Today, let’s discuss the RF amplifier model structure with you~

This article will explore some RF amplifier model structures that combine linear S-parameter data with nonlinear data such as noise figure, IP3, P1dB, and PSAT. System-level simulation results will also be presented to assess how accurately they model the actual characteristics.

Table 1. Typical Sys parameter data set

S-parameter data sets are by far the most widely used RF simulation models. They are standardized tabular data sets that include input return loss, gain, reverse isolation, and output return loss at different frequencies, all in vector format. The data is generally collected under small signal conditions where the drive signal is well below the signal compression point. S-parameters are often used for cascaded gain simulations, the design of input and output matching networks, and the evaluation of stability. However, S-parameters do not contain information about the noise, compression, or distortion characteristics of the device.

Table 1 lists a portion of the sys-parameter data set for the ADPA7002, an 18 GHz to 44 GHz, 0.5 W power amplifier. This sys-parameter device model structure was defined by Keysight for use with its PathWave RF Synthesis (Genesys) and PathWave System Design (SystemVue) RF circuit and system simulators. The tabular structure of the data set includes S-parameter data at different frequencies along with the corresponding noise, third-order intermodulation, and 1 dB compression data. These data sets provide enough information to support simulations of RF signal levels, cascaded gain, and reverse isolation. However, the inclusion of IP3, P1dB, and noise figure data opens up the possibility of RF power sweeps and signal-to-noise ratio simulations. In addition, simulations of higher-order signal characteristics, such as ACLR and EVM, can be performed over the device’s operating frequency range.

Analog Devices maintains an extensive library of RF amplifier and mixer sys-parameters that is available for download and is also included in the Keysight Genesys and SystemVue installers. Figure 1 shows a screenshot of Keysight Genesys. Analog Devices’ sys-parameter library is easily accessible through the device selector. The sys-parameter device model for each device contains the data shown in Table 1, as well as additional information included in the model properties window. This additional data includes power supply information and default offsets for _PSAT and OIP2 relative to OP1dB.

Figure 1. Keysight Genesys screenshot showing a typical sys-parameter model.

To evaluate the accuracy of the sys-parameter model, we will now perform a series of comparisons between measured results and simulations. Figure 2 shows the measured and simulated results of a power sweep at 10 GHz for the HMC788A (10 MHz to 10 GHz RF gain block). As can be seen, the simulated power sweep is very close to the measured data. The simulator uses the gain and OP1dB data of the device and the P _SAT _Delta to generate the graphs shown. In this case, the P _SAT _Delta is 2 dB. This results in a P _SAT value that is 2 dB higher than the OP1dB level, which is the typical default value for GaAs RF amplifiers.

Figure 2. Measured and simulated power sweeps of a gallium arsenide (GaAs) RF amplifier.

Figure 3. Simulation and measurement of AM-to-AM and AM-to-PM distortion.

Figure 4. Simulated and measured power sweeps for the HMC1114 (3.2 GHz, 10 W GaN amplifier).

To examine the simulated compression behavior in more detail, we can look at the AM-to-AM and AM-to-PM distortion. The measured and simulated results shown in Figure 3 are for the HMC930A. The measured AM-to-AM distortion is very close to the simulation. However, the simulation results do not show the AM-to-PM distortion, which is incorrect. This is because the device model and data set only contain small-signal phase information (i.e., S21). While the simulator can use the OP1dB and P _SAT _Delta data from the device model to estimate the AM-to-AM distortion, it does not have any large-signal S-parameter data to use. In this case, using a more detailed model, such as in the X-parameter format (the X-parameter model has built-in level-dependent S-parameters), would be appropriate.

Figure 4 shows a power sweep of the HMC1114LP5DE, a 10 W gallium nitride (GaN) RF amplifier, at 3.2 GHz. GaN RF amplifiers tend to have much milder compression characteristics than GaAs devices. This requires adjustment of P _SAT _Delta, which is the difference between the 1 dB compression point and the saturation point. In this case, the delta has been set to 7 dB based on observed measurements. While the simulator will generate warnings in some cases due to the large delta, it will still simulate correctly and produce results very close to the measured performance.

As we move from CW signal measurement and simulation to modulated signals, the value of sys-parameter data sets becomes greater. While information about device gain, compression, IP3, and noise figure is readily available in device data sheets, curves showing performance under modulated signals are unlikely to be found in data sheets for devices designed for general use. Additionally, metrics such as ACLR and EVM are not easy to predict without simulation or measurement.

Figure 5 shows the simulation results of a power sweep of a 0.25 W ADL5320 driver amplifier at 2140 MHz driven by a 5 MHz wide carrier. The simulated carrier consists of 11 evenly spaced subcarriers, and the ACLR is measured at a 5 MHz carrier offset.

Figure 5. ACLR simulation.

The simulation shows that the ACLR reaches its optimum value at an input power of –15 dBm. Below this input power, the ACLR decreases with input power at a rate of 1 dB/dB. This region of the curve is dominated by the noise figure data. As the input power increases above –15 dBm, the rate at which the ACLR decays is closely related to the IP3 of the device. It is important to note that the results of this simulation rely on both the noise figure data (at low power) and the IP3 data (at high power) to produce an ACLR sweep that is accurate over a wide power range.

The plot also includes measured data (blue). It does not reach the same optimal level for an input power level of –15 dBm, which is due to limitations of the measurement setup. It is noteworthy that the measured ACLR drops faster as the input power level increases. This is because the OIP3 of the device degrades slightly with input/output power levels (ideally, it should not change). The IP3 in the device model data set is a single data set and does not vary with power level; it can be considered the small signal IP3 of the device. This is another example where an X-parameter model and its more detailed modeling of level dependencies may produce a more accurate simulation.

The sys-parameter models can also be used to reliably perform EVM simulations. Figure 6 shows the measured and simulated EVM versus RF power sweep with a 1 MSPS, 16 QAM carrier driving a 50 MHz to 4 GHz gain block ADL5602. This shows excellent correlation between measurement and simulation at both low and high power levels.

The default sys-parameter data set in the ADI library contains only ambient temperature data. However, the model can be extended by adding additional worksheets to the data set that contain temperature data. Figure 7 shows a data set for the ADPA7007, an 18 GHz to 44 GHz, 1 W power amplifier. This data set has multiple worksheets containing gain, noise, and distortion data at –55°C, +25°C, and +85°C. These three data points can be used by the Genesys and SystemVue simulators to generate interpolated data at other temperatures, as shown in Figure 7.

The sys-parameter data set is native to Keysight Genesys and SystemVue, but not to Keysight ADS. There is a workaround to import the sys-parameter data set into ADS for noise, distortion, and compression simulations. This requires the use of an Amplifier2 model. The Amplifier2 model is native to Keysight ADS and provides similar functionality to the sys-parameter model. Figure 8 shows an ADS schematic that includes the Amplifier2 model. The schematic also contains two data access devices: DAC1 and DAC2. These DACs are used to associate the sys-parameter data with the Amplifier2 model. The noise figure, OIP3, and OP1dB data are formatted into a text file and associated with the Amplifier2 model through the DAC1 device. The DAC2 device is used to associate the S-parameter data with the Amplifier2 model. This will produce an Amplifier2 model in ADS that can be used to perform all the simulations discussed above, but in Keysight ADS.

Be careful when using this method. When an RF power sweep is performed and the Amplifier2 model is driven into compression, the simulated performance often differs significantly from the measured performance observed. In addition, creating an Amplifier2 model that uses S-parameter data along with noise, distortion, and compression data is appropriate for devices that have good baseline input and output return losses (S11 and S22), which is the case for most ADI RF amplifiers that do not require external RF matching components. A simpler Amplifier2 model can be created by adding scalar gain to the DAC1 device and omitting the S-parameter data (that is, omitting DAC2).

Figure 6. Simulated and measured EVM power sweeps for a wideband gain block.

Figure 7. Simulated gain and noise figure vs. temperature for the ADPA7007, an 18 GHz to 44 GHz, 1 W power amplifier.

Sys-parameter data sets represent a novel and useful tool for RF amplifier simulation. They are more powerful than S-parameters, which cannot model noise, distortion, and compression. They are not as sophisticated as X-parameter models, which can improve model-level dependent characteristics such as AM to PM distortion and ACLR. However, sys-parameter models have a simple tabular structure and can be easily created by combining S-parameter data with noise figure, OIP3, and OP1dB data. Comparison of simulated and measured data shows excellent agreement. Although sys-parameter models cannot be used in ADS, a relatively simple process can be used to migrate the data sets to use the Amplifier2 model structure native to ADS.

Analog Devices is committed to maintaining and expanding its sys-parameter model library. As new models are added to the library, we will add support for temperature simulations.

Figure 8. Using sys-parameter data in Keysight ADS using the Amplifier2 model.