宽是不同的，因为它们可以使用谐波平衡和线性模拟进行测量。
对于线性模拟，增益为dB（S21），这可以作为频率的函数来改变，以定义带宽但是独立的信号幅度，这可以使用线性模型或非线性模型来完成。
对于谐波平衡，增益可以测量为Pout / Pin，但可以定义为不同的频率（谐波），也可以表示为输入信号幅度的函数，这显然是P1dB测量的一部分，但需要进行任何非线性模拟。
模拟功率相关性能必须使用非线性样式模型。
亲爱的先生，用于测量不同参数的线性（S参数）模型或非线性模型的模型选择是不同的。
我怀疑在S11，S22和S12测量中选择哪种型号。
我确信线性（S-PARAMETER）模型适用于测量S11，S22和S12。
请确认它是测量S11，S22和S12的正确型号，因为我是初学者。编辑：btv_murthy@rediffmail.com于2014年6月29日凌晨2:35



     以下为原文

  > {quote:title=sdgrant wrote:}{quote}
> Choice of Models:
> 
> There is no definite answer to this question in all cases. As has been mentioned to you before some analysis are performed using linear s-parameter simulations some using nonlinear harmonic balance simulations. Each type of model can be used sometimes in one style, sometimes in the other and in certain cases in both.
> 
> For example IP3 and P1dB compression are large signal measurements, the response of the circuit depends on the characteristic of the input signal(s), i.e. amplitude, DC offset etc., and these can be done using Harmonic Balance simulation. This must use a nonlinear style model.
> 
> Noise Figure, Group Delay (Figure of Merit) and stability are typically small signal measurements, where the characteristics of the input signal(s), except frequency, are not considered in the calculations, and this is done using Linear (S-Parameter) simulation. For these is is possible to use either style model, linear or nonlinear. Most often though linear model are used.
> 
> Gain and Bandwidth are different as they can be measured using both Harmonic Balance and Linear simulations. For Linear simulation the gain is dB(S21) and this can change as a function of frequency to define the bandwidth but is independant signal amplitude and this can be done using either the linear model or the nonlinear model. For Harmonic Balance the gain can be measured as Pout/Pin but can be defined at different frequencies (harmonics) and can also be expressed as a function of input signal amplitude and this is obviously part of the P1dB measurement but any nonlinear simulation that required to simulate power dependant performance must use a nonlinear style model.
Dear Sir,
             Choice of model either linear (S-parameter) model or nonlinear model for measuring different parameter are different. I had doubt on which model to choose for S11,S22 and  S12  measurement.
               I am sure that linear (S-PARAMETER) model is suitable for measuring S11,S22 ,and S12.
               Please confirm it is correct model for measuring S11,S22, and S12, because , I am beginner.

Edited by: btv_murthy@rediffmail.com on Jun 29, 2014 2:35 AM

对于布局：1。您没有“制造”线性设备或非线性设备。
布局使用打包设备的轮廓。
该封装器件可以是模拟线性或非线性模拟，但它只是布局的轮廓。
2.布局不需要偏置源。
但是，要在物理电路中连接这些源，您需要为这些互连创建适当的布局。
这些互连将导致布局寄生效应，并且所有这些寄生效应需要在设计中正确建模，以便针对BOTH线性s参数模拟和非线性谐波平衡模拟进行适当模拟。
3A。
谐波平衡仿真和使用非线性模型的线性s参数仿真始终要求将直流偏置源连接到直流馈电连接。
3B。
使用线性s参数模型的线性s参数模拟从不要求DC偏置源连接到DC馈电连接，根据配置，它们可以由0V DC电源或短路/开路电路代替，但由于这些DC导致的寄生效应
必须包括所有情况下的馈送连接。



     以下为原文

  For Layout:

1. You are not 'fabricating' a linear device or a nonlinear device. The layout uses the outline of a packaged device. That packaged device can be models as linear or nonlinear for simulation but it is just an outline for layout.

2. Layout does not need the bias sources. However, to connect those sources in the physical circuit you need to create the appropriate layouts for these interconnects. These interconnects will result in layout parasitics and all these parasitics need to be correctly models in the design so that they are simulated appropriately for BOTH linear s-parameter simulations and nonlinear Harmonic Balance simulations.

3a. Harmonic Balance simulation and linear s-parameter simulations using nonlinear models ALWAYS requires DC bias source to be connected to the DC feed connections.

3b. Linear s-parameter simulation using linear s-parameter models NEVER require DC Bias sources to be connected to the DC feed connections, they can be replaced by 0V DC sources or short/open circuits depending on the configuration, but the parasitics due to these DC feed connections must be included in all cases.

sdgrant先生，非常感谢你的耐心，精心设计，知识全面，深刻的解释，先生。
我有很多疑惑和困惑，现在已经清除了。
先生，我很感谢你。



     以下为原文

  sdgrant Sir,
                        Thank you very much for your patience, elaborate, knowledge full ,  deep explanatory answer from you sir. I had many doubts and confusions , now it is cleared. I am grateful to you sir.

首先简要说明连接线。
在你的两个图像中，Nonlinear_Inductor_connectionbetween_two_wires.png（在S1）和Inductor_at_Source_S1.png（在地面连接到S2）你有断开连接。
针仍然显示为红色钻石。
这意味着每根电线未连接到该位置的符号。
请确保正确捕捉每个引脚的导线连接。
现在接通电线。
两个源极引脚内部连接在一起。
如果真的有必要，你可以将一个断开连接，但这可能不是一个好主意。
在内部，GaAs FET晶体管芯片通常在芯片的相对端具有两个源极连接，栅极和漏极连接也彼此相对并且与源极正交。
请参见此处的典型结构：http：//www.microwaves101.com/encyclopedia/fets.cfm设备的每一侧都连接到包装上的两个选项卡之一。
因此，如果您正在执行除将源直接连接到地之外的任何操作，则必须将所需组件连接到两个引脚。
您展示的两个示例的接地电感连接到其中一个引脚，另一个引脚直接接地或电感的接地端。
不要那样做!!
你正在缩短电感器。
正如我所说，你真的需要将两个引脚连接在一起，然后将电感连接到地。
在实践中，虽然在创建布局时不需要这样做。
因此，一种选择是将单独的电感器连接到每个源极引脚接地。
这两个电感器并联有效连接，因此2x 2nH电感器代替1x 2nH电感器。
如果你直接将这两个光源接地，那么无论如何你都要做的是将每个光源直接连接到顶侧接地层或通过过孔连接到背面接地层。



     以下为原文

  First a brief note on connecting wires. In 2 of your images Nonlinear_Inductor_connectionbetween_two_wires.png (at S1) and Inductor_at_Source_S1.png (at ground connected to S2) You have disconnects. The pins are still showing as red diamonds. This means that each wire is NOT connected to the symbol at that location. Please make sure you snap the wire connections to each pin correctly.

Now on to the wiring. The two source pins are internally connected together. You could leave one disconnected if really necessary but it is probably not a good ideas. Internally the GaAs FET transistor chip typically has two source connections at opposite ends of the chip, the gate and drain connections are opposite each other also and orthogonal to the sources. See a typical structure here:

http://www.microwaves101.com/encyclopedia/fets.cfm

Each side of the device is connected out to one of the two tabs on the package. So if you are doing anything other than connecting the source directly to ground then the required component has to be connected to the both pins together. Two of the examples you show have a grounded inductor connected to one of the pins and the other pin grounded directly or the grounded end of the inductor. DON'T DO THAT!! You are shorting out the inductor. As I said you really need to connect the two pins together and then connect the inductor to ground from that. In practice though that is not convenent to do when you create the layout. So one option would be to connect a separate inductor to each source pin to ground. These two inductors are the effectively connected in parallel, so 2x 2nH inductors instead of 1x 2nH inductor. If you were directly grounding the two sources this is what you would do anyway but connecting each to a top side ground plane directly or through vias to a backside ground plane.

btv_murthy，除了sdgrant为您提供的培训材料外，安捷伦还提供现场培训课程。
我建议您访问http://edocs.soco.agilent.com/display/support/ADS+Support+Home联系您当地的技术支持工程师，他们可以提供其他培训选项方面的帮助。
ADS论坛是一个社区，旨在让您直接与世界各地的其他ADS用户进行交流，旨在提供一个询问和回答其他用户的问题，交流想法以及分享使用ADS的提示和技巧的地方。
ADS论坛不是联系Agilent EEsof获得培训或技术支持帮助的最佳或主要方式。
注册用户可以访问ADS支持主页访问ADS支持和文档。
Joe Civello ADS产品经理



     以下为原文

  btv_murthy,

In addition to the training materials sdgrant has provided you, Agilent also offers live training classes. I suggest you visit http://edocs.soco.agilent.com/display/support/ADS+Support+Home to contact your local tech support engineer, who can help with with other training options.

The ADS Discussion Forum is a community created to allow you to communicate directly with other ADS users around the world and is designed to be a place to ask and answer other users’ questions, exchange ideas, and share tips and tricks for using ADS. The ADS Discussion Forum is not the best or primary means to contact Agilent EEsof for assistance with training or technical support. Registered users may access ADS Support and Documentation by visiting ADS Support Home. 

Joe Civello
ADS Product Manager